Home
  By Author [ A  B  C  D  E  F  G  H  I  J  K  L  M  N  O  P  Q  R  S  T  U  V  W  X  Y  Z |  Other Symbols ]
  By Title [ A  B  C  D  E  F  G  H  I  J  K  L  M  N  O  P  Q  R  S  T  U  V  W  X  Y  Z |  Other Symbols ]
  By Language
all Classics books content using ISYS

Download this book: [ ASCII | HTML | PDF ]

Look for this book on Amazon


We have new books nearly every day.
If you would like a news letter once a week or once a month
fill out this form and we will give you a summary of the books for that week or month by email.

Title: Observations of a Naturalist in the Pacific Between 1896 and 1899, Volume 1 - Vanua Levu, Fiji
Author: Guppy, H. B. (Henry Brougham)
Language: English
As this book started as an ASCII text book there are no pictures available.
Copyright Status: Not copyrighted in the United States. If you live elsewhere check the laws of your country before downloading this ebook. See comments about copyright issues at end of book.

*** Start of this Doctrine Publishing Corporation Digital Book "Observations of a Naturalist in the Pacific Between 1896 and 1899, Volume 1 - Vanua Levu, Fiji" ***

This book is indexed by ISYS Web Indexing system to allow the reader find any word or number within the document.



                          Transcriber's Notes

When italics were used in the original book, the corresponding text has
been surrounded by _underscores_. Subscripted characters have been
preceded by _ and surrounded by {}. In some cases, ditto marks have been
replaced by the text they represent. Some corrections have been made to
the printed text. These are listed in a second transcriber’s note at the
end of the text.



                  OBSERVATIONS OF A NATURALIST IN THE

                     PACIFIC BETWEEN 1896 AND 1899

[Illustration: [Image: Publisher]

[Illustration: NA RARO (2,420 feet) from the south-west, a peak of acid
andesite.]

[Illustration: NDRANDRAMEA (1,800 feet) from the south-east, a peak of
acid andesite rising about a thousand feet from its base.]

                                                        [_Frontispiece._



                            OBSERVATIONS OF
                            A NATURALIST IN
                          THE PACIFIC BETWEEN
                             1896 AND 1899



                                   BY
                      H. B. GUPPY, M.B., F.R.S.E.



                                VOLUME I

                           _VANUA LEVU, FIJI_
   _A description of its leading Physical and Geological characters_



                                 London
                       MACMILLAN AND CO., LIMITED

                    NEW YORK; THE MACMILLAN COMPANY
                                  1903

                         _All rights reserved_



                    RICHARD CLAY AND SONS, LIMITED.
                      BREAD STREET HILL, E.C., AND
                            BUNGAY, SUFFOLK.



                               Dedication

                          TO THE FIJIAN PEOPLE



                                PREFACE


DURING a sojourn in the Pacific, which covered a period of rather over a
year in Hawaii (1896-97), and of two years and three months in Fiji
(1897-99), my attention was mainly confined to the study of
plant-distribution and to the examination of the geological structure of
Vanua Levu.

With Hillebrand’s “Flora of Hawaii” always in my hands I roamed over the
large island of Hawaii, ascending the three principal mountains of Mauna
Kea, Mauna Loa, and Hualalai, and in the case of my second ascent of
Mauna Loa spending twenty-three days alone on its summit. Similarly in
Fiji, Seemann’s “Flora Vitiensis” was my counsellor and guide in the
matter of plants.

In Hawaii I was in a land of active sub-aerial volcanoes, and I paid my
devotions at all the altars of “Pele,” their presiding deity. In Fiji I
trod upon the surface of submarine volcanoes that emerged ages since
from the ocean and still retain their coverings of sea-deposits. Both in
Hawaii and Fiji I lived much among the people; and though my chief
interest lay in the comparison of these two types of volcanic islands, I
could not but be drawn to the kindly natives whose hospitality I so long
enjoyed.

Destiny led me to Vanua Levu in the following fashion. With the relief
party to take me down from Mauna Loa there arrived a well-known German
naturalist who, like myself, had been interested in coral-reef
investigations. We discussed this warm topic at an elevation of nearly
14,000 feet above the sea, with the thermometer at 20° F. As we sipped
our hot coffee and listened to the occasional “boom” from the bottom of
the great crater, at the edge of which we were camped, I remarked to my
friend that I was thinking of spending some months in Samoa. To this he
good-humouredly replied that I might leave Samoa to his countrymen and
describe one of the large islands of Fiji. International rivalry over
that group of islands was then rather keen. However, Dr. K. went to
Samoa, and I have now completed this volume on the geology of Vanua
Levu, Fiji.

It will not be necessary to lay stress here on the difficulties and
hardships connected with the exploration of little known tropical
regions. Many will be familiar with all that these imply, where the
rainfall ranges from 100 to 250 inches, where the forests are dense,
where tracks are few and swollen rivers are numerous, and where the
torrent’s bed presents often the only road.

The only extensive geological collections made in Fiji previous to my
visit were those of Kleinschmidt in 1876-78, which together with a small
collection previously made by Dr. Gräffe were examined by Dr. A.
Wichmann. These rocks were obtained from Viti Levu, Kandavu, Ovalau,
etc., but not from Vanua Levu. Dr. Wichmann’s paper of 1882, descriptive
of these collections, presents us with the results of one of the
earliest studies by modern methods of research of the volcanic rocks of
the Pacific Islands. It is to this investigator that we are indebted for
the establishment of the occurrence of plutonic rocks, such as granites,
gabbros, diorites, in Viti Levu.

Although, as far as I can ascertain, few, if any, rocks have been
specially described from Vanua Levu, this island was visited by Dana in
1840 when attached to the United States Exploring Expedition under
Wilkes. His observations on its geology were published in his volume on
the geology of the expedition. Although not extensive they are valuable
from their reference to his discovery of trachytic and rhyolitic rocks
as well as acid pumice-tuffs in the island. It is singular that his
observations have apparently been overlooked by all his successors.
Wichmann with this discovery unknown to him remarked on the seeming
absence of quartz-bearing recent eruptive rocks from the South Seas.

When the “Challenger” Expedition visited the group in 1875 some
geological collections were made which were described by Prof. Renard in
the second volume on the “Physics and Chemistry” of the expedition. No
collections, however, were made in Vanua Levu. In 1878 Mr. John Horne,
Director of the Botanic Gardens at Mauritius, made some important
observations on the geological structure of this island and of other
parts of the group, which he published in his account of the islands
given in “A Year in Fiji.” No collections were obtained by him; but
prominence is given to his observations by Dr. Wichmann and others. Like
Dana in the case of the acid volcanic rocks, Mr. Horne has forestalled
me in his conclusion that Vanua Levu amongst the other larger islands
has been formed mainly of the products of submarine eruptions.

The visit of Prof. A. Agassiz to Fiji in 1897-98 gave a fresh impetus to
its geological investigation. We are indebted to him not only for his
own extensive memoir on the islands and coral reefs of this group, but
also for the subsequent important explorations of Mr. E. C. Andrews and
Mr. B. Sawyer in Viti Levu and the Lau Islands. These two gentlemen have
since published a short paper on the caves of these islands. Mr. Eakle
has described the volcanic rocks collected during the visit of Prof.
Agassiz. It is, however, noteworthy that, although the collections were
made in Viti Levu, Kandavu and in many other of the smaller islands,
Vanua Levu is not represented. Mr. Eakle’s conclusion that basic
andesites and basalts are the characteristic rocks of the region, the
augite-andesites predominating, would apply to Vanua Levu in great part.
This island possesses also in fair amount hypersthene-andesites and
dacitic or felsitic andesites, which are very scantily represented in
the collections examined by Mr. Eakle. In connection with the
quartz-porphyries and trachytic rocks which also occur in Vanua Levu, it
should be observed that Mr. Andrews describes a rhyolite from Suva in
Viti Levu. Unlike Viti Levu, Vanua Levu displays but a small development
of plutonic rocks.

In conclusion it should be pointed out that much remains to be done in
the geological exploration of this island, and that I would have spent a
third year in this task much to my profit. Still I hope that a period of
two years devoted to its investigation will be regarded as some excuse
for a certain over-confidence in the expression of my opinions.

To enumerate all those from whom I received much kindness in these
islands would be a lengthy task. My indebtedness is very great to Bishop
Vidal, Father Rougier, and to various other members of the Roman
Catholic Mission, and I experienced similar favours at the hands of Mr.
Williams and other Wesleyan Missionaries in Vanua Levu. Mr. F. Spence
and Mrs. Spence showed me great kindness, and from Dr. Corney I received
valuable assistance on my arrival in the group. To the planters my debt
is equally great, more especially to Mr. Barratt, Mr. Dods, and Mr.
Mills.

In conclusion I would suggest the foundation of a “Fijian Society” for
the investigation of the islands, for the gathering together of all that
has been written about the group and its people, and for the advancement
of science.

                                                   HENRY BROUGHAM GUPPY.

_June, 1903._

  _Note._—A type set of my geological collections representing the
  massive rocks from this island has been kindly accepted by the
  Curator of the Geological Museum, Jermyn Street.



     LIST OF SOME OF THE PRINCIPAL AUTHORITIES QUOTED IN THIS BOOK


DANA, J. D., on the Geology of Fiji in vol. x, Geology, United States
  Exploring Expedition Reports, Philadelphia, 1849.

KLEINSCHMIDT, T., “Reisen auf den Viti-Inseln,” Journal des Museum
  Godeffroy, heft 14, Hamburg, 1879.

HORNE, J., “A Year in Fiji,” London, 1881.

WICHMANN, A., “Ein Beitrag zur Petrographie des Viti-Archipels,
  Mineralogische und Petrographische, Mittheilungen,” band v, heft 1,
  Wien, 1882.

RENARD, A., on andesites from Kandavu, “Report on the Petrology of
  Oceanic Islands,” vol. ii of “Physics and Chemistry,” Challenger
  Expedition, 1889.

AGASSIZ, A., “The Islands and Coral Reefs of Fiji,” Bulletin, Museum of
  Comparative Zoology, Harvard College, vol. xxxiii, 1899, Cambridge,
  Mass.

EAKLE, A. S., “Petrographical Notes on some rocks from the Fiji
  Islands,” Proceedings, American Academy of Arts and Sciences, vol.
  xxxiv, no. 21, May, 1899.

ANDREWS, E. C., Notes on the limestones and general geology of the Fiji
  Islands, with special reference to the Lau Group. Based upon surveys
  made for Alexander Agassiz. With a Preface by T. W. Edgeworth David.
  Bulletin, Museum of Comparative Zoology, Harvard College; vol.
  xxxviii, Cambridge, Mass. 1900.



                                CONTENTS


                               CHAPTER I

      GENERAL INTRODUCTORY REMARKS ON SOME OF THE LEADING PHYSICAL
                         FEATURES OF THE ISLAND

Its remarkable shape, 1.—Its building up, 2.—Study of its profile,
  3.—Mount Seatura.—Regions of acid andesites.—Basaltic
  tablelands.—Great ridge-mountains, 5.—Boundary of the regions of basic
  and acid rocks, 6.—Its primary features, the dacitic peak, the
  basaltic plateau, and the ridge-mountain

                                                             _Pages_ 1-6


                               CHAPTER II

     ON THE EVIDENCE OF EMERGENCE OR OF UPHEAVAL AT THE SEA-BORDERS

Elevated coral reefs scantily represented, 7.—Apparent absence of coral
  reefs in the early stages of the emergence, 8.—Elevated reefs confined
  to the coast and its vicinity.—Detailed examination of the
  sea-borders, 9.—Silicified corals and siliceous concretions the only
  evidence in many localities of the upraised reefs, 13.—The relations
  of the mangrove-belt to the reef-flat, 14.—Indications of a very
  gradual movement of emergence in our own time, 15.—The rate of advance
  of the mangroves, 16.—Conclusions, 19

                                                            _Pages_ 7-20


                              CHAPTER III

                     THE HOT SPRINGS OF VANUA LEVU

The thermal springs of other parts of the group, 21.—The hot springs of
  the Wainunu valley, 22.—The boiling springs of Savu-savu, 25.—Analyses
  of the water, 28.—The hot springs of other localities,
  31.—Distribution of the springs, 35.—The algæ and siliceous deposits,
  37.—The cold and thermal springs of Hawaii and Etna, 38.—Infiltration,
  the source of the springs, 39.—A view negatived by Prof. Suess.—List
  of the hot springs of Vanua Levu, 40.—Summary of the chapter, 42

                                                           _Pages_ 21-42


                               CHAPTER IV

     DESCRIPTION OF THE GEOLOGICAL AND GENERAL PHYSICAL FEATURES OF
                               VANUA LEVU

Naivaka, 43.—Korolevu Hill, 45.—Bomb formation of Navingiri,
  46.—Remarkable section near Korolevu, 48.—Wailea Bay to Lekutu,
  50.—Mount Koroma, 51.—Mount Sesaleka, 53.—The Mbua-Lekutu Divide,
  55.—The Mbua and Ndama plains, 55.—The shell-bed of the Mbua river,
  58.—Lekumbi Point, 60

                                                           _Pages_ 43-60


                               CHAPTER V

      DESCRIPTION OF THE GEOLOGICAL AND GENERAL PHYSICAL FEATURES
                             (_continued_)

Mount Seatura, 61.—Its eastern slopes, 63.—Its western slopes, 64.—Its
  northern slopes, 65.—Ascents to the summit, 66.—The Ndriti Basin,
  67.—A huge crateral cavity, 68.—Its dykes of propylite, 69.—Seatura a
  basaltic mountain of the Hawaiian order, 72.—The Seatovo Range,
  73.—Solevu Bay, 75.—Koro-i-rea, 77.—Nandi Bay, 78.—Na Savu Tableland,
  79

                                                           _Pages_ 61-81

                               CHAPTER VI

      DESCRIPTION OF THE GEOLOGICAL AND GENERAL PHYSICAL FEATURES
                             (_continued_)

The basaltic plateau of Wainunu, 82.—Its margins covered by pteropod and
  foraminiferous ooze-rocks, 86.—The hill of Ulu-i-ndali, 87.—Kumbulau
  Peninsula, 90.—The basaltic flow of Kiombo Point, 92.—Soni-soni
  Island, 93.—Yanawai coast, 95

                                                           _Pages_ 82-97


                              CHAPTER VII

      DESCRIPTION OF THE GEOLOGICAL AND GENERAL PHYSICAL FEATURES
                             (_continued_)

The Ndrandramea district, 98.—Its mountains and hills of acid andesites,
  100.—Ngaingai, 101.—Ndrandramea, 102.—Soloa Levu, 103.—The underlying
  altered acid andesites, 106.—Section of the district, 107.—The
  magnetic peak of Navuningumu, 108.—The Mbenutha Cliffs and their
  pteropod and foraminiferous beds, 109

                                                          _Pages_ 98-112


                              CHAPTER VIII

      DESCRIPTION OF THE GEOLOGICAL AND GENERAL PHYSICAL FEATURES
                             (_continued_)

Mount Vatu Kaisia and district, 113.—The Nandronandranu district,
  117.—Nganga-turuturu cliffs, 119.—Ndrawa district, 120.—Tavia ranges,
  121.—Na Raro, 123.—Its Ascent, 125.—Na Raro Gap, 127

                                                         _Pages_ 113-127


                               CHAPTER IX

      DESCRIPTION OF THE GEOLOGICAL AND GENERAL PHYSICAL FEATURES
                             (_continued_)

The basaltic plains of Sarawanga, 129.—Tembe-ni-ndio and its
  foraminiferal limestones, 131.—The basaltic plains of Ndreketi,
  132.—The Nawavi Range, 135.—Nanduri, 136.—Tambia district, 137.—The
  basaltic plains of Lambasa, 138

                                                         _Pages_ 128-139


                               CHAPTER X

      DESCRIPTION OF THE GEOLOGICAL AND GENERAL PHYSICAL FEATURES
                             (_continued_)

The Va Lili Range, 140.—Its Nambuni spur, 144.—Originally submerged and
  covered with palagonite-tuffs and agglomerates, 145.—The Waisali
  Saddle, 146.—Narengali district, 147.—Nakambuta, 148.—The valleys of
  the Ndreke-ni-wai, 150.—Their origin, 151

                                                         _Pages_ 140-152


                               CHAPTER XI

      DESCRIPTION OF THE GEOLOGICAL AND GENERAL PHYSICAL FEATURES
                             (_continued_)

The Korotini Range, 153.—Traverse from Waisali to Sealevu, 154.—Traverse
  from Mbale-mbale to Vandrani, 156.—Traverse from Vatu-kawa to
  Vandrani, 160.—Traverse from Nukumbolo to Sueni, 161.—The Sueni
  valley, 163.—General inference concerning the range, 164

                                                         _Pages_ 153-165


                              CHAPTER XII

      DESCRIPTION OF THE GEOLOGICAL AND GENERAL PHYSICAL FEATURES
                             (_continued_)

The Koro-mbasanga Range, 166.—The Sokena Ridge, 169.—Lovo valley,
  169.—Mount Mbatini, 172.—The Vuinandi Gap, 175.—The Thambeyu or Mount
  Thurston Ranges, 176.—Structure of Thambeyu, 177.—The Avuka Range, 179

                                                         _Pages_ 166-180


                              CHAPTER XIII

      DESCRIPTION OF THE GEOLOGICAL AND GENERAL PHYSICAL FEATURES
                             (_continued_)

The Valanga Range, 181.—Its western flank, 183.—Ngone Hill, 183.—Valley
  of Na Kula, 184.—The Mariko Range, 185.—Savu-savu Peninsula,
  189.—Naindi Bay, 192.—The Salt Lake, 194

                                                         _Pages_ 181-196


                              CHAPTER XIV

      DESCRIPTION OF THE GEOLOGICAL AND GENERAL PHYSICAL FEATURES
                             (_continued_)

The Natewa Peninsula, 197.—Viene district, 198.—Lea district,
  199.—Waikawa Mountains, 201.—Ndreke-ni-wai coast, 203.—Waikatakata,
  203.—Mount Freeland or the Ngala Range, 204.—Traverse from Tunuloa to
  Ndevo, 205.—Coast from Ndevo to Mbutha Bay, 205

                                                         _Pages_ 197-206


                               CHAPTER XV

      DESCRIPTION OF THE GEOLOGICAL AND GENERAL PHYSICAL FEATURES
                             (_continued_)

The north-east portion of the island from Mount Thurston to Undu Point,
  207.—Coast between Vuinandi and Tawaki, 208.—The corresponding inland
  region, 209.—The gabbro of Nawi, 211.—Uthulanga Ridge, 211.—Ascent of
  Mount Vungalei or Ndrukau, 213.—Nailotha, 214.—Exposure of altered
  trachytes and quartz-porphyries at its base, 215.—From Nandongo to
  Vanuavou, 216.—From Ngelemumu to Wainikoro, 217.—Sea border between
  Lambasa and Mbuthai-sau, 218.—Coast between Mbuthai-sau and the
  Wainikoro and Langa-langa Rivers, 219.—Coast between the Langa-langa
  River and Thawaro Bay, 221.—The Globigerina clay of Visongo,
  221.—Vui-na-Savu River, 222.—Some General inferences, 223

                                                         _Pages_ 207-223


                              CHAPTER XVI

      DESCRIPTION OF THE GEOLOGICAL AND GENERAL PHYSICAL FEATURES
                             (_continued_)

The Wainikoro and Kalikoso Plains, 224.—Vaka-lalatha Lake, 225.—Its
  floating islands, 226.—A region of acid rocks, 227.—Silicified corals
  and limonite, 228.—Tawaki district, 229.—Thawaro district, 230.—Mount
  Thuku, 231.—Undu Point, 232.—General characters of the Undu
  Promontory, 233

                                                         _Pages_ 224-234


                              CHAPTER XVII

                    THE VOLCANIC ROCKS OF VANUA LEVU

Their varied character, 235.—Their classification, 236.—Descriptive
  formula, 237.—Synopsis, 239.—Orders of the Olivine-Basalts,
  241.—Orders of the Augite-Andesites, 245.—Orders of the
  Hypersthene-Augite-Andesites, 247.—Description of the Plutonic Rocks,
  249

                                                         _Pages_ 235-251


                             CHAPTER XVIII

             THE VOLCANIC ROCKS OF VANUA LEVU (_continued_)

The Olivine Basalts

                                                         _Pages_ 252-265


                              CHAPTER XIX

             THE VOLCANIC ROCKS OF VANUA LEVU (_continued_)

The Augite-Andesites

                                                         _Pages_ 266-284


                               CHAPTER XX

             THE VOLCANIC ROCKS OF VANUA LEVU (_continued_)

The Hypersthene-Augite-Andesites

                                                         _Pages_ 285-292


                              CHAPTER XXI

             THE VOLCANIC ROCKS OF VANUA LEVU (_continued_)

           THE ACID ANDESITES, TRACHYTES, QUARTZ-PORPHYRIES.

The Hornblende-Andesites of Fiji, 293.—Occurrence of Dacites in Fiji,
  294.—Suggestion of “felsitic andesite” as a rock-name, 295.—The Acid
  Andesites of Vanua Levu, 295.—The Hypersthene-Andesites, 296.—The
  Hornblende-Hypersthene-Andesites, 298.—The Quartz-Andesites or
  Dacites, 302.—Tabular comparison of the Acid Andesites, 304.—The
  characters of the Rhombic Pyroxene, 306.—Magmatic Paramorphism,
  306.—The Oligoclase Trachytes, 308.—Quartz-Porphyries and Rhyolitic
  rocks, 309

                                                         _Pages_ 293-311


                              CHAPTER XXII

             THE VOLCANIC ROCKS OF VANUA LEVU (_continued_)

Basic pitchstones and basic glasses, 312.—Volcanic Agglomerates, 314

                                                         _Pages_ 312-316


                             CHAPTER XXIII

         CALCAREOUS FORMATIONS, VOLCANIC MUDS, PALAGONITE-TUFFS

General Character, 317.—Coral Limestones, 318.—Foraminiferal Limestones,
  319.—Pteropod-oozes, 320.—Foraminiferous Volcanic Muds, 321.—Samples,
  322.—Altered kinds, 324.—Submarine Palagonite-tuffs of mixed
  composition, 326.—Samples, 330.—Altered Basic Tuffs, 332.—Submarine
  Basic Pumice Tuffs, 333.—“Crush-tuffs” formed of basic glass and
  palagonite, 334.—Zeolitic Palagonite-Tuffs, 334.—Palagonite-marls,
  335.—Acid Pumice Tuffs, 336

                                                         _Pages_ 317-336


                              CHAPTER XXIV

                               PALAGONITE

Its abundance in a fragmental condition in Vanua Levu, 337.—Its
  occurrence in deep-sea deposits, 338.—Modes of formation _in situ_,
  338.—In the upper portion of a basaltic flow, 339.—In the groundmass
  of hemi-crystalline basaltic rocks, 339.—In veins in a basic
  tuff-agglomerate, 340.—In the fissures of a basaltic dyke, 341.—In the
  matrix of pitch-stone agglomerates, 349.—In “crush-tuffs,”
  341.—Regarded as a solidified magma-residuum of low fusibility,
  342.—Its connection with crushing, 342.—Bunsen’s experiment,
  343.—Rosenbusch and Renard, 344.—The Nandua series of beds,
  345.—Suggested explanation of the origin of palagonite, 346.—Type of
  basalt associated with palagonite, 347.—Hydration and disintegration
  of palagonite, 348

                                                         _Pages_ 337-349


                              CHAPTER XXV

                  SILICIFIED CORALS, FLINTS, LIMONITE

Mode of occurrence of the silicified corals, 351.—Their character and
  structure, 352.—Flints, nodules of Chalcedony, Agates, etc.,
  353.—Other siliceous concretions, 354.—Jasper, 355.—Deposits of
  Limonite, 356.—Magnetic Iron-sand, 357.—Suggested explanation of the
  silicification of the corals, 358.—Note on a silicified Tree-fern, 360

                                                         _Pages_ 350-360


                              CHAPTER XXVI

                             MAGNETIC ROCKS

Previous observations, 361.—Magnetic Polarity usually caused by
  atmospheric electricity, 362.—Displayed by both acid and basic rocks,
  364.—Very frequent in Vanua Levu, 365.—Its relation to specific
  weight, 366.—The influence of locality, 367.—Frequently observed in
  mountain peaks, 367.—Description of the peaks, 368.—Measurement of the
  polarity of rocks, 370

                                                         _Pages_ 361-371


                             CHAPTER XXVII

                  SOME CONCLUSIONS AND THEIR BEARINGS

Vanua Levu, a composite island formed during a long period of emergence,
  372.—The submarine plateau probably produced by basaltic flows,
  373.—The distribution of the volcanic rocks, 374.—Comparison with
  Iceland, 374.—The mountain-ridges, 375.—The emergence of the Fiji
  Islands, 376.—Wichmann’s view of the early continental condition not
  supported, 376.—Age and character of the emergence, 377.—The evidence
  of the Lau Group and of the Tongan Islands, 378.—Two principal stages
  of the emergence, 379.—Relative antiquity of the Hawaiian, Fijian, and
  Tongan Islands as indicated by their floras, 379.—Islands have always
  been islands, 380.—The hypothesis of a Pacific continent not yet
  needed, 381.—The great dilemma, 381.—Much remains to be learned of the
  possibilities of means of dispersal in the past and in the present,
  382

                                                         _Pages_ 372-382


                               APPENDIX.

  (1) Note on microscopical examination of stone-axes.

  (2) Note on the ascent of the tide in the Ndreketi River.

  (3) Note on the “talasinga” districts.


  INDEX      385



                         LIST OF ILLUSTRATIONS


                              PLATES


                                                                 TO FACE
                                                                   PAGE

 Na Raro (2,420 feet) from the south-west, a peak of acid    }
   andesite                                                  }
                                                             } _Frontis-
 Ndrandramea (1,800 feet) from the south-east, a peak of     } piece_
   acid andesite rising about a thousand feet from its base  }

 The Ndrandramea District from the westward                           98

 Mount Tavia (2,210 feet) from Vatu Kaisia                    }
                                                              }      108
 The magnetic peak of Navuningumu (1,931 feet) from the south }

 Mbenutha Cliffs, showing volcanic agglomerates overlying tuffs and
   clays, containing shells of pteropods and foraminifera, which are
   raised 1,100 feet above the sea                                   111

 Duniua Lagoon, representing an old mouth of the Ndreke-ni-wai       153


                             LITHOGRAPHS

 Vanua Levu, Fiji Islands                                              1

 Fiji Islands                                                        373


                              FIGURES

                                                                    PAGE

 Profiles of Vanua Levu as Viewed from the South. Graphically
   Represented on a Horizontal Scale of about 16 miles to the inch     4

 Korolevu Hill (800 feet) from Wailea Bay                             46

 Profile and Geological Section of the western end of Vanua Levu from
   the Wainunu estuary across the summit of the basaltic mountain of
   Seatura to the edge of the submarine platform off the Ndama coast
   as limited by the 100-fathom line                                  62

 Profile, looking north from off the mouth of the Wainunu River       83

 Rough plan of the Ndrandramea district in Vanua Levu; made with
   prismatic compass and aneroid by H. B. Guppy                       99

 Profiles of Ngaingai and Wawa Levu from Nambuna to the south-west.
   Both are dacitic mountains                                        101

 Profile and Geological Section of Vanua Levu, across the island
   from the Sarawanga (north) coast to the Yanawai (south) coast     107

 Profile-sketch of the Vatu Kaisia district from S.S.E.              113

 Section of the Vatu Kaisia district                                 115

 Profiles of Na Raro                                                 124

 Profile-sketches of the Va-Lili Range                               141

 Profile-sketch of the mountainous axis of Vanua Levu                167

 Koro-mbasanga from the north-north-east                             167

 Mount Mbatini from the top of Koro-mbasanga                         173

 View from Muanaira on the south coast of Natewa Bay                 173

 Ideal Section of Thambeyu                                           177

 Diagram illustrating the two sets of felspar-lathes in a dyke       238

 Magma-lakelet, ·25 mm. in size, magnified 290 diameters, from a
   basalt at Navingiri                                               339

 Showing fragments of glass with eroded borders and of plagioclase
   with more even edges in a matrix of palagonite traversed by
   cracks                                                            342

 Diagram showing the succession of deposits below the Nandua
   tea-estate                                                        345

[Illustration:


  VANUA LEVU,

  FIJI ISLANDS.

  DRAWN ON A SCALE OF 25 MILES TO 3 INCHES BY H. B. GUPPY, M.B.

  _Based on the Admiralty Surveys, but most of the topographical
  details of the interior have been supplied from the author’s
    observations
  with the aneroid and prismatic compass in 1897-99. It is
  merely intended to illustrate his general account of the physical
  and geological characters of the island and is very far from
  complete. (see introduction.)_
]



              OBSERVATIONS OF A NATURALIST IN THE PACIFIC



                               CHAPTER I

          GENERAL INTRODUCTORY REMARKS ON SOME OF THE LEADING
                    PHYSICAL FEATURES OF THE ISLAND


THE remarkable shape of this island at once attracts the attention: and
indeed it is in its irregular outline and in the occurrence over a large
portion of its surface of submarine tuffs and agglomerates that will be
found a key to the study of its history. With an extreme length of 98
miles, an average breadth of 15 to 20 miles, and a maximum elevation of
nearly 3,500 feet, it has an area, estimated at 2,400 square miles,
comparable with that of the county of Devon.

Whilst its peculiarly long and narrow dimensions are to be associated
with the narrowing of the submarine basaltic platform, from which it
rises together with the other large island of Viti Levu, its extremely
irregular shape is closely connected with the composite mode of its
origin. We have here exemplified the process of the building up of a
continental island in the great area of emergence of the Western
Pacific, that region which displays at various heights above the sea the
ancient reefs and the underlying deposits of the Solomon Islands, New
Hebrides, Fiji, Tonga, &c. But this process of construction has never
been completed, and is at present suspended; yet it is in its incomplete
condition that Vanua Levu possesses its importance for the investigation
of this subject.

This island has in fact been formed by the union of a number of smaller
volcanic islands during a long protracted period of emergence. These
original islands are indicated approximately by the 1,800-feet
contour-level in the accompanying map. There is, however, no reason for
supposing that the movement of emergence has altogether ceased. In the
course of ages the extensive submarine plateau, from which it rises,
will be laid bare; and the small surrounding islands that are situated
upon it, such as Yanganga, Kia, Mali, Rambi, Kioa, &c., will be included
in the area of Vanua Levu.[1]

Excluding for the moment the effects of denudation, which have been very
great, we shall be able to discern some of the stages of the building-up
of the island during the emergence or upheaval by looking at the map and
reversing the process in imagination. A subsidence of only 50 feet would
cause the Natewa Peninsula to be isolated by a sea-passage along the
line of the Salt Lake; whilst several islands would be formed along the
northern and southern coasts, and the Naivaka Peninsula would become
detached. If the subsidence extended to 300 feet, the sea would flow
over a large portion of the island, where it would regain what was not
many ages since its own, an area of basaltic plains, which by their
prolongation under the sea form the great submarine plateau. A
subsidence of 1,000 feet would break up the remaining elevated axis of
the island into a number of lesser portions; and after a total lowering
of 1,800 feet there would exist only a few scattered islands, the
arrangement of which would show but little relation to the present form
of Vanua Levu. At either end of the area there would arise from the sea
the isolated volcanic peaks of Seatura and Ngala (Mount Freeland); and
between them would be situated four or five long narrow islands,
together with a group of small islands and islets where Na Raro and the
other acid andesite mountains of the Ndrandramea district now lie.

As might be partly expected, there is in the surface-configuration of
the interior of Vanua Levu an absence of that simplicity of contour
which exists in a volcanic island of supra-marine formation, as for
instance in the large island of Hawaii where the three great volcanic
mountains of Mauna Kea, Mauna Loa and Hualalai together with the older
Kohala range, determine the form of the whole island’s surface.[2] Here
in Vanua Levu there is, on the contrary, but little order amongst its
physical features. The rivers often run obliquely with the sea-border,
whilst mountains frequently rise at the coast and plains lie far inland,
and the view of the elevated interior, as obtained from one of the
peaks, presents in many parts a series of mountain-ridges running
athwart the island’s axis.

A study of the profile of the island is an important preliminary step to
its more detailed examination. One may ramble over a particular region
of it for weeks, as I have done, without getting any satisfactory idea
of the true configuration of the surface. In a locality densely wooded
and occupied by steep mountain ridges and deep gorges, the field of view
is often very limited; but seen from the deck of a passing ship the main
features of the island assume their true proportions and relations, and
much that was uncertain is in this manner made plain. The profile here
given has been constructed from a number of others, and represents in a
graphic fashion Vanua Levu as viewed from the southward. I have here
sacrificed smaller details and occasionally some degree of accuracy in
small matters in order to bring out the principal features of the
island.

At and near the extreme western extremity rise the conspicuous hills of
Sesaleka (1,370 feet), Naivaka (1,651 feet) and Koroma (1,384 feet), all
of them formed of basic volcanic materials.[3] Naivaka, which is
connected with the main island by a narrow isthmus, only about 30 feet
in height, is probably one of the most recent additions to the island’s
area; and it is at the same time one of the most recent of the numerous
volcanic vents that once existed. The leading feature, however, of this
end of Vanua Levu is the great mountain of Seatura (2,812 feet), which
occupies a large part of the Mbua province and monopolises most of the
landscape whilst largely determining the form of the western extremity
of the island. It is a basaltic mountain of the Mauna Loa type, its long
eastern slope descending gently at an angle of three or four degrees for
about ten miles to the mouth of the Wainunu River. In its deeply eroded
radial valleys and gorges, and in other respects, it is not unlike the
island of Tahiti, as described by Dana.

The Ndrandramea region to the eastward, which I have named after one of
its best known peaks, has a profile of a very different character. Its
broken outline indicates the existence of numerous mountains and hills
of acid andesites, occasionally dacitic. Although some of them attain a
height of 2,000 feet and over their tops alone are seen from seaward.
Between the foot of these mountains and the south coast extends a great
plateau of columnar basalt, incrusted at its borders with submarine
deposits, which descends coastward with a very gentle slope, the fall in
about five miles being only about 300 feet (1,100 to 800 feet). It
terminates abruptly opposite the elevated headland of Ulu-i-ndali, a
range, composed mainly of grey olivine-basalts, which is not indicated
in the profile.

      PROFILES OF VANUA LEVU AS VIEWED FROM THE SOUTH. GRAPHICALLY
    REPRESENTED ON A HORIZONTAL SCALE OF ABOUT 16 MILES TO THE INCH.

[Illustration: From NAITHOMBOTHOMBO Point to UNDU Point, representing
the length of the island (98 miles).]

[Illustration: The NATEWA Peninsula from the Salt Lake to Kumbalau
Point.]

The two conical peaks of Vatu Kaisia (1,880 feet) and Na Raro (2,420
feet), which rise up so unexpectedly in the region immediately east of
the Ndrandramea district, are also of acid andesitic rocks, in the last
case approaching the dacitic type. They lie within the borders of the
area of basic tuffs, basic agglomerates, and basic massive rocks, that
here begins and extends eastward to Mount Thurston and a little beyond.
East of Na Raro there is a gap or break in the profile, where the
greatest elevation is probably not over 800 feet; and on its farther
side rises up the mountain of Va Lili (2,930 feet), a lofty inland ridge
that lies towards the southern coast. Palagonite-tuffs and agglomerates
are the prevailing surface-formations in this district.

Eastwards from Va Lili extends for eight or nine miles a lofty,
level-topped, and almost peakless range, which I have called the
Korotini Table-land, after the towns once situated on its southern
slopes. Its outline is shown in the background of the view facing page
153. It is, however, not so level-topped as it appears to be; but the
gradual variations in elevation between 2,000 and 3,000 feet, when
spread over a length of some miles, are more or less lost in the general
outline of the range as viewed from the coast. Basic agglomerates are
principally exposed on the lower slopes; whilst higher up, reaching
often to the summit of the table-land, occur palagonite-tuffs containing
tests of foraminifera and molluscan shells, massive basic rocks being
exposed in places.

The level profile of the Korotini tableland gives place, as one proceeds
eastward, to the broken outline of the several lofty peaks of Mariko
(2,890 feet), Mbatini (3,437 feet), Thambeyu (3,124 feet) and others.[4]
Each of these peaks marks one of the bold mountain-ridges that form such
a striking feature in the surface-configuration of this part of the
island. On the slopes of these ridges, and often also on their summits,
appear basic agglomerates and palagonitic tuffs and clays often
inclosing tests of foraminifera; whilst exposed in the gorges and
protruding at times through the tuffs and agglomerates on the crests of
the ridges are displayed massive basic rocks of the type of the
hypersthene-augite andesites.

East of Thambeyu the level sinks to about 1,000 feet above the sea, and
beyond rises an irregular group of hills and mountains which attain
their greatest height in Nailotha,[5] 2,481 feet above the sea. We are
now near the limit of the area of basic rocks. Following the profile as
it slopes away, marked by occasional peaks and breaks, towards Undu
Point, we pass at first over a district where basic rocks are mixed with
those of more acid type; but before we reach Mount Thuku we enter the
district of oligoclase-trachytes, quartz-porphyries, and rhyolitic
tuffs, that extends to the extremity of the Undu promontory.

There remains to be noticed the profile of the Natewa Peninsula. As
shown in the diagram, this level begins at a few feet above the sea in
the vicinity of the Salt Lake; and as it proceeds eastward it attains a
level of 1,960 feet in Ngalau-levu and of 1,540 feet in the Waikawa
promontory, finally culminating, as it nears Kumbulau Point, in a
mountainous district which attains its greatest elevation of 2,740 feet
above the sea in the lofty ridge of Ngala, the Mount Freeland of the
chart. Altered basic rocks prevail in this peninsula; but more acid
andesites also occur, and foraminiferous tuffs and clays are exposed on
the slopes, reaching to over 1,000 feet above the sea.

I will conclude this reference to the profile of the island with the
remark that if I had neglected to indicate here the close connection
that exists between the nature of the surface-configuration and the
character of the prevailing rocks I should have ignored a means of
investigation which has proved of the greatest value. The rock and
surface characters go together. The inland plateau now upheaved 1,000
feet above the sea, was built up by submarine flows of basaltic lava.
The isolated conical peak that so unexpectedly intrudes itself into the
view is the dacitic core of some submarine volcano long since stripped
of most of its fragmental coverings. The lofty mountain-ridges that run
athwart the island’s breadth, with their summits usually in the
rain-clouds received their coverings of tuffs and agglomerates ages ago
when they were submerged; and now they rise to heights of over 3,000
feet above the sea. Bound up with the mysterious origin of these great
ridges is the history of the island of Vanua Levu.

These preliminary remarks are only intended to serve as a general
introduction to the detailed description of the island and its
formations. The closing chapter is devoted to a summary of the principal
results of my investigations.



                              CHAPTER II.

    ON THE EVIDENCE OF EMERGENCE OR OF UPHEAVAL AT THE SEA-BORDERS.


ONE would have expected that in an island where submarine muds and tuffs
are of such common occurrence at the surface, extending from the
sea-border to elevations of 2,000 feet and over, upraised coral reefs
would be also frequent and extensive. But it is remarkable that the
uplifted masses of reef-limestone, so characteristic of the islands of
the Lau Group, are here very scantily represented. It is certainly true
that the fossiliferous volcanic muds that form the foundations of coral
reefs are often exposed at and near the coast; but the elevated reefs
that ought to be found reposing on them are rarely to be observed.

It is not to be inferred, however, that in a region so remarkable for
the great development of reef-formations coral reefs did not then thrive
in these localities, but rather that such a long period has elapsed
since the emergence of the present sea-border that the upraised coral
reefs at and near the coast have long since been in a great part
stripped off by the denuding agencies. Notwithstanding this, it is
evident that coral reefs could never have been very extensive at the
sea-border during the last stages of the emergence; whilst they do not
appear to have existed at all during the early periods of the history of
the island.

In this connection it may be observed that hard compact limestones of
any kind are rarely to be found, and only in a scanty fashion. The
extensive development of dolomites and hard limestones, described by Mr.
Andrews and others in the valley of the Singatoka in Viti Levu, is not a
character of Vanua Levu. The foraminiferous and pteropod clays, which
exist in the interior and often in the heart of the island, are not
overlaid by ancient reef-limestones, but by great masses of volcanic
agglomerate and coarse fossiliferous tuffs, the foraminiferous muds in
their turn covering the core of massive volcanic rocks. There were no
signs of coral-reef fragments in the volcanic agglomerates in any
locality examined, notwithstanding that these agglomerates are so
intimately associated with the fossiliferous tuffs and clays that their
submarine origin could not be doubted.

The conditions for reef-formation evidently did not exist in that early
stage of the island’s history, when the foraminiferous tuffs and clays,
now occurring at elevations of 2,000 feet and over, were being deposited
on the sea-bottom. At some time or other, however, these high
mountain-slopes, previous to their emergence from the sea, must have
been within the limits of the zone of reef-building corals. If reefs had
been formed along those ancient coasts, or on the original shoals, they
would have been in some cases preserved, as in the case of the
foraminiferous tuffs and clays, by a covering of volcanic agglomerate.
These soft submarine deposits have been in this manner saved from the
destructive effects of denudation over a large part of the island
whether on the higher slopes or at the lower levels; but no trace of
reef-formation ever came under my notice in the higher regions of the
interior. This is a puzzling point that will have to be considered in
connection with the origin of the great mountain ridges, one of the most
difficult problems in the history of the building-up of Vanua Levu.

I will now refer to the evidence of the latest stage of the upheaval of
the island as indicated at and near the sea-border by the scantily
occurring upraised reefs. The elevated reefs are mostly to be found on
the south coast between Fawn Harbour and Na Viavia Islet off Harman’s or
Savu-savu Point. Na Viavia Islet itself is 300 or 400 yards in length
and is formed of much honeycombed reef-limestone, which is raised 10 or
12 feet above the high-water line. Proceeding eastward along the south
coast of the Savu-savu promontory we next come upon uplifted reefs in a
curiously isolated hill that rises on the coast between Naithekoro and
Naindi Bay. This hill is about 250 feet in height and is composed in the
mass of coral limestone. About 100 feet above the sea-level it exhibits
an erosion-line, above which it rises precipitously to the summit. The
west point of Naindi Bay is formed of reef-limestone reaching to a
height of 40 to 50 feet and displaying in position massive corals,
“Fungiæ,” and “Tridacna” shells. Near its base, four to five feet above
the present high-water level, it shows an erosion line. This limestone
overlies a rock in which blocks of volcanic rocks, five to six inches
across, are imbedded in a calcareous matrix.

Raised coral limestone occurs at intervals on the coast between Naindi
Bay and the mouth of the Salt Lake Passage, usually forming low islets,
of which the smaller about 12 feet in height often assume, through the
erosion of the sea at their base, that peculiar mushroom-shape, so
characteristic of upheaval on reef-bound coasts. The passage into the
Salt Lake lies in a slightly elevated reef-mass; and the islet which
rises up in its centre to about a foot above the water-level is mainly
formed of coral blocks, although I did not find any remains of coral on
the low neck of land intervening between the Salt Lake and Natewa Bay.
Eastward from the Salt Lake Passage to Nanutha in the vicinity of Fawn
Harbour low cliffs of coral limestone, six to eight feet high and
occasionally displaying massive corals in position, most frequently
constitute the sea-border, rarely, however, extending more than a few
paces inland or attaining there a greater elevation than 12 or 15 feet.

This limitation of the upraised reef-belt to the immediate vicinity of
the coast is true of all this district. It is only when the sea-border
is low and swampy that it is found 100 or 200 yards inland; and in any
case as one follows it inland it soon gives place to the fossiliferous
mud-rocks and tuffs of the interior. It should be noted that the
upraised reefs of this region were rarely observed at greater heights
than 20 feet above the sea, in fact usually at a much lower level. The
exceptional occurrence in mass of reef-limestone at a height of 250 feet
in a coast hill between Naithekoro and Naindi therefore lends colour to
the idea that the elevated reefs formerly extended farther inland and
that they have been stripped off by denudation.

On the north coast of the Natewa Peninsula elevated reefs are of very
rare occurrence. I walked along the whole of that coast from the head of
Natewa Bay to within four miles of Kumbulau Point and only found them in
the locality, one to one and a half miles west of the mouth of the river
Ndreke-ni-wai. Here there were two islets, 20 to 25 feet high and lying
close to the shore, which were formed entirely of coral-rock, massive
corals occurring in position in their lower part. Although, however,
upraised reefs are so scantily to be found on this coast, other proofs
of upheaval are to be observed in the fossiliferous tuffs exposed
occasionally by the beach. On the east coast of this peninsula, between
Ndevo and Loa, submarine tuffs and sandstones, at times fossiliferous,
were alone noticed.

Upraised reefs are also very rare on the north coast of Natewa Bay. Here
again I traversed the whole coast from the head of the bay to Undu
Point, a distance as the crow flies of about 50 miles; but I find no
record in my notes of any elevated reef-formations. However the
calcareous nature of the volcanic tuffs exposed in places at the coast
indicate emergence. The extreme rarity, if not the absence, of upraised
reefs on this long stretch of coast, which is usually bordered by
shore-reefs, is very remarkable, more especially since there is
extensive evidence of upheaval in the plains of Kalikoso in the
interior, as indicated in the succeeding paragraph.

On the other side of Undu Point, between that headland and Lambasa,
elevated reefs did not come under my observation, although in the
low-lying inland district of the Kalikoso lake silicified corals are
scattered about in quantity at an elevation of 20 or 30 feet above the
sea. But the emergence of the sea-border is shown in the occurrence of a
“Globigerina” sedimentary tuff near Visongo at a height of 200 feet (see
page 221), and by the occasionally calcareous character of the
pumice-tuffs that mainly compose the coast cliffs. Near Nukundamu these
tuffs of the shore cliffs inclose subangular fragments of massive corals
of the size of a walnut; whilst in a cutting between Mbuthai-sau and
Lambasa, about 50 feet above the sea, I observed bits of coral limestone
in a basic tuff. Mr. Horne refers to seams or layers of coral limestone
occurring in the volcanic agglomerate of the coast cliffs between
Lambasa and Tutu Island.[6] Since his experience of this coast was
mostly confined to a passage in a canoe along the shore, it is very
probable that he only saw the beds of white pumice-tuffs that prevail in
places on this coast. I found no beds of coral limestone in the
shore-agglomerates of this coast, nor does Dana in his description of
the pumiceous formation of the cliffs of Mali Point make any reference
to them.[7]

Along the stretch of 50 miles of coast between Lambasa and Naivaka
upraised reefs are of infrequent occurrence. However between Lambasa and
Wailevu, coral limestone is extensively exposed in a low range of hills
a mile or two inland but not over 100 feet above the sea. No elevated
reefs came under my notice between the mouth of the Wailevu river and
Nanduri Bay. That a small upheaval has been recently in progress in this
part of the coast is indicated by two circumstances. In the first place
an erosion-line about a couple of feet[8] above the high-water line, and
a few paces removed from it, is displayed in the volcanic tuff of the
point bordering the reef-flat on the east side of Nanduri Bay. In the
next place there exist at different places in the midst of the
mangrove-belt extensive bare mud-flats, sometimes several hundred yards
across, which are only covered by the higher tides. These flats are
quite bare of mangrove or any other vegetation and are often cracked on
the surface and sun-dried and firm to walk upon.[9] These naked
mud-flats in the midst of the mangrove tracts are peculiar to this part
of the coast. Their general level must be between one and two feet above
that of the mangrove belt in other parts of the island; and I infer that
a slight upheaval or emergence has led to the death of the mangroves in
these situations.

I know little of the coast between Nanduri Bay and the mouth of the
Ndreketi River. At two localities where I landed no elevated
reef-formation was observed. Dana referring to the coast opposite
Mathuata Island alludes only to the volcanic agglomerates. The low
mangrove-bordered coast between the mouths of the Ndreketi and Lekutu
rivers was not actually visited by me; but I traversed the region behind
the broad mangrove-belt, and found occasionally in the tuffs and muds
exposed in the river-banks marine-shells and foraminiferous tests,
indicating an elevation of a few feet. I examined much of the coast
between Lekutu and the extremity of the Naivaka peninsula, but came upon
no upraised reef-rocks. In the low isthmus, 20 to 30 feet high, which
connects this peninsula with the main island only volcanic rocks came
under my notice. A palagonitic tufaceous sandstone exposed in the cliffs
on the north coast of Naivaka contains a little carbonate of lime, and
being probably a submarine deposit it implies an emergence of the
sea-border.

Although I have been able to produce but scanty evidence of uplifted
reefs on the north coast of Vanua Levu, it is probable, judging from the
heights given in the Admiralty Sailing Directions, that such formations
exist in a few of the numerous low islands and islets that front this
coast. Some of these islands and islets, which are often not much more
than reef-patches largely reclaimed by the mangroves, will be noticed
below when considering the question of the extension of the mangrove
belts since the survey of Commodore Wilkes in 1840.

Neither on the south coast of the peninsula of Naivaka nor on the west
coast of the Sesaleka promontory did upraised reefs come under my
observation; but my acquaintance with the last locality is very scanty.
The emergence of the Sesaleka promontory is however indicated by the
occurrence inland at heights of at least 700 feet of palagonitic tuffs,
occasionally containing foraminifera.

With the long tract of coast between Naithombothombo Point and Solevu
Bay, I am fairly well acquainted. However, with the doubtful exception
of Lekumbi Point, no elevated reef-formations were observed. Evidence of
an emergence of a few feet, and of a very extensive seaward advance of
the land-surface in recent times, is afforded by a curious bed of marine
shells exposed in the banks of the Mbua River, nearly two miles inland
and in the vicinity of the Wesleyan Mission station. This is described
on page 58. The submergence at some period of the watershed between the
Mbua and Lekutu districts is indicated by the presence of microscopic
foraminifera in the hyalomelan tuffs that are exposed in the dividing
ridge.

Along the whole coast between the mouth of the Mbua River and Solevu
Bay, there are but few if any traces of upheaval. Even volcanic tuffs
are of rare occurrence, and there is only the case of the formation of
Lekumbi Point to be here referred to. This singular low cape is
described on page 60. Here it is sufficient to remark that it is
monopolised by the mangroves except at the outer part where the swampy
ground passes into the dry sandy soil of a reef-islet, occupied by the
usual littoral vegetation, and raised only a foot or two above the
high-water level. It exhibits on the beach the bedded sand-rock so often
found on coral islets, but this in itself is no evidence of emergence.

Neither on the shores of Wainunu Bay nor in the Kumbulau peninsula were
upraised reefs observed, although the presence in places of submarine
tuffs inland and near the coast affords evidence of elevation. The same
remark applies to the coasts of Savu-savu Bay.

I have little doubt that the absence of elevated reefs on the coasts of
by far the greater part of the island is the result largely of
denudation. In this case we have to explain why an island in a region of
coral reefs exhibits on the surface of its interior submarine tuffs and
clays in most localities, whilst uplifted reefs are very rarely to be
found at the coast or in fact anywhere. This view receives support from
the existence of traces of old elevated reefs in different parts of the
island. These traces are afforded by the occurrence on the surface in
different localities of silicified fragments of coral associated with
concretions of chalcedony, bits of flints and hornstones, jasper, impure
siliceous nodules, &c. The localities may be at the coast or a mile or
two inland, and are not usually more than 100 or 200 feet above the sea.
This subject is treated with some detail in Chapter XXV. Here I may say
that such localities are confined mostly to the open, low, undulating
districts on the north side of the island. Silicified corals are not
always present with the fragments of chalcedony and other siliceous
concretions that are found so frequently in these situations; but from
their association in the plains of Kalikoso, where the silicification of
corals may almost be observed in operation, the previous existence of
corals may be more than suspected in localities where only the other
siliceous materials are observed.

I pass on now to some general considerations regarding the relations of
the mangrove-belt to the sea-border and the character of the slope of
the land-surface as compared with that of the submarine platform. An
accurate conception respecting these matters will help one to avoid some
pitfalls in forming an estimate of the character of the movement of
emergence which this region has experienced.

Beginning with the mangrove-belt, some curious preliminary reflections
arise, when we endeavour to look back into the past stages of the
history of a mangrove tract in an area of emergence. We might perhaps
expect to find the remains of such a belt in the upraised sea-borders;
or if no traces existed, we ought to find in some places an extension
inland of the reef-flat on which the mangroves at one time flourished.
If a rapid movement of emergence is now in progress, the mangroves ought
to cover the whole or greater part of the reef-flat; and in the mangrove
tract of an emerging area we might look for signs of central decay and
marginal growth, the mangroves dying in the middle of the tract and
flourishing at the advancing margins.

When, however, we look at the mangrove-belt, as it at present exists
around much of the coast of this island, we find that, except in the
vicinity of the mouths of rivers, there extends beyond it a considerable
extent of bare reef-flat, varying usually between 200 and 1,000 yards in
width, and covered by the rising tide. There is no evidence of recent
emergence in this condition of things. This relation between the
mangrove-belt and the reef-flat indicates a state of equilibrium which
might have been established long ago. It is the normal relation that
exists between reef and mangrove growth; and it excludes all but very
gradual movements of upheaval or emergence of the sea-border. It is not
always easy to see why there should be this fine adjustment between the
rapidly-growing mangrove and the slowly-growing reef. Under normal
conditions, however, that is to say, when the land is stationary or when
the change of level is of a very gradual nature, the reclaiming agency
of the mangrove receives a check, and this relation between the
mangrove-belt and the outer reef-flat is maintained.

Actual acquaintance with such localities soon forced me to the
conclusion that whilst a gradual emergence or upheaval of 3 or 4 feet in
a century would not materially affect the relation between the
mangrove-belt and the reef-flat, a sudden or rapid change of level of
that amount would destroy the mangroves around the whole island. There
is some evidence, however, of there having been a rapid upheaval of this
kind in different parts of the coast: and it follows, therefore, if this
movement was general, that the present mangrove-belts date only from the
last upheaval. But this elevation may have occurred ages ago; and the
equilibrium between mangrove-belt and reef-flat may have been long since
established. Accordingly, the breadth of the mangrove-belt can afford no
indication of the period that has since elapsed. From data referred to
below, it is evident that the mangrove-belt, taking its average width,
away from the estuaries, at about 500 yards, might have been formed in
two or three centuries, whilst a thousand years or more may have passed
since it assumed its present relation to the reef-flat. If, therefore,
upheaval is in progress, it must be of a very gradual character, since
the normal relation of mangrove-belt to reef-flat now prevails.

There are indeed signs of such a gradual movement of emergence or of
elevation being in operation on the north coast of Vanua Levu at the
present time. I have before referred (page 11) to the extensive bare
mud-flats in the midst of the mangrove-belt between Nanduri and Lambasa,
which are well represented on the Tambia coast and in Nanduri Bay. They
are only covered by the higher tides, and in the intervals their
surfaces are dried and cracked by exposure to the sun. Here we have the
central decay and the marginal growth which would be expected in a
mangrove tract situated in a gradually rising area.

An indirect indication of such a slow upheaval on the north coast is to
be found in the circumstance that the great submarine platform, which
reaches seaward to the line of barrier-reefs, 15 to 20 miles away,
passes gradually, as it extends landward, into the low-lying plains that
constitute the sea-border between Lekutu and Ravi-ravi Point. As shown
in the profile-section on p. 62, these low coast districts are prolonged
inland, with an average rise of between 20 and 30 feet in a mile, to the
heart of the island; and we have here an extension inland of the slope
of the submarine platform. These broad inland plains, and I may here
include those behind Lambasa, are covered over much of their surface
with submarine tuffs and clays in such a manner that we may almost trace
their continuity at the coast with similar deposits now in actual
formation beyond the low-water level on the surface of the submarine
platform.

A glance at the map of the island, where these inland plains are
indicated by the 300 feet of the contour-line, will make this point more
clear. These plains are traversed by the Sarawanga, Ndreketi, Wailevu,
and Lambasa rivers; and so slight is the fall that cutters usually
ascend the rivers for several miles, whilst the tide extends for a
considerable distance up their courses. That the emergence of the inland
plains of Kalikoso in the eastern part of the island is comparatively
recent there can be but little doubt. In that locality as described on
page 224, the low marshy land, surrounding the fresh-water lake of
Vakalalatha, although five miles inland, is only elevated 20 to 30 feet
or less above the sea, and silicified corals are scattered over its
surface.

There is one other method of ascertaining the character and amount of
elevation that may be still in progress in this island namely the
comparison of the results of surveys of the coasts at different periods.
In this manner data may be obtained as regards the growth of the
mangrove belt, changes in size of the low reef-islets and islands, and
alterations in depth. For this purpose I have employed the charts of the
north and west coasts of the island made by Commodore Wilkes in 1840[10]
and the Admiralty charts 379 and 382 as completed from the survey of
these coasts by Commander Combe in 1895-96.

It was not easy to make many good comparisons in the case of the advance
of the mangrove-belt of the main coast. There certainly has been no
great advance seaward of the margin of the mangroves in this half
century. The average amount probably lies between the estimate obtained
for the coast opposite Mathuata Island, where there has either been no
change or an advance of only 100 yards or so, and that for the advance
seaward of the mangrove promontory of Lekutu which amounts to 500 or 600
yards. In this last case, however, much of the extension may be due to
the advance of the mangroves on the mud brought down by the Lekutu
river, so that, as far as these data show, the average advance of the
belt of mangroves on this coast between 1840 and 1895 would appear to be
slight.[11]

On the other hand, the mangrove-borders of the several low islands and
islets, mainly formed of reef-_débris_, that lie off the coast, have
often extended themselves during this period in a marked degree. The
results of my comparisons are given below, the rate of advance being
obtained by halving the increase in length or breadth as measured
between the mangrove-borders, the breadth being used in the long
islands.

 _Advance of the Mangrove-Borders of Low Islands on the North Coast of
                   Vanua Levu between 1840 and 1895._

 Thukini, or Gibson Island of Wilkes    700 to 800 yards
 Nangano, or Piner’s Island of Wilkes   300 to 400   "
 Nandongo, or Nuvera of Wilkes                 500   "
 Talailau (two new islands)             400 to 900   "
 Nukunuku or Clark’s Island of Wilkes } Not much change.
 Thakavi, or Day’s Island of Wilkes   }

It will be noticed that the islands of the Talailau Reef are not marked
in the chart of 1840; they are both low mangrove islands, the largest
being slightly under a mile long and the smallest a little under half a
mile. In Nukuira Island, the Vatou of Wilkes, there has been a decrease
of about two-thirds of a mile during this period. The difference between
Thukini in 1840 and in 1895 is very noticeable. In the time of Wilkes
the mangroves only occupied about one-third of the reef-patch. Now they
occupy about two-thirds, the area of the reef-patch remaining much about
the same. Taking the minus and plus values of all the islands here
measured, the average rate of the advance of the mangrove-margins during
this half-century may be placed at about 250 yards in the case of these
reef-islands, which would amount to a mile in 400 years.

It is probable that a long island like Ndongo, which is about four miles
in length, has been formed by the union of smaller mangrove islands.
Therefore, taking half its maximum breadth of a mile as a guide, it
would at this average rate of growth require two centuries for its
formation. But since the extension of the mangroves depends on the
growth of the reef-patch, which takes place on the average at a much
slower rate, it follows that this can only be a minimum limit for the
age of this island. We can only assume that if the reef-patch had
suddenly appeared 200 years ago, Ndongo Island could by this time have
acquired its present dimensions. It does not follow that the mangrove
border has been continuously advancing. A hundred years ago there may
have been a state of equilibrium between the growth of the mangrove and
the reef-patch, which does not now exist. All we can say of some of
these low islands is that the mangroves have been rapidly extending
their margins during the last half century, and that the normal
adjustment between reef-growth and mangrove-growth, which must have once
existed, does not now prevail.

There is evidence of the shoaling of the ship channel amongst these
islands to the extent of about a fathom during this period.[12] The
usual depth immediately around the patches, on which the islands have
been formed, is 8 to 10 fathoms. If, therefore, the shoaling is a
general process, it is to be inferred that although the outward growth
of the reef-patches would be usually very slow, probably not over fifty
yards in a century, there must be times when, in shallowing depths, the
growth of the reef-patch would be comparatively rapid; and it is at such
times that the adjustment between the relations of mangrove and
reef-patch would be upset so that the advance of the mangroves would be
for a time unrestricted.

It is, therefore, apparent that the rate of growth of one of these low
islands is not to be determined by the rate of growth of the
mangrove-tract occupying the surface. The subject is a complicated one;
but I think enough has been said to show that the destructive agencies
do not prevail on this great submarine platform on the north coast of
Vanua Levu.

If the data here adduced of the increase of the low islands, of the
shoaling of the channels, and of the advance of the delta of the Lekutu
river,[13] are well founded, all the islands, islets, and reef-patches
that lie along this north coast will be united to each other and to the
main island within a thousand years.

The facts here produced do not directly indicate a movement of upheaval
but they are quite consistent with the conclusion that the great
movement of elevation which has built up Vanua Levu by the union of
several smaller islands is still in operation at its coasts. To assume
that there is now in progress at the sea-border the same process of
island-building which has produced Vanua Levu, as we now see it, is to
assume a uniformity in nature’s methods which is disregarded by the
hypothesis that the great submarine platform, from which the large
islands of Viti Levu and Vanua Levu now arise, represents the work of
marine erosion into the flanks of the upheaved islands since the last
elevation. The origin of this submarine platform is dealt with in
Chapter XXVII. Here it may be remarked that I regard it as older than
the islands that rise from it.

However, this movement of upheaval is so gradual that the utmost one can
expect to do by the comparison of surveys made half a century apart is
to show the lack of evidence of the destructive agency of erosion. As
far as the comparison admits of judging, there seems to have been no
important change on the coasts of the western end of the island during
this period. The low neck of land connecting Naivaka with the main
island, if we take the low-water line in the Admiralty chart as the
limit, had much the same breadth at the time of both surveys. The depths
in Mbua Bay remain about the same, with perhaps a shoaling of less than
a fathom in places. There are two cays awash in the Admiralty plan of
this bay which were described as sand-spits in the time of Wilkes. The
promontory of Lekumbi could scarcely have been expected to show any
extension during this time, since there are depths of 10 to 16 fathoms
close to its extremity; and there is in fact no difference of critical
importance indicated in the charts.

Some of the principal points of this chapter may be thus summed up:—

(1) Upraised reef-limestones are of very limited occurrence. They occur
at and near the coast and do not extend higher than 300 feet. Their
scarcity at the sea-border is to be attributed to the denuding agencies.

(2) Since foraminiferous muds and sedimentary tuffs with marine organic
remains occur at all elevations up to over 2000 feet, it is assumed that
the absence of reef-limestones in the elevated interior indicates the
paucity or absence of reef-growths in the early stages of the history of
the island. The overlying agglomerates have often preserved from
destruction the soft sedimentary deposits beneath; but they seem to have
never covered over a coral reef.

(3) The relation between the mangrove-belt and the reef-flat indicates a
state of equilibrium which might have been established long ago. If the
movement of emergence is still in progress, it must therefore be of a
very gradual nature, since the normal relation between the mangrove-belt
and reef-flat now prevails.

(4) From the circumstance that the submarine platform passes with a
uniform slope into the low-lying plains, covered with submarine
deposits, it may be inferred that a very gradual emergence is now in
operation.

(5) A comparison of the charts of Wilkes and of the British Admiralty
shows that on the north coast of the island during the last half century
the destructive agencies of marine erosion have not prevailed.

(6) The results of the comparison of the charts, whilst they do not
directly imply a change of level, are quite consistent with the
conclusion that the movement of emergence, which has been in operation
probably since the later Tertiary period, is not suspended.


_Note._—The extensive evidence of emergence presented by this island is
treated in Chapter XXVII. in connection with the whole group. It is not
always possible to avoid in such a discussion the use of terms such as
“upheaval” and “subsidence,” although there is much to be said for the
terms “negative” and “positive” employed by Suess. In the present
chapter, however, I have avoided committing myself definitely to any
view relating to the stability either of the land or of the sea,
reserving the consideration of the subject for Chapter XXVII.



                              CHAPTER III

                     THE HOT SPRINGS OF VANUA LEVU


THE abundance of hot springs in Vanua Levu, and in fact in the group
generally, is not commonly known. In the earlier accounts of these
islands those of Savu-savu are often alone referred to, not only for
this island but for the whole archipelago. The United States Exploring
Expedition under Wilkes spent six months in 1840 in making a survey of
the whole group. Yet Dana, who was attached to the expedition, remarks
that “the only trace of actual volcanic heat which the islands appear to
contain is found at Savu-savu Bay.”[14] Horne in his excellent account
of the group, which he visited in 1878, was among the first to direct
attention to the abundance of hot springs there; but he does not
enumerate many. Although he travelled extensively over Vanua Levu, he
refers to only three in that island, namely, at Savu-savu, Wainunu, and
Vunisawana.[15] It will be shown below that most of the thermal springs
discovered by me might easily have been overlooked.

Before dealing with those of Vanua Levu I will mention the other
localities in the group in which thermal springs are from various
sources known to me. They probably form but a small proportion of those
that actually exist; but the list can be readily extended by those
acquainted with special parts of the archipelago. In Viti Levu they
occur amongst other places at Wai Mbasanga, on the Singatoka river
(Horne) and at Na Seivau on the Wai Ndina, where Macdonald in 1856 found
temperatures of 106° and 140° Fahr. in two different springs.[16] Mr.
Thiele in more recent years referred by hearsay to some hot springs on
the Wai Ndina.[17] Kleinschmidt in 1876 visited a hot spring near the
village of Nambualu in the island of Ono which rose up in the midst of a
brook and had a temperature of about 100° Fahr.[18] The same naturalist
in July of that year, when accompanied by Dr. Max Büchner, came upon a
hot spring issuing among the mangroves at the coast about a mile from
the village of Ndavingele in Kandavu. He did not take the temperature;
but he says that Colonel Smythe (about 1860) observed the temperature to
be 144° Fahr.[19] Different writers refer to extensive hot springs on
the island of Ngau. They are placed near the beach, and close to an
ordinary cool spring. Miss Gordon Cumming in _At Home in Fiji_ gives an
illustration of them. Horne mentions a hot spring on the island of
Rambi. Andrews describes two others that bubble up through the limestone
near the tidal zone in the southern part of Vanua Mbalavu. Both these
springs are in close proximity to the junction line between the intruded
andesite and the old reef rock. One of them, though not boiling, was hot
enough to scald the skin.[20] This list is no doubt capable of being
much extended, especially for Viti Levu and the Lau Group.

A description of the several systems of thermal springs of Vanua Levu
will now be given.


1. THE HOT SPRINGS OF THE LOWER VALLEY OF THE WAINUNU RIVER.—This is one
of the most extensive systems of the kind in the island. The temperature
of the various springs during my sojourn in this district in 1898 ranged
from 100° to 130° Fahr. Those known to me are mostly situated in the
lower part and at the mouth of the Ndavutu Creek, one of the tributaries
of the Wainunu. They open usually on the river-bank, either close to the
water or a few feet above it, but some of them find an exit under water
at the bottom of the river. Natives allege that hot springs occur at
intervals on the left bank and at the river-bottom along the whole
length of the river below Ndavutu Creek. There is certainly a hot spring
on the right side of the river’s mouth near Mr. Dyer’s house. It issues
from the reef-flat and can only be observed at exceptionally low tides.
There is also a hot spring which rises up at the edge of the stream at
Thongea (Cogea) nearly a mile above Ndavutu. If the above statement of
the natives is correct, as I believe it is, then these thermal springs
issue along a line quite four geographical miles in length extending
inland from the mouth of the Wainunu.

All the springs are situated in the tidal part of the river-valley, with
the exception of that of Thongea, which is just above this limit. They
are but little elevated above the sea-level, those exposed being usually
not more than ten feet above the river and often much less. This is a
region of basalt, the valley of the Wainunu lying, as described on page
82, in the fold between two great basaltic flows, and probably
representing a line of weakness, along which the hot springs issue
either from among loose blocks, or from the soil, or from a tufaceous
sandstone. They deposit little if any of the siliceous sinter which is
often found in the thermal waters of this island. This is due probably
to their scanty exposure and to their low temperature. The density of
the water is near that of fresh water, being not over 1001. The
following temperatures may be useful for comparison with future
observations:

 Thongea, when not covered by the stream             July, 1898, 127° F.
 Ndavutu, bath-spring at Mr. Barratt’s house             Usually 100° "
  " on left bank of the creek near the landing place June, 1898, 126° "
  " on left bank of creek near mouth                 Dec.     "  127° "
  " pool in foot-path on left bank                 { June 2,  "  112° "
                                                   { July 27, "  111° "
  " at bottom of main river in depth of 3 feet,  }
     close to the left bank and just above the   }   July     "  122° "
     mouthof the Ndavutu creek, self-registering }
     Six thermometer used                        }


2. The Hot Springs of Natoarau and its Vicinity.—This thermal system
lies in the lower valley of the Mbale-mbale branch of the river
Ndreke-ni-wai. The principal springs are situated at Natoarau, a village
about half a mile in a direct line from Mbale-mbale, about three miles
from the coast, and only about fifty feet above the sea. They bubble up
in pools near brooks, and extend at intervals over an area probably
several hundred yards across. Five springs came under my notice; but
there are doubtless several others in the low-lying and often swampy
land of this district. No deposits were noticed, but the mode of
occurrence and low temperature of the springs serve to explain this
fact. The following temperature observations were made by me in March,
1899:—

 A. Pool 4 feet across, with sides of stone, close to village 126° F.
 B. Pool 10 feet wide, a few paces from pool A                114° "
 C. Pool 12 feet wide, 100 yards from village, near the river 103° "
 D. Pool on the road to Mbale-mbale, mixed with surface water 100° "

The natives and others often state that the thermal springs here and in
other localities are much hotter in dry than in rainy weather. This is
correct in a sense, because in wet weather the surface water would
usually find access to the pools; but there is no reason to believe that
the temperature of the water at the hole of exit varies at all from this
cause. The temperature of pool A was taken at the bottom where the water
bubbled up; and probably it represents the true degree of heat of these
springs, since in the other cases observation of this point was not so
easy. The weather was dry during this visit; but, three months before, I
tested the temperature of this pool after heavy rain, when the district
was flooded, and then I got a reading of 127° at the exit-hole of the
spring.

Another thermal spring, which is distant about a mile from Natoarau, is
known as Waitunutunu, that is, Warm Water. It lies about a third of a
mile from the village of Nambuniseseri, between Mbale-mbale and Waisali,
and is quite four miles inland and about 100 feet above the sea. The
springs bubble up into a pool, about 12 feet across, which is close to a
brook and had a temperature in March, 1899, of 109-112° F.


3. THE HOT SPRINGS OF NUKUMBOLO.—The village of Nukumbolo, where the
springs are situated, lies on the banks of a tributary of the Vatu-kawa
branch of the river Ndreke-ni-wai, and is distant as the crow flies
about six miles inland from the river’s mouth. The springs issue on a
hill-slope from several places a few steps apart, and are removed about
a hundred yards from the river, and from 20 to 30 feet above it. Their
elevation above the sea would be about 130 feet. The temperature taken
in the two hottest places was 157° F, in November 1898, and 158° in the
following February. As in the case of the springs of Savu-savu and a few
other localities, the rocks are coated with siliceous sinter mixed with
carbonate of lime, and a gelatinous incrusting alga grows on the borders
of tiny hollows bathed often in water of a temperature 137-140°, but
thriving most where the temperature is 115-120°. The water runs down the
slope into a series of pools made by the natives for bathing, the
temperature of the lowest pool being 103-105° and of the highest 120°.
This is one of the best localities I have seen in the island for the
erection of thermal baths. The rock pierced by the springs is apparently
a basic agglomerate-tuff. Large blocks of a hard and somewhat altered
palagonitic tuff lie around the bathing pools.


4. THE BOILING SPRINGS OF SAVU-SAVU.—These springs figure in all the
descriptions of the group, and they are also famous amongst the natives.
Since they were described by Wilkes, who visited them in 1840, in his
narrative of the United States Exploring Expedition, many accounts of
them have been written by subsequent visitors; not infrequently they
have been sketched as well as described; and several analyses of their
waters have been made.[21] The accounts of these springs that lie before
me extend at intervals over a period of nearly sixty years; but I shall
allude to them only so far as they throw light on the history of the
springs during this period.

The principal springs are situated in a slight hollow in a more or less
level tract extending in from the beach, and are distant about 150 yards
from the shore and about ten feet above the sea-level. They are five or
six in number, and at the time of my visits in July and November, 1898,
they were boiling briskly, the thermometer readings being 208-210° F.,
but the mercury probably fell two or three degrees in withdrawing the
thermometer. When, as was the case when Wilkes visited this locality in
1840, there is but a slight appearance of boiling, brisk ebullition is
produced by covering them over with leaves. The natives call this
locality Na Kama, which signifies “the burning place,” and employed the
springs extensively for cooking their food. Just as Wilkes describes, a
freshwater brook runs past the springs and receives their outflow. The
temperature of the brook immediately above the springs is that of an
ordinary freshwater stream 75-76° F.; but below it is scalding. The
account given by Wilkes of the spring and of the brook in 1840 applies
to them in our own time. The small stones lying in the effluent channels
of the springs are incrusted with siliceous sinter, and a green alga
lines the sides, bathed generally in the steam but sometimes partially
immersed in water only a few degrees below the boiling point. It is
noteworthy that this alga which was flourishing in July was all dead in
November.

The scalding water also oozes through the sand of the adjacent beach in
abundance for a distance of at least some hundreds of yards. It is even
stated that as far as Ndaku, a mile to the westward, the hot springs
issue at intervals through the beach.[22] There are evidently also
extensive submarine springs close to the beach; and probably Wilkes was
not far from the truth when he remarked that the “whole area of
half-a-mile square seems to be covered with hot springs.”

Off the beach, a few hundred yards to the westward of the springs, is a
batch of dead reef formed of massive corals and only approachable from
the shore at extreme low-tide when it is a little exposed. From numerous
small holes and cracks in the dead-coral hot water issues almost at the
boiling point (210° F). It is apparent that these springs have appeared
at this particular spot since the corals grew. But it is remarkable that
this has been apparently going on since the visit of Wilkes in 1840. He
refers to a coral rock, distant one-third of a mile from the springs and
150 feet from the beach, through which boiling water was issuing in
several places. This rock which was then 10 feet wide and 20 feet long,
was at his visit exposed for three feet at low-tide and covered at
high-tide.[23]

The geological characters of this locality are described on page 191. I
may here remark that if these thermal springs occupy the position of an
old crater, it would require much imaginative power to restore it now.
The off-lying small island of Nawi might by its situation appear to
countenance this idea, but I found no special indication, when I
examined it, in support of this view. From the geological character of
the district, I would infer that if a crater once existed here it was
submarine and that it has been long since obliterated by marine and
aërial denudation. The boiling springs come up through apparently a
rotten volcanic agglomerate. The slight hollow of three or four feet
deep, in which they lie, was considered by Kleinschmidt to be an old
crater cavity; but it is only 40 or 50 feet across, and in the earlier
descriptions the hollow is described as surrounded by a mound of earth.
As shown below, the natives themselves may be held responsible for many
changes in the surface around the springs. There is, in fact, no trace
of a crateral cavity in this district now.

I will now briefly notice the history of the boiling springs since 1840,
when they were visited by Commodore Wilkes. At that time there were five
springs, situated in a basin 40 feet across, and possessing a
temperature of 200-210° F. Although there was scarcely any appearance of
boiling, rapid ebullition could be excited by covering the springs with
leaves and grass. The natives alleged that the springs had always been
in the same condition. In 1863, when the Chief of Wainunu (Tui Wainunu)
came to fight the Savu-savu people, he endeavoured but without success
to choke up the springs by heaping earth over them. I was informed of
this circumstance by Mr. A. H. Barrack, the owner of the springs. Miss
Gordon Cumming also refers to it in her book _At Home in Fiji_. When
this lady visited the springs in August, 1876, they were intermittent in
their action, the highest making a fountain two to three feet high.
According to the description of Kleinschmidt they were in the same
intermittent condition in May of the same year. There were then four
springs situated in a bowl-shaped hollow. The two larger springs were
not constantly bubbling up, but displayed periodic ebullitions of about
twenty minutes’ duration, the waters disappearing in the intervals. The
other two springs were not then active. Horne, who visited this locality
in 1878, refers to three or four principal springs situated in the
centre of a hollow, which was surrounded by a mound of earth, the water
boiling up to the height of about a foot.

About this time the springs entered for a while into a new phase of
action and assumed the form of geysers. According to information
received from Mr. A. H. Barrack and other old residents in Savu-savu,
the waters spouted up to a height of from 40 to 60 feet, not vertically
but at an angle. Each outburst, which lasted for ten or twenty minutes,
was followed by a similar interval of repose, during which the springs
dried up. This continued for a month or two, after which the springs
gradually resumed their normal level. When I visited the springs in July
and November, 1898, they were boiling briskly, attaining a height of a
few inches, and showed no signs of intermittent action.

I come now to the different analyses that have been made of the water of
these thermal springs of Savu-savu. Specimens have been analysed at
different times by chemists in various parts of the world, in America,
in Germany, in Australia, etc., and the results as far as known to me
are now appended.


    A. _Analysis by Dr. C. T. Jackson of Boston, U.S., of the water
       obtained in 1840 by the Wilkes Exploring Expedition._[24]

Specific Gravity 1·0097. Temperature 57° F.

The evaporation of a quantity equal to 1000 grains of distilled water
gave 7·2 grains of salt, thus composed:—

 Chlorine                                            3·577
 Sodium                                              1·665 or Soda 2·238.
 Magnesia                                            0·440
 Lime                                                0·366
 Silica and iron with a trace of phosphate of lime   0·200
 Carbonic acid                                       0·493
                                                     -----
                                                     6·741
 Organic matter and loss                             0·459
                                                     -----
                                                     7·200


        B. _Analysis by Dr. Oscar Pieper of Hamburg of the water
            obtained by Mr. Kleinschmidt in May, 1876._[25]

The report stated that the water was clear, neutral in reaction and
salt-bitter in taste, brown flakes of hydrated iron oxide occurring in
it after long standing. The dissolved salts amounted to “8·48 g. per
litre,” and the remark is made that “the concentration is therefore not
so great as in sea-water.” The solid constituents consisted in by far
the greatest part of Natrium and Calcium chlorides. A quantitative
determination, which on account of the small quantity of the water was
confined to “eine Chlor und Kalkbestimmung,” gave this result:—

 Chlor (Chlorine)      4·79 g. per litre.
 Kalk (Lime)           2·31 "   "

Reckoned as Chlornatrium (Kocksalz) and Chlorcalcium, these results were
obtained:—

 Chlorcalcium (Calcium chloride)      4·55 g. per litre.
 Chlornatrium (Sodium chloride)       3·09 "  "

Amongst other constituents found in small quantities were Sulphuric
acid, Silicic acid (Kieselsäure), Potash, and Iron oxide. Iodine,
Bromine, Nitrates, and Borates were completely wanting. “If this water,”
says Dr. Pieper, “has healing properties, it does not owe them to its
chemical composition.”


      C. _Analysis by Mr. H. Rocholl of sample obtained by Mr. H.
            Stonehewer Cooper probably in 1877 or 1878._[26]

   Total solids at 212° F.        ·8796 per cent.
     "     "    ignited           ·7726  "   "

 The residue consisted of—
   Free Sulphuric Acid (SO_{3})   ·0049  "   "
   Calcium sulphate               ·0260  "   "
   Calcium chloride               ·4355  "   "
   Magnesium chloride             ·0021  "   "
   Potassium chloride             ·0415  "   "
   Water                          ·1070  "   "
   Sodium chloride                ·2641  "   "
                                 -------
                                  ·8811


D. _Analysis by Prof. Liversidge of the Sydney University of a sample of
  the water collected by Dr. Bromlow, R.N., about 1879._[27]

The specific gravity was 1·0064 at 60° F. The total solids in solution
were 582·4 grains per gallon; but when heated to a dull red heat, the
residue was 546·9 grains per gallon, the combined water having been
driven off. Iodine and bromine were carefully sought for, but in vain.
Four pints of the water were examined.

                              COMPOSITION.

 +--------------------------+------------+-----------------+----------+
 |                          |Per cent. in|    Parts per    |Grains per|
 |                          |  residue.  |million of water.| gallon.  |
 +--------------------------+------------+-----------------+----------+
 |Silica, insoluble         |   1·681    |      133·3      |   9·20   |
 |Silica, soluble           |    ·074    |        5·8      |    ·40   |
 |Alumina and traces of Iron|            |                 |          |
 |  sesquioxide             |    ·534    |       41·7      |   2·92   |
 |Aluminium chloride        |   1·646    |      128·6      |   9·00   |
 |Phosphoric acid           |   traces   |     traces      |  traces  |
 |Calcium chloride          |  46·754    |    3,652·9      | 255·70   |
 |Calcium sulphate          |   4·770    |      372·7      |  26·09   |
 |Magnesium chloride        |    ·154    |       12·0      |    ·84   |
 |Sodium chloride           |  42·171    |    3,294·8      | 230·64   |
 |Potassium chloride        |   1·756    |      137·2      |   9·60   |
 |Carbonic acid             |   traces   |     traces      |  traces  |
 |Loss                      |    ·460    |       34·0      |   2·52   |
 |                          +------------+-----------------+----------+
 |                          | 100·000    |    7,813·0      | 546·91   |
 +--------------------------+------------+-----------------+----------+

Looking at the general character of these thermal springs of Savu-savu
we may quote the remarks of Prof. Liversidge and Dr. Pieper that the
salts in solution consist for the most part of chlorides, the chlorides
of calcium and sodium largely prevailing.

    COMPARISON OF THE ANALYSES OF THE WATER OF THE SAVU-SAVU THERMAL
          SPRINGS, STATED IN GRAINS PER THOUSAND OF WATER.[28]

 +----------------+-------+------+-------+--------+-------+-------+------+--------+
 |                | Date  |Chlo- |       |        |Calcium|Natrium|Total |        |
 |                | col-  |rine. |Sodium.|Calcium.|chlor- |chlor- |salts.|Density.|
 |                |lected.|      |       |        | ide.  | ide.  |      |        |
 |----------------+-------+------+-------+--------+-------+-------+------+--------+
 |Dr. Jackson     | 1840  | 3·57 | 1·66  |  0·36  |  —    |  —    | 7·20 | 1·009  |
 |Dr. Pieper      | 1876  | 4·79 |  —    |  2·31  | 4·55  | 3·09  | 8·48 |   —    |
 |Mr. Rocholl     | 1878  |  —   |  —    |   —    | 4·35  | 2·64  | 8·81 |   —    |
 |Prof. Liversidge| 1879  |(4·50)|(1·29) | (1·42) | 3·65  | 3·29  | 7·81 | 1·006  |
 +----------------+-------+------+-------+--------+-------+-------+------+--------+
 |Sea-water, tropics      |19·46 | 11·08 |  0·46  |  —    |  —    |35·00 | 1·02   |
 +------------------------+------+-------+--------+-------+-------+------+--------+

It is to be inferred from the above that the quantity of salts in
solution remains about the same, the proportion varying only in the four
analyses, which extended over a period of forty years, between 7·2 and
8·8 grains per thousand grains of water. This is considerably less than
the salts in solution in sea-water, namely 35 grains per thousand. The
relative proportions of the salts, excepting those of calcium, do not
vary more than we should expect in the case of analyses made by varying
methods and probably with a varying degree of exactness.

Dana[29] considered from Dr. Jackson’s analysis that the water of the
Savu-savu springs is probably of marine origin; but the absence of
bromine and iodine, as especially remarked by Dr. Pieper and Prof.
Liversidge does not support this view. We might also expect the
proportion of the salts to each other to show a greater similarity to
that in sea-water than they do. On the other hand the total volume of
water discharged, not only by the springs proper but for several hundred
yards along the beach, and also between the tide-marks and beyond, must
be far greater than could be supplied by the rainfall of this portion of
the Savu-savu peninsula, which is only one and a half to two miles
across and 800 feet high. We must look, I think, for the source of these
waters in deep subterranean streams or artesian basins that would be fed
by the rains precipitated in the mountainous districts where the
rainfall amounts to at least 200-300 inches in the year. This matter is
further discussed in my general remarks on the hot-springs of this
island (page 38).


5. THE HOT SPRINGS NEAR RAVUKA.—These springs rise up in the centre of
the breadth of the island about nine miles direct from the coast. They
are about 200 feet above the sea and are situated on the Ndrawa branch
of the Ndreketi River some two miles below the hamlet of Ravuka. They
are on a small scale and ooze through a bed of rounded blocks and
pebbles close to the water on the left bank. Their temperature in
August, 1898, was 148° F. They are covered by the river when it is
swollen by the rains, and very probably other hot-springs issue along
the river-bottom. The conditions are not suitable for the formation of
deposits.


6. THE HOT SPRINGS OF VUINASANGA.—These thermal springs are also
situated in the heart of the island on a tributary of the Ndreketi some
three or four miles westward from Va Lili and about 150 feet above the
sea. On each bank of the river four or five paces from the water and
three or four feet above it, there is a small pool two to four feet
wide. In June, 1899, the pool on the right bank had a temperature of
131° F., and that on the left bank of 134°. There were no deposits.


7. THE HOT SPRINGS ON THE SOUTH SIDE OF THE NAWAVI RANGE.—These springs
also lie within the borders of the valley of the Ndreketi. They may be
“located” by describing them as lying a few miles inland from the north
coast fronted by Mathuata Island. I did not visit them and have only
learned of them from Mr. Thomson’s Mathuata paper.[30] That gentleman
refers to them as two in number and situated at the back of the coast
range about four miles inland from the village of Nangumu; but no
particulars are given.


8. THE HOT SPRINGS OF VATULOALOA.—These springs lie on the Mathuata
Coast in the neighbourhood of Mathuata Island. I have not seen them, but
am indebted to Mr. Thomson for the particulars here given, which are
taken from his paper above quoted. Mr. Thomson, who discovered them in
1880, named them the “Graçie” springs. They issue below high-water mark
at Vatuloaloa, and had a temperature in 1880 of about 140° F. They are
said to possess many valuable healing qualities.


9. THE HOT SPRINGS OF NAMBUONU.—These springs are situated on the same
part of the Mathuata coast as those of Vatuloaloa above referred to. I
learned from Mr. Bulling of Undu Point that they issue from swampy
ground half a mile inland. They were discovered accidentally by a
Japanese who put his foot into them, the temperature being sufficiently
high to scald the feet, but not at the boiling-point, probably about
140° F.


10. THE HOT SPRINGS NEAR TAMBIA.—These extensive springs, situate 1½ to
2 miles inland, and rather under 100 feet above the sea, lie near the
Mathuata or north coast of the island, some four miles west of the
Wailevu river. They rise up in the midst of level country about a mile
from the town of Tambia, and near the village of Ngovungovu. Although
situated in the valley of the Tambia river, these springs are not
adjacent to the river, and in this respect they differ from nearly all
the inland hot springs. The hottest spring bubbles up into a pool 5 or 6
feet across, which had a temperature of 180° F., in March, 1899. Near by
is a large deep pool, some 20 feet or more across, with a temperature of
100°. It receives the overflow from the smaller pool, and apparently hot
water also bubbles up at the bottom. Around the smaller hottest pool
there is a considerable deposit of what is mainly siliceous sinter. It
incrusts the stones and also the oyster-shells lying about the pool in
quantities, where they have been left by the natives after their
contents had been cooked and eaten. Some of the shells are almost
decayed away, the sinter for the most part alone remaining.


11. THE HOT SPRINGS OF VANDRANI.—These springs occur in the heart of the
island, about 8 miles from the coast in a straight line, and about 270
feet above the sea. This is the greatest elevation, as far as I know, at
which a hot spring exists in this island. Here they rise up near the
base of the central mountain range, close to the head-waters of the
Wailevu river which opens into Lambasa bay. The springs bubble up into a
pool, a foot deep, on the left side of the river, four or five paces
away from the water’s edge, and scarcely raised above it. They are
covered over when the river is in flood. In February, 1899, the
temperature recorded by my thermometer was 100° F.; but probably it was
a few degrees higher at the bottom of the pool. I noticed no deposits.


12. THE HOT SPRINGS OF NA KAMA ON THE WAILEVU RIVER.—These boiling
springs, which are of an extensive character, come up in half-a-dozen
places on either bank of the river, and are from 5 to 6 miles inland,
and about 90 feet above the sea. They are close to the water, and from 1
to 10 feet above it. The temperature of one small pool, where the water
bubbled up briskly, was 204° F. in February, 1899. In another it was
194°. The water was probably at the boiling-point in some cases as it
entered the pools, and in the others it could have been only a few
degrees below it. The rocks of the district are agglomerates and tuffs.
I have no recollection of deposits of any extent around the springs.


13. THE HOT SPRINGS OF VUNIMOLI ON THE LAMBASA RIVER.—A few minutes’
walk from Vunimoli, and about 100 yards from the left bank of the river,
these hot springs issue in a place named Vunimbele from the
foraminiferous clay rock (soapstone) of the district. They are on the
side of a ditch which communicates with the river. The natives have cut
out of the soft rock small square basins which receive the waters. The
temperature of the hottest spring in August, 1899, was 155° F. That of
others was 140°. The conditions are not favourable for the formation of
deposits. These springs lie about 8 miles inland and are rather over 100
feet above the sea. They are, however, small and unimportant, and the
locality in which they occur is now overgrown with vegetation and not
easy to discover.


14. THE HOT SPRINGS OF MBATI-NI-KAMA ON THE NGAWA RIVER.—These springs
are situated in the Lambasa district about 7½ miles from the coast, and
rather over 100 feet above the sea. They issue copiously from the
volcanic agglomerate at a temperature of 161° F. (August, 1899), and are
only removed a few paces from the river, and a foot or so above it. Algæ
flourish in the water, and siliceous sinter incrusts the rocks.


15. THE HOT SPRING OF NANDONGO ON THE HEAD-WATERS OF THE WAI-NI-KORO
RIVER.—A few hundred yards from the village and elevated about 180 feet
above the sea there is a small pool in the clay of the river bank, 2 or
3 feet above and close to the water, which in September, 1899, had a
temperature of 97° F.


16. THE HOT SPRINGS OF NATUVO ON THE NORTH COAST OF NATEWA BAY.—About a
mile east of Mbiagunu and near the village of Natuvo, there are two hot
springs of small size which I visited in August, 1899. One that issued
on the reef-flat from the coral-rock at a temperature of 136° F. was
covered over towards high-tide. The other issued near by at a
temperature of 131° from swampy ground a few paces among the trees.


17. THE HOT SPRINGS OF NDAKU-NDAKU ON THE NORTH COAST OF NATEWA BAY.—At
this place about 2 miles north of Vuinandi some hot springs rise through
the reef-flat, which are only exposed at low tide. At the time of my
visit they were covered over by the rising tide. The natives described
them as not very hot and like the neighbouring hot springs of Natuvo.


18. THE HOT SPRING OF NAVAKARAVI, NATEWA BAY.—The coast village thus
named lies about one and a half miles north of Were-kamba. The hot
spring is about a mile inland and not over 30 to 40 feet above the sea.
It is reached after traversing a low and often swampy tract. The spring
in August, 1899, issued from a little rise at a temperature of 133°
Fahr., and formed a rivulet 18 inches across.


19. THE HOT SPRINGS OF VUNISAWANA AT THE HEAD OF NATEWA BAY.—Mr. Horne,
who was in this locality in 1878, refers to these springs in his book _A
Year in Fiji_. They had at one time, he remarks, a wide reputation for
their curative qualities; but the people around became so poor on
account of the hospitality that custom compelled them to extend to the
numerous visitors that they buried up the springs. Mr. Horne was shown
the site at the bottom of a muddy creek. I saw it in 1898. It lies 300
or 400 yards in from the beach and only a few feet above the sea. There
were no signs of heat then; but I was told that when the stream close by
is very low it sometimes is a little warm.


20. THE HOT SPRING OF NDREKE-NI-WAI ON THE SOUTH COAST OF NATEWA
BAY.—This small spring issues between the tide-marks from an old
reef-patch close to the shore and is only to be seen at low-water. Its
temperature in May, 1898, was 130-135° Fahr.


21. THE HOT SPRING OF WAIKATAKATA ON THE SOUTH COAST OF NATEWA BAY.—This
important spring lies about four miles east of the town of Natewa. It
issues on a hill-slope about 400 yards from the beach and is some 25 or
30 feet above the sea; but it is so beset by undergrowth that the source
is not easy to reach. Boulders and blocks of a basaltic rock lie about
on the slope; and it is from under a huge boulder of five or six tons in
weight that the spring emerges at a temperature of 148° Fahr. (April,
1898). There is a good volume of water, and a series of bathing pools of
varying temperature could be readily made. Unlike most of the inland hot
springs, it is not in connection with a stream or river.


22. THE HOT SPRING OF NDEVO ON THE COAST OPPOSITE TO RAMBI.—I did not
hear of any spring when in the locality; but I learned afterwards that
near a stream on the beach there is a hot spring which is covered at
high tide.


23. THE HOT SPRING OF NAVUNI NEAR FAWN HARBOUR.—This small spring is
situated in a hilly district in a region where olivine-basalts prevail.
I was indebted to Mr. Pickering for showing me its locality. It lies
about three-quarters of a mile inland and about 100 feet above the sea.
It issues from the volcanic agglomerate a few paces from the right bank
of the Navuni stream and five or six feet above its level. In May, 1898,
it had a temperature of 112-113° Fahr.


                   GENERAL REMARKS ON THE HOT SPRINGS

This island is therefore remarkable for the number of its hot springs.
In the list given on page 40 I have enumerated 23 localities where they
occur; but, as shown below, their number will probably in time be
extensively increased.

On referring to the map it will be observed that the distribution of
these springs is fairly general over two-thirds or three-fourths of the
island. Taking this area at about 1,500 square miles and dividing it
into squares with sides of eight miles, we should, if the springs were
quite evenly dispersed, find a thermal system in every square. Even
amongst the Fijians and among the white residents the number of hot
springs will cause surprise. Only those of Savu-savu, Wainunu,
Nukumbolo, Mbatini-kama, and Na Kama on the Wailevu river have been up
to this time generally known. The reason of this is that most of them
are insignificant, and with a temperature far below the boiling-point,
and ooze up in unlikely and out-of-the way places, as by the water-side
in little visited river-valleys, on the reef-flats of not much
frequented coasts, and in swampy situations where they are likely to be
overlooked. The natives only recognise as “Na Kama” the boiling or very
hot springs; and it was only after much questioning that I could get
them to tell me of some unimportant “wai katakata” (hot water) which
they deemed to be far beneath my notice. The natives were keenly
interested in my botanical and geological investigations; but they
considered it to be beneath the dignity of a man who had seen the
wonders of Na Savu-savu to spend some time looking for a half-forgotten
thermal spring in a swamp. From this cause alone I no doubt failed to
find several springs. All the boiling springs and those of very high
temperature are probably known; but as is pointed out below it is more
than likely that a large number of unimportant springs remain to be
discovered in many a deserted inland valley and between the tide-marks
along the very extensive reef-bound coasts.

As above remarked the hot springs did not come under my notice in all
parts of the island. They are to all appearance wanting in the western
or Mbua portion, and also in the Undu portion north of Natewa Bay.
Taking the first-named region, it will be noticed that no hot springs
are indicated in the map west of the Ndreketi and Wainunu rivers. I made
inquiries wherever I went, but with no result. On my writing to Mr.
Wittstock, of Mbaulailai, who is well acquainted with the Mbua
peninsula, he informed me that if hot springs existed in that part of
the island he would probably have known of them. In that portion of the
island which ends in Undu Point I could neither discover nor hear of any
thermal springs east of Lambasa on the north side, and of Lakemba on the
south or Natewa Bay side; nor could Mr. Bulling, who has resided at Undu
Point for many years, tell me of any springs in his neighbourhood.

On looking at the general map it will be observed that the hot springs
are confined to the area of basic rocks, although they do not occur all
over that area, not being indicated in the map to the west of the
Ndreketi and Wainunu rivers. They are not known to occur in the region
of dacites and acid andesites, as in the case of the Drandramea
district; and they have not been found in the area of rhyolitic and
trachytic rocks that extends from Undu Point to Mbuthai-sau on the north
coast and to near Tawaki on the Natewa Bay side. The region of hot
springs would be limited on the east by a line joining the Mbati-ni-kama
springs with those of Nandongo on the Wainikoro river and Natuvo on the
north shore of Natewa Bay. Such a line, though lying within it, roughly
indicates the limit between the regions of basic and acid rocks.

The situation of the hot springs in the lower levels, and their
non-discovery at elevations exceeding 300 feet above the sea, are facts
of importance. In more than half the cases they arise close to and often
on the banks of streams and rivers, occasionally indeed at the
river-bottom; and no doubt numerous unknown thermal springs issue under
water from the river beds. In about a third of the known cases the
springs come up on the coast between the tide-marks, usually rising
through the reef-flat. At times even they are to be observed below the
low tide level; and one can scarcely doubt that there are a large number
of undiscovered springs that are never exposed at the lowest tides. It
is also very likely that a number of hot springs issuing between the
tide-marks are still to be discovered without much difficulty.

The same may be said of inland hot springs. Looking at the insignificant
character of many of them and noting their occurrence in places where
they might easily be overlooked, it is highly probable, as before
remarked, that a number of springs exist inland, which, though once
known to the natives, are now forgotten. The interior of the island is
very sparsely inhabited now; but there is evidence of a much more
populous condition in old times. The present natives are fast losing the
knowledge of the interior of the island which their forefathers
possessed; and many tracts in the mountain districts are far removed
from existing paths. From the haphazard manner in which I lighted upon
thermal springs beside the head-waters of the Ndreketi, Wailevu, and
Wai-ni-koro rivers, I cannot doubt that many more exist in similar
localities not visited by me.

With regard to the distribution of the springs as respecting
temperature, I cannot find any marked arrangement either in their
grouping or in the amount of elevation. It is noticeable, however, that
the three systems of hottest springs, that of Savu-savu (210°), that of
Na Kama on the Wailevu river (204°), and that of Tambia (180°) are all
less than 100 feet above the sea. Although the springs of highest
temperature are confined generally, with the exception of those of
Savu-savu, to the main mass of the island, it would seem that adjacent
systems of springs may differ much in temperature. The springs of
Vunimoli, for instance, have a maximum temperature of 155°, which is
nearly 50° lower than that of Na Kama, three miles to the westward. Hot
springs are more numerous in the region around Lambasa than in most
other districts. Lastly, I may add that earthquakes are apparently more
frequent in the Mbua district, where no thermal springs are known, than
in any other part of the island.

With regard to the deposits formed around the springs, it may be
observed that the circumstances are not usually suitable for their
formation, as for instance when they rise through the reef-flat or in
swampy localities. In those springs, however, where the temperature is
over 150° F., and where the water spreads over a surface so as to
facilitate evaporation, deposits of white sinter associated with algæ
occur, as at Savu-savu, Tambia, and Nukumbolo. Its composition varies a
little in different localities. At Savu-savu it is compact and laminated
and formed almost entirely of hydrated amorphous or colloid silica. At
Mbati-ni-kama the siliceous sinter is more friable, with a tendency to
form opal. The sinter of the Nukumbolo springs resembles that of
Savu-savu; but it also contains a good proportion of carbonate of lime
(20 per cent.) in a granular form, and that of Tambia has the same
characters. It is not unlikely that this lime is derived from the
decayed shells, such as I have referred to in the case of the Tambia
springs.... It may be here observed that Mr. Weed and others, who have
studied the origin of siliceous sinter in the Yellowstone region and
elsewhere, regard it as the secretion of algæ, mosses, &c., that grow in
hot waters (_American Journal of Science_, vol. 37, 1889).

I come now to some general considerations respecting the hot springs of
Vanua Levu. In the first place there is the singular fact that the
inland hot springs nearly always make their appearance along the present
lines of surface-drainage. But I do not gather that the hot springs are
of more recent origin than the rivers and streams, by the side of which
they rise. On the contrary the hot springs are probably far older. The
conditions of subterranean drainage that favour the formation of springs
at the surface, whether cold or thermal, would no doubt often determine
the direction of surface drainage in a newly-formed land. Those familiar
with modern volcanoes will recall the absence or rarity of streams and
rivers, and the frequency often of cold and thermal springs at and near
the coast, which are sometimes of such bulk at the exits that the
expression “subterranean river” would be nearly appropriate. The
presence of artesian reservoirs may also in some localities be safely
assumed. I will here draw a little on my own experience of volcanic
regions.

On the lava-bound coasts of the riverless southern portion of the large
volcanic island of Hawaii, the subterranean waters issue as cold and
thermal springs at numerous localities. At Punaluu, and at Ninoli, a
mile to the westward, there are extensive freshwater springs at and near
the beach which have a temperature of 64° F. all through the year,[31]
those at Ninoli issuing as a large subterranean stream. East of Punaluu
and at intervals along the Puna coast, springs of water, sometimes fresh
and cold with a temperature occasionally as low as 64°, at other times
mineral and thermal, but with a temperature not usually above 95°, issue
at the surface or at the bottom of deep fissures in the old lava
flows.... In Oahu, another island of the Hawaiian group, where the
volcanic forces have been long extinct, artesian wells have been in
extensive use for some years in the irrigation of the sugar-cane
plantations. The last water-bearing strata are reached at depths of 400
to 500 feet.[32] The subterranean or artesian reservoirs are evidently
therefore on a large scale; yet Oahu is scarcely one-third the size of
Vanua Levu in Fiji.... Lastly, I will refer to the numerous subterranean
streams that issue forth, as cold and thermal springs, from beneath the
lavas near and at the Etna coast, as for instance in the vicinity of
Acireale. The Etna slopes are in great part deforested, and in
consequence soakage is relatively small, and after heavy rains much of
the water runs off in the torrents. Whilst in this locality I was
impressed with these facts, and I formed the opinion that in ancient
times when Etna was well wooded the discharge of subterranean streams at
the coast was far greater than at present.

For these reasons and on other grounds, amongst them notably the absence
of recent crateral cavities, I infer that the numerous hot springs are
the outflows of subterranean streams, fed originally by the “soakage”
arising from a rainfall of at least 200 to 300 inches in the mountainous
portions of the island. Such subterranean streams run probably at
considerable depths, emerging, it is likely, as often under the sea as
they do on the land.

Since writing the above I have read in the Journal of the Royal
Geographical Society (November 1902), an abstract of a lecture by Prof.
Suess on the subject of hot springs and volcanic phenomena. Thermal
springs, he holds, are supplied by hypogene waters and do not receive
their salts from the sea. Such springs, according to this view, being
the survivals of volcanic activity, originate in the depths of the
earth’s crust and bring water to the surface for the first time, not
deriving it from infiltration. It seems almost impertinent to suggest a
view opposed to that of such a high authority; but it appears to me that
the frequent situation of the Vanua Levu thermal springs along the lines
of surface-drainage requires an explanation that does not altogether
exclude the agency of infiltration.


            LIST OF THE HOT SPRINGS OF VANUA LEVU, 1898-99.

 ---------+----------+----------+----------+----------+---------+--------
 Locality.| Coast or | Height   |Near or   |Siliceous |Nature   |Tempera-
          |  inland. |  above   |far from  |  sinter. |of the   |ture.
          |          |sea-level.|streams.  |          |surface  |
          |          |          |          |          |at the   |
          |          |          |          |          |exit.    |
 ---------+----------+----------+----------+----------+---------+--------
 1. Wai-  |Coast to  |Sea-level |River     |Little or |Soil-cap,|100°-
 nunu     |four      |to 20 or  |-side and | none     |tuffs,   |130° F.
          |miles     |30 feet   |under     |          | &c.     |
          |inland    |          |water     |          |         |
          |          |          |          |          |         |
 2. Nato  |Three to  |50 to 100 |Near      |Little or |Soil-cap |110°-
 -arau and|four      |feet      |brooks    |  none    |         |126° F.
  vicinity|miles     |          |and       |          |         |
          |from coast|          |streams   |          |         |
          |          |          |          |          |         |
 3. Nukum-|Six miles |130 feet  |About 100 |Mixed with|Agglomer-|157° F.
 bolo     |inland    |          |yards     | lime     |ate-tuff |
          |          |          |from      | carbonate|         |
          |          |          |river     |          |         |
          |          |          |          |          |         |
 4. Savu  |Beach and |Sea-level |Inland    |In fair   |Rotten   |208°-
 -savu    |150 yards |to 10 feet|springs   |  quantity|volcanic |210° F.
          |inland    |          |near a    |          |agglomer-|
          |          |          |brook     |          |ate      |
          |          |          |          |          |         |
 5. Ravuka|Nine      |200 feet  |River     |None      |Pebble   |148° F.
          |miles     |          |-side     |          |bed      |
          |inland    |          |          |          |         |
          |          |          |          |          |         |
 6. Vuina-|Ten miles |150 feet  |River     |None      |Soil-cap |131°-
 sanga    |from      |          |-side     |          |         |134° F.
          |north     |          |          |          |         |
          |coast     |          |          |          |         |
          |          |          |          |          |         |
 7. Foot  |Four      |Not known |Not known |Not known |Not known|Not
 of Nawavi|miles     |          |          |          |         |known.
 Range    |inland    |          |          |          |         |
          |          |          |          |          |         |
 8. Vatu- |Coast     |_Nil._    |          |          |         |140° F.
 loaloa   |below     |          |          |          |         |
          |high-tide |          |          |          |         |
          |mark      |          |          |          |         |
          |          |          |          |          |         |
 9. Nambu-|Half-mile |Probably  |Not known |Not known |Swampy   |140° F.
 onu      |inland    |slight    |          |          |         |(proba-
          |          |          |          |          |         |bly).
          |          |          |          |          |         |
 10. Tam- |Inland    |90 feet   |Not near  |Abundant  |Soil-cap |180° F.
 bia      |one and   |          |a stream  |          |         |
          |a-half to |          |          |          |         |
          |two  miles|          |          |          |         |
          |          |          |          |          |         |
 11. Vand-|Eight     |270 feet  |River     |None      |Old      |100° F.
 rani     |miles     |          |-side     |          |river-bed|
          |from      |          |          |          |         |
          |coast     |          |          |          |         |
          |          |          |          |          |         |
 12. Na   |Inland    |90 feet   |River     |Probably  |Agglomer-|194°-
 Kama,    |five to   |          |-side     | little   |ates and |204° F.
 Wailevu  |six       |          |          |          |tuffs    |
          |miles     |          |          |          |         |
          |          |          |          |          |         |
 13. Vuni-|Inland    |120 feet  |About 100 |None      |Foramini-|{140° F.
 moli     |eight     |          |yards     |          |ferous   |{155° F.
          |miles     |          |from      |          |clay-rock|
          |          |          |river     |          |         |
          |          |          |          |          |         |
 14. Mbati|Inland    |130 feet  |River     |In fair   |Volcanic |161° F.
 -ni-kama |seven and |          |-side     |  quantity|agglomer-|
          |a-half    |          |          |          |ate      |
          |miles     |          |          |          |         |
          |          |          |          |          |         |
 15. Nan- |Inland    |180 feet  |River     |None      |Clay     |97° F.
 dongo    |four      |          |-side     |          |         |
          |miles     |          |          |          |         |
          |          |          |          |          |         |
 16. Nat- |Coast and |_Nil_     |          |None      |Reef-flat|{131° F.
 uvo      |between   |and a few |          |          |and      |{136° F.
          |the tide- |feet      |          |          |swampy   |
          |marks     |          |          |          |ground   |
          |          |          |          |          |         |
 17. Nda- |Between   |_Nil_     |          |None      |Reef-flat|Not
 ku-ndaku |the tide- |          |          |          |         |known.
          |marks     |          |          |          |         |
          |          |          |          |          |         |
 18. Nava-|A mile    |30 to 40  |Swampy    |None      |Soil-cap |133° F.
 karavi   |inland    |feet      |ground    |          |         |
          |          |          |          |          |         |
 19. Vuni-|From      |A few feet|Near a    |Not known |Soil-cap |Not
 sawana   |beach 300 |          |brook     |          |         |known.
          |or 400    |          |          |          |         |
          |yards     |          |          |          |         |
          |          |          |          |          |         |
 20. Ndr- |Coast     |_Nil_     |          |None      |Old reef-|130°-
 eke-ni   |between   |          |          |          |patch    |135° F.
 -wai     |the tide- |          |          |          |         |
          |marks     |          |          |          |         |
          |          |          |          |          |         |
 21. Wai- |Inland    |25 or 30  |Not near  |None      |Rises    |148° F.
 katakata |400 yards |feet      |a stream  |          |beneath a|
          |          |          |          |          |boulder  |
          |          |          |          |          |of basalt|
          |          |          |          |          |         |
 22. Ndevo|On coast  |_Nil_     |Near a    |Not known |Probably |Not
          |below     |          |stream    |          |the      |known.
          |high-     |          |          |          |reef-flat|
          |water     |          |          |          |         |
          |level     |          |          |          |         |
          |          |          |          |          |         |
 23. Nav- |Inland    |100 feet  |Near a    |Little or |Volcanic |112°-
 uni      |three-    |          |stream    |  none    |agglomer-|113° F.
          |quarters  |          |          |          |ate      |
          |of a      |          |          |          |         |
          |mile      |          |          |          |         |

Summary of the previous remarks on the hot springs of Vanua Levu.

(1) Hot springs have been recorded from 23 localities, but there are
probably many undiscovered or forgotten.

(2) They are distributed over much of the island; but have not been
observed in the Mbua or Western end and in the Undu extremity east of
Lambasa and Lakemba.

(3) They are confined to the areas of basic rocks and are not known in
the districts of dacites and other acid andesites or in those of
quartz-porphyry and trachyte.

(4) They are always found at low elevations, never exceeding 300 feet.

(5) Whilst more than half are situated along river and stream courses,
nearly all the remainder lie between the tide-marks.

(6) In only two localities is the temperature at or near the
boiling-point. In one place it is 180° F., and in most of the other
springs it ranges between 100° and 150°.

(7) Siliceous sinter is formed where the temperature is over 150°.

(8) As exemplified by the water of the Savu-savu springs the proportion
of salts in solution (8 per 1000) is constant over many years; whilst in
this fact and in the relative amounts of each salt there is a sharp
distinction from the composition of sea-water.

(9) The hot springs are older than the streams and rivers, along which
they are so frequently found.

It would appear that they are largely supplied from the “soakage” of the
heavy rainfall in the mountains.



                               CHAPTER IV

DESCRIPTION OF THE GEOLOGICAL AND GENERAL PHYSICAL FEATURES OF VANUA LEVU.


IN this chapter the detailed description of the island is commenced,
beginning with the western extremity and proceeding eastward. Most of
the petrological details are dealt with under their respective sections;
but it has been found necessary also to frequently refer to them in this
connection.

THE NAIVAKA PENINSULA.—This mountainous peninsula forms the conspicuous
feature of the western extremity of Vanua Levu. Amongst all the
mountains of the island its appearance from a distance gave most promise
of displaying the products of recent volcanic eruptions; but as shown
below it affords evidence of an antiquity nearly as great as that of the
rest of the island, although there are reasons for believing that its
eruptions took place during the last stage of the emergence.

Naivaka is connected with the adjacent relatively little elevated part
of the main island by a low and narrow neck a little less than a mile in
breadth. In its highest part, where it is only raised between 20 and 30
feet above the sea, this isthmus is formed of the basic volcanic rocks
of the district; but about three fourths of its width are occupied by
mangrove-swamps which are especially extensive on the south side.

Viewed from some miles to the eastward the mountain has a regular
conical outline; but from the south, when seen from Ruku-ruku Bay, it
has an elongated and a much more irregular profile, descending rapidly
on the east side, but displaying a gradual and a fairly regular slope of
about 10 degrees on the west side. The upper part of the mountain is in
the form of a curve with the concavity facing south, the crest being
more or less broken up into five or six peaks showing often precipitous
and at times vertical rocky faces having a drop of from 100 to 300 feet,
the highest peaks ranging from 1500 to 1658 feet above the sea.

All around the mountain, except on the upper steep portion on the south
side where it is well-wooded, the slopes have the usual character of the
“talasinga” districts, being occupied only by grass, ferns, cycads, and
the ordinary scanty vegetation of such regions. Whilst on most sides the
surface configuration is fairly regular and the ascent to the summit is
more or less regular, on the south side bold spurs with valleys between
them descend to the coast, and the central mass rises abruptly in the
middle of the peninsula from a height between 300 and 500 feet above the
sea. It is on this side that Naivaka has the appearance of having been
originally a crateral mountain, of which, however, only the north
segment in a much degraded condition now remains, whilst the other
two-thirds have disappeared.

The prevailing rocks are a blackish compact olivine-basalt, having as a
rule much smoky glass in the ground-mass and possessing a specific
gravity of 2·92-2·94. They are referred to in the description of genus
25 of the olivine-basalts given on page 259.

These rocks compose the agglomerate and the agglomerate-tuffs that form
the eastern portion of the summit and probably most of the elevated part
of the mountain. Similar agglomerates occur along most of the north
coast, the rock being in a few places scoriaceous or amygdaloidal; and
they occur in huge fallen masses on the south side near the foot of the
precipitous portion. The blocks in the agglomerate of the summit are
usually six to eight inches across.

On the south-west side the massive rocks exposed are less basic with a
specific gravity of 2·76 to 2·79. They are also more altered, the
olivine being infrequent and the interstitial glass scanty. They differ
besides in the parallel arrangement and in the length of the
felspar-lathes (·18 mm.), which are on the average half as long again as
those of the prevailing olivine basalts (·12 mm.). They are placed in a
different order of these rocks and belong to genus 37 described on page
262.

Tuffs did not come frequently under my notice. At one part of the north
coast the cliffs are formed of a palagonitic tuff-sandstone,
effervescing with an acid, which is described on page 330. Although no
organic remains are to be noticed, it is probably a submarine deposit.

On a spur on the south-west side, at an elevation of 600 feet, there is
exposed a hard red palagonitic tuff dipping away from the summit at an
angle of 40°. It is mainly composed of the palagonitised _débris_ of a
vacuolar basic glass and incloses broken and entire crystals of
plagioclase, augite, and olivine.

The augite crystals, which attain a length of five or six mm., project
from the weathered surface and are easily detached, lying about in
quantities on the ground in places. Although they are now imbedded in
evidently a submarine tuff, these pyroxene crystals could only have been
ejected as such from a subaerial vent; and it would therefore appear
that they fell into the sea around the shores of a volcanic island in a
state of activity. These crystals are often cracked and are as a rule
not so perfect as those I have gathered from the slopes of Vesuvius,
Stromboli, and Etna. They exhibit an unusual tabular form arising from
the great development of the clinopinakoid at the expense of the
orthopinakoid faces.

On the whole it may be inferred that the Naivaka volcano was submerged
at the time of its origin, but that the eruptions continued after it
began to show itself above the sea. In many of its features, especially
in the character of the agglomerate that forms its upper portion, and in
the palagonitic nature of the tuffs, Naivaka differs only from other
elevated districts of the island, where organic remains occur, in the
absence of such remains. Its form bears testimony to the extreme
degradation we find in other districts, and the occurrence of
foraminiferous tuffs high up the neighbouring slopes of Mount Sesaleka
affords additional evidence of the original submergence of this
district.

THE HILL OF KOROLEVU.[33]—About three miles east of Mount Naivaka there
rises to a height of 800 feet, about a mile inland from the shores of
Wailea Bay, the singular flat-topped hill of Korolevu. It displays
vertical cliff-faces, with a drop often of 200 or 300 feet, which have
become so deeply furrowed or fluted by the eroding atmospheric agencies
that they appear at a distance to be made of columnar basalt. The hill
is, however, formed in mass of a compacted tuff or agglomerate tuff
built up of materials of a hyalomelan basic glass that has undergone
partial conversion into palagonite. In the upper thirds these rocks show
no bedding, but in the lower slopes on the seaward side they are bedded
and dip to the north away from the summit at an angle of 15° or 20°. The
form of this hill is well shown in the sketch attached, and there is
little doubt that we have here an old volcanic “neck,” the remains of a
submarine vent.

A specimen of the tuff from the summit is made up of compacted
fragments, in size ranging up to one third of an inch, of a bottle-green
vacuolar glass, which fuses readily in a lamp-flame and is not dissolved
by hydrochloric acid. This glass is usually isotropic, but much of it is
also palagonitic and feebly refractive, the vacuoles or steam-holes,
which are often elongated, being in the last case filled with the same
palagonitic material. Plagioclase crystals occur macroscopically in the
glass; they are much eroded and contain numerous large inclusions both
of the clear isotropic glass and of its palagonitised form.

[Illustration: Korolevu Hill (800 ft.) from Wailea Bay.]

About a third of a mile west of the Korolevu hill rises the hill of
Ngangaturuturu, 450 feet high, which presents a precipitous cliff-faced
summit in which are exposed basic tuffs showing pyroxene crystals
projecting from the weathered surface.

THE BOMB FORMATION OF NAVINGIRI.—A mile north-west of Korolevu Hill,
where the coast road crosses a spur at the back of Navingiri, a very
curious formation is exposed at an elevation somewhat under 200 feet
above the sea. Here there are to appearance a number of large more or
less spherical volcanic bombs, two to three feet across and formed of a
semi-vitreous scoriaceous basalt, imbedded in a hyalomelan-tuff
displaying the same microscopical characters as in the case of the tuff
forming the adjacent hill of Korolevu.

The ash is light grey in colour and rather friable; but where in contact
with the bombs it becomes darker and is hardened. The steam pores of the
bombs are round and not elongated; and as is usual with these bodies
they increase in size from the outside, where they are very small (1
millimetre and less), to the centre, where they vary from two to five
millimetres across. A vitreous border, about an inch in breadth, forms
the outer shell of the bomb where it is in contact with the tuff. Some
of the bombs are only two or three inches apart; and one of them shows
evidence of fracture, fragments of the outer vitreous shell lying
imbedded in disorder in the surrounding tuff.

Before entering into more detail it may be at once observed that the
contiguity of some of the bombs to each other makes it at first
difficult to view them as having been formed in the manner volcanic
bombs are supposed to originate. Those who have seen the huge bombs
lying scattered about on the summit of Vulcano in the Lipari Islands
will appreciate the difficulty of imagining how these bombs can occur in
such a close arrangement without having often shattered each other to
fragments. However, Mr. Wittstock of Mbaulailai in a letter to me
describes even larger bombs that came under his notice exposed on the
surface in the Mbua district, their outer crust when broken looking
“like the slag of a blast-furnace.”

The bomb-rock is a semi-vitreous basaltic andesite. It displays
microporphyritic plagioclase in a ground-mass formed mainly of a smoky,
almost isotropic glass, in which numbers of felspar microliths (·1 mm.)
are developed, the augite being but slightly differentiated. Scattered
about in the glass are little irregular patches, or “lakelets,” of
residual magma composed of a yellowish feebly refractive material that I
cannot distinguish from palagonite.

The ash, in which the bombs are imbedded, is a somewhat friable
hyalomelan-tuff composed of fragments of basic glass often partially
palagonitised, and usually 2 or 3 mm. in size. In it occur pumiceous
lapilli of the same material up to 2 centimetres in diameter. The glass
is markedly vacuolar, the cavities being either filled with gas or with
alteration-products. The vacuoles are often drawn out into tubes, giving
the glass a fibrillar appearance. The numerous plagioclase phenocrysts
inclosed in the glass are much honeycombed and contain large inclosures
of the glass, both altered and unchanged.

Although the line of contact is well defined in a hand-specimen, the two
rocks cannot be separated along the junction. In a thin section, in
which the union of the vitreous shell of the bomb with the ash is well
shown, there is no defined line of demarcation, the non-vacuolar
isotropic glass of the bomb being there broken up into fragments, with
the interspaces filled with the partially palagonitised pumiceous ash.
In the vitreous shell the felspar microliths are much less developed
both in size and number than in the central portion of the bomb.
Numerous cracks communicating with the round steam-pores, which are much
larger than the vacuoles of the ash-glass, are filled with the same
yellowish magma-exudation referred to in the case of the rock forming
the centre of the bomb. Through the cracks this palagonite-material has
found its way into the steam-pores.

It would appear from the above that the bombs were but partially
consolidated when they fell into the bed of ash. The tuff is somewhat
“baked” where it is in contact with the bombs; and there is evidence of
a collision between the bombs in the fragments of the vitreous shell
imbedded in the ash. Although the ash itself contains no organic
remains, there occur, not many hundred yards away and at an elevation
100 feet higher above the sea, foraminiferous tuffs of basic glass which
are described below. There is no indication of a crateral cavity in this
locality; whilst the ancient “neck” represented by Korolevu Hill is a
mile away. These bombs most probably after being ejected from some
sub-aerial vent fell into the sea around, on the floor of which much
basic pumice-ash had been previously deposited. Such masses as they sank
would lose most of their original momentum.

REMARKABLE SECTION NEAR KOROLEVU HILL.—Between the hills of Korolevu and
Nganga-turuturu, at an elevation of about 300 feet above the sea, there
is a singular exposure of tuffs horizontally stratified and forming a
low escarpment or line of cliff about 15 feet high on the hill-side.
These beds display the passage from basic tuffs below to relatively acid
tuffs above, and they establish that in this locality the period of acid
andesites followed that marked by the eruption of basalts and basaltic
andesites. From their horizontal and undisturbed position, it may be
inferred that these deposits began to be formed under the sea when the
activity of the submarine basic vents was on the wane. In their
composition and in the various degrees of coarseness of their materials,
we can plainly discern the history of volcanic action in this locality.

A hard compacted palagonite-tuff makes up the lower half of the
thickness of beds exposed, 15 feet in all. The greater portion of it has
the uniform texture of a sedimentary rock, fine-grained below where the
fragments are ·1 to ·3 mm. in size, and becoming coarser above where the
larger measure 1 to 2 mm. It is composed of more or less angular
fragments of a basic vacuolar isotropic glass, and of plagioclase and
augite with much fine palagonitic _débris_. There is no effervescence
with an acid; but in the upper part there are a few casts of
foraminifera of the “globigerina” type, as indicated in the thin
sections. Above this lies a bed of a similar basic tuff, having however
a banded appearance from the arrangement of materials of different
degrees of coarseness, the finer being ·1-·2 mm. in size, the coarser
·4-·8 mm. There is little or no carbonate of lime; but occasional tests
of foraminifera of the type above mentioned occur in the slide. The
basic tuffs here abruptly terminate. They represent the quiet deposition
in water comparatively deep of the products of marine erosion, and of
the finer ejectamenta of some distant subaerial vent.

Above the basic tuffs lie a series of tuffs, about 5 feet in thickness,
and composed mainly of the debris of acid andesitic rocks of the
hornblende-andesite type, such as occur in the Ndrandramea district.
They mark a period of active eruption on the part of some neighbouring
acid andesitic vent in this neighbourhood, which the subsequent explorer
may be able to identify with some volcanic “neck.”

These tuffs are composed partly of fragments of a hemicrystalline
hornblende-andesite and partly of crystals, broken and entire, of
plagioclase, hornblende, rhombic pyroxene, and augite. The plagioclase
is tabular, zoned, and glassy, and gives extinctions of
oligoclase-andesine (6 to 12°). The hornblende is bottle green, markedly
pleochroic, and gives extinctions up to 14°. The rhombic pyroxene has
the characters described on page 301, in the case of the Ndrandramea
rocks. The augite is less frequent, but the two pyroxenes are sometimes
associated as intergrowths.

These acid tuffs do not effervesce with an acid, nor can any tests of
foraminifera be observed in them; but since these organisms are
represented in the basic tuffs below, it is highly probable that the
whole series of these horizontal beds is submarine. The first or lowest
bed of the acid tuffs indicates a somewhat violent volcanic outbreak in
this neighbourhood, following the deposition of the basic tuffs. It is
composed of loosely compacted subangular fragments, 1 to 3 millimetres
in size, in which the macroscopic prisms of the rhombic pyroxene are
especially frequent. It passes upward without interruption into a
regularly grained sandstone formed of rounded and subangular fragments
measuring ·3 to ·7 mm. across. Above this lies a quite distinct bed, a
few inches thick, of a fine compact clay rock, where the mineral
fragments measure only ·05 to ·12 mm. in diameter, hornblende being well
represented, although the rhombic pyroxene is very scanty. Up to this
time these beds of acid tuffs indicate a gradual defervescence of the
volcanic activity that began with some violence, as shown by the
characters of the lowest bed. Now another outbreak occurred, and
overlying the clay-like bed we find a coarse tuff made up of fragments 2
to 5 millimetres across, and approaching in texture and appearance a
subaerial tuff, but in other respects similar to those below it. It is
the last and uppermost of this series of acid tuffs, and with it
terminates an interesting record of the past in this region, the chief
features of which may thus be summarised.

A prolonged period of quiet deposition of submarine basic tuffs, the
products partly of marine erosion and partly of distant eruptions, was
abruptly followed by the outbreak of a neighbouring vent during which
tuffs formed of the debris of acid andesites were deposited. The gradual
decrease in the degree of activity is plainly shown in the gradual
diminution in size of these tuffs, until they acquire the fineness of a
clay. Then another burst of activity from the same vent or vents
occurred, and the record ends. Since that time there has been apparently
an upheaval to an elevation of 300 feet above the sea. As, however, the
beds are quite undisturbed, the emergence may have been due to the
lowering of the sea-level, a subject which is discussed in Chapter
XXVII.

COAST BETWEEN WAILEA BAY AND LEKUTU.—The hills here often approach the
coast, their spurs running down to the beach. In the low range, 250 to
300 feet high, east of Wailea Bay, are exposed palagonite-tuffs dipping
gently north-east and composed of fragments of a vacuolar basic glass,
more or less palagonitised, and of minerals (plagioclase, etc.) not
exceeding 2 mm. in size. These deposits are apparently non-calcareous
and show no organic remains.

Farther along the coast towards Nativi basic tuffs and agglomerates
appear at the surface; but the underlying rock, exposed in position in
the stream-courses and prevailing along much of the sea-border to Nativi
and a mile or so beyond, is a vesicular semi-ophitic basaltic andesite
with coarse doleritic texture and containing much interstitial smoky
glass. (It belongs to the non-porphyritic group of genus 9 of the
augite-andesites described on page 273.) Such rocks evidently represent
ancient flows. They give place as one proceeds east to porphyritic
semi-ophitic doleritic rocks of the same genus and to semi-vitreous
basic rocks. About half a mile west of Nukunase a vesicular doleritic
basaltic andesite forms a spur protruding at the coast. It is
semi-ophitic and contains in the smoky glass of the groundmass little
irregular cavities filled with a yellowish residual magma like
palagonite in character. (It is referable to genus 12 of the
augite-andesites, described on page 275.) A few paces west of this spur
a vertical dyke, 20 feet wide and trending N.W. and S.E., appears on the
beach. It is formed of a bluish scoriaceous basaltic andesite containing
much glass in the groundmass and showing imperfectly developed felspar
lathes. It is included in genus 4 of the augite-andesites described on
page 270.

A little east of the spur there is another dyke apparently vertical and
formed of a vesicular rather than a scoriaceous basaltic andesite
referred to genus 1 of the augite-andesites (page 267). It differs from
the rock of the previous dyke in the presence of small plagioclase
phenocrysts which contain abundant magma-inclusions; but it resembles it
in the characters of the groundmass. This dyke is about 40 feet in
thickness and trends N.E. and S.W.

It may be inferred from the foregoing remarks that there was at one time
a volcanic vent in the district west of Nukunase. The lines representing
the trend of the two dykes above noticed would if extended meet at a
common focus a little way inland. The rocks of the dykes differ
conspicuously from the prevailing doleritic rocks that form, as before
remarked, the ancient flows, the average length of the felspar-lathes in
the former being ·1-·2 mm., in the latter ·3-·4 mm. Both, however,
belong probably to the same vent of which now the exact situation would
not be easy to discover, on account of the re-shaping of the surface
through the denuding agencies.

MOUNT KOROMA.—The highest peak of the hills lying inland between Wailea
Bay and Lekutu is named Koroma and attains a height of 1,384 feet. I did
not ascend its slopes higher than 900 feet, and approached it from the
Mbua or south side. Extensive plains, covered with the usual “talasinga”
vegetation, reach inland from the shores of Mbua Bay to the foot of this
range without attaining a greater elevation than 100 feet. This low
district is drained by the Mbua river and its tributaries, the rock
usually exposed at its surface being a decomposing porphyritic basaltic
andesite. It is again referred to on page 56 in connection with the
low-lying level region of this portion of the island of which it in fact
forms a part.

A basic non-calcareous fine-grained tuff-sandstone is exposed in a
stream at the foot of the south slope of Mount Koroma. Whilst crossing
some low wooded outlying hills in this locality, I came suddenly upon
what seemed like a desert in miniature, quite bare of vegetation and
occupying an area of some acres. Here a porphyritic basic rock, from
some cause unknown to me, has decomposed in the mass to a depth of 20
feet and more; and the result is a surface of white crumbling rock
scored deeply by the rains and carved out by the denuding forces into
miniature hills and dales. It is not improbable that a small crater in
its last solfatara-stage once existed here; but the whitened
disintegrated rocks alone remain, and we can now only hazard a
conjecture as to the cause.

I found a variety of basic rocks exposed on the hill slopes up to 900
feet. The most frequent of the deeper-seated rocks which occurred in
mass at this elevation, and as large blocks on the lower levels, is a
dark grey rather altered hypersthene-augite-andesite, referred to genus
1 of that sub-class as described on page 286. The specific gravity is
2·73, whilst the groundmass displays a little greenish altered glass.
Another of the deeper rocks, exposed 500 feet up the slopes, is placed
in the same sub-class, augite and rhombic pyroxene being porphyritically
developed, separately and as intergrowths. The groundmass displays short
stout felspars, augite, and a little altered glass. The rock is
therefore referred to the orthophyric order described on page 290. Spec.
grav. 2·78.

Evidence of more recent surface lava-flows here exists. In one place I
came upon such a bed 12 feet thick, compact in its upper half and slaggy
or scoriaceous in its lower half. The rock is an aphanitic
augite-andesite (spec. grav. 2·77) and belongs to species B, genus 16,
of the augite-andesites, as described on page 281. Its groundmass
displays felspar-lathes in flow-arrangement with a little interstitial
glass. Slaggy lava is not uncommon on these slopes. One specimen beside
me is a semi-vitreous form of the deeper hypersthene-augite-andesites of
this range.

There appears to be better evidence of sub-aerial lava-flows on the
lower slopes of Mount Koroma than I found in any other part of the
island. It should have been before remarked that one of these flows lies
upon a bed of a hard reddish compact tuff, which appears in the thin
section as an altered palagonite-tuff, containing fragments of minerals
including both rhombic and monoclinic pyroxene, but showing neither lime
nor organic remains. The larger fragments are 2 mm. in size. It seems
likely that this flow ran into the sea during the emergence of this part
of the island.

The prevalence of rocks of the hypersthene-augite-andesite type in Mount
Koroma distinguishes this range from the surrounding regions of
olivine-basalts and basaltic andesites. This district is well worth a
detailed examination, and perhaps the remains of a crateral cavity may
yet be found.

THE COAST BETWEEN NAIVAKA AND KORO-NI-SOLO AT THE FOOT OF THE NORTH
SLOPE OF THE SESALEKA RANGE.—Basaltic andesites, and olivine-basalts of
the Naivaka type occur on this coast. A rock of more acid character,
light grey and much altered, is exposed at the surface where the track
crosses the headland projecting into Ruku-ruku Bay. It is one of the
propylites referred to in my description of the second genus of the
augite-andesites (p. 269). The felspars of the groundmass give the small
extinctions of oligoclase; and in this respect it differs from the other
augite-andesites. Besides the altered plagioclase phenocrysts there is
much microporphyritic augite but slightly changed. Calcitic and other
alteration products occur in the interstitial glass.

MOUNT SESALEKA.—This is the name of the highest peak, 1,370 feet, of a
remarkable ridge-shaped range, which is very precipitous on the east and
north-east sides, where there is a sheer drop apparently of 500 or 600
feet, whilst on the other sides the slope is more gradual, especially on
the north where there is a gentle descent to the sea. The actual summit
is bare, rocky, and narrow. There is a curious native legend relating to
a pond on the top of this hill. From what Mr. Wittstock tells me, it
seems probable that there is a spring near the summit. Close to the top
are the remains of an old “koro-ni-valu” or war-town; whilst numbers of
shells of species of Cardium, Cypræa, and Strombus, such as would be
used for food, lie about. Many years ago there was a prolonged siege of
this stronghold, which is referred to here as indicating that the
defenders had some independent water-supply.

In ascending from Koro-vatu on the west side basic agglomerates and
agglomerate-tuffs were found exposed as far as half-way up. In the upper
half occurred at first fine-grained calcareous tuffs, bedded and dipping
gently down the slope, composed of palagonite-debris, mineral fragments
and calcitic material and displaying a few macroscopic tests of
foraminifera. These tuffs became non-calcareous and coarser as one
approached the summit. A specimen obtained from the top is
coarse-grained, being composed of fragments of basic glass, usually
palagonitised, much augite, a little plagioclase and fresh olivine, but
no tests of foraminifera, the size of the fragments being usually
·5-1·5 mm. Massive rocks were rarely exposed on this side; but half-way
up in a stream course I came upon an exposure of a porphyritic
olivine-basalt containing a fair amount of devitrified interstitial
glass. Its specific gravity is 2·85 and it is referred to genus 25 of
the olivine basalts (page 259). I descended by a gentle slope to the
north, coarse basic tuffs and agglomerates containing amygdaloidal
fragments being displayed on the surface. In a stream at the foot, close
to Koro-ni-solo, were blocks of a heavy compact olivine-basalt with
specific gravity 2·96.

DISTRICT BETWEEN MOUNT SESALEKA, THOMBO-THOMBO POINT, AND VATU-KAROKARO
HILL.—This is a broken country with several abruptly rising lesser
hills. Starting from Koro-vatu and crossing the Thombo-thombo
promontory, I reached the coast of Mbua Bay near Navunievu. Basic tuffs
and agglomerates prevailed on the way, the last containing blocks of a
scoriaceous basaltic lava bearing olivine. The massive rocks exposed
belong in some cases to genus 13 of the olivine-basalts as described on
page 256, being dark grey and having a specific gravity of 2·88, and in
other cases to genus 16, species B, of the augite-andesites when they
are lighter in colour and have a specific gravity of 2·77. In both cases
the interstitial glass is scanty.

I ascended Vatui, one of the numerous small hills of the district. It is
450 feet high and is capped by a bare mass of tuff-agglomerate, 40 to 50
feet high and containing fragments of vesicular basic lava. This mass is
pierced by a dyke, 18 inches thick, which is inclined to the N.NE. at a
high angle of 60 or 65 degrees with the horizon. This dyke is composed
of a compact olivine-basalt which is remarkable for the prevalence of
small augite prisms in the groundmass. It is described on page 265 under
genus 44 of the olivine-basalts. Hand-specimens are magnetic and display
polarity, which is due, as pointed out in Chapter XXVI., to the exposed
situation of the peak.

Vatui in its characters is evidently typical of the other lesser
hills around, which, as viewed from below, possess bare tops and
precipitous declivities of the same formation. All the hills in the
district including Sesaleka are capped by these basic tuffs and
tuff-agglomerates; and doubtless as in the case of Sesaleka these
deposits are all submarine. This is true also of Vatu-karokaro, a
hill 600 feet high, overlooking Mbua Bay and about two miles east of
Sesaleka. In the lower part of this hill is exposed a dark compact
basaltic andesite, referred to genus 13, species B, of the
augite-andesites (sp. gr. 2·83), whilst blocks of a black
olivine-basalt (sp. gr. 2·91) occur in the agglomerate of the
summit. These hills may all be regarded as “volcanic necks” or the
stumps of volcanic cones, probably submarine.

THE DIVIDING RIDGE BETWEEN THE MBUA AND LEKUTU PLAINS.—A level rolling
“talasinga” district intervenes between Mbua Bay and the dividing ridge.
The upper part of this ridge, which attains a height of about 500 feet
above the sea, is composed of a hard grey sandstone-like tuff,
effervescing feebly with an acid, which on examination proves to be
formed in great part of fragments, ·07-·1 mm. in size, of a dark basic
glass occasionally vacuolar. The rest of the deposit consists of
similar-sized fragments of plagioclase and other minerals, and includes
also a few tests of foraminifera of the “Globigerina” type.

The mass of the ridge, however, is composed of coarse tuffs and
agglomerates of a different kind which have been covered over by the
foraminiferous deposit just described. Thus there are exposed on the
lower slopes, tuffs and agglomerates of a basic pitchstone formed of a
brown glass containing a few felspar and pyroxene microliths. In the
tuff the fragments are three to six mm. in size and show evidence of
crushing _in situ_, the interstices being filled with debris of the same
material more or less palagonitised,[34] but there is no carbonate of
lime. Large masses of an agglomerate made up of blocks of an acid
andesite occur higher up the slopes. The component rock belongs to an
unusual type of hypersthene-andesite, specially noticed on page 297.

The interesting feature in this ridge lies in the testimony it affords
that the extensive Mbua and Ndama basaltic plains, on which I was unable
to discover any submarine deposits, were at one time submerged.

THE MBUA AND NDAMA PLAINS.—These rolling plains are a striking feature
in the western end of Vanua Levu. They have an arid barren look, are
clothed with a scanty and peculiar vegetation, possess a dry crumbling
soil often deeply stained by iron oxide, are traversed by rivers without
tributaries descending from the wooded uplands of the interior, and in
fact have well earned the name given to them by the natives of
“talasinga” or sun-burnt land. Both Seemann and Horne have remarked on
the South Australian aspect of these regions, which are characteristic
of the lee and drier sides of the larger islands of the group. Covered
for the most part with grass, ferns and reeds, these low-lying districts
are dotted here and there with Casuarinas, Pandanus trees and Cycads,
whilst such other trees and shrubs as Acacia Richii and Dodonæa viscosa,
add to the variety and peculiarity of the vegetation. The origin of
these “talasinga” districts is discussed in the last chapter.

The Mbua and Ndama plains form a continuous region extending three to
five miles inland to the foot of the great mountain of Seatura, to the
watershed between Mbua and Lekutu, and to the base of Mount Koroma;
whilst it reaches along the sea border from the vicinity of Navunievu
about four miles west of the Mbua River to beyond Seatovo a few miles
south of the Ndama River. Their extent is defined in a general sense by
the 300 feet contour line in the map. Their elevation, however, above
the sea does not generally exceed 200 feet and is usually only 50 or 100
feet; but at the foot of Seatura they rise to between 300 and 400 feet.
Whilst on the one side these plains form a continuation of the lower
slopes of the great Seatura mountain, on the other side they are
extended under the sea as the broad submarine platform, the edge of
which, as defined by the 100-fathom line, lies eight to ten miles off
the coast. It is pointed out on page 372 that this continuity of
surface, both _supra_-marine and submarine, extends probably to the
geological structure and that the submarine platform represents the
extension under the sea of the basaltic flows of the plains.

The whole region of the plains is occupied by olivine-basalts and
basaltic andesites, such as are found on the neighbouring lower slopes
of the Seatura mountain. They are as a rule much decomposed, even at a
depth of several feet below the surface. Typically, they are neither
vesicular nor scoriaceous, and in this respect they possess the
character of submarine lava-flows. The rolling surface of the plain is
varied occasionally by small “rises” or hillocks marking apparently some
secondary cone, of which the much degraded “wreck” alone remains. Here
and there fragments of limonite, approaching hæmatite in its compact
texture, lie in profusion on the soil, representing doubtless small
swamps long since dried up, some of which still occur in the hollows of
the plain. Mingled with these fragments are often pieces of siliceous
rocks and concretions, such as are found in the other “talasinga”
districts of the island, the description of which is given on pages 128,
132, &c.

I will now refer more in detail to some of the points alluded to in this
short description of these plains. With reference first to the compact
limonite, it should be remarked that it occurs on the surface either as
fragments of hollow nodules two or three inches across, or as portions
of flat “cakes” half to one inch thick. It is especially abundant in the
district lying a mile or two on either side of the Navutua stream-course
between Ndama and Mbua. Here the subsoil is charged with ferruginous
matter, and the water of the series of stagnant pools in the bed of the
stream is stained blood-red by iron-oxide, a circumstance that has
naturally given rise to native legends of a corresponding hue. These
fragments of iron ore, which lie between 100 and 150 feet above the sea,
represent the final stage of a process which is now no doubt in
operation on the bottom of the neighbouring pools and small swamps.
Their presence on the surface goes to indicate that this open country
has been for ages a land-surface free from forest, as it is in our own
time.

In a similar manner, the extensive disintegration of the basaltic rocks
that form these plains affords evidence of the great antiquity of these
“talasinga” plains in their present unforested condition. The extent to
which these rocks have weathered downwards is very remarkable. Between
Ndama and Mbua they are decomposed to a depth often of eight or ten feet
below the surface. This is well exhibited in the sides of deep channels
excavated by the torrents during the rains. Here the spheroidal
structure is well brought out in the disintegrating mass, all stages
being displayed in the formation of the boulders that are scattered all
over these plains.

In one locality, near the lower course of the Ndama river, a thickness
of 25 feet of decomposed rock was exposed in a cliff-face. In this case
the rock was a porphyritic basaltic andesite, the disintegrating process
having affected the whole thickness of the large spheroidal masses with
the exception of a hard central nucleus of the size of the fist. In one
of these nuclei by my side it is apparent that during the extension of
the weathering process the phenocrysts of glassy plagioclase become
opaque long before the groundmass is affected. In this specimen the
stage of disintegration as affecting the felspar phenocrysts is at least
one and a half inches in advance of that affecting the groundmass.

This great disintegration of the basaltic rocks, which as pointed out on
page 64 is also in progress on the slopes of the adjacent spurs of Mount
Seatura, is more characteristic of the porphyritic basaltic andesites
than of the olivine-basalts. It is to the spheroidal weathering that we
must look for an explanation of the rounded boulders so frequent in
these districts. It may also be inferred that the soil produced from
this extensive decomposition of the rocks is often very deep. At the
Wesleyan Mission Station at Mbua, on level ground nearly a hundred feet
above the river, a well has been sunk to a depth of 20 feet in soil of
this description; and away to the westward a similar thickness of soil
produced by the same cause is in places to be observed.

Coming to the characters of the basaltic rocks of the Mbua and Ndama
plains, it may be remarked that the prevailing rocks are the porphyritic
basaltic andesites, having a specific gravity of 2·77 to 2·81, which are
in most cases to be referred to genus 13 (porphyritic sub-genus) of the
augite-andesites described on page 278. They possess large phenocrysts
of plagioclase and but little interstitial glass. The other rocks are
olivine-basalts with specific gravity 2·88 to 2·90 and showing only a
few small plagioclase-phenocrysts. They display a little residual glass
and belong for the most part to genus 37 of the olivine basalts
described on page 262. In both these basaltic rocks the felspar-lathes
are in flow-arrangement; but in the basaltic andesites they average
·11 mm. in length, whilst in the olivine-basalts they average ·18 mm.

The low mound-like “rises” in these plains, to which previous reference
has been made, are not usually elevated more than 50 feet above the
general surface. One of these hillocks that lies near the track from
Mbua to Navunievu, about two miles from the Wesleyan Station, is
composed of a remarkable semi-vitreous pyroxene-andesite passing upward
into a rubbly rock of the same nature. The rock of this old volcanic
neck is of an unusual type and is referred to the prismatic order of the
hypersthene-augite andesites described on page 289. Both the felspar and
pyroxene prisms of the groundmass are in flow-arrangement. One of these
mounds near the Mbua Wesleyan Station is apparently formed of the
decomposing basaltic andesite of the district. On its surface are
fragments of earthy limonite and siliceous rocks.

The rarity of submarine tuffs and clays on these plains is somewhat
singular; but in the occurrence of foraminiferous tuffs high up the
slopes of Sesaleka and on the crest of the Mbua-Lekutu dividing ridge we
have evidence of the original submergence of all these lower regions. It
is probable enough that the ages of exposure that have since witnessed
the reduction of the solid basaltic rock to a crumbling mass several
feet in depth were more than sufficient for the stripping off of most of
the overlying submarine deposits. Such deposits are, however, common on
the surface of the extensive “talasinga” plains that constitute much of
the north side of the island.

THE SHELL-BED OF THE MBUA RIVER.—Rather curious evidence of an emergence
of a few feet and of a considerable advance of the delta of the Mbua
river in comparatively recent times is afforded by a bed of marine
shells exposed in the right bank of this river, about 200 yards below
the boat-shed of the Wesleyan Mission Station and about two miles in a
straight line from the sea. This bed, which is about a foot in
thickness, is exposed for a distance of 70 or 80 yards. It slopes
gradually seaward as one descends the river, being raised two or two and
a half feet at its upper end above the river level at low tide, whilst
at its lower end it is at about the water-level. The river-bank is here
15 or 16 feet high, and is composed in its upper half of a fine gravel
of volcanic rocks mixed with earth, which below passes abruptly into a
friable non-calcareous black mud-rock (not bedded and looking like
consolidated swamp mud), in which the layer of shells is contained.
These shells are, therefore, covered by deposits, 13 or 14 feet in
thickness, of which the upper eight feet are formed of gravel and earth,
and the rest of mud-rock. They are evidently gathered together on the
slope of an old mud-flat.

The shells are all large marine bivalves, belonging to the genera
Ostræa, Meleagrina, Cardium, Arca, &c., no freshwater shells occurring.
They are often much decayed and have lost the ligaments. The valves are
generally separate; but in some cases they are still in apposition, the
cavity being then filled with the same black mud in which the shells are
embedded. They lie about in all positions, some vertical, some
horizontal, and are often piled on each other. In some cases large
borers have perforated one or both of the valves; and here and there
valves may be noticed with smaller oyster-shells attached to the inner
surface. No vegetable remains were discovered with the exception of a
single “stone” of the fruit of the Sea tree,[35] which is common in
these islands, its empty almost indestructible stones occurring
frequently in the drift stranded at the mouths of rivers.

At first sight one would look to human agency for the explanation of
this shell-bed; but many of its features are inconsistent with such a
view. If the shells had been originally collected by the aborigines for
food, the absence of those of marine univalves of the genera Turbo,
Strombus, Cypræa, &c., such as are much appreciated as food by natives,
is inexplicable. The extent of the bed and its uniform thickness are
characters that give no support to such an explanation. It represents,
as I apprehend, an ancient shell-bank formed on a muddy bottom in
comparatively shallow water near the mouth of a river. Since that time
the Mbua River has cut through its old deposits, and the margin of its
delta is now two miles to seaward, the intervening new land being formed
of extensive mangrove-swamps in its lower part, whilst nearer the
shell-bed there is much level land raised a few feet above the sea, on
which the native town and different villages now stand. The amount of
emergence here indicated since the time when this bank of shells was
forming under the sea does not probably exceed a couple of fathoms.

LEKUMBI POINT.—This singular long and low promontory is between three
and three-and-a-half miles in length and rather less than a mile in
average width. It is monopolised by mangroves, except at the extremity
where the swampy ground passes into the dry sandy soil occupied by the
characteristic vegetation of coral beaches. This terminal portion, which
is about a third of a mile in length and raised a couple of feet above
high-water mark, was originally a reef-islet. The outer third of the
cape, however, is cut off from the remainder by a narrow winding passage
in the mangroves, which being 25 or 30 feet wide can be traversed by
boats at and near high-water, and is often used to shorten the journey
down the coast. The flowing tide rushes in at both entrances, and when
the tide is ebbing it finds its way out at both exits, the passage
presenting the readiest way of the filling and emptying of the interior
swamps with the flow and ebb of the tide.

Before explaining the origin of this low tongue-shaped promontory of
Lekumbi, it should be observed that it lies on a long projecting patch
of coral reef which is continuous with the neighbouring shore-reefs.
Depths of seven and eight fathoms are found off the sides and of 11 and
12 fathoms off the end of the reef-patch. This reef in its turn must
have been built up on a submarine bank protruding from the coast. Such a
bank may have originally been produced by the deposits brought down by
the Ndama River which finds an exit through the mangroves near the base
of the cape. With the exception, however, of the Lekutu River, none of
the other Vanua Levu rivers have given rise to such tongues of land at
their mouths. I am more inclined to hold that the submarine shoal, which
underlies the present low cape of Lekumbi, indicates an old lava-flow
from the great crateral valley of Seatura, opposite the mouth of which
it lies. Traces of such flows are still to be found in that locality.



                               CHAPTER V

      DESCRIPTION OF THE GEOLOGICAL AND GENERAL PHYSICAL FEATURES
                             (_continued_)


THE SEATURA MOUNTAIN.—In my description of the profile of this part of
Vanua Levu, reference has already been made (p. 3) to the great mass of
this mountain which occupies five-sixths of the breadth of the island.
Viewed from seaward it looks like a huge table-topped mountain-ridge,
and as such it is represented in the Admiralty charts; but when its true
contours are distinguished it appears, when defined by the 300-feet
level in the map, as a somewhat rounded mass, measuring 12 miles in
length and 10 miles in breadth and attaining a maximum height of 2,812
feet. Seen from the deck of a passing ship it displays more or less
regular volcanic slopes, especially on the east, where there is a
gradual descent at an angle of 3 or 4 degrees for some 10 miles, and on
the north towards the Lekutu lowlands. It also shows a fairly regular
descent towards Mbua Bay on the west. (See profile, p. 62.)

On the west side, however, there is a great gap in the mountain-mass
(the Ndriti Gap), marking, as I hold, an old crateral cavity of large
dimensions, and now occupied by the Ndama River and its tributaries.

The adjacent Seatovo Range to the southward obscures the profile of the
mountain on the south; and it is in fact not at all easy for this reason
to get a view with all the slopes displayed. It is only at times, when
viewed in its complete mass with uninterrupted outlines, as from off the
mouth of the Ndreketi River to the north-east, or when the symmetry of
its long eastern slope is observed from Wainunu Bay that Seatura
displays itself as a gentle-sloped mountain-mass of the Mauna Loa type.
Dense forest clothes the greater part of it, except on the north and
north-west, where it lies within the limits of the scantily vegetated
“talasinga” region.

[Illustration: Profile and Geological Section of the western end of
Vanua Levu from the Wainunu estuary across the summit of the basaltic
mountain of Seatura to the edge of the submarine platform off the Ndama
coast as limited by the 100-fathom line.]

The slopes of this mountain are deeply furrowed by river-valleys which
radiate like the spokes of a wheel from its central elevated mass. Down
its northern slopes flow the Lekutu River and its tributaries and the
principal tributaries of the Sarawanga River. The large western
affluents of the Wainunu River descend from its eastern side, whilst the
Korolevu, Tongalevu, and other small rivers flow south into Wainunu Bay,
and the Ndama River drains its western slopes. In all these cases,
excepting that of the Ndama River, the rivers have worn deep valleys
into the mountain-mass, valleys of denudation that represent the work of
ages. That of the Lekutu is a deep cut almost into the heart of the
mountain; at Nandroro in this valley, which lies 6 to 7 miles inland and
800 feet above the sea, the hills rise steeply on either side of the
river to an elevation of 1,100 and 1,200 feet and more. Some of the
large tributaries of the Sarawanga and the Wainunu flow through
gorge-like valleys 200 to 300 feet in depth. On the western slopes north
of the Ndama river, the mountain-side presents an alternating series of
lofty spurs and deep broad valleys. In fact, all around Seatura its
slopes are deeply furrowed through the denudation and erosion of ages.

The rocks of this ancient volcanic mountain are almost all of the
massive basic type, and except at the mouth of the Ndriti Gap hardly
ever display a scoriaceous character. It is also noteworthy that no
detrital rock, whether agglomerate, tuff, or tuff-clay came under my
observation. The rocks exposed on the surface are mostly blackish brown
olivine-basalts and porphyritic basaltic andesites, the former much
prevailing. In the northern portion, however, grey olivine basalts of a
different type occur. In the great crateral hollow, which I have named
the Ndriti Gap, are displayed numerous dykes formed of highly altered
basaltic rocks that may be classed among the propylites.

The dense forest that clothes the greater part of this mountain offers
many serious hindrances to geological exploration. Except in the
northern portion, views of the surroundings are very limited, and one
has often to rely mainly on the aneroid and the compass to obtain
correct ideas of the contours and general configuration. During most of
the time spent in the southern part of the mountain, my work was greatly
impeded by heavy rains, and from this cause and from the frequent
necessity of following up the stream-courses and of crossing rivers in
flood, I was usually wet through all the day.

_(a) The Eastern Slopes of Seatura._—The basaltic flows, of which this
mountain is principally composed, are best observed on the eastern side
where the original volcanic slopes are preserved. Although the rivers
have worn such deep valleys into the mountain sides, it is however not
often that any great exposure of rock occurs, on account of the dense
forest-growth over much of this region. It is only occasionally that the
columnar structure of these old basaltic flows is displayed. It is
especially well exhibited in the face of a waterfall, distant about two
miles in a straight line from Tembenindio and elevated about 700 feet
above the sea. Here there is an exposure to the extent of 25 feet of
huge basaltic vertical columns, four to five feet across, and pentagonal
in form. The rock is a blackish basalt with scanty olivine and a
specific gravity of 2·87. It is referred to genus 25 of the
olivine-basalts which is described on page 259. Micro-phenocrysts of
plagioclase and a few of augite occur, the olivine being mostly replaced
by pseudomorphs. The felspar-lathes of the groundmass average ·18 mm. in
length, and there is a little brown opaque interstitial glass. Boulders
and fragments of a closely similar basalt, with a specific gravity of
2·9, lie about on the surface in this region. The Seatura slopes here
abut on the plateau of Na Savu, formed largely of volcanic agglomerates,
to be subsequently described.

On the south-eastern slopes of the mountain between Ndawathumi (inland)
and Korolevu (at the coast), somewhat similar basalts with scanty
olivine are exposed (sp. gr. 2·86-2·91). Some of them show the
felspar-lathes of the groundmass arranged in a plexus (genus 25), whilst
others exhibit flow-structure (genus 37), the average length of the
lathes varying in different localities between ·15 and ·21 mm. All
display scanty residual glass. On the shores of Wainunu Bay between the
Wainunu and Korolevu rivers occur porphyritic basaltic andesites with a
considerable amount of glass in the groundmass. There is exposed on the
right side of the mouth of the last-named river a highly basic variety
of olivine-basalt with a specific gravity of 3·07. It is referred to
genus 15 (described on page 258), which includes the most basic rocks in
my collection. There are in this rock no plagioclase phenocrysts and the
felspar-lathes of the groundmass are relatively infrequent, whilst
olivine and augite occur in abundance. There is little or no residual
glass. In the district of Tongalevu blackish olivine-basalts and
basaltic andesites of the usual character are found. In the Na Suva
range, which lies two miles inland from the shores of Nasawana Bay and
forms the southerly extension of the mountain, a somewhat compact
variety of olivine-basalt (sp. gr. 2·92) prevails up to the summit,
1,550 feet above the sea. It is included in genus 37 of the
olivine-basalts. In the length of the felspar-lathes (·15 mm.) it
belongs to the Seatura type of these dark basalts.

_(b) The Western Slopes of Seatura._—Here overlooking the plains north
of the Ndama River the same olivine-basalts and porphyritic basaltic
andesites occur. The vegetation is of the scanty “talasinga” character,
and since there is little or no soil-cap the disintegration of the rocks
has been very great, often extending to a depth of 10 or 12 feet. It is
remarkable that this disintegration is most marked in the “talasinga”
and similar scantily wooded districts of the mountain. On the densely
wooded eastern and southern sides where there is a thick soil-cap, it is
by no means so evident. Here on the western slopes have been carved out
deep broad valleys and lofty spurs, the last in their turn furrowed on
their flanks, without any apparent sufficient cause. The shallow streams
at the bottom of the valleys appear quite incompetent to produce such
great erosion; and doubtless these results are partly due to the action
on the crumbling rock-surface of temporary torrents formed during the
rains.

_(c) The Northern Slopes of Seatura._—Here within the scantily vegetated
“talasinga” region the conformation of the land is well displayed.
Broad, deep and nearly parallel valleys, separated by level-topped spurs
and occupied by the Lekutu and its tributaries, score the mountain’s
slopes. The prevailing rocks are blackish-brown olivine-basalts and
porphyritic basaltic andesites, such as occur around the other parts of
Seatura; but grey olivine-basalts also occur, possessing opaque
plagioclase-phenocrysts and looking like porphyrites. They are
essentially holocrystalline and are probably more deeply situated than
the other basaltic rocks. They are referred to genera 26 and 38
described on pages 261, 263, and have a specific gravity of 2·75-2·83.
Dark doleritic basalts distinct from all the others are exposed in
places.

A good idea of this region may be obtained by following the road
westward from Tavua on the head-waters of the Sarawanga River to Wailevu
on the westernmost tributary of the Lekutu River, a distance of about 6
miles. Leaving Tavua one at once begins to ascend and cross the long
spur that descends from Seatura and divides the valleys of these two
river-systems. On its slopes are exposed much decomposed blackish
basalts possessing scanty olivine and showing large porphyritic crystals
of plagioclase. They have a specific gravity of 2·84 and are assigned to
the porphyritic sub-genus of genus 25 (page 259). At the summit, 800
feet above the sea, occur blocks of a grey holocrystalline basalt with
scanty olivine and semi-opaque plagioclase-phenocrysts referred to genus
26 and having a specific gravity of 2·76. It appears to form the axis of
the spur. Descending to the main Lekutu River, just below Kavula, where
the elevation is about 300 feet above the sea, one observes exposed in
mass in the river-bed a dark semi-ophitic doleritic basalt similar to
the doleritic rocks without olivine prevailing on the coast between
Wailea Bay and Lekutu (see page 50), but differing in the absence of
felspar-phenocrysts. It displays a considerable amount of opaque
interstitial glass and is assigned to genus 12 of the augite-andesites
(page 275). The specific gravity is 2·78, but there are a few minute
irregular cavities in its substance.

On leaving Kavula one crosses another of the Seatura spurs at a level of
650 feet, descending then into a smaller river-valley occupied by a
tributary of the Lekutu, on the banks of which lies the village of
Nawai, 350 feet above the sea. Then another spur is crossed at an
elevation of 450 feet and the descent is made into the valley of the
Wailevu tributary of the Lekutu. Crossing the valley, which at the town
of Wailevu is elevated 300 feet, one rises to a height of 700 feet and
then descends into the Mbua plains. These three almost parallel valleys
of the Lekutu and its two tributaries are worthy of a detailed
examination.

The rocks on the surface between Kavula and Wailevu vary in character.
Nearer Kavula there appears a blackish compact olivine-basalt (spec.
grav. 2·88), showing a little microporphyritic plagioclase and
belonging to genus 37 of the olivine rocks. Further on is exposed one
of the holocrystalline grey olivine-basalts with porphyritic
plagioclase-phenocrysts and specific gravity 2·83. It belongs to the
type described in genus 38 of the rocks on page 263. Nearer Wailevu
there occurs a blackish porphyritic basalt with scanty olivine and
specific gravity 2·81. It contains but little residual glass and is
referred to the porphyritic sub-genus of genus 25. In some cliffs at
the river-side close to Wailevu, there is displayed a semi-vitreous
basaltic andesite, showing large porphyritic plagioclase crystals, 3
to 8 mm. Its low specific gravity (2·68) is to be attributed to the
large amount of glass in the groundmass. There is a loose mesh-work of
felspar-lathes, but the augite is not differentiated. Westward of
Wailevu commence the decomposing basaltic rocks of the Mbua plains.

_(d) Traverse of the Northern Part of the Summit of Seatura from Kavula
South-West to Narawai._—The track first lay up the picturesque valley of
the Lekutu River to Nandroro, 2½ miles distant and 800 feet above the
sea. On the way blackish basaltic rocks of the prevailing Seatura type,
with or without scanty olivine, were displayed often in a decomposing
condition. At one place a characteristic grey olivine-basalt, showing
opaque porphyritic plagioclase (sp. gr. 2·87), and looking like a
porphyrite, was exposed. On account of the abundance of the olivine, it
is placed in genus 2 of the olivine-rocks. After Nandroro the path lay
up the steep mountain-side to a height of 1,500 feet: and afterwards
across the summit of the northern part of Seatura, which is here about
two miles in breadth. This elevated region is well wooded with here and
there a patch of “talasinga” land; but it is by no means level, its
elevation varying between 1,400 and 1,800 feet, and it soon became
evident that we were crossing the heads of valleys, sometimes 200 or 300
feet in depth, that could only have been excavated by the torrential
rains. These streamless valleys afford another indication of the
denudation to which this ancient mountain has been subjected.

The rocks prevailing in this elevated northern portion of Seatura, at
heights of 1,500 to 1,800 feet above the sea, are: (_a_) blackish
basalts with scanty olivine, a little interstitial glass, and belonging
to the porphyritic and non-porphyritic sub-genera of genus 25 of the
olivine-rocks: (_b_) grey olivine-basalts with porphyritic opaque
plagioclase, containing but little residual glass, but varying greatly
in the amount of olivine and belonging to the genera 2 and 26 of the
olivine-basalts; they would be classed, as far as appearance goes, as
porphyrites; their specific gravity ranges 2·85 to 2·90. The rock
exposures were, however, scanty; and but little information could be
obtained of the mode of occurrence. No scoriaceous rocks were found
except in the instance of a compact dark basalt without plagioclase
phenocrysts, apparently a dyke rock, and belonging to genus 40 of the
olivine-basalts.

_(e) Ascent to the Summit of Seatura from Ndriti._—The town of Ndriti
lies in the great gap in the south-west side of the mountain which has
been previously mentioned as probably an old crateral cavity. After
traversing a district of highly altered basic rocks or propylites, to be
subsequently described, and reaching an elevation of about 400 feet
above the sea, I came to the long slope that leads up to the summit. A
dense forest hid everything from view, so that the compass and aneroid
had alone to be relied on.

At first one traversed a series of step-like alternations of level
ground and steep “rises,” until the old site of the village of Seatura,
about 1,200 feet above the sea, was reached. There are some strange
legends connected with this old mountain-village, which is now only
indicated by little piles of stones and the debris of a wall, and was
evidently abandoned long ago. We finally reached the summit by following
up a spur or ridge in a northerly direction from Seatura. There was a
precipitous descent on either side of the ridge with evidently a broad,
deep valley to the eastward. The summit was rounded; but on account of
the forest no view could be obtained. There was never any extensive
exposure of rock noticed during the ascent; but all the way up
occasional small blocks of a blackish olivine-basalt were observed on
the surface, of the same general type as that found all around the
mountain and referred to genus 37 in the synopsis.

_(f) The Ndriti Basin or Gap._—This great hollow in the side of Seatura,
which I have named after the town in its midst, is apparently a crateral
cavity now drained by the Ndama river, and its tributaries, and covered
with dense forest to such a degree that a general view of the whole is
impracticable. The glimpses, however, that one obtains of the mountain
scenery are very grand, the town of Ndriti lying in the midst of
mountains that rise almost on all sides of it except on the west. This
great cavity is contracted at its mouth a little below the town and
expands in its interior, where it must be two or three miles in width.
Its floor is fairly level and is elevated only about 200 feet above the
sea;[36] whilst its mountainous sides rise to 2,000 feet and over.

As shown in the map there are two breaks in the outline of this ancient
crater, the one on the west through which the Ndama river flows, the
other on the south where the dividing ridge, separating it from the
Nandi Valley is under 700 feet in elevation. The Nandi Gorge, as I will
term the last-named, is a narrow picturesque ravine leading through the
mountains from Nandi to Ndriti. One follows up a rocky stream-course
hemmed in by precipitous sides until the top of the gorge is reached,
when the watershed is crossed, and the descent is then made to Ndriti by
one of the tributary stream-courses of the Ndama river.

Two or three large rapid streams, after draining its mountainous slopes,
unite within the basin to form the Ndama river, which, as it issues from
its mouth, becomes a comparatively placid stream rolling sluggishly
along to the sea, some five or six miles away, with an average drop of
about thirty feet in a mile. In the course of ages the original
configuration of this great hollow has doubtless been extensively
modified by the denuding agencies. The rainfall on the mountain-slopes
must be very great, probably not under 250 inches in the year[37]; and
Ndriti, though only 200 feet above the sea, is in all probability on
account of its situation one of the wettest places in the island. The
rivers have evidently been important factors in reshaping the original
cavity.

Nearly all the rocks exposed _in situ_ in the beds of the rivers and
streams in the floor of the great Ndriti basin, and for 300 or 400 feet
up its sides are more or less highly altered basic rocks, to which the
old and the new names of greenstone and propylite may be fitly applied.
They often sparkle with pyrites, and not uncommonly effervesce with an
acid, so that one is apt to imagine one’s self in a region of limestone.
The degree of alteration varies considerably, those most altered being
light-coloured and greenish, whilst the others are darker, the specific
gravity ranging from 2·69 to 2·79. In spite of these differences almost
all of them appear to belong to the same eruptive series, being as a
rule sharply distinguished from the prevailing unaltered surface
basaltic rocks of the slopes of Seatura by the size of the felspars of
the groundmass, which average about ·3 mm. in length, whilst those of
the basaltic rocks just alluded to average only ·17 or ·18 mm. long.
These rocks are also well displayed in the sides of the Nandi Gorge; and
from their mode of exposure by river-erosion, as well as from their
relatively coarse crystalline texture, and from their alteration, it may
be inferred that they are older and more deeply situated than any of the
Seatura rocks before referred to. Whether these rocks, which extend over
an area of some square miles, have been altered by solfataric action or
contact-metamorphism,[38] I will not now say. The fact remains, however,
that they are best exposed wherever the streams have worn deeply into
the floor, and lower slopes of the great basin, or have cut down into
the mountain-mass as in the case of the Nandi Gorge. The rocks that lie
in loose blocks on the surface either at the bottom of the basin or on
its slopes extending even to the very summit of the mountain (see page
67), are characteristic blackish olivine-basalts of the type prevailing
around the mountain’s slopes. These propylites are most frequently
exposed as dykes in the beds of the rivers at the bottom of the basin.
Such dykes vary from 4 to 6 feet in thickness, and they are very
conspicuous when they stretch across the river’s breadth projecting more
or less above the water. From their frequency it may be inferred that in
many other small exposures, ill suited for displaying the mode of
occurrence of the rock, we have also to deal with dykes. Judging from
four dykes that were particularly examined, they are all vertical or
nearly so, and all run in much the same direction, namely, N.N.W.—S.S.E.
or N.W.—S.E., whether on the north or south side of the great basin. In
one instance, a rudely columnar structure across the thickness of the
dyke was observed. From their exposure in river-beds it was rarely
possible to ascertain much more than is given above. However, in the bed
of a river, a mile above Ndriti, there was an extensive exposure of a
highly altered greenish rock which was crossed by a vertical dyke, 4
feet thick, formed of a dark grey less altered rock. I have referred
these two propylites to two different genera of the augite-andesites,
the dyke-rock to genus 2, and the other to genus 4. In the case of the
dyke the rock is a little vesicular; whilst in the other it is densely
charged with pyrites. Both have been subjected to the same alteration;
but in a different degree; and it would thus seem that solfataric
influences were here in operation before and after the intrusion of the
dyke.

With reference to the characters of the alteration of these rocks of the
Ndriti basin, it may be remarked that where the change is greatest the
felspars of the groundmass are alone recognisable. The plagioclase
phenocrysts are quite disguised by alteration products, and chlorite,
viridite, epidote, calcite, pyrites, &c., occupy much of the groundmass.
Other rocks are less affected and in a few the change is only slight.

With regard to the prevailing types of the propylites of the Ndriti
Basin, it has already been observed that in most of them the
felspar-lathes of the groundmass are unusually large, the average
length being ·3 mm. From the rare occurrence of olivine in some of
the rocks that are but slightly changed, it is to be inferred that
most of them belong to the augite-andesites, and might be termed
doleritic basaltic andesites. But in other respects they differ
considerably, both as regards the presence or absence of
flow-arrangement of the felspar-lathes, and in the occurrence and
size of the plagioclase-phenocrysts, some having large porphyritic
crystals, others small phenocrysts, and others none at all. Many of
them contained a little interstitial glass. In my classification of
the augite-andesites they are assigned to genera 2, 4, 16, &c., and
additional particulars concerning their characters are given in the
description of those genera. Judging from the average large size of
the felspar-lathes it may be held that, although in other features
they often differ, some of the general conditions under which they
were produced were the same.

On the right bank of the Ndama river, opposite Ndriti, there is a
singular association of a vertical dyke of a bluish-grey basic andesite
with a reddish scoriaceous lava, apparently a flow. The dyke is about 4
feet thick and runs N.W. and S.E., like the other dykes of the basin,
exhibiting also a rudely columnar structure across its breadth. Where
the two rocks are in contact, the dyke has a vitreous border half an
inch thick, and an offshoot of the dyke, four inches wide, has
penetrated the lava, acquiring at the same time a more glassy texture.
The small size of the felspar-lathes of both rocks distinguishes them
from the dyke rocks of the basin, where the felspars are twice as long.
Both rocks show some degree of alteration.[39]

In following the valley of the Ndama River from Ndriti to Telana, about
three miles farther down, one traverses a picturesque region. Emerging
from the great basin the river flows through the rolling plains of the
“talasinga” district. Near Ndriti, and occasionally on the way to
Telana, is exposed a scoriaceous grey basaltic rock; and between two and
three miles below Ndriti there is to be observed in the river-bed
evidence of a comparatively recent flow of a highly basic scoriaceous
lava from the ancient crater of the Ndriti basin. The rock, which is
dark and fresh-looking, shows large porphyritic crystals of augite and
olivine but no plagioclase, whilst the groundmass contains a little
brown interstitial glass. Its characters will be found described under
genus 3 of the olivine-basalts (p. 255). Its specific gravity,
notwithstanding its large empty steam-pores, is 2·91. It differs
markedly from the basaltic rocks of the Seatura slopes and the Mbua and
Ndama plains, in the great porphyritic development of augite and
olivine, in the large size of the felspars and augite of the groundmass,
and in its numerous steam-holes. But in the coarseness of its small
felspars it belongs to the same type as the altered or propylitic basic
rocks of the Ndriti basin. It is probably by some such lava flow from
the old Ndriti crater that the submarine bank was formed off the
adjacent coast on which the low Lekumbi promontory has been built up.

In the numerous dykes of the Ndriti basin and in the great alteration
which their rocks have frequently undergone, we have evidence in support
of the view that this is an old crateral cavity, an opinion that is
supported by the indications of lava-flows that have issued, apparently
in later times, from the mouth of the basin. Reference has already been
made to the locality where a dyke-rock and the rock-mass, into which it
has been intruded, are both propylitic; and from this and other facts,
such as the varying degrees of alteration in different parts of the
basin, it is to be inferred that in the last stage of the activity of
this vent its bottom and sides were extensively affected by solfataric
influences. Since that period, the configuration of the crater-basin has
been greatly modified through the denuding agencies.

The absence, or at least the great rarity, of tuffs and agglomerates in
the case of Seatura is remarkable. The mountain has evidently been built
up in the mass by flows of basic lava; and from this source have no
doubt in an important degree been derived the basaltic flows of the
Ndama, Mbua, and Sarawanga plains, great streams of basalt that further
seaward have helped to form the submarine platform extending several
miles from the coast. The submarine tuffs and agglomerates that occur at
various elevations, reaching as high as 1,200 feet above the sea, in the
Sesaleka, Lekutu, Sarawanga, and Ndrandramea districts lying to the
north-west, north, and east, did not come under my notice on the Seatura
slopes. On the other hand, except in the few localities, where
scoriaceous rocks occur, the general type of the basalts is such as we
would expect to find in submarine flows. In no part of the island,
however, is the antiquity of the land-surface so well attested by the
disintegration of the basaltic flows, which extends here to depths of
ten and even twenty feet. This is in favour not only of the sufficiency
of time, but also of the ability of the denuding agencies to strip off
the surface-deposits.

However this may be, it is evident that the mountain of Seatura
possesses a history quite independent of that of the rest of the island.
I have pointed out in Chapter I. that it represents a mountain of the
Tahitian type. In its radiating valleys and in its basaltic character it
much resembles the mountainous island of Tahiti, which Dana describes as
a gently sloping cone of the Hawaiian order that through the erosion of
ages has become a dissected mountain.[40]

THE SEATOVO RANGE.—This remarkably situated mountain-range, which I have
named after a town at the foot of its western slope, extends from the
valley of the Ndama River to Solevu Bay. It attains a maximum height of
about 1,800 feet, and varies between this elevation and 1,500 feet until
in the vicinity of Solevu, where it descends as a mountainous headland
to the coast. Its summit is narrow and ridge-shaped, and although the
whole range is not interrupted by gaps it has a composite origin. At its
north end, where it is cut off from the Seatura Range by the Nandi Gorge
it helps to close in the large Ndriti basin. Towards the south an
offshoot proceeds eastward and shuts in Solevu Bay. But, although
apparently all the rocks are basic, considerable variety prevails, and
there are many puzzling points in the geological structure of this
region.

At the place where this range abuts on the Ndama valley, below Ndriti,
the grey scoriaceous basalt, before referred to, is exposed at its foot.
However, the usual blackish basaltic rocks, often carrying a little
olivine, form in mass the mountainous southern headland that culminates
in Solevu Peak (Ulu-i-matua); and the same rocks prevail in the lower
regions on the west side of the range from Vuia Point to the valley of
the Ndama River. The southern portion will be described in the account
of Solevu Bay; and I will now give the results of my journey across the
summit of the range about half a mile south of the Leading Peak of the
chart.

The eastern slopes are steep and often precipitous, whilst on the
western side there is a more or less gentle descent to the lower levels,
suggestive of a volcanic slope; and it is remarkable that whilst the
rocks exposed on the precipitous eastern side for the lower two-thirds
are sometimes markedly altered, on the western side they are
comparatively unchanged. These facts at once suggest that we have here
the western rim of a large crateral cavity, though the topography of
this district is not sufficiently well shown in the chart to enable one
to define its original limits. This inference is also supported by the
occasional scoriaceous character of the rocks below referred to.

The most frequent rocks in the upper two-thirds of the range are grey
porphyritic olivine-basalts, displaying opaque plagioclase phenocrysts
and more or less hematised olivine, the specific gravity being about
2·9. They approach in characters the grey porphyritic olivine-basalts of
the northern part of Seatura (pages 65, 66); but differ amongst other
features in the greater abundance of the olivine and in exhibiting
flow-structure. They are usually almost holocrystalline, and are
assigned for the most part to genus 14 of the olivine-basalts. They are
extensively exposed in the stream-courses on the west side between 500
and 900 feet; and huge masses of the same rocks, but containing less
olivine and more glass, and displaying much calcite, viridite, and other
alteration products, are found near the base of the eastern slopes. The
semi-vitreous condition of these rocks is represented in the large
masses of a dark very scoriaceous porphyritic lava, possessing quite a
cindery appearance, that occur on the narrow ridge-shaped summit. The
groundmass shows a few scattered felspar microliths; but it is in the
main composed of a dark opaque glass. Small cube-like crystals of
chabazite line some of the cavities.

Other basic rocks are not infrequent and apparently represent dykes.
Thus on the eastern side at 800 feet is exposed a dark-grey semi-ophitic
doleritic rock (sp. gr. 2·77) assigned to genus 12 of the
augite-andesites (page 275). The felspar-lathes average ·3 mm. in
length, and there is a little interstitial glass containing viriditic
and calcitic alteration products, the same materials filling small
rounded vesicular cavities. On the same slope between 1,000 and 1,200
feet, there are displayed fresh-looking compact non-porphyritic basaltic
andesites (sp. gr. 2·84), where the felspar-lathes average ·2 mm. and
the interstitial glass is scanty. They are referred to genus 16, species
C, of the augite-andesites. On this side also between 600 and 800 feet
occur blocks of a highly altered slightly vesicular augite-andesite
showing a little microporphyritic plagioclase. It is assigned to genus
13, species B, of the augite-andesites. In one place where it is in
position it is scoriaceous, the steam-holes being round, empty and one
to five mm. in size. In the less glassy rock it displays numerous small
irregular cavities either filled with fibrous viridite or calcite or
showing concentric zones of the two minerals. The felspar-lathes are
·15-·2 mm. in length. In blocks near the foot of the eastern slope occur
a blackish olivine-basalt (sp. gr. 2·88) of the prevailing Seatura type,
possessing a little interstitial glass and felspar-lathes with an
average length of ·2 mm. It belongs to genus 25 of the olivine-rocks....
On the western slopes at a height of 500 feet occurs a dark compact rock
(sp. gr. 2·89) with abundant olivine which is referred to genus 1 of the
olivine basalts. There is a little residual glass, the felspar-lathes
averaging only ·08 mm. in length. A similar-looking rock is exposed at
1,400 feet, which displays felspar-lathes averaging ·2 mm. long (sp. gr.
2·9). It belongs to genus 37 of the same olivine class. Here also is
assigned an aphanitic basalt, with a few scattered large plagioclase
phenocrysts and felspar-lathes averaging ·15 mm. long, which is
displayed near the base of the slope.

I could not satisfy myself as to the presence of tuffs on the slopes of
this range. Some fine argillaceous rocks exposed half-way up on either
side show no lime and contain no organic remains. One specimen beside me
is certainly a disintegrated basic rock. No agglomerates came under my
notice. In the absence or rarity of detrital rocks this part of the
range resembles the adjacent mountain of Seatura.

Although olivine-basalts prevail in this part of the Seatovo Range there
is great variety in their characters; and it does not appear possible to
explain such a diversity except to assume that we have here an old
crateral ridge which has again and again been penetrated by dykes and
has since been greatly denuded. We have here one of those singular
mountain-ridges that characterise the central portion of the island, but
differing in this respect that the submarine tuffs and agglomerates,
which there occur on the surface, even in the higher levels, are here
absent.

SOLEVU BAY.—There are few localities in the island where so many kinds
of basic rocks are displayed as around Solevu Bay. In addition to the
prevailing blackish porphyritic basalts and basaltic andesites, there
are grey porphyritic basalts, grey non-porphyritic basalts, black
basalts with abundant large crystals of olivine, &c., all of which have
their distinctive characters.

This picturesque bay is surrounded by hills. On the west side it is
inclosed by the promontory forming the southern extension of the Seatovo
range which, culminating in Ulu-i-matua, or the “Head-of-the-Strong”
peak, descends at first steeply and then gradually to the coast, where
it projects as Vulavulandre Point. On the east side is a broken line of
hills, of which Koro-i-rea, the hill known to the natives as the “Town
of the Albinos,” is the most conspicuous. Beyond it stretches the
eastern point of the bay, which the Fijians call “Ua-nguru,” that is,
“the noise of the waves.” On the shores lie the village of Nawaindo,
“the running-stream,” and the once populous town of Solevu, which has
given its name to the bay. Solevu, as its name indicates, is the place
of the “great assembly.” In the background rises the three-peaked
mountain of Koro-tolutolu, “the three towns,” which forms a continuation
inland of the eastern arm of the bay, and joins the Seatovo Range at the
head of it. Between these two ranges inclosing the bay lies the valley
of Solevu, down which descends the Solevu River to the sea. In ascending
this valley from the shore, one rises only about 100 feet above the sea
for the first mile or two.

The promontory, which in the even-topped Ulu-i-matua or Solevu Peak,
attains a height of 1,100 feet above the sea, displays on its summit and
on its eastern slopes descending to the Solevu river, and on its western
slopes reaching down to the coast at Vuia, more or less porphyritic
blackish olivine-basalts of the usual type with specific gravity
2·88-2·90. These basaltic rocks contain scanty olivine and only a little
interstitial glass. The felspars of the groundmass vary in different
localities from ·11 to ·15 mm. in average length. The rocks belong to
genus 37 of the olivine class which is described on page 262.

They are in the lower regions often decomposed to a considerable depth,
the spheroidal structure being well displayed during the weathering
process. Where this promontory terminates in the low Vulavulandre point,
these rocks give place in part to grey porphyritic olivine-basalts, with
specific gravity 2·79-2·83, which from the abundance of the macroscopic
opaque felspar look like porphyrites. They come near to the rocks
exposed on the north slopes of Seatura and in the Seatovo Range. At the
end of the point they become scoriaceous and more vitreous; but with
this exception they contain but little glass. They vary somewhat in
character and are referred to genera 2 and 38 of the olivine-class.

The prevailing rock in the interior of the Ua-nguru promontory to the
south of Koro-i-rea is the blackish porphyritic basalt, containing a
little olivine, and often much decomposed; but at the point and on the
east shores of Solevu Bay, there is a considerable variation in the
character of the basic rocks, of which the two following are the most
conspicuous. Near the village of Nawaindo, there is an apparent
intrusion of a black lava-like basalt of high basicity (specific gravity
3·01) showing abundant large olivine crystals, five or six mm. across,
with some porphyritic augite, but no macroscopic felspar. At the point
the rock is somewhat scoriaceous, with calcite occasionally filling the
cavities, whilst the olivine is so thoroughly hæmatised that it glistens
like brown mica. The compact rock contains a little devitrified
interstitial glass, the felspar-lathes being unusually small, their
average length being only ·07 mm. It belongs to genus 15, the most basic
of the genera of the olivine class represented in the island. The second
rock to be noticed is a slightly altered compact basalt without olivine
forming apparently a dyke near the coast about half way between the
village of Solevu and Ua-nguru Point. It has a specific gravity of 2·84,
the felspar lathes (·15 mm.) presenting a marked flow-arrangement,
whilst there is a fair amount of altered residual glass in irregular
spaces, a millimetre in size. The rock, on account of its
joint-structure, could be easily worked as a building-stone. It is
referred to genus 16, species B, of the augite andesites.

The hill of Koro-i-rea, which rises on the east side of the bay to a
height of 850 feet, has a ridge-shaped summit. Its upper half is
composed of a bluish-grey rock looking like a phonolite and usually
compact, except at the top of the hill, where it is a little
scoriaceous. It has, however, a specific gravity of 2·91 or 2·92, and is
in fact a pretty grey olivine-basalt studded with small olivine crystals
about a millimetre in size and showing no other phenocrysts. This type
of olivine-basalt occurs also at Ulu-i-ndali on the east side of Wainunu
Bay, but is rare in the island. It differs amongst other features from
the porphyritic olivine-basalts of the northern part of Seatura and of
the Seatovo range in the absence of plagioclase phenocrysts. There is
apparently no interstitial glass, whilst the average length of the more
or less parallel felspar-lathes is ·13 mm.[41] On the lower slopes of
the hill the common blackish porphyritic basalt or basaltic andesite is
exposed. In the grey-basaltic upper portion of this hill we have
probably an old volcanic “neck.”

Following the line of hills inland from Koro-i-rea, we cross the
intervening saddle 450 feet above the sea, and ascend the slopes of
Koro-tolutolu, a ridge-shaped mountain backing Solevu Bay, and having,
as its name indicates, three peaks, of which the highest is 1,280 feet
above the sea. My observations indicate that this mountain is formed in
mass of the common blackish-basalts described under genus 37, their
specific gravity being 2·88 to 2·94. But Koro-tolutolu has also the
peculiarity that it appears to be in mass magnetic. The rocks obtained
from its summit, half-way up its western slopes, and near its foot on
the same side, all display polarity, a character also of the rocks of
the neighbouring hills of Ulu-i-matua and Koro-i-rea, but in their cases
seemingly confined to the higher levels.[42]

Neither tuffs nor agglomerates came under my notice at Solevu Bay. This
appears to be an ancient corner of the island, from which denudation has
stripped off nearly everything that could guide us in speculating as to
its past. Although the hills of Koro-i-rea and Koro-tolutolu doubtless
represent old volcanic necks, the relation of Ulu-i-matua to the very
differently composed northern part of the same range, as described on
page 73, is extremely puzzling. Then again in the opposite sides of
Solevu Bay we see exposed the remains of lava-flows that bear no
relation to the present configuration of the surface. We may suspect,
however, that most of the volcanic energy was displayed under the sea.

NANDI BAY.—Lying north of Solevu Bay, this bay is situated between
spurs, descending to the coast from the mountainous interior. The valley
extends a long distance inland without much change of level, the
elevation 1½ miles from the coast being not over 100 feet above the sea.
At its head is the Nandi Gorge, which leads into the Ndriti Basin, the
great crateral cavity of Seatura. There are some remarkable lofty,
isolated hills in this valley that would be well worth examining.

That the bay represents the site of an old volcanic centre is indicated
by the occurrence on the shore of two basaltic dykes, one on either side
of the village of Na Savu and 300 to 400 yards apart. The eastern dyke
is perhaps 30 feet thick, whilst that to the west is scarcely half this
thickness. They exhibit an imperfect columnar structure, the columns,
which are 6 to 12 inches across, being inclined at an angle of 15° or
20° from the vertical in such a way that it may be inferred that the
molten material was ejected from some subterranean focus lying to the
northward (or inland) at an angle of 15° or 20° above the horizon. The
basalt is a compact bluish-black rock with specific gravity 2·95-2·99.
It contains abundant olivine but no other phenocrysts and very scanty
interstitial glass, whilst the felspar-lathes average ·1 mm. in length.
It is referred to genus 16 of the olivine basalts, and is remarkable for
the flow arrangement not only of the felspar-lathes but also of the
smaller olivine crystals.

Blackish basaltic rocks of the prevailing type are exposed on the
surface of the broad spur, not over 500 feet in height, that divides the
Nandi and Nasawana valleys and descends to the coast between the two
bays thus named. They belong to genus 37 of the olivine-basalts and
display a few small plagioclase phenocrysts. The felspar-lathes average
·2 mm. in length, and there is a little interstitial glass. Entering
Nasawana Bay we find ourselves on the southern slopes of Seatura, of
which the high Na Suva range that backs the bay is the southern
extension.

THE TABLE-LAND OF NA SAVU.—This remarkable plateau has an elevation
varying usually between 700 and 800 feet above the sea and a maximum
breadth of four or five miles. It is an area of basic agglomerates and
basic tuffs and lies in the hollow between the basaltic mountain of
Seatura and the acid andesitic hilly region of Ndrandramea. For the
convenience of description I have named it after the picturesque falls
of Na Savu[43] at its southern edge. These falls are celebrated in
Fijian tradition; and from the brink in old time the native desirous of
ending his life leapt into the gorge below.

After flowing sluggishly along on the surface of the table-land, the
Mbutu-mbutu River arrives suddenly at the edge of a line of cliffs of
volcanic agglomerate, that here form the southern border of the plateau,
and with a volume 30 to 40 feet across, it plunges down into the ravine
150 feet below. As shown in the view from the gorge below, there is a
break in the middle of the descent. These falls, however, are not easily
accessible. They are best approached by proceeding from Wainunu to
Ndawathumi and thence up the gorge of the Mbutu-mbutu River.

The surface of the plateau of Na Savu is densely wooded. In places it is
marshy, and here thrives the Giant Sedge (Scirpodendron costatum). The
Makita tree (Parinarium laurinum) also flourishes in the wet districts;
and in the drier localities occur the Ndakua (Dammara vitiensis) and the
Ndamanu (Calophyllum-burmanni) together with a palm of the genus
Veitchia. Here on this level watershed between the basins of the Wainunu
and Sarawanga rivers, the sluggish streams flow aimlessly along in but
slightly eroded channels; and it is not always possible to determine the
side of the island to which they ultimately direct their course. In
their beds are pebbles and irregularly formed concretions of an impure
reddish flint which I have described on page 354. On the north and south
sides the table-land is much excavated by the tributaries of the
Sarawanga and Wainunu rivers. On the west where it meets the foot of the
Seatura slope portions of columns of basaltic rocks appear on the
surface, and deep gorges are worn by the large streams descending from
the mountain. On the east towards Nuku-ni-tambua and Tambu-lotu, the
surface is also much cut up. The preservation of this table-land in a
region, where the denuding agencies are very active in their operations
all around it, is to be attributed to its being a level watershed, where
the head-waters of the Wainunu and Sarawanga rivers in part take their
rise but have little or no eroding power.

It is not easy to obtain a good general view of the district of the
falls on account of the dense forest-growth. When making the traverse
from Tambu-lotu to Ndawa-thumi, it is observed that there is here a
singular hollow, about half a mile in length, which receives the falls
at the western end. The river crosses this hollow and is at once
received into the gorge below, but there is no stream to explain the
origin of the cavity. On its north side the cliffs of agglomerate rise
to a height of 150 to 200 feet from their base, but on the south the
sides are much lower. Here there seem to be the remains of the crater of
the ancient vent from which all the tuffs and agglomerates of the
district were derived. We must look for their origin in the vicinity,
and the only evidence of a crateral cavity is this streamless hollow
extending east from the falls of Na Savu.

With reference to the basic tuffs and agglomerates of this plateau it
may be observed that they cover the massive basic rocks and are probably
not over 100 or 150 feet in maximum thickness. They are well exposed
where the streams cut into the borders of the plateau. The tuffs are
sometimes bedded and slightly inclined, and they may be fine or coarse
grained. They are more or less palagonitised hyalomelane-tuffs, being
composed mainly of fragments of a basic glass, often finely vesicular
and even fibrillar, the vacuoles being filled with different materials,
whilst the palagonitisation is well advanced. Sometimes they have a
brecciated appearance, and in that case when the alteration of the basic
glass is very extensive we find angular fragments, 1 to 2 inches across,
of a greenish palagonite imbedded in a pale matrix of palagonitic
debris, the whole rock having a soapy feel and a steatitic appearance.
This is well shown on the sides of the stream-course at Ndawathumi which
lies at the border of the table-land. These tuffs effervesce but
slightly with an acid.

The basic agglomerate is displayed in the face of the falls and in the
gorges. The blocks are as a rule composed of semi-vitreous basaltic
andesites of varying type, showing no olivine and containing a fair
amount of smoky glass in the groundmass. At times they are scoriaceous
and display amygdules of calcite or a zeolite. In places the rock shows
large phenocrysts of plagioclase and a semi-ophitic groundmass, when it
is referred to the porphyritic group of genus 9 of the augite-class. In
a few of the scoriaceous blocks the augite of the groundmass is for the
most part prismatic and rarely granular (genus 5).

The massive rocks underlying the agglomerates in the vicinity of Na Savu
are aphanitic augite-andesites, differing in important characters from
the rocks of the agglomerates. They probably represent ancient lava
flows of the Na Savu vent. They are compact (sp. gr. 2·72-2·76), and
display a groundmass formed of a felt of felspar-lathes, averaging ·05
or ·06 mm. only in length, and in flow-arrangement. That occurring just
below the falls is almost aphanitic, but is referred to genus 13,
species A, sub-species _a_, of the augite-andesites. The rock from the
gorge below is of the same character, but on account of its opaque
plagioclase phenocrysts it is referred to genus 14, and is described on
p. 279.

In one place on the plateau a tuff-agglomerate is penetrated by veins, a
few inches thick, formed apparently of a finely brecciated tuff of basic
glass fragments in a palagonitic matrix. It is, however, pointed out on
p. 340 that they were originally veins of basaltic glass which have been
subjected to crushing, and that the palagonite has since been produced.

In concluding this description of the table-land of Na Savu, it may be
inferred that the source of its basic tuffs and agglomerates is to be
found in the same locality; and probably the original vent is now
represented by the hollow extending eastward from the falls. With the
exception of a large block of silicified coral found in the vicinity of
Ndawathumi and of the impure flints of the surface of the plateau, which
are described on pages 354, &c., no direct testimony of its submarine
origin offered itself to me. The palagonitic characters of the tuffs
afford, however, indirect evidence in this connection; and indeed the
occurrence of submarine tuffs and limestones in the vicinity of
Tembenindio on its lower northern slopes (see page 131), and the
existence at elevations of several hundred feet above the sea of
fossiliferous tuffs and clays in the Wainunu and Ndrandramea districts
to the eastward, afford strong presumptive evidence that the tuffs and
agglomerates of the table-land were deposited under the sea, and I may
add in a period subsequent to that of the formation of the great
basaltic flows of Seatura and Wainunu.



                               CHAPTER VI

      DESCRIPTION OF THE GEOLOGICAL AND GENERAL PHYSICAL FEATURES
                             (_continued_)


THE BASALTIC PLATEAU OF WAINUNU.—This table-land extends for a distance
of seven miles from the base of the Ndrandramea mountains in the heart
of the island, where it is elevated 1,100 to 1,200 feet above the sea,
to the valley immediately north of the hill of Ulu-i-ndali, where within
a short distance of its termination it still retains a height of 700 to
800 feet. Limited on the west by the valley of the Wainunu River and on
the east by that of the Yanawai River, its breadth varies usually
between four or five miles. It is best seen in profile when viewed from
the south-west on the western shores of Wainunu Bay, between Korolevu
and Nasawana, when it presents itself to the eye as a table-land,
descending with a very gradual slope from the interior towards the
coast. From such a point of view the two great basaltic slopes of
Seatura and Wainunu may be seen together, the former descending eastward
to the Wainunu valley at an angle of 3 or 4 degrees, the latter
descending at right angles to it to the southward with a similar small
gradient of 2 or 3 degrees.

In the profile of the island attached to this work the Seatura slope is
well shown; but that of the Wainunu table-land being seen from the south
is represented only by a level contour-line at the base of the
Ndrandramea mountains. The two great series of basaltic flows, though
closely approaching in a direction at right angles to each other, do not
come into actual contact, and the intervening space is now occupied by
the valley of the Wainunu River. In the accompanying rude outline-sketch
of this region, as seen from off the mouth of the Wainunu estuary, the
relation of this valley to the two great series of basaltic flows is
clearly shown. On the left is the foot of the Seatura basaltic slope; on
the right is the Wainunu basaltic table-land; and between them lie the
estuary and valley of the Wainunu, at the back of which appears the “Na
Savu” table-land, formed of basic tuffs and agglomerates. Behind all
there rise up suddenly the Ndrandramea mountains formed of acid
andesites; whilst in the foreground to the right is the hill of
Ulu-i-ndali, which is composed in the mass of a grey basalt of a type
quite different from the blackish basaltic rocks of the Seatura slope
and of the Wainunu table-land. It was from this view off the mouth of
the estuary that I received my first lesson in studying the structural
formation of the island. I kept it always in my mind’s eye, and for
months in an almost unmapped region it was my only guide.

[Illustration: Profile, looking north from off the mouth of the Wainunu
River.]

The gradual slope of the Wainunu table-land from an elevation of 1,100
or 1,200 feet in the interior to 700 or 800 feet near the coast has
already been referred to. Beyond this lower limit it descends much more
rapidly and within less than a mile it terminates at Masusu in a
steep-sided declivity 300 feet high opposite Ulu-i-ndali, and in a
gentler slope on the eastern side in the Ndranimako district. Its
somewhat undulating surface is well wooded; but on account of the small
gradient the small streams on the table-land do not excavate deep
channels, but flow slowly along in shallow courses and often stagnate in
swampy land where the interesting “Scirpodendron costatum,” the
giant-sedge, flourishes. In their beds occur reddish flinty concretions,
up to 3 inches across in size, and magnetic iron sand in great
abundance. A sample of this sand roughly washed on the spot contains 77
per cent. of magnetic iron.[44]

Basaltic rocks, often exhibiting a columnar structure, are exposed at
intervals on the surface and slopes of this table-land all over its
area. Now and then when traversing this region one comes upon a tract
strewn with large blocks, amongst which occur fragments of huge columns
3 to 4 feet in diameter; but it is on the steep southern slopes of the
plateau in the vicinity of Ndavutu and Masusu that the most extensive
exposures of columnar basalt are to be found. Here there have been large
clearings made for the tea-plantations, and portions of columns 2 to 3
feet in thickness are scattered all over the slopes and surface of
Masusu.

A very interesting exposure occurs on the southern edge of the Masusu
flat facing Ulu-i-ndali. Here there is displayed in the face of a
waterfall a mass of basalt about 40 feet deep, formed of regular
cross-jointed columns, 3 to 4 feet in diameter and often pentagonal in
shape, which are almost perpendicular, being inclined about five degrees
from the vertical. But in the upper portion of the fall the columns are
smaller (2 to 3 feet across) and become arched and nearly horizontal.
This was the only section of the inner mass of the basaltic flows that I
found, and here the columns are almost vertical. In this locality
several other exposures of the columnar basalt occur; but they are all
at the surface and the columns are nearly horizontal or very much
inclined from the vertical, being often pentagonal in form, 2 to 3 feet
across, and sometimes curved with joints 10 to 20 feet in length.

Neither vesicular nor scoriaceous rocks came under my notice in this
region, and the presence of pteropod-ooze deposits and of foraminiferous
clays and tuffs on the slopes of the basaltic tableland indicates that
the flows were submarine. The common character of a sub-aërial basaltic
flow, where there are large vertical columns below and smaller radiating
columns above, did not present itself; and it is probable that the
singular arrangement of the columns in the upper portion of these flows
may be connected with the conditions of depth under which the flows took
place.

It is apparent from the description given by Dana of the columnar basalt
of Tahiti[45] that it was formed under different conditions from those
under which the basaltic flows of Wainunu and Seatura were formed. The
columns composing a cliff 500 feet high in the Matavai valley were 10 to
20 inches across. A bluff, 200 to 300 feet high, in another part of the
valley, was made up of columns 5 to 8 inches in width. The tallest cliff
displayed in places converging and curved columns, which is attributed
to the unequal cooling of the interior of the mass; but it is evident
from a diagram given by the author that the columns were not inclined at
a large angle from the perpendicular.[46] He also refers to some prisms
of a grey basalt exposed just below the Wailuku Falls near Hilo in the
large island of Hawaii which were 8 feet in diameter and were surmounted
by others only 1 to 4 feet across.

The basalts of the Wainunu table-land are blackish and non-vesicular,
with a density of 2·87 to 2·90. They all carry olivine and
microporphyritic plagioclase, and display a little interstitial glass,
and the felspar-lathes are usually in plexus-arrangement, being stout
and often showing twin lamellæ. But the rocks exhibit important
variations in different localities as regards the amount of olivine, the
length of the felspar-lathes, the presence or absence of the ophitic
character, &c., and they are grouped in different genera of the olivine
class (1, 13, 25, 33). Probably the type of genus 25, with scanty
olivine and granular augite, would prevail.

From the varying size of the felspars of the groundmass it is apparent
that the flows are not all of the same character. At Masusu, where the
rock is doleritic in texture, they average from ·25 to ·3 mm. in length.
A mile further north, they are about ·17 mm. long, and two miles more to
the north they average only ·1 mm. in length. It is probable that a
semi-vitreous basaltic andesite (spec. grav. 2·73), that shows no
olivine and is referred to the porphyritic sub-genus of genus 9 of the
augite-andesites, which is exposed in the stream-courses near the base
of the dacitic mountains of the interior, is the product of a later
eruption. Occasionally one finds, as at Thongea in the Wainunu valley, a
basalt rich in olivine (spec. grav. 2·95), the felspars of the base
averaging ·1 mm. in length. It may be remarked here that one cannot draw
a sharp distinction between the basalts of this region and those of the
adjacent eastern slope of Seatura. Their specific gravity is about the
same (2·87 to 2·90); but the coarse texture of the Masusu basalts did
not come under my notice in the last locality, where the felspars of the
groundmass average ·18 mm. in length or about two-thirds the length of
those of the Masusu rocks.

By referring to the section across this part of the island, it will be
observed that the basaltic lavas of this table-land must have issued
from some fissure near the south side of the base of the Ndrandramea
mountains. In crossing the head of this plateau on the way from Nambuna
to Ndrawa one passes from the region of the acid andesites into that of
the basalts. The track first skirts the base of Mount Wawa-Levu, where
the prevailing altered dacitic rocks are exposed in a much decomposed
condition in the stream-courses. Then there is a gradual ascent through
somewhat broken country to reach the western slope of the table-land,
and here are at first displayed the semi-vitreous basaltic andesites
just referred to.

The Wainunu table-land is bisected in a singular fashion by the Ndavutu
River. Since, however, the deep and often gorge-like channel of the
river displays submarine deposits incrusting the basaltic slopes on its
sides, it is evident that the break in the basaltic table-land existed
in part at least before the emergence.

With regard to the total thickness of the basaltic flows of this plateau
I have only a few data. In the bed of the Ndavutu River opposite
Vunivuvundi, and about 400 feet above the sea, there is exposed a
greyish porphyritic rock showing pyrites, apparently an altered
andesite. If this is the bed-rock, the basaltic plateau in that locality
would be 300 to 400 feet in thickness. This is rather over the thickness
of the end of the table-land at Masusu.

I pass on now to consider briefly the submarine deposits that overlie
the marginal slopes of this basaltic table-land in places. They are for
the most part pteropod and foraminiferous ooze-rocks and are extensively
represented on the surface and slopes of the Nandua flat to the north of
Ndavutu, where they occur at all elevations up to 500 feet above the
sea. They are also displayed on the eastern slopes overlooking the
Yanawai but at rather lower heights; and little patches of them occur
here and there in different places but not exceeding 500 feet in
elevation. These friable clayey rocks, which contain from 30 to 40 per
cent. of carbonate of lime, are described in detail on page 320. It may
however be remarked here that these deposits are but partly derived from
the degradation of the submerged basaltic table-land or from the
washings of a basaltic coast. They were formed in a clear sea-way, but
probably at no great depth, at a time when the basaltic plateau was
submerged below the level of breaker-action.

It is remarkable that these deposits do not repose directly on the
basaltic rock. In one place below the Nandua tea-plantation, where there
is a steep descent to the river of about 250 feet, the pteropod
ooze-rock, which is exposed in the upper half, passes down into a
chocolate-coloured marl that contains 5 per cent. of carbonate of lime
and is horizontally bedded. It is composed in the main of fine
palagonitic debris, with some fragments of minerals, &c., and contains a
few microscopic tests of foraminifera. This deposit passes down into
apparently a rock of pure palagonite. The succession of these beds and
their characters are described more in detail on page 344; and as
indicated in the diagram there given it is to be inferred that a very
extensive formation of palagonite has taken place on the surface of a
submarine basaltic flow.

On a similar slope of the Nandua district, and about half a mile nearer
Ndavutu, the pteropod ooze-rock overlies a coarse zeolitic
palagonite-tuff composed in great part of fragments of a highly altered
vacuolar basic glass, but without organic remains. These tuffs are
horizontally stratified. Tuffs precisely similar occur on the northern
slopes of Ulu-i-ndali three miles to the south. They are all described
in detail on page 335.

Some miles up the valley of the Ndavutu River on the steep slope
descending from Vunivuvundi to the river, and on the sides of the river
lower down, are exposed dark palagonitic and sometimes calcareous clays
and tuffs. I traced them as high as 450 feet above the sea where they
were bedded and dipped gently to the west. In the river-channel they
were mostly confined to the right bank, the slope on the other side
being strewn with large fragments of columnar basalt. At the mouth of
the Ndavutu River, there are exposed tufaceous sandstones and a
tuff-conglomerate, probably in great part formed of palagonitic
materials, but I have kept no specimens.

There is much that is puzzling about the tuffs of the region between
Ndavutu and Vunivuvundi. The surface pteropod and foraminiferous
ooze-rocks, that are found here and on the Yanawai or eastern border of
the basaltic plateau and in other localities, offer no difficulties; but
the origin of the palagonitic tuffs that in places lie beneath them is
not so easy to explain. At Mr. Simpson’s old estate on the Nandua flat
one finds numbers of huge blocks of columnar basalt scattered about on
the slope descending to the river; and in places there is exposed in a
small stream, up to a height of 500 feet, a fossiliferous ooze-rock
containing marine shells. The ooze-rock is evidently an incrusting
deposit; but when one goes down to the river-side, which is there about
200 feet above the sea, one finds displayed _in situ_ in the river-bed
an amygdaloidal basic lava with coarse tuffs and agglomerates a little
lower down.

THE HILL OF ULU-I-NDALI.—The meaning of the name of this hill is “Head
of the rope.” It is noted on account of the dense growth of tall forest
trees that clothes its surface, such as the Vesi (Afzelia bijuga), the
Ndamanu (Calophyllum burmanni), the Ndakua (Dammara vitiensis), the
Wathi-wathi (Sterculia sp.) &c.; and it may be that its name is
connected with the launching of the large canoes that were at one time
constructed on its slopes.

Ulu-i-ndali, which has a broad level summit 1,100 to 1,150 feet in
height, rises on the left side of the mouth of the Wainunu estuary. Its
relation to the surrounding region is partly shown in the rough sketch
given on page 83. It is separated from the basaltic table-land to the
north by a deep and wide valley, the bottom of which is raised only a
few feet above the sea; the small stream known as Ndawa-ndingo, that
apparently flows through it, is merely a branch of the Wainunu estuary,
the tide ascending it for some distance. This singular valley, like the
main valley of the Wainunu, dates back in great part to the period
preceding the emergence of this region. The steep basaltic slopes of
Masusu, strewn with fragments of large columns, bound it on the north.
On its south side are the lower slopes of Ulu-i-ndali which are composed
of volcanic tuffs.

A long spur descends to the south from Ulu-i-ndali to form the rocky
promontory of Vatu Vono or “Stone turtle,” so-named from the fanciful
resemblance of the large rounded blocks of basalt on the shore to the
backs of turtles. To the south-east extend the low tuff-formed Ravi-ravi
plains which are but slightly elevated above the sea. The Ulu-i-ndali
range is apparently connected by a “col” with a range of similar height
to the eastward, the highest peak of which is about 3 miles distant.

A more or less coarse doleritic grey olivine-basalt forms the mass of
this hill and is chiefly exposed in its upper portion. Around its
slopes, extending from the coast usually halfway up the hill, are
blackish-brown olivine-basalts; they differ amongst other points from
the grey basalts—which are practically holocrystalline, in their greater
amount of interstitial glass, to which, doubtless, is due their dark
colour. These dark basalts also occur scantily on the summit; but from
their greater prevalence on the lower slopes and from some other of
their characters, it may be inferred that they are in the main formed at
the surface. Outside all, on the north and south sides of the hill, are
exposed coarse tuffs composed of fragments of palagonitised vacuolar
basic glass and containing much secondary zeolitic and calcitic
materials. They are purely of eruptive origin, and although containing
no organic remains were doubtless, as in the case of the precisely
similar tuffs of the neighbouring district of Nandua, deposited under
the sea. A description of their characters is given on page 335. Such
tuffs extend as high as 300 feet above the sea on the north-west slopes,
where there are exposures, 10 to 12 feet in thickness, in the dry stream
courses; and here they may be seen overlying the basalt and rudely
bedded, dipping away from the summit at an angle of 15 degrees.

The grey olivine-basalts of Ulu-i-ndali, which often look like
clinkstone, range generally in specific gravity from 2·9 to 2·95. They
contain microporphyritic olivine in abundance, which is usually more or
less hæmatised and in extreme cases of the change looks like brown mica.
Most of them are referred to genus 16 of the olivine class and their
characters will be found described on page 258. The felspar-lathes are
stout and show sometimes lamellar twinning, and on account of their
large size (·2 to ·5 mm in average length) the rock acquires a doleritic
texture. They display as a rule a flow arrangement around the olivine
crystals. Augite granules occur in great abundance, and there is rarely
any interstitial glass.

These grey olivine-basalts are as a rule non-vesicular, but rocks with
minute irregular cavities, though without glass, occur scantily on the
upper slopes. They come near to the grey olivine-basalts of the hill of
Koro-i-rea in the Solevu district, as described on page 77; but they
differ in their doleritic or coarser texture, the felspar-lathes in the
last-named locality being much smaller, their average length being
·12 mm.

The blackish basalts, mostly characteristic of the lower slopes of
Ulu-i-ndali, vary somewhat in character; but they may on the whole be
regarded as surface forms of the more deeply situated grey basalts which
are practically holocrystalline. The rock of this kind that prevails on
the south and west sides has a specific gravity of 2·96. It is referred
to the same genus (16) as the grey basalts, but differs from them in the
circumstance that the microporphyritic olivine is serpentinised and not
hæmatised, and in the occurrence of a fair amount of devitrified
interstitial glass, to which probably the dark colour of the rock is
due.... The dark aphanitic basalt, with flinty fracture and a specific
gravity of 3·00, that is displayed in Vatu Vono Point, is merely a
compact surface variety of the more coarse-textured grey basalts, being
referred to the same genus. Here there is a great abundance of
microporphyritic olivine in a groundmass of parallel felspar-lathes and
augite grains; but the felspars are unusually small, averaging ·1 mm. in
length; and there is a much larger amount of fine magnetite than in the
grey basalts. There seems to be no interstitial glass; and the olivine
when not fresh is usually serpentinised but occasionally hæmatised.

The dark basalts of Ulu-i-ndali when they occur on its upper slopes
become ophitic. A specimen lying beside me has a specific gravity of
2·91. Allowing for the structural differences, it appears as an ophitic
surface variety of the deeper seated grey basalts. A description of it
is given under genus 12 on page 256, of which it forms the type.

From the data above given, the hill of Ulu-i-ndali is to be regarded as
the basal portion of a submarine volcano still retaining part of its
ash-coverings. The grey doleritic basalts probably represent the core
and the dark fine-grained basalts represent the flows of this ancient
vent.

THE KUMBULAU PENINSULA.—South-east of Ulu-i-ndali stretches a remarkable
“talasinga” district which for convenience I will call the peninsula of
Kumbulau. Its south or seaward border is broken and hilly, and presents
an irregular line of hills 300 to 470 feet in height, extending from
Kumbulau Point to Soni-soni Island, which is almost connected with the
coast. The rest of the peninsula is a low-lying and often marshy plain,
which, though elevated in some places 20 to 25 feet above the sea, is
usually much lower. On the north-east side of the isthmus is the narrow
Nandi inlet, bordered by low mangrove-belts, which represents the broad
channel that in a very recent period of the island’s history cut through
the present neck of the peninsula between the head of the Nandi inlet
and Ravi-ravi.

Stratified and often steeply inclined tuff-sandstones and clays, more or
less basic and palagonitic in character, form together with basaltic
agglomerates the prevailing rocks of the peninsula, whether in the hilly
portion or in the plains. They belong to the basic tuffs of mixed
composition described on page 330; and though the agency of eruptions
can be recognised in their components they are also the products of
marine erosion.

Some of the hills represent volcanic “necks”; whilst the low narrow
promontory between Kiombo and Soni-soni Island has been formed by an old
basaltic flow.

I will begin the description of this peninsula with the eastern
extremity north of Kumbulau Point, the interior of which is cut up into
ridgy hills 300 to 350 feet in height. On its eastern coast are exposed
volcanic agglomerates, composed of large blocks, which from their
dimensions given below would weigh between one-third and two-thirds of a
ton, a size indicating the immediate vicinity of the vent, now
obliterated, from which they were originally ejected. Near Kumbulau
Point the blocks, which are made of basaltic andesite, measure five or
six cubic feet. Further north in the vicinity of Vatu-Ndamu, the
precipitous coast cliffs are composed of agglomerates, the large blocks
of which, often ten cubic feet in dimension, are formed, not of the
prevailing basaltic andesites, as in other parts of the peninsula, but
of a grey hornblende-andesite. This singular appearance of an acid
andesite in a region of basic rocks has no doubt given rise to the
native name of Vatu-Ndamu, “the red or brown stone.” It belongs to the
second order of the hornblende-hypersthene-andesites, and is described
on page 298.

Proceeding along the south coast westward from Kumbulau Point, before
arriving at the village of Na Tokalau we pass from the district of
agglomerates into that of the bedded tufaceous sandstones and clays
which are exposed all along the coast to Kiombo about three miles away.
The transition is indicated by the agglomerates becoming interstratified
with the tuff-beds. These sedimentary tuffs are as a rule steeply
inclined at angles of 20 to 40 degrees, the prevailing direction of the
dip being to the north-east, its uniformity for such a length of coast
being noteworthy. These beds however are occasionally “crumpled”; and
here and there a globular structure is developed.

The hills of this region of sedimentary tuffs between Na Tokalau and
Kiombo are the highest of the peninsula. They usually attain a height of
400 feet, but do not reach 500 feet. From each of them descends to the
coast a spur terminating in a rocky point; whilst between these points
lie low sandy flats, where the native villages of Levuka, Kiombo, &c.,
are situated. The tuff-rocks extend to the top of the hills behind Na
Tokalau, and probably this will be found true of most of the other
hills. Agglomerates are not common in the district. In the point west of
Na Tokalau, however, they are overlaid by basaltic agglomerates, some of
the blocks being scoriaceous. In the point east of Levuka, a
chocolate-coloured somewhat calcareous tuff-clay occurs interstratified
in thin beds with the coarser deposits.

The general characters of these tuff-sandstones and tuff-clays have
already been briefly referred to. The former are much more prevalent and
non-calcareous; the latter are sometimes a little calcareous and look
like marl, and may perhaps contain a few tests of foraminifera. Both are
formed of the debris of basic rocks and are more or less palagonitic.
The coarser deposits are described as sample A on page 330. At times
these tuffs are composed of much coarser fragments of the same
materials, some of them a centimetre in size. A type of tuff
intermediate in character is not uncommon.

The promontory that lies between Kiombo and Soni-soni Island has been
formed by a remarkable basaltic flow. The low tongue, about 50 feet high
and 200 to 300 yards across, in which it terminates, was originally
severed by a passage worn by the sea from the main portion; but it is
now joined by a low tract only 2 or 3 feet above the beach and partly
occupied by mangroves.

The structure of the flow is well exhibited in the shore-flat and
coast-cliffs west of Kiombo, and extending to the end of the point. The
waves have here cut into its mass and exposed its structure. Its lower
part, as exposed in the shore-flat, is made of a compact hemicrystalline
basalt; whilst its upper portion, as displayed in the cliffs, 30 or 35
feet in height, is composed of vitreous and semi-vitreous forms of the
same rock looking like pitchstone. The upper vitreous part is sometimes
massive; but usually it is rubbly, with a tendency to form spheroidal
masses. All transitions can there be traced between the hemicrystalline
rock of the shore-flat and the vitreous rock of the cliffs.

The rock of the shore-flat, which has a specific gravity of 2·83, is a
blackish porphyritic basalt with scanty olivine, and on account of the
semi-ophitic character of the augites of the groundmass it is placed in
genus 33 of the olivine class. The plagioclase phenocrysts are 3 to
5 mm. in size. About half of the groundmass is made up of felspar-lathes
(·17 mm. long) and large augites (·11 mm.), the rest consisting of a
smoky devitrified glass containing a few irregular “lacunæ” filled with
the residual magma in the form of a reddish-brown opaque palagonite-like
material. The rock intermediate between the lower and upper portions of
the flow is also intermediate in character, having a specific gravity of
2·77, whilst quite three-fourths of the groundmass are of smoky glass.

The vitreous rocks of the cliffs, though usually rubbly in appearance,
have also the aspect in places of brecciated pitchstone tuffs with the
interstices filled with waxy palagonite; but the microscopical
examination shows that we have not to deal with a rock of detrital
origin. We have here the effects of the breaking up and crushing _in
situ_ of a dark-brown isotropic basic glass[47] carrying porphyritic
plagioclase. The interspaces then became partially filled with the finer
fragments of the glass and of the crushed felspar; but they were in the
main occupied by a still liquid magma which penetrated into the cracks
of the glass-fragments and into those of the felspars, where the
fractured portions in some cases remained _in position_. There it has
become devitrified and often palagonitised. Whether this liquid magma
was produced by a partial remelting resulting from the heat developed
during the crushing of the glassy upper portion of the flow during the
contracting process, or whether it was squeezed upwards from the less
consolidated lower portion, I cannot determine, although the last
supposition seems more probable. At all events the edges of the
glass-fragments are peculiarly eroded as if by the magma. (The bearing
of these facts on the origin of palagonite is discussed in Chapter
XXIV.)

I infer that this flow has descended from the hills west of Kiombo. Huge
masses of agglomerate are exposed in the lower third of the hill marked
“470 feet” in the chart, and immediately north of the town. Fine clayey
tuffs are exposed in the hill at the back and to the westward of this
place; but the locality requires a more detailed examination. The
absence to all appearance of vesicular and scoriaceous rocks in the case
of this basaltic flow is remarkable. This would not have been expected
in the case of a supra-marine flow; and indeed the testimony of the
tuffs of this peninsula sufficiently indicates that during their
deposition the whole district was submerged.

The future inquirer will doubtless discover some old volcanic “necks” in
the hills of this peninsula. One such hill overlooks the Soni-soni inlet
about a mile west of Kiombo. It is a singular isolated hill which I have
named Bare-poll Peak for descriptive purposes. In my notes its height is
stated as 120 feet, but it appeared to me to be rather higher than this.
It is capped by two huge masses, 14 or 15 feet high, of a dark grey
slightly scoriaceous augite-andesite with a cryptocrystalline
groundmass, which apparently form the uppermost portion of a volcanic
“neck” or pipe. According to the size of these rock-masses the “neck”
would have a circumference of 80 or 90 feet. These masses are in part
incrusted with agglomerate.

The adjacent island of Soni-soni, which is almost joined by the
mangrove-belt to the adjoining coast, probably represents one of the
numerous small vents that were once active in this region. Its single
peak is 460 feet in height. As there did not seem much prospect of
finding rocks exposed on its upper part, its slopes being densely
covered with tall reeds, my examination was confined to the lower
portion during a walk around the island. On its east and north sides
occur rocks of much the same character as those exposed in the
neighbouring low promontory to the east of it. In addition to
agglomerates and basaltic andesites occurred a rubbly pitchstone
composed of fragments, up to a centimetre in size, of an opaque brown
glass displaying a few phenocrysts of plagioclase and pyroxene, the
interstices being filled with crushed fragments of the phenocrysts and
finer glass debris. This rock is allied to the “crush-tuffs” described
on page 334. It may be added that the basic tuffs are more frequent on
the west and south sides of the island.

The low island of Na Vatu in the midst of the Soni-soni inlet is about
250 feet across and only 3 or 4 feet above the ordinary high-tide level.
In 1898, when I visited it, this tiny island possessed about 20 houses
and a population of 60 or 70 persons, and I gather from Hazlewood’s
account of these islands that Na Vatu was crowded with houses more than
half a century ago. It was apparently in the first place a sand-key, and
is protected against the wash of the waves by a low sea-wall formed of
large blocks of stone.

An interesting exposure of bedded tuffs and clays is displayed at
Ravi-ravi on the west side of the peninsula. A broad shore-flat has been
formed by the marine erosion of a line of coast composed of these
deposits. The strike is well exhibited, the dip being about 30 degrees
N. by W. Here there are alternating beds, a few inches thick, of coarse
and fine tufaceous sandstones, sometimes calcareous, with marls or
calcareous clays. The mineral fragments of the coarser rocks are
composed of plagioclase, augite and rhombic pyroxene, the last being
abundant and giving a more acid character to these deposits. The
calcareous fragments appear to be principally shell debris. The marl is
in part composed of much finer detritus of the same minerals. The other
materials of these deposits are derived from the degradation of basic
andesitic rocks, and include also a little palagonite. To the westward
of Ravi-ravi these beds show signs of disturbance, being steeply tilted
to the N.W. Agglomerates also occur in the disturbed area.

The history of the Kumbulau peninsula is evidently the history of the
eruptive phases of a number of more or less submerged small vents and of
the periods of great marine erosion that followed during the emergence
of this part of the island. The absence or rarity of dykes is
remarkable; but most of the hills would represent volcanic “necks”
whether of massive rock, tuff, or agglomerate.

THE DISTRICT BETWEEN THE KUMBULAU PENINSULA AND THE YANAWAI
RIVER.—Between Nandi Inlet and the village of Rewa the sea-border is low
and often swampy, whilst occasional spurs descend from the inland range
into the swamps without reaching the coast. Pebbles of “soapstone”
(foraminiferous mud-rock) occur in streams and are no doubt derived from
the incrusting deposits of the neighbouring hill slopes. In one
stream-bed in the swamps is exposed _in situ_ a remarkable
chocolate-coloured rock that looks like a greasy pitchstone or a
palagonite-rock. It is however of detrital origin, and is composed in
mass of minute fragments of a basic, sometimes vacuolar, glass in great
part converted into palagonite; whilst there are a number of broken
crystals of olivine and plagioclase. Through the palagonitic alteration
the fragmental character is somewhat obscured, zeolites being
extensively developed in the interstices. A little lime occurs and there
is a suspicion of foraminifera. The deposit belongs to the group of
palagonite marls described on page 335. The deeper rocks of the district
are represented in a spur by an altered augite-andesite, originally
hemicrystalline and containing much granular epidote.

Proceeding northward from the village of Rewa, one crosses another spur
descending from the inland range. It is formed in mass of a dark
doleritic olivine-basalt (spec. grav. 2·91) characterised by the length
of the felspar-lathes (·28 mm.), possessing a little interstitial glass,
and referred to genus 25 of the olivine class. It probably represents an
ancient flow. Its surface is incrusted, as high as the road ascends,
nearly 200 feet above the sea, by fine and coarse palagonite-tuffs;
whilst the pebbles of foraminiferous mud-rock in the stream indicate the
existence of incrusting marine deposits further up the slopes. The road
then leads down into a low-lying undulating district that forms the sea
border as far as the mouth of the Yanawai, and reaches about two miles
inland without exceeding an elevation of 100 feet, although low hills
occur here and there. This region is fronted by mangrove swamps and is
traversed by the Matasawalevu and Ndranimako streams. It is a district
of basic tuffs and foraminiferous clays, which, as shown below, extend
up the slopes of the basaltic Wainunu table-land that lies behind. The
soil in all the low country between Rewa and the Yanawai is red, heavy,
wet, and clayey; and affords a contrast to the dry friable soil of the
Kumbulau and Kiombo region to the southward.

The Navakavura plain lying north of Rewa deserves especial mention. It
is a low, swampy district which a mile inland is raised only 20 or 30
feet above the sea, and is mostly occupied by casuarina and pandanus
trees. Red argillaceous rocks, representing more or less decomposed
palagonite coarse and fine tuffs, are exposed in the banks of the
streams. Some of them were originally made up of fragments of basic
glass which after being palagonitised became much disintegrated. A
typical specimen by my side has a soapy feel and looks like a lump of
red clay. Microscopical examination shows that it is composed in mass of
palagonite, but in an extreme stage of the alteration process.

After traversing the Navakavura plain, one crosses a low hill rather
over 100 feet above the sea before descending to Ndranimako. On the hill
are exposed reddish clay-rocks, much weathered, but showing vegetable
remains and a few univalve and bivalve shells. Extensive submarine
deposits occur in the inland district west of Ndranimako. They are the
usual foraminiferous clay-rocks or “soapstones,” and in places they
contain pteropod shells. They are well displayed in river-banks, and in
the hill-slopes on either side; but they are probably of no great
thickness since in one locality named Na Savu, nearly two miles west of
Ndranimako, the underlying basaltic rock is exposed in the bed of a
gully, the sides being of “soapstone.” These deposits were formed in
comparatively deep water.[48] The greatest elevation at which they were
observed was about 100 feet; but this was as high as I reached in the
ascent of the river. According to the natives, who are very observant in
such matters, these submarine deposits extend up the slopes of the
adjacent Wainunu plateau. On page 86 reference is made to their
occurrence on the slopes of this basaltic table-land, 1½ or 2 miles
farther north.

In the district between the Ndranimako and the Yanawai rivers basic
tuffs and “soapstone” prevail. In this locality, and especially in the
vicinity of Ndranimako, siliceous concretions 2 to 3 inches across,
occur in places on the surface. Their nature is described in Chapter
XXV.

From the foregoing remarks it may be inferred that the sea-border
between the Kumbulau Peninsula and the Yanawai River is formed of
submarine deposits overlying basic rocks which probably represent
ancient flows. Some of the deposits are largely formed of glassy erupted
materials, which have been converted into palagonite. Others again are
more characteristic sedimentary formations accumulated in relatively
deep water.



                              CHAPTER VII

      DESCRIPTION OF THE GEOLOGICAL AND GENERAL PHYSICAL FEATURES
                             (_continued_)


                        THE NDRANDRAMEA DISTRICT

THIS hilly region of acid andesites is a continuation of the mountainous
backbone of the island, being separated from the basaltic mountain of
Seatura by the saddle formed by the Na Savu table-land. These acid
andesites exhibit in nearly all cases a felsitic groundmass and
phenocrysts of plagioclase and rhombic pyroxene; whilst many of them are
characterised by brown hornblende more or less pseudomorphosed in the
manner described on page 306, and a few display porphyritic quartz.
Although these rocks have a common facies, they vary considerably among
themselves; and it is difficult to find a term that would strictly
include them all. A general description of their characters is given in
the chapter on the Acid Andesites.

In this interesting region a number of hills or mountains formed in mass
of acid andesites rise up abruptly without any regular arrangement
within an area measuring 5 by 6 miles, and elevated 600 to 1,000 feet
above the sea. Of these hills, thirteen in all, nine range in height
between 1,600 and 2,500 feet above the sea, none of the others rising
less than 1,000 feet above that level. But the actual height of each
hill above the country at its base is much less than this. The height of
the hill-mass, in five or six of the largest, ranges between 900 and
1,200 feet, whilst in the smaller hills it varies between 400 and 800
feet. (See accompanying plan.)

[Illustration: The NDRANDRAMEA District from the westward. The hills and
mountains are of acid andesites and dacites. The foreground is elevated
about 450 feet above the sea.]

                                                          [_Face p. 98._

[Illustration: Rough plan of the Ndrandramea district in Vanua Levu;
made with prismatic compass and aneroid by H. B. Guppy.]

These hills have sometimes a rounded profile, when their summits are
usually wooded. Others again terminate in conical bare rocky peaks,
either pointed or truncated. They have often precipitous slopes and
display vertical cliff-faces high up their sides. Their arrangement is
rather singular. To the south and apart from the others lies Soloa Levu
(1,600 feet). Navuningumu (1,930 feet) is similarly isolated on the
north. On the east rises Ngaingai (2,430 feet), the highest of the
peaks, with Wawa Levu (2,000 feet), Vatu Kerimasi (1,900 feet), Vatu
Vanaya (1,600 feet), and Mbona Lailai (2,100 feet) closely clustered by
its side. On the west there is another group of hills, of which
Ndrandramea (1,800 feet) is the highest and best known. Associated with
it are Kala-Kala (1,600 feet), Mako-mako, Thoka-singa (1,300 feet), Vatu
Mata (1,050 feet), and another unnamed peak (1,400 feet) lying west of
Ndrandramea.

The districts between and among the hills are much cut up into lesser
hills and ridges, the result of the very extensive denudation to which
this region has been subjected. The greater part of this area is drained
by the Tambu-lotu tributary of the Wainunu; but in the northern part we
cross the watershed between the Wainunu and Ndreketi basins, and to
reach Navuningumu we cross the valley of one of the tributaries of the
Ndreketi. To the east of the Ndrandramea region extends a broken
country, elevated rather more than 1,000 feet above the sea, and from it
there rise one or two hills with bare cliff-faces, which are probably
composed of similar acid andesites.

Although for the most part composed of these acid andesites, each hill,
as far as my observations show, has as a rule its own type of the rock,
differing from the others in specific weight, in the texture of the
groundmass, and in the relative proportion of the porphyritic
constituents. The petrological characters will be found more fully
discussed in Chapter XXI.; and only some of the more distinctive
features will be noticed here in the following description of the
district.

THE NGAINGAI GROUP OF HILLS.—Within a space less than a mile square rise
Ngaingai, Wawa Levu, and the other three hills above named, so closely
clustered together that the collective name of “Hen and Chickens” might
be aptly applied to the group.

The peculiar form of Ngaingai is shown in the accompanying
profile-sketch. It is the Nangorongoro of the Admiralty chart. The
height of the mountain from its base is 1,100 to 1,200 feet. Its ascent,
which is not difficult, may be made from the west side. Above its wooded
slopes rises its bare rocky peak, from which a magnificent panoramic
view of the western half of Vanua Levu can be obtained. Characteristic
dacites with porphyritic quartz came under my notice all the way up from
the foot to the summit, being occasionally exposed in perpendicular
cliff-faces. Specimens taken from the upper and lower portions are
uniform in character, and have a specific gravity of 2·57. No other
rocks were observed on its slopes. The whole hill-mass is in great part
if not entirely formed of these acid andesites.

The contrast between the narrow crested peak of Ngaingai and the
dome-shaped summit of Wawa Levu is seen in the sketch; and this is the
more remarkable because it is not associated, as far as I could
ascertain, with any important difference in geological character. Wawa
Levu rises precipitously to a height of 900 or 1,000 feet above its
base, and displays often perpendicular cliff-faces on its sides. Its
broad level soil-covered summit is mostly covered with young wood, few
of the trees having trunks more than 4 inches in diameter, whilst they
are usually clothed with damp moss, and are often decayed and
rotten.[49] True dacites, closely similar to those of the neighbouring
Ngaingai and having a specific gravity of 2·61, were displayed often in
slab-like blocks from the base to near the top. The rudely columnar
structure to be observed in some of the other hills is rarely exhibited.
No other rocks came under my notice. The remains of the stone walls of
two old “war-towns,” one of them named “Ndaku-i-tonga,” occur on its
south and south-east slopes.

[Illustration: Profiles of Ngaingai and Wawa Levu from Nambuna to the
south-west. Both are dacitic mountains.]

The other three hills of the Ngaingai group were not ascended by me.
They show the same bare cliff-faces and have to all appearance the same
geological character. Mbona Lailai and Vatu Kerimasi are two
blunt-topped conical hills with precipitous slopes that rise
respectively about 900 and 700 feet above the country at their base.
Vatu Vanaya, about 500 feet in height, has a rounded summit.

THE NDRANDRAMEA GROUP OF HILLS.—A view of these hills from the westward
is given in the accompanying illustration. They have a lower elevation
than the hills of the Ngaingai group, none of them rising to over 1,800
feet above the sea, whilst their height from the base is also less,
ranging between 400 and 900 feet. They rise, as the illustration shows,
in the midst of a densely-wooded broken country.

Ndrandramea, which is 1,800 feet above the sea, has an individual height
of about 900 feet. Fijians in distant parts of the island are familiar
with the name of this remarkable peak. It has a legendary fame; and like
Wawa Levu in the old time it served as a mountain stronghold in times of
war. The remains of a stonewall of a “koro-ni-valu” or “town of war,”
known as Mata-mei-ndami-ndami, occur on its side, 300 or 350 feet below
its summit; whilst among the wild lemon trees that cover the slopes
below large ovoid sling-stones 4 or 5 inches in length may still be
found. Viewed from the south-east, as shown in the frontispiece,
Ndrandramea has the shape of a woman’s breast; and evidently the origin
of its name is connected with this resemblance. But seen from the west
and south-west, as in the other general view of the district (page 98),
it has a broadly truncated conical outline, its form being indeed
somewhat elongated or elliptical.

This hill presents precipitous slopes, and on the south side it shows
bare rocky faces. As seen in the illustration, it might appear
inaccessible; but the ascent is not difficult on the west side. It is
composed in mass of an acid andesite allied to the dacites of Ngaingai
and Wawa Levu, but differing in the hemicrystalline character of the
groundmass (except at the base), in the porphyritic development of
rhombic pyroxene, and in the absence of porphyritic quartz. As remarked
on page 301, the rock becomes more basic as one descends the hill. At
the top its specific weight is 2·44, about 300 feet below it is 2·58, at
700 feet from the top it is 2·68, and at the base of the hill where it
is holocrystalline and has a dioritic appearance it is 2·71. That it
possesses a rudely columnar structure is shown by the occurrence here
and there on the slopes and at the base of the hill of portions of
prostrate columns, 3 to 4 feet broad and sometimes 20 to 25 feet long,
which have a rounded surface and look like fossil tree-trunks. Masses of
agglomerate of the same andesitic rocks lie about in places on the lower
slopes, the included blocks, which are a few inches across, being
sometimes rounded.

The neighbouring hills lying south and west of Ndrandramea are, as far
as my observations show, of the same acid type of andesite. It is
connected with those nearest by a saddle, 1,100 feet above the sea,
where the same holocrystalline form of the rock occurs, having a
specific gravity of 2·7 and being often rudely columnar in structure.
Kala-kala, about 1,600 feet above the sea, is an imposing-looking hill
with perpendicular cliff-faces on some of its sides. I did not ascend
it, but found at its base a rock of the same andesitic type, differing
from that of Ndrandramea in the more crystalline character of the
groundmass, and having a specific gravity of 2·61. West of Kala-kala is
the outlying hill of Vatu Mata with a flat top and rising only about 400
feet from its base. It has all the appearance of being composed of the
same andesitic rocks. It is shown on the left-hand in the illustration.

Lying south of Kala-kala are the two peaks of Mako-mako and Thoka-singa,
rising respectively 1,400 and 1,300 feet above the sea. I ascended the
last-named, which has a rounded summit covered with trees. Approaching
it from Nambuna on the east, I found at its foot a large mass of
pitchstone-agglomerate, formed of fragments of vitreous basic rocks,
such as occurs around the lower part of Soloa Levu on the other side of
the valley. The slopes of Thoka-singa, between 200 and 450 feet below
the summit, are strewn with masses of another kind of agglomerate made
up of blocks 3 to 8 inches across, occasionally rounded, and composed of
the same felsitic andesite, of which the mass of the hill is formed.
This last-named rock is exposed in bulk in the upper part, but on the
summit the agglomerate reappears. It has a granitoid appearance, and is
distinguished from the acid andesites of the other hills of the
Ndrandramea district by its greater specific gravity (2·72 to 2·74), by
its holocrystalline texture, and by the coarse grain of the mosaic of
its felsitic groundmass, which is probably quartz-bearing but is
relatively scanty. It is, however, referable to the same group of
felsitic andesites, but is to be placed at the basic end of the series.
(Its description is given on page 302.) In Thoka-singa we have therefore
a hill which is evidently formed in mass of these holocrystalline
felsitic andesites but covered in places with an agglomerate of the same
materials. I have already referred to this feature in the structure of
Ndrandramea. Since the blocks are sometimes rounded, such agglomerates
may represent the result of marine erosion during the emergence of this
part of the island. In the case of Navuningumu, where they lie abruptly
on calcareous clays containing tests of foraminifera and shells of
pteropods, a different explanation appears to be needed.

THE HILL OF SOLOA LEVU.—This isolated hill, which presents another type
of these acid andesites, has a broad rounded summit; and though elevated
about 1,600 feet above the sea, the hill itself rises only 800 or 900
feet above the country at its base. It is not easy to obtain a view of
the profile of this hill and to ascertain its relation to its
surroundings; and it was only when I viewed it from near the top of Vatu
Kaisia six miles to the eastward that I was able to understand its
position. Looking from that standpoint across the basaltic table-land of
Wainunu one observed Soloa Levu rising as a dome-shaped hill at the
western margin of the table-land and apparently not separated from it.
The examination of the district shows that on the east and south-east
sides this hill was in part surrounded by the great basaltic flows by
which the table-land was built up. Basic tuffs and agglomerates,
however, occur on the lower slopes on the north-west, west, and
south-west sides, so that Soloa Levu in fact lies in the midst of an
area of basic rocks.

The type of acid andesite which is displayed in the upper two-thirds of
the hill is distinguished from those of the other hills of the
Ndrandramea district by its orthophyric groundmass. Instead of a fine
mosaic, the matrix displays as a rule an arrangement of short stout
plagioclase prisms; but in one of my slides the two forms of groundmass
are associated. In their general characters as described on page 296,
they cannot be separated from the acid andesites of the Ndrandramea
district. Their specific weight ranges between 2·54 and 2·62, and like
most of the other acid andesites they contain little, if any,
interstitial glass. Huge blocks of these rocks lie about on the slopes,
often assuming a columnar form, the fragments of such columns being
sometimes 5 or 6 feet in diameter, and 12 to 15 feet in length. I found
one such block standing erect like a solitary obelisk.

The best way to observe the basic rocks that invest the lower slopes of
Soloa Levu is to follow the track that skirts it on the south side on
the way from Tambu-lotu to Vunivuvundi. Palagonitic tuffs containing in
places a little lime[50] and composed of fragments of basic glass of
varying size and more or less palagonitised extend from Tambu-lotu and
Nuku-ni-tambua (two villages lying about a mile to the westward) to the
west and south-west slopes of Soloa Levu. A pitchstone-agglomerate,
formed of fragments of a basic glass inclosing large crystals of
plagioclase felspar one-third of an inch in length, is associated with
these tuffs on the lower north-west, west, and south-west slopes of the
hill. The tuffs are formed of the same materials as the
pitchstone-agglomerates, but differ in their character of being more or
less palagonitised. However, on the north-west side the latter have also
undergone this change. On page 312 will be found a description of the
basic glass of these agglomerates in its fresh and in its altered
condition. Huge blocks of these rocks strew the surface on the
south-west slopes of Soloa Levu, and in one place the underlying acid
andesite that forms the mass of the hill is exposed in a stream-course.

These pitchstone-agglomerates and palagonitic pitchstone-tuffs are
elevated between 600 and 750 feet above the sea. As one proceeds on the
road to Vunivuvundi and skirts the south-east side of the hill one
ascends the western border of the basaltic Wainunu table-land which,
however, is much cut up by rivers in this locality. Here the tuffs and
agglomerates give place to a basaltic andesite, and on reaching an
elevation of 1,000 feet we arrive at the top of the table-land from
which an ascent of Soloa Levu is easily made. The road then lies on, but
parallel to, the border of this plateau for some distance until it
descends into a deep valley worn by one of the tributaries of the
Wainunu River.

This hill of Soloa Levu is in fact a mass of acid andesite situated in
the midst of an area of basic rocks. I found basaltic rocks exposed in
the stream courses to the north and similar rocks prevail on the
north-west on the way between Nambuna and Tambu-lotu. It has been above
remarked that on the east and south it has been in part surrounded by
the basaltic flows of the Wainunu table-land, and that pitchstone-tuffs
and agglomerates cover its lower slopes on the west and south-west, yet
it is not easy to find any trace of the vent from which they flowed or
were ejected.

It may be here remarked that the occurrence here and there of basic
rocks in the midst of this region suggests the vicinity of dykes. For
instance, in a deep gulley about half a mile south-west of Kalakala,
where a dacitic rock was exposed _in situ_, I came upon a single large
mass of an aphanitic augite-andesite of the type described under genus
16, species A, of the augite-andesites.

THE ALTERED ACID ANDESITES OF THE NDRANDRAMEA DISTRICT.—One of the most
important features of the geological structure of this district lies in
the fact that the bed-rock exposed in the lower region between the hills
is a highly altered acid andesite of the type found in the hills around.
By referring to the map of this locality, it will be observed that
between the Ndrandramea hills on the west and the Ngaingai hills on the
east is the valley of the Tambu-lotu river and its tributaries, an open
broken country deeply eroded by the streams, and elevated 600 to 700
feet above the sea. These altered rocks are well exposed in the deep
gorge-like channel of the river between the village of Nambuna and the
foot of Ndrandramea, and in fact in all places in this district where
the streams have worn deeply into the surface.

They have a coarse felsitic groundmass, and are described under the
felsitic order of the hypersthene-andesites on page 297. They present
all degrees of change from the hard dark grey mottled rocks, in which
the phenocrysts of plagioclase and rhombic pyroxene are in part replaced
by calcitic, viriditic, and chloritic materials, to those where the
pseudomorphism and alteration is complete, when the decomposition
products give their character to a pale yellowish rock, which sparkles
with pyrites and often effervesces briskly with an acid. After this
comes the final stage of disintegration, and we get a whitish rotten
stone, often full of pyrites, the last condition of which is shown in a
kaolin-like material exposed in the river-side.

The extensive alteration of these rocks is also indicated by the
occurrence amongst the gravel of the river-bed and small stream courses
near Nambuna of fragments of clear quartz prisms, half an inch across,
and of nodules, three inches in size and sometimes hollow in the centre,
formed of radiating quartz crystals that once filled cavities in the
altered rock. Small masses of vein-quartz also occur in these streams,
formed in a fissure by the growth of the crystals from the sides towards
the centre. I was unable to find the source of the quartz; but it is
probable that it was produced near the line of contact between the
basaltic flows to the eastward and the older felsitic rocks of the
district. The great alteration of the acid andesitic rocks exposed as
the bed-rocks in this region may in all probability be attributed to the
vicinity of these basaltic rocks. The two formations apparently come
into contact about a mile east of Nambuna. In traversing this district
on the road to Ndrawa one first observes _in situ_ in the streams the
decomposed felsitic bed-rock with occasional loose blocks of a
quartzitic rock that displays in the thin section a mosaic of irregular
grains of quartz. Afterwards, as one rises gradually to the top of the
basaltic plateau, basaltic rocks are alone exposed in position.

In the character of the fine river sand a clue may be found to the exact
locality of the contact. In the midst of the andesitic area between
Nambuna and Ndrandramea, the sand, besides containing much magnetic
iron, is also composed to a large extent of rhombic pyroxene prisms,
clear quartz grains, and fragments of plagioclase, all derived from the
porphyritic crystals of the dacites, &c. Near the basaltic district we
find that the quartz and rhombic pyroxene have disappeared, the sand
being largely made up of magnetic-iron grains mixed with fragments of
plagioclase.

[Illustration: Profile and Geological Section of Vanua Levu, across the
island from the Sarawanga (north) coast to the Yanawai (south) coast.]

THE EXTENT OF THE AREA OF ACID ANDESITE ROCKS IN THE NDRANDRAMEA
DISTRICT.—By referring to the map of this locality it will be observed
that this region of andesites extends northward to the Navuningumu
Range, and that on the south it would be separated from the district of
tuffs and agglomerates, named the table-land of Na Savu, by a line
joining the hills of Soloa Levu and Thokasinga. On the east it is
bounded by the basaltic area of the Wainunu table-land. On the west it
extends at the surface, with an occasional overlying patch of submarine
tuffs and clays, for a distance of at least two or three miles from the
base of the hills, and sometimes, as in the direction of Sarawanga, more
than half way to the coast. I have endeavoured to show the relation of
these acid rocks to the basalts and to the sedimentary deposits in the
geological section.

When taking the track from Sarawanga to Nambuna by way of Ndrandramea
one soon enters the region of these acid andesites. The prevailing rock
exposed on the surface, where it is usually much decomposed, is a
bluish-grey hypersthene-andesite with a specific gravity of 2·54, and
displaying in a cryptocrystalline groundmass, where the felsitic texture
can be recognised, abundant phenocrysts of plagioclase and rhombic
pyroxene. As high as 500 feet above the sea it is occasionally capped by
patches of palagonitised clays and tuffs scantily foraminiferous, and at
one place I noticed a patch of agglomerate, the subangular blocks six to
eight inches across being formed of the same acid andesite. In the same
way by taking the road from Tembe-ni-ndio to Nambuna, passing the hill
of Kala-kala on the way, we leave behind the foraminiferous tuffs and
limestones of the lower coast regions; and when about 400 feet above the
sea we enter the inland district of felsitic andesites which begin about
two miles from Tembe-ni-ndio.

THE NAVUNINGUMU RANGE.—By following the track from Nambuna to
Navuningumu one skirts the bases of Wawa Levu and Ngaingai, where
dacitic rocks are exposed. After passing the watershed[51] between the
Wainunu and Ndreketi rivers, the track descends into the deep valley of
one of the western tributaries of the Ndreketi, where a characteristic
holocrystalline type of these felsitic andesites is exposed. Approaching
Navuningumu one finds exposed at its base agglomerates, composed of
scoriaceous and amygdaloidal semi-vitreous basic rocks, overlying a dark
tufaceous sandstone which on examination proves to be a basic pumiceous
tuff of the type described on page 333, and scantily foraminiferous.

We stand now in a region of basic rocks on the south-east side of the
range, and before us rises abruptly the weird-looking magnetic peak of
Navuningumu, which is well represented in the accompanying illustration.
In the wet season its summit is usually enveloped in the thunder-clouds.
Its elevation above the sea is 1,930 feet, but estimated from its base
its height is 1,000 to 1,100 feet. The natives also name this peak Na
Seyanga, after a town that once existed in this locality. It is the
summit of a range that extends a mile or more to the north where it
terminates in a lesser peak known as Mumu.

Ascending the peak of Navuningumu from the south-east one finds
exposed in its lower part, up to 1,200 feet above the sea,
pitchstone-agglomerates (composed of fragments of a vitreous basic
rock) and white tufaceous sandstones (containing a few tests of
foraminifera), such as are described below in the case of the
neighbouring Mbenutha Cliffs. Between 1,300 and 1,500 feet there is
displayed in position a typical dacite of the type described on page
303.

[Illustration: MT. TAVIA (2,210 feet) from VATU KAISIA. It is probably
formed of an acid andesite.]

[Illustration: The magnetic peak of NAVUNINGUMU (1,931 feet) from the
south. The summit represents a basaltic neck.]

                                                         [_Face p. 108._

The peak itself is formed of a dark-brown slightly vesicular
semi-vitreous basaltic andesite, of which, in fact, for the upper 200
feet, the summit is composed. The rock is somewhat rubbly; and where it
is exposed on the bare peak it is powerfully magnetic, displaying
polarity in a marked degree, and rendering the compass useless (see page
368). A specimen of the magnetic rock, which is a little vesicular, has
a specific gravity of 2·82. It is referred to genus 1 of the
augite-andesites described on page 267. It displays in the slide
porphyritic plagioclase, with a little augite, in a groundmass formed of
a plexus of minute felspar-lathes (·06 mm. in length), and exhibiting a
large amount of a brown opaque glass in which grains and rods of
magnetite with a few pyroxene granules are developed. The magnetite in
the groundmass, although abundant, is not in greater quantity than is
usually found in semi-vitreous basaltic rocks without polarity.... This
terminal mass of basic lava-rock evidently forms the “plug” of a
volcanic pipe that pierces the acid andesitic rocks of the district; and
from this ancient vent were doubtless ejected the basic tuffs and
agglomerates that now cover the lower slopes of the mountain.

The conditions under which this volcano displayed its activity are
further illustrated in a remarkable section exhibited on the east side
of the mountain half a mile or more north of the summit. Here there is a
line of bold cliffs, in which, as shown in the illustration, a bed of
agglomerate, 60 or 70 feet thick, overlies a series of foraminiferous
clays and tufaceous sandstones, which are elevated about 1,100 feet
above the sea. The locality is named “Mbenu-tha” or “Rubbish-heap.” It
is well known to the natives on account of its caves, which serve as a
half-way resting-place on the road from Nambuna to Ndreketi. These caves
have been produced by the more rapid weathering of the underlying clays
and sandstones. The line of cliff extends northward to Mumu, the peak at
that end of the range, and preserves there the same structure. The clays
and tuff-sandstones are more or less stratified, and dip generally to
the west or south-west at an angle perhaps of 20 degrees; but in more
than one place they show signs of great disturbance, being contorted and
steeply tilted.

The foraminiferous clays form a more or less compact rock and contain 15
or 16 per cent. of lime. They inclose pteropod shells in places and show
many minute foraminiferous tests of the pelagic type. Their composition
is given on page 323; but it may be here remarked that the residue is
made up mainly of palagonitic debris, fine clayey material and minerals.
The mineral fragments form about 20 per cent. of the mass, and consist
principally of glassy plagioclase, with some rhombic pyroxene, and
magnetite, their size averaging ·1 mm. The tuff-sandstones
interstratified with the clays contain only 2 or 3 per cent. of lime,
and show only a few scattered microscopic tests of foraminifera. About
two-thirds of the rock consist of fragments of a bottle green basic
glass, vacuolar and but little altered, the rest being composed chiefly
of glass debris, plagioclase, and a little pyroxene, the larger mineral
and glass fragments averaging ·3 to ·5 mm. in size. They are in fact
submarine hyalomelane tuffs very similar to those first met with at the
foot of the mountain, which are referred to on page 108. (They are
described on page 333.)

These interbedded clays and tufaceous sandstones of the Mbenu-tha cliffs
were deposited under somewhat different conditions. The clays represent
the quiet deposition in fairly deep water of fine materials derived from
the degradation of acid andesites as well as of basic rocks. The
hyalomelane tuff-sandstones were formed more rapidly by the accumulation
of fine volcanic ash consisting of fragments of a basic glass ejected
from some neighbouring volcano that rose above the sea-surface.

Submarine hyalomelane-tuffs with basic agglomerates appear to be of
common occurrence around the base of the Navuningumu mountain. As we
leave the range behind and begin to descend the long spur that slopes
northward to Ndreketi, we find for the first mile or two these
agglomerates. But where the deeper rocks are exposed at an elevation of
600 feet, near the village of Singa-singa, there are displayed fine
basic pumiceous tuffs and compact palagonitised clays containing little
if any lime, the last, however, containing a few casts of microscopic
foraminifera. The tuff is made up of minute fragments, the largest less
than ·1 mm. in size, of a basic hyalomelane glass, which is vacuolar,
and often fibrillar like ordinary pumice, and in places shows the early
stage of alteration into palagonite. The clay principally consists of
more or less palagonitised debris of the same basic glass, together with
minute fragments of plagioclase and rhombic pyroxene. These tuffs and
clays represent the two conditions of deposition above referred to, the
last indicating a period of quiescence when the fine materials resulting
from the degradation of both acid and basic andesites were slowly
accumulating in deep water, the first denoting the activity of a
neighbouring supra-marine vent from which fine dust and ash formed of
basic pumice were ejected.

[Illustration: MBENUTHA Cliffs showing volcanic agglomerates overlying
tuffs and clays, containing shells of pteropods and foraminifera, which
are raised 1,100 feet above the sea.]

                                                         [_Face p. 111._

The bed of agglomerate, 60 to 70 feet thick, which overlies the
foraminiferous tuffs and clays exposed in the line of cliff extending
from Mbenu-tha to Mumu, is made up of subangular blocks, not usually
over 6 inches in diameter, of an acid andesite of the general type found
in the Ndrandramea region, but possessing a semi-vitreous
groundmass.[52] By clambering up the steep slope on the south side of
these cliffs, it will be observed that this thick bed of agglomerate is
covered by bedded foraminiferous clays and tuffs similar to those that
underlie it. It is therefore without doubt submarine, and presents the
result of the more violent outbursts of some neighbouring vent. That
this vent is now represented by the “plug” of basic lava forming the
peak of Navuningumu is highly probable. It is, however, noteworthy that
these beds of agglomerates, tuffs, and clays, as shown in the photograph
of the cliffs, are all inclined at an angle of 20° towards the axis of
eruption or to the westward. The tuffs and clays underlying the
agglomerates are, as already remarked, much disturbed in places. It
would seem that all the beds here exposed were originally horizontal,
and were tilted up during the disturbances accompanying the outbursts of
volcanic activity.

The natural section, which the Mbenu-tha cliffs present, is doubtless
due to landslips. Similar exposures, displayed by cliffs of basic
agglomerate with submarine tuffs and clays at their base, are common on
the mountain-slopes of other parts of the island. Water oozes through
the underlying soft deposits, and the result is seen in the occurrence
of huge masses of agglomerate on the slopes below.

From the details here given respecting Navuningumu and its surroundings,
it is apparent that there have been two stages in the history of this
volcanic mountain. The first was submarine and was characterised by the
discharge of acid lavas which consolidated around the vent and were
afterwards covered over with deposits of foraminiferous clays. The
second was in the last part supra-marine. With the renewal of activity,
the overlying acid andesites were broken through and basic materials
were discharged from the new vent. The bed of acid agglomerates exposed
in the Mbenu-tha cliff belongs to that period of the second stage when
the explosive agencies were most violent. It represents the extensive
destruction of the overlying rocks. The foraminiferous tuff-sandstones
are submarine accumulations of the finely comminuted fragments of basic
pumice that constituted the dust and fine ash discharged from a
supra-marine vent. The scoriaceous and amygdaloidal blocks of the basic
agglomerates overlying these tuffs around the base of the mountain have
had a similar origin. The original ash-cone that at one time rose above
the surface of the sea has long since been destroyed by the denuding
agencies; and its situation is alone indicated by the “neck” of basic
lava-rock that forms the peak of Navuningumu.

A very long period must have elapsed since this last stage in the
activity of the vent. The clays containing pteropod-shells and tests of
foraminifera, with which the basic pumice tuffs and the acid
agglomerates were interstratified, are now about 1,100 feet above the
sea, and are situated in the centre of the island. During the emergence
the denudation of the new land-surface was no doubt very great; and
these submarine clays and tuffs, as displayed in the cliffs, owe their
preservation in great part to the protection of the overlying mass of
agglomerate.

Much light is thrown on the history of the whole Ndrandramea region of
acid andesites by the examination of this old volcano of Navuningumu.
Some of the hills, as in the case of Ngaingai and Wawa Levu, seem to
have been stripped of everything that could give information to the
geologist. Others again, like those of Thoka-singa and Ndrandramea,
display here and there on their slopes agglomerates of the same
materials, the rounded forms of some of the blocks being in part
indicative of marine erosion during the emergence of this region from
the sea. In Soloa Levu, however, we have one of these hills partially
surrounded by later basaltic flows and covered in places on its lower
slopes by basic tuffs and agglomerates, probably submarine. In
Navuningumu the original mass of acid andesite is only scantily exposed.
It is for the most part buried beneath submarine clays which are in
their turn covered by the tuffs and agglomerates of later basic
eruptions.



                              CHAPTER VIII

      DESCRIPTION OF THE GEOLOGICAL AND GENERAL PHYSICAL FEATURES
                             (_continued_)


                     MOUNT VATU KAISIA AND DISTRICT

THIS peak, 1,880 feet in height, starts up suddenly in the mountainous
interior of the island. Being situated in the valley of the Yanawai
river, which opens to the south, it forms a conspicuous landmark for
vessels off the south coast; but from most other points of view, on
account of its peculiar situation, it is usually difficult and often
impossible to obtain even a glimpse of it.[53] From its remarkable
blunt-topped conical shape it has received the not very appropriate name
of Marling Spike in the Admiralty charts. The natives name it Vatu
Kaisia, the first word signifying “rock,” whilst the second is the name
of a demon.

[Illustration: Profile-sketch of the Vatu Kaisia district from S.S.E.]

Some idea may be formed of its situation and of the character of the
neighbouring country from the profile-sketch and photograph here
produced. I was unable for reasons given below to take a photograph of
the mountain itself, as it was either too near or too far away. Vatu
Kaisia is approached either from Ndrawa on the north or from Ndawara on
the south, the ascent being best made from the west side. The regions
traversed on the way are so densely wooded that the mountain does not
become visible until the traveller is right upon it. He becomes suddenly
aware that there is some huge mass close to him looming above his head
through the trees; and it is with a feeling of awe that he first looks
upon a mountain that although only a few hundred yards away nearly
escaped his search. He is startled by its proximity, and wonders what
strange forces have been at work to place it there; but his view is
transitory, and whether proceeding north or south he sees it no more,
unless he essays to climb its slopes.

Vatu Kaisia lies not in the centre but towards the west side of the
Yanawai valley the river flowing as an impetuous stream around the foot
of its eastern slope. In the profile-sketch the mountain itself conceals
the peculiar feature of its position, which is, however, shown in an
exaggerated form in the geological section below. On its west side rises
a broad ridge running south which in places is not much higher than the
basaltic plateau of Wainunu to the west of it. This ridge is only
separated from Vatu Kaisia by a dark narrow gorge not many hundred yards
in width, across which my natives were able to make themselves heard
when near the summit. The mountain rises 1,100 or 1,200 feet above the
gorge on its west side, which is 700 feet above the sea, and some 1,400
or 1,500 feet above the Yanawai river on the east, which is 300 or 400
feet above the sea. It possesses two peaks, of which the western one is
smaller and lateral and has a height of 1,600 or 1,650 feet, whilst the
eastern is the main peak and rises to 1,880 feet. The saddle between the
peaks has an elevation of about 1,500 feet. It is very difficult to
obtain a distant view of the two peaks, which lie about N.W. and S.E.
with each other. They are either merged into one as in the view from the
south, or else the highest portion of the main peak is alone visible.

On the lower slopes of the mountain as high as 1,100 or 1,200 feet is
exposed a porphyritic doleritic basalt showing semi-ophitic augite and
abundant interstitial glass. Its specific gravity is 2·8, but there is
no olivine. It belongs to a type of basalt described under genus 9,
sub-genus A, of the augite-andesites. The upper double-peaked portion
rises precipitously, displaying bare rocky cliff-faces with a drop of
100 or 150 feet, and formed in mass of a grey andesitic rock with a
specific gravity of 2·71 and showing abundant small porphyritic crystals
of hornblende and rhombic pyroxene. It represents a type of the
hornblende-hypersthene-andesites described on page 301. I was unable,
through want of a rope-ladder, to accomplish the last hundred feet of
the summit; but the general uniformity of structure was evident. No
detrital rocks came under my observation.

That the porphyritic basalt represents a later flow around this old
andesitic mountain is indicated amongst other things by this absence of
tuffs and agglomerates. Vatu Kaisia is undoubtedly the core of an
ancient cone of hornblende-andesite, and as in the case of Mount Soloa
Levu, which is formed of somewhat similar andesites (see page 103), it
has been more or less completely surrounded by later basaltic flows.
Vatu Kaisia and Soloa Levu occupy similar positions with respect to the
great basaltic table-land of Wainunu, the first lying just within its
eastern border, the second lying partly within its western margin.

The structure of the ridge immediately west of Vatu Kaisia lends support
to this view of the formation of this region. The ridge is here, it is
true, elevated a hundred feet or so above the table-land which is about
1,000 feet above the sea; but whilst on its slopes facing the mountain
the same porphyritic basalt prevails, there is a limited exposure on its
top of the same rock (sp. gr. 2·68), differing only in the larger size
of its porphyritic crystals of hornblende and rhombic pyroxene.

[Illustration: Vatu Kaisia]

The narrow gorge isolating the mountain on the west is occupied by a
tributary of the Yanawai River. It has a depth of 400 feet below the
ridge; and as illustrated in the section below it has evidently been
largely formed by the eroding agency of the stream. However, at the
bottom of the gorge there is exposed a heavy aphanitic basalt showing no
olivine and having a specific gravity of 2·85. Though of much finer
texture, the felspar microliths only measuring ·05 mm. in length, it
differs conspicuously from the overlying porphyritic basalt in
possessing little or no interstitial glass. It is referred to genus 16,
species A, sub-species 1, of the augite-andesites (page 280).

The probable structure of this district is shown in the geological
section here given. It is assumed from the limited exposure of the same
rock on the top of the ridge that the basaltic flows which surrounded
the lower portion of Vatu Kaisia at the same time covered over another
similar peak lying immediately west of it. Through stream-erosion Vatu
Kaisia has now been isolated on its west side; and since the basaltic
rocks rise to about the same height on both sides of the gorge thus
produced, the original surface was probably as indicated by the dotted
line in the diagram.

By following the summit of the ridge, as it runs south on the right side
of the Yanawai valley towards Ndawara, some interesting rocks are
observed. For the first mile from the camping-place opposite Vatu Kaisia
the elevation increased from 1,100 to 1,300 feet, and blocks of a
blackish basaltic andesite (sp. gr. 2·76) lay on the ground. About a
mile further on fragments of white quartz-rock appeared on the surface
having been thrown out of a shaft close to the track which had been sunk
to a depth of 15 or 20 feet by a gold miner[54] a few years before. I
could not descend the shaft to examine it: but the specimens picked up
are evidently a white vein-quartz, some of them having a striated
“slickenside” surface on one side.[55] There is evidently a “contact” in
this locality, probably of a basaltic rock with an acid andesite.

Leaving the shaft, the track proceeds southward and eastward, and one
descends gradually from a height of 1,100 feet down to the Yanawai river
where the elevation is only about 150 feet above the sea. Occasional
blocks of basaltic rocks lie on the surface of the ridge, and in one
locality there is exposed a curious-looking agglomerate formed of
fragments of a greenish altered augite-andesite, somewhat scoriaceous,
the cavities being filled with a zeolite. At the crossing of the river a
black basalt (sp. gr. 2·82) occurs _in situ_; whilst loose blocks of
basalt and of an acid andesite occur in the river-bed. Continuing the
journey from the Yanawai crossing to Ndawara near the mouth of the
river, one follows the track across a range of hills, 500 to 600 feet in
height, basaltic rocks prevailing on the surface.


                      THE NANDRONANDRANU DISTRICT.

Lying north-west of Vatu Kaisia is an elevated district which I have
named after its highest summit, a square-topped peak rather higher than
Vatu Kaisia and probably about 2,100 feet above the sea. Koro-ni-yalewa,
which signifies “town of the women,” is another name of this peak. It is
shown in the sketch given on page 113, and is situated about two miles
north-west of Vatu Kaisia. I did not ascend this mountain, which from
its form would seem to be made of an acid andesite like the Ndrandramea
peaks. Much of this elevated region varies between 1,000 and 1,500 feet
in elevation. It is connected with the Ndrandramea district by somewhat
broken country not much over 1,000 feet in height, which is the “divide”
between the river systems of the Ndreketi and Wainunu. A long
tongue-like extension of similar elevation projects to the north-west
between the Ndrawa and Navuningumu branches of the Ndreketi. This
elevated region is continuous to the eastward with the Tavia Range which
is described below. For convenience the valleys of the upper course of
the Ndrawa river have been included in this district as their geological
features can in this connection be best explained.

This region is well distinguished from most of the other districts of
the island by the prevalence of aphanitic augite-andesites. These rocks
have also supplied the agglomerates of the locality, and the
palagonite-tuffs which are in places extensively represented are
evidently in great part derived from vitreous forms of the same rocks.
We seem to get nearer to supra-marine eruptions in this region than in
most others. The palagonitic-tuffs and agglomerates appear to have
rapidly accumulated in shallow water, and there is reason for regarding
one exposure of the aphanitic augite-andesites as at all events a
shallow-water lava-flow. The aphanitic character of the massive rocks,
however it may have arisen, is here, as I take it, associated with the
shallow-water habit of the tuffs and agglomerates.

(1) EAST SIDE OF THE NANDRONANDRANU DISTRICT.—By following the track
leading from the ridge on the west side of Vatu Kaisia northward to
Ndrawa one rises gradually to a more elevated region. The rocks exposed
on the surface for the first mile are for the most part altered
hypersthene-augite-andesites possessing a micro-felsitic groundmass.
When a height of about 1,400 feet was attained, the track could not have
been far from the peak of Nandronandranu, but on account of the wood no
view was obtainable. In this locality between 1,300 and 1,400 feet soapy
palagonitic clay-rocks and coarser palagonite-tuffs are displayed on the
surface. No organic remains are to be noticed in the specimens collected
here, but they are much affected by hydration. Judging from the
fossiliferous character of similar deposits over a large part of the
island, it is highly probable that these tuffs and clays are also
submarine.

Afterwards a descent was made to an undulating region about 1½ miles
across and elevated between 750 and 850 feet. The blocks there displayed
on the surface are composed of a dark rather compact augite-andesite
with a specific gravity of 2·75 (see genus 13) and of an altered
greenish aphanitic augite-andesite with a specific gravity of 2·59 in
which calcite occurs as an alteration product (genus 16). Aphanitic
rocks of this character as shown below, are very prevalent in the
north-west and north parts of the Nandronandranu district, but are not
usually altered.

(2) THE NORTH-WEST PART OF THE NANDRONANDRANU DISTRICT.—The best route
to follow here is to take the track from Nambuna to Ndrawa. After
crossing the upper portion of the Wainunu table-land one reaches the
headwaters of the Ndavutu River and then ascends the watershed between
the Ndreketi and Wainunu river-systems, reaching Savulu, about 1,050
feet above the sea, where a solitary house marks the site of an old
mountain town. This region is much cut up in deep valleys usually 200 to
300 feet deep, which are occupied by affluents of the Ndrawa branch of
the Ndreketi, flowing north. The valley of the main affluents is from
400 to 500 feet in depth; and this constant ascent and descent of steep
and often slippery valley sides makes the journey very tedious.

At Savulu one stands within the Nandronandranu district. Behind lies the
Wainunu table-land with its olivine basalts; but here aphanitic
augite-andesites prevail and extend to Ndrawa and beyond. They are
exposed in position in the stream-courses and furnish most of the blocks
and pebbles found in the bed of the main Ndrawa River for miles down its
course towards the sea. They are dark, compact, and non-porphyritic
rocks and are all referred to genus 16 of the augite-andesites as
described on page 279. They vary, however, in certain features, as in
the specific gravity, the amount of glass, &c. The residual glass is,
however, usually small; but in a stream-course east of Savulu I found in
position at an elevation of 750 feet a semi-vitreous scoriaceous variety
of these rocks, in which the steam-pores had been drawn out into long
tubular cavities half an inch and more in length. The scoriaceous
character is infrequent; but reference should here be made to another
exposure of a slaggy semi-vitreous rock showing abundant steam-pores in
the tuff-district of the river valley above Ravuka. It differs in some
respects from the prevailing rock, since it displays prismatic augite as
well as felspar microliths in its glassy groundmass, and is for this
reason referred to genus 20 of the augite-andesites. In the elevated
region east of Savulu the aphanitic augite-andesites are in places
overlain by tuffs and agglomerates formed of the same materials. There
is a very good exposure of the tuffs in the Nganga-turuturu cliffs about
2 miles west of Savulu.

(3) THE NGANGA-TURUTURU CLIFFS.—These picturesque cliffs, 50 to 70 feet
in height, rise up at the head of the Liwa-liwa valley between Savulu
and Ndrawa. They are elevated about 1,200 feet above the sea; and
probably derive their name from a small waterfall which, after
descending over their face, drops into the valley below. At its bottom
is situated the hamlet of Liwa-liwa, which is about 600 feet above the
sea. This is the Fijian word for “cold,” and doubtless it has allusion
to the coolness of the valley. On account of the more rapid weathering
of the tuffs in the lower part of the cliffs, there is a rude shelter
afforded by the overhanging portion which is the main feature of
interest that the cliffs present from a native’s point of view.

The tuffs composing the cliffs are horizontally bedded and overlie the
prevailing aphanitic augite-andesite exposed on the valley-slopes below.
Originally grey in colour, they have been largely affected by the
hydration accompanying the weathering process. They are fine in texture
and somewhat friable, but contain no lime, and are chiefly made up of
the palagonitised fine detritus of vitreous varieties of the aphanitic
augite-andesites of the district. No organic remains came under my
notice. Some of the beds contain a number of lapilli of basic pumice, 1
to 3 centimetres in size, which are often in the last stage of the
disintegration produced during palagonitisation. It would seem probable
that these lapilli, after having been ejected from some supra-marine
vent, were deposited with the tuffs in the sea around. It should,
however, be not forgotten that vesicular and pumiceous materials may be
discharged during a submarine eruption. When I visited the museum at
Catania, Prof. Platania showed me portions of a bomb, highly vesicular,
that had been thrown up in a submarine eruption off Vulcano in the
Lipari Islands.

(4) THE UPPER VALLEYS OF THE NDRAWA RIVER.—The two valleys of Liwa-liwa
and Ndrawa meet at Ravuka, where their two streams unite to form the
main Ndrawa River. The former is the largest; and its large impetuous
stream, during its descent of about two miles from Liwa-liwa past
Lutu-kina to Ravuka, which is between 200 and 250 feet above the sea,
has a drop of 300 or 350 feet. The main stream flows with a gentle
gradient to the coast about ten miles away. I did not descend its course
for more than two miles below Ravuka, where some hot springs well up
through the gravel on the left bank (see page 31.)

This is a region of palagonite-tuffs which like those of the
Nganga-turuturu cliffs are mainly derived from vitreous and
semi-vitreous aphanitic augite-andesites. They do not effervesce with an
acid, and neither foraminiferous tests nor other organic remains occur.
The palagonitic material is usually vacuolar, the vacuoles being filled
with palagonitic glass or with a zeolite as in the more altered rocks.
Where bedding is shown, the beds are generally horizontal. These tuffs
are extensively displayed in the sides and beds of the rivers from
Liwa-liwa and Ndrawa to Ravuka and as far as I went down the main river,
namely to the hot springs. They are associated with agglomerates, formed
of the aphanitic augite-andesites, below Ravuka and in the Ndrawa
valley.

(5) THE VICINITY OF NDRAWA.—The village of Ndrawa, which is not elevated
more than 300 feet above the sea, is situated in the heart of the island
in a deep valley more or less hemmed in by the mountains. This is one of
the wettest localities in Vanua Levu, and probably, as in the case of
that of Ndriti in the Seatura basin, the rainfall is not far under 300
inches in the year. In the river-gorge descending westward to Ravuka are
displayed horizontally bedded palagonite-tuffs and agglomerates above
referred to in the description of the Ravuka district, and the same
rocks are exposed on the mountain-slopes to the south of the
village.[56]

Immediately to the north lies a broken hilly country, about 800 feet
above the sea, which has to be crossed on the way to Mbatiri and is much
cut up by streams descending from the vicinity of Na Raro to join the
Ndrawa River below Ravuka. The prevailing rocks are tuff-breccias and
agglomerates. The first are made up chiefly of angular fragments, less
than an inch in size, of aphanitic augite-andesites, some of them being
more or less vitreous and in different stages of palagonitisation,
whilst the finer material derived from the same rocks contains some
carbonate of lime. The agglomerates are composed of the same type of
these augite-andesites, with however but little interstitial glass. It
should be added that pebbles of a kind of jasper or iron-flint occur in
the stream-beds in this locality. (The microscopical characters are
described on page 355.)

By following up the valley that extends to the east from Ndrawa, one
enters after about a mile into the region of Na Raro, which is described
on page 123.


                           THE TAVIA RANGES.

North of Vatu Kaisia the elevated Nandronadranu district divides into
two ranges, one of which stretches eastward to the south of Na Raro as
far as the gap of that name, whilst the other extends southward on the
east side of the Yanawai valley. Near the angle of bifurcation is
situated Mount Tavia, a remarkable pyramidal peak marked 2,210 feet in
the Admiralty chart and lying 1½ miles north-east (N33°E) of Vatu
Kaisia. It is shown in the view facing page 108. All this region is
densely wooded, and I had chiefly to rely on “course-and-distance,” and
on my aneroid, to determine the surface-configuration.

(1) RANGE ON THE EAST SIDE OF THE YANAWAI VALLEY.—No ascent of these
hills was made. They vary from 1,500 to 1,800 feet in height, and
judging from the loose blocks and gravel in the bed of the Yanawai River
below Vatu Kaisia they would seem to be mainly formed of basaltic rocks,
acid andesites being also represented. However, I crossed the southern
end of the range, where it is 500 to 600 feet in height, to the north of
Ndawara, and found basaltic andesites prevailing at the surface.

(2) RANGE EXTENDING EASTWARD FROM MOUNT TAVIA ON THE SOUTH SIDE OF NA
RARO.—Mount Tavia, which has the appearance of a dacitic peak, was not
ascended; but the range was crossed in two places in going from Ndrawa
to Vatu-vono and from Valeni to Nareilangi, its usual height varying
between 1,200 and 1,500 feet, the extreme height being about 1,700 feet.

In making the traverse from Ndrawa to Vatu-vono, one first passes
through a part of the hornblende-andesite region of Na Raro, which is
described in a later page. Afterwards while ascending the north slopes
of the range, basaltic andesites, often doleritic in texture and
referred to genus 1 of the augite-andesites, are usually found as far as
the summit 1,200 to 1,300 feet above the sea. On descending the south
slopes one finds coarse and fine palagonite-tuffs and clays at 900 to
1,100 feet up, similar to those prevailing near the sea-border. They are
probably submarine, but my specimens are weathered and give no
effervescence with an acid. In the bed of the river above Vatu-vono,
about 400 feet above the sea, there occurs in position an aphanitic
augite-andesite (spec. grav. 2·77), referred to genus 16, species A;
whilst blocks of a coarser grained basaltic andesite lie loose in the
stream.

In my traverse across the range from Valeni to Nareilangi I noticed
about a mile from Valeni and not much over 100 feet above the sea an
agglomerate formed of blocks of an altered acid andesite possessing a
micro-felsitic groundmass and showing microporphyritic rhombic pyroxene
with dark alteration borders (spec. grav. 2·5). It is distinct from the
Na Raro rocks; and its presence in an agglomerate seems to indicate the
vicinity of some old acid andesite peak buried beneath later basic
eruptive products. Ascending the south slopes of the range, I found
decomposing basaltic andesites and basic tuffs, the prevailing rocks up
to an elevation of 1,300 feet; but in one locality (800 feet) occurred
large masses of what seemed to be a disintegrating dacitic rock
penetrated by quartz veins less than an inch thick. An aphanitic
augite-andesite, of a somewhat exceptional character (spec. grav. 2·63),
was displayed at the top of the ridge, 1,500 feet above the sea.[57]
Basic rocks were exposed in the spur running northward on the east side
of Na Raro.

THE SEA-BORDER EXTENDING EAST FROM THE YANAWAI RIVER TO THE LANGO-LANGO
RIVER.—In this district is included the area between the foot of the
slopes of the Tavia Ranges and the shores of Savu-savu Bay. This
undulating country, two to three miles in breadth, does not attain a
greater elevation inland than 300 or 400 feet. Fine and coarse
palagonite-tuffs, some of them with the texture of sandstone, are the
characteristic rocks. They at times contain a little lime and probably a
few tests of foraminifera. The palagonitised glass is often vacuolar,
the vacuoles being filled with the same material. In places where they
are well displayed these tuffs generally show bedding, as in a
hill-slope just east of Vuni-evu-evu, where there are fine and coarse
tuffs interstratified and dipping gently W. by S. Basic agglomerates
also occur in this district.

In the promontory named Yanutha Point in the map there is displayed an
old flow of basaltic lava, showing a columnar structure at the end of
the point. The columns are 20 inches in diameter, and are inclined about
20 degrees from the vertical in such a direction that it may be inferred
that the original flow, doubtless submarine, descended at that angle
from N.N.W. The dark grey rock of the columns (spec. grav. 2·76) has a
fair amount of interstitial glass, whilst a blackish compact rock (spec.
grav. 2·78) that represents apparently a more superficial part of the
flow has an abundance of smoky glass in the groundmass. These rocks are
basaltic andesites and are neither vesicular nor scoriaceous, and come
near the basalts of the Kiombo flow which, however, contain some olivine
(see page 92). They are semi-ophitic and are referred to genus 21 of the
augite-andesites which is described on page 283.


                                NA RARO.

In Na Raro we have one of the most interesting of the isolated
hornblende-andesite mountains of Vanua Levu. Unlike Vatu Kaisia, which
often eludes the observation, Na Raro is visible from most points of
view. It is double-peaked, the two peaks lying in a north and south line
and rising precipitously. It is this feature that gives the mountain
such a variety in its profile. From the north and south it appears as
shown in the accompanying sketch as a sharp conical peak. From the
north-east and south-east, as illustrated in the two other sketches, it
has the form of a blunt or square-topped mountain; and its true shape is
only shown when it is seen from the east or west. In the photograph here
reproduced which was taken about 1½ miles to the south-west, the two
peaks are with difficulty distinguished. (See frontispiece.)

Not many ascents have been made of these precipitous peaks. Mr. A.
Barrack, who kindly supplied me with some information about it, made the
ascent some years ago; and Mr. Blyth (?), a magistrate, also reached the
top. There are stories of some big officials being hauled up in baskets;
and the natives told me of a white man who was seized with a
shivering-fit when he arrived at the summit. It is certainly a rather
hazardous climb; but the safest plan is to resign oneself into the hands
of the natives, who “bundle” one up in an expeditious, if not in a very
ceremonious, fashion. Nareilangi, near the foot of the mountain on the
north side, is a convenient starting-point, and half a dozen stout
Fijians will not prove too many to assist the climber in the difficult
parts of the ascent. Since the top usually becomes clouded as the day
progresses, it is best to spend a night in a cave about 1,400 feet above
the sea from which the ascent can be made in the early morning. The view
from Na Raro is panoramic and extends over a large part of the island
from Naivaka to Savu-savu.

[Illustration: PROFILES OF NA RARO.

From the south off Kumbulau Point.

From the north in the Ndreketi Plains.

From the North-east.

From the East-south-east.]

Na Raro rises up to a height of 2,420 feet in the midst of a region of
basic rocks. Agglomerates and coarse tuffs formed of aphanitic
augite-andesites prevail in the broken country on the north and west
sides towards Nareilangi and Ndrawa. Immediately south rises the Tavia
Range with its basaltic andesites and overlying palagonite-tuffs; whilst
on the east lies a spur of this range.

Nareilangi, the village from which the start is made, is about 2½ miles
distant from Na Raro, and though situated in the heart of the island it
is only about 100 feet above the sea. The track first passes through a
district of foraminiferous tuffs and clays reaching up to 200 or 250
feet. Afterwards a broken country extending up to 800 feet is traversed.
Here prevail agglomerates and tuff-agglomerates derived from aphanitic
augite-andesites.[58] One then descends into a valley about 600 feet
above the sea, and from this place the ascent of the mountain proper
begins.

The ascent at first is fairly steep, dacitic tuffs prevailing up to
1,000 or 1,100 feet above the sea and forming in places precipitous
cliff-faces. Large masses of hornblende-andesite lie on the slopes. The
dacitic tuffs distinguish Na Raro from all the other peaks of
hornblende-hypersthene-andesite rocks that I examined. They seem
generally to have been stripped off by the denuding agencies; and only
at times, as around the slopes of Ndrandramea and Thokasinga, are to be
found the remains of agglomerates of the same formation. In the case of
Na Raro, however, the tuffs differ somewhat in their components from the
rocks forming the mountain mass. The tuffs are derived from a
hornblende-andesite of dacitic type; whilst the massive rocks of the
mountain are of hornblende-hypersthene-andesites, without porphyritic
quartz, but approaching the dacitic habit.

The tuffs of Na Raro, which are sometimes compacted and at other times
rather friable, do not display bedding. They contain a little lime; but
I found no tests of foraminifera. They are composed of fragments, up to
a centimetre in size, of a dacite displaying brown hornblende,
plagioclase, and quartz in a microfelsitic groundmass, together with a
few fragments of a semi-vitreous basic andesite.

Above 1,100 feet the tuffs give place to the massive
hornblende-hypersthene-andesite. At an elevation of 1,450 feet, a
shoulder of the mountain is reached, near the top of which is the cave
above mentioned. Crossing the shoulder one descends for 100 or 150 feet
into a gap, thus reaching the foot of the precipitous northern peak,
which rises up like a wall for a height of from 900 to 1000 feet
overhead. It is in mass of the andesite just mentioned, many of its
faces presenting inaccessible cliffs displaying seemingly no structure.
This peak is somewhat lower than the southern peak. I placed its height
at 2,270 feet, which, taking the total elevation of the mountain at
2,420 feet, as given in the chart, makes the difference 150 feet. A deep
and broad cleft, that goes half-way down the mountain, separates the two
peaks. The southern one, which appears to be inaccessible, is evidently
formed of the same acid andesite.

These hornblende-andesites, with or without porphyritic quartz, appear
to be for the most part restricted to the immediate vicinity of Na Raro,
except to the south-west, where at a distance of about a mile and a half
from the mountain at an elevation of 500 feet occur a rubbly
hornblende-andesite and agglomerates of the same materials. Though the
rock is of the Na Raro type, its presence here is suggestive of a
distinct vent of small size, of which most of the traces have been swept
away during the emergence of the island. About half a mile south-east of
this locality at an elevation of 450 feet occur some singular banded
palagonite-tuffs which, although they do not show foraminifera in the
section examined, contain a little calcite and are probably of submarine
origin.... In this locality I found a large white mass, measuring 4 × 4
× 5 feet, formed of a siliceous rock appearing in thin sections as
granular chalcedonic quartz (see page 355).

The hornblende-andesite of Na Raro, as in the case of the rocks of most
of the other peaks of acid andesites, has its peculiar characters. It
differs, for instance, from that of Vatu Kaisia in the larger grain of
the felsitic groundmass (N. R. ·021 mm.; V. K. ·013 mm.), in the absence
or rarity of phenocrysts of rhombic pyroxene, in its lower specific
gravity (2·6 N. R.: 2·7 V. K.), in the presence of a little interstitial
glass, and in other particulars. Both, however, belong to the sub-class
of hornblende-hypersthene-andesites, and are described on page 301. In
the Na Raro rock the rhombic pyroxene is represented in the groundmass.

With regard to the relative age of Na Raro I am inclined to think that
it is the most recent of the acid andesite peaks of the island. Neither
vitreous nor vesicular rocks came under my notice in its vicinity;
whilst the tuffs that clothe its lower slopes are non-pumiceous, though
of dacitic origin, but containing also a few fragments of a semivitreous
basic andesite showing tiny felspar lathes and augite-granules. Since
the everywhere prevailing submarine palagonite-tuffs and foraminiferous
clays do not extend over its area, we may assign to it a later date. It
is evidently also posterior in time to the basaltic andesites and
aphanitic augite-andesites around, which are covered by these submarine
deposits. Relatively recent as it apparently is, this mountain bears the
impress of a high antiquity. There is nothing to indicate that this
“core” of a volcanic mountain belonged to a subaerial vent. Na Raro has
shared in all the later stages of the submergence and emergence of the
island. Though it presents the final page in the history of the
hornblende-andesite volcanoes, that chapter has been for unknown ages
closed.

THE NA RARO GAP.—Between the Tavia and Va-lili Ranges there is a break
in the mountainous backbone of the island, to which I have given this
name. The greatest elevation is probably not over 800 feet. It is from
the south side of this watershed that the Lango-lango river takes its
rise.



                               CHAPTER IX

      DESCRIPTION OF THE GEOLOGICAL AND GENERAL PHYSICAL FEATURES
                             (_continued_)


            THE BASALTIC LOWLANDS OF SARAWANGA AND NDREKETI.

ONE of the most striking features of the north side of the island is the
extensive undulating plain that stretches from the Lekutu river to near
Sealevu on the head-waters of the Ndreketi, a distance of almost 30
miles. In its western half this plain slopes gradually to the sea-coast,
where it is bordered by a broad belt of mangroves. In its eastern half,
from the mouth of the Ndreketi eastward, the lofty Nawavi coast range
intervenes between it and the sea-shore. Its breadth varies usually
between 4 and 6 miles, and its elevation, though it reaches a maximum of
about 300 feet, is as a rule between 100 and 200 feet above the sea.

Over nearly all its area it presents the dried-up and scantily vegetated
appearance of the “talasinga” regions. It is an open country mostly
clear of forest; and it is to this character as well as to its peculiar
vegetation that it in some measure owes its barren look. Amongst the
bracken, grass, and tall reeds (Eulalia japonica) that clothe much of
its surface flourish the Pandanus, the Casuarina, and the Cycad, which
give a special physiognomy to the whole area; whilst several sea-side
plants, as Ipomea pes capræ, Morinda citrifolia, Cerbera odollam, &c.,
have spread themselves far and wide over its extent. It is traversed by
the rivers Ndreketi, Sarawanga, and Lekutu, the two first named being
navigable for several miles, as the tide ascends a long way from the
coast.

In its essential characters this region corresponds with the Mbua and
Ndama plains at the west end of the island, which have been previously
described. Wherever the rivers have worn channels of any depth, basaltic
rocks, sometimes columnar in structure, are exposed; and over most of
its surface the same rocks are displayed, often much decomposed and
developing a spheroidal character, or lying in large blocks all around.
Overlying the basaltic rocks in various localities occur foraminiferous
clays and other submarine deposits. This great region of plains is
partially divided into two by the projecting mass of the dacitic
district of Ndrandramea, the slopes of which descend to within 3 or 4
miles of the coast between the Sarawanga and Ndreketi rivers. For
convenience of description I will deal with these two sub-regions
separately under the names of the Sarawanga and Ndreketi plains.

THE BASALTIC PLAINS OF SARAWANGA.— These plains extend about 6 miles
inland to the village of Tembe-ni-ndio on the head-waters of the
Sarawanga river. The prevailing type of basalt in this region is a
porphyritic olivine-basalt showing a few large crystals of glassy
plagioclase and having a specific gravity of 2·84 to 2·9. They are
neither vesicular nor scoriaceous and are referred to genera 25 and 37
of the olivine class. The felspar-lathes of the groundmass average
·2 mm. in length, and there is a little interstitial glass. They cannot
often be distinguished in their characters from the olivine-basalt
displayed in vertical columns, 4 to 5 feet in diameter, on the lower
slopes of Seatura at the back of Tembe-ni-ndio (page 63). It is highly
probable that most of the basalts of these plains belong to lava-flows
that descended from the great Seatura vent. In the lowlands it is much
decomposed, and a spheroidal structure is frequently developed during
the disintegrating process, just as has been noticed in the case of the
Mbua and Ndama plains on the west side of Seatura. The rounded blocks
that commonly occur on the surface may be regarded in each instance as
the nucleus of a weathering spheroidal mass. When this rock is exposed
unaltered in the streams it is usually massive or non-columnar.

There is a less common type of basalt in this region which perhaps may
represent the upper portion of these basaltic flows. I found it exposed
in the bed of the Selesele river about half-way between Lekutu and
Sarawanga and about 2 miles inland, where it formed vertical columns 1½
feet across. It differs principally in the presence of a few small
amygdules and in the greater amount of interstitial glass. The columnar
basalt that Dana in the “Geology of the United States Exploring
Expedition” describes as occurring at the mouth of the same river
probably belongs to the same flow. He remarks that a few hundred yards
back from the “Watering-place” there is an exposure of columnar basalt,
the columns being vertical, 1 to 2½ feet in diameter, and usually
six-sided.

The incrusting submarine deposits found in patches over these plains are
generally calcareous clay-rocks containing tests of formaminifera and
often also univalve, bivalve, and pteropod shells. They are referred to
the foraminiferous mud-rocks described on page 321. Such deposits are
properly dark-coloured; but as exposed at and near the surface they have
often lost by hydration most of their lime, and have acquired by the
removal of the iron oxides a whitish or pale-yellow appearance, whilst
they have a peculiar soapy “feel,” on account of which they are
generally known as “soapstone” amongst the whites. Streams flowing
through such districts have a somewhat milky colour. These deposits are
extensively represented on the slopes of the Sarawanga valley, and
especially to the east of the town of that name. They are well displayed
on the way from Sarawanga to Tembe-ni-ndio, and are also to be seen on
the surface of the plains between Lekutu and the Mbua-Lekutu watershed
to the southward.

In the vicinity of Sarawanga they attain an elevation of 200 feet above
the sea; but they may be traced in patches up to 500 feet on the
adjacent slopes of the acid andesite region of Ndrandramea. Near the
river, and less than 100 feet above the sea, these deposits are in one
place overlain by an agglomerate formed of large blocks, 1 to 2 feet
across, of these Ndrandramea andesites and dacites. In another place,
near the town of Sarawanga, I found them exposed in the river-bank,
where they were covered over by a coarse palagonitic bedded tuff,
dipping gently eastward and somewhat calcareous. From the character of
the shells of marine univalves inclosed in this tuff, it appears to have
been formed in shallow water.

A very interesting display of these surface marine deposits occurs in
the upper part of the Sarawanga valley in the vicinity of Tembe-ni-ndio.
Here we have fine and coarse calcareous palagonitic tuffs, containing
tests of foraminifera, associated with impure foraminiferal limestones.
They occur up to elevations of 300 feet above the sea on either side of
the Sarawanga valley above this town, incrusting on the north side the
lower dacitic slopes of the Ndrandramea district, and on the south side
the lower basaltic slopes of Seatura. At the bottom of the valley, as in
the rising ground between Tavua and Tembe-ni-ndio, they conceal in part
the basaltic rocks of the district.

Near the last-named place, on the right bank of the Tembe-ni-ndio branch
of the Sarawanga river, the foraminiferal limestones are displayed in
low cliffs 15 to 20 feet in height. They are sometimes earthy when they
contain about 25 per cent. of lime, and at other times more compact with
about 45 per cent. of lime, the residue being composed of palagonitic
materials, tiny fragments of minerals and of a basic rock, &c.[59] Large
shells, of Ostræa and Cardium are also contained in these limestones,
the valves being detached from each other. The oyster shells project
from the weathered surface; and it is probable that the name of
Tembe-ni-ndio, which signifies “the shell of the oyster,” may be thus
explained. Underneath the foraminiferal limestones in this locality
occur bedded coarse tufaceous sandstones, slightly inclined E.N.E., and
inclosing waterworn gravel and pebbles. These low limestone cliffs,
although about six miles inland, are not more than 120 or 130 feet above
the sea. In their face there is evidence of an old erosion-line of the
river 10 or 11 feet above its present level.

By following up this branch of the river for a little distance I came
upon an exposure of nearly horizontal bedded palagonitic tuffs on its
floor and sides. Here a coarse tuff, of which the larger fragments
composing it range between 3 and 5 mm. in size, passes upward into a
chocolate-coloured compact tuff-clay formed of the same materials, the
larger averaging ·2 or ·3 mm. in size. These tuffs are made up chiefly
of a palagonitised vacuolar basic glass, the vacuoles being filled with
the alteration products. The lower coarse tuffs contain very little
lime, probably not over 1 per cent., and exhibit no organic remains in
the slide. The upper fine tuffs have 3 or 4 per cent. of lime, and
inclose numerous minute tests of foraminifera of the globigerina type,
their cavities being generally filled with palagonitic material.

Further up the valley about a mile above Tembe-ni-ndio, and about 250
feet above the sea, the impure foraminiferal limestones again appear;
but they here exhibit an important difference in texture. In the
groundmass of those of the lower locality, the calcite is granular and
loosely arranged, or displays in an obscurely indicated mosaic the
commencement of recrystallization. In the case of those of the upper
locality the calcitic material of the groundmass has more completely
recrystallized, and shows a fairly clear mosaic; whilst in one place the
rock was overlain or rather incrusted above by a layer, 3 inches thick,
of a white crystalline limestone, looking like statuary marble, and
inclosing portions of a material like that of the rock beneath it. This
last, when examined in the slide, exhibits itself as formed in mass of
crystalline calcite, displaying a regular mosaic, and inclosing small
fragments of palagonitised materials and of minerals (pyroxene) such as
are abundant in the rock below. In places the grains of the mosaic are
bordered by brown and black iron oxide. It would, therefore, appear that
a metamorphism has been in operation here, and that the process which
began with the recrystallization of the matrix in the lower rock is
almost completed in the overlying thin layer where even most of the
non-calcareous materials have disappeared. No evidence suggestive of
contact-metamorphism came under my notice in this locality. These
foraminiferal limestones are surface formations, and it was in the
uppermost portion that the metamorphism was most complete. We here
witness in operation the transformation of a rock containing 46 per
cent. of carbonate of lime (the residue of minerals, palagonite, &c.),
into a marble or crystalline limestone. I gather that as in the instance
of several of our old British limestones the change is a purely
interstitial one, and is not connected with thermal metamorphism.

These remarks on the basaltic plains of Sarawanga and on their
incrusting submarine deposits may be concluded with a brief reference to
the siliceous concretions, 2 or 3 inches across, the silicified portions
of corals, and the fragments of clay iron-stone and limonite resembling
hæmatite, that occur frequently on the surface. They are common on the
plains south of Lekutu and between Lekutu and Sarawanga, and up to
elevations of 200 feet in the foraminiferous clay district east of
Sarawanga, where fragments looking like portions of the silicified
branches of Madrepores are to be found; but they are not limited to such
localities, and may occur also where the surface is formed of decomposed
basaltic rock. (These matters are generally discussed in Chapter XXV.)

THE BASALTIC PLAINS OF THE NDREKETI.—This low-lying region of rolling
“talasinga” country now serves as the basin of the Ndreketi river, the
largest of the rivers of Vanua Levu. It is usually elevated between 100
and 300 feet above the sea, and its limits are well defined by the 300
feet contour line in the map of the island. On the east it is separated
from the basin common to the Wailevu and Lambasa rivers by the Sealevu
Divide, which is described on p. 136. On the west, as before observed,
it is only in part distinguished from the basin of the Sarawanga by the
spur descending from the dacitic mountains of Ndrandramea. It meets the
coast in the vicinity of the mouth of the Ndreketi; but for two-thirds
of its length it is cut off from the sea by the great Nawavi range. It
supports the characteristic vegetation of the “talasinga” or sun-burnt
land. Whilst the Pandanus and the Casuarina are most conspicuous amongst
the trees, bushes, herbs, grasses and ferns predominate. Here the native
Ginger and the native Turmeric with species of Tacca are frequently to
be recognised, and the waste-land bushes of Dodonæa viscosa and Mussænda
frondosa are abundantly to be found.

As in the Sarawanga plains, the basaltic rocks are here often overlain
or incrusted by submarine deposits, the former exposed in all the deeper
river-beds, the latter frequently displayed in the sides of their
tributaries.

I will deal first with the basaltic rocks. In the places where the
surface deposits have been stripped off, these rocks are generally
exposed as decomposing boulders, the spheroidal structure being well
developed in the weathering process. Not infrequently, however, a rudely
columnar structure is exhibited where the rivers have cut deeply into
the basalt. The columns that I observed were usually vertical. In the
river-bed at the landing-place at Mbatiri, for instance, the columns are
from 2½ to 3 feet across and vertical. As exposed in the river-crossing
about a mile above this town they are 12 to 15 inches in diameter and
also vertical. However, at Na Kalou, a coast village about 1½ miles east
of the mouth of the Ndreketi, where there is an unexpected exposure of
basalt, the columns, about a foot in diameter, are inclined at an angle
of about 20° from the vertical and face to the north.

These rocks are, as a rule, compact, only showing a typical scoriaceous
structure in the case of specimens obtained near the foot of Nakambuta,
an isolated hill about three miles to the southward of Natua, which
probably represents a vent of more recent times. Often, however, they
have a pseudo-vesicular appearance, from the occurrence in the midst of
the patches of interstitial glass of minute irregular cavities that seem
to have been formed during the last stage of consolidation of the magma.

The prevailing type of basalts is a blackish, doleritic, semi-ophitic
rock without olivine, with specific gravity 2·78 to 2·80. They are
characterised by the length of the felspars of the groundmass
(·22-·35 mm.), by the large size of the augite granules (·1-·3 mm.), and
by the quantity of dark interstitial glass. They present two forms, one
with and the other without plagioclase phenocrysts. The first kind is
referred to genus 9 of the augite-andesites (page 272), some of the
specimens being referred to the porphyritic sub-genus, and others to the
non-porphyritic sub-genus, according to the size of the plagioclase
phenocrysts. The second kind, without felspar phenocrysts, belongs to
genus 12 of the same class (page 275). A good example of the porphyritic
rocks is afforded in the large blocks lying in the stream-beds during
the first half of the way from Ndreketi to Sarawanga.

It may be pointed out here that these doleritic, semi-ophitic basaltic
andesites of the Ndreketi plains differ conspicuously from the
prevailing type found on the slopes of Seatura, on the Sarawanga and
Mbua plains, and on the Wainunu table-land. There we have, as a rule,
olivine-basalts, having a specific gravity of 2·86 to 2·90, and showing
but scanty interstitial glass, the felspars of the groundmass being on
the average not over ·2 mm. in length, whilst the augite granules are,
as a rule, only ·02-·03 mm. in diameter, and the ophitic structure is
infrequent.

The submarine deposits, consisting of foraminiferous clays and coarser
tuff-sandstones, the former being usually beneath, are found at
intervals all over this area. They occur inland as far as Vuinasanga and
Nareilangi, near the base of the mountains of Va Lili and Na Raro,
reaching as high as 300 feet, their place being taken on the mountain
slopes by coarser tuffs and agglomerates. When not weathered they are
more or less calcareous, and contain occasionally marine molluscan
shells, whilst palagonitic debris enter largely into their composition.
The foraminiferous clays, often much bleached by hydration, are well
represented around Mbatiri and in the districts between that town and
Natua and Nareilangi. They are relatively deep-water deposits, and
belong to the type described on page 323. Others, again, as exposed in
the banks of the river at Natua, are chocolate coloured and of the kind
referred to in detail on page 335. These foraminiferous clays in the
region between Natua and Mbatiri are overlain in places by coarse,
almost brecciated, tuffs, formed in part of the debris of acid
andesites, such as compose the not far distant mountain of Na Raro.

Since the massive basaltic rocks are exposed in all the deeper rock
channels of these plains, it is apparent that the overlying submarine
deposits can possess no great thickness. Probably they are never 100
feet thick, and usually far less. In many places, through their
denudation, the underlying basaltic rocks are exposed, and in a
decomposing condition largely form the surface. These deposits as a rule
display bedding, the beds being horizontal or at least only inclined 2
or 3 degrees. This horizontality is a nearly constant feature of these
submarine beds, as they overlie the basaltic rocks of the plains; and it
is a feature we should expect to find where there has been emergence
rather than upheaval.

Siliceous concretions and silicified coral fragments, so characteristic
of the surface of some of these plains of Vanua Levu, did not frequently
come under my notice here. They, however, occur occasionally, as in the
district between Nanduri and Natua.


                           THE NAWAVI RANGE.

With this remarkable coast range, which fronts the Mathuata sea-border
for a distance of 12 or 13 miles between Ravi-ravi Point and Nanduri, I
have unfortunately but scant acquaintance. It attains its maximum
elevation in Mount Nawavi of 2,238 feet, and is described by Mr. J. P.
Thomson,[60] who surveyed this coast, as broken in two nearly opposite
Niurua, the pyramidal mountain of Koro Navuta rising in the gap. Various
other peaks, besides that of Nawavi, are marked in the latest Admiralty
chart; they vary in height from 1,000 to 1,700 feet. As this range lies
only a mile or less back from the beach, it gives to the sea-border a
bold and often precipitous appearance, which is well shown in an
illustration in Wilkes’ _Narrative of the United States Exploring
Expedition_ (iii. 226).

Basic rocks probably prevail in this range. When I ascended its eastern
spurs from Nanduri, and reached a height of 800 feet, only basic tuffs
and agglomerates came under my notice. From Dana’s remarks[61] it is to
be inferred that the “frowning bluffs” opposite Mathuata Island are of
similar formation; and it would seem that the rugged black stones,
described in the _Admiralty Sailing Directions_[62] as topping the hills
behind Ravi-ravi Point, are of the same basic character. From its
contour and profile I would gather that, as in the great mountainous
ridges that constitute the backbone of the island between Va Lili and
Mount Thurston, palagonitic tuffs and clays of submarine origin will,
together with volcanic agglomerates, be found far up the slopes of this
range, and that the axis will prove to be largely composed of massive
basic rocks.

The hot springs referred to by Thomson and others as occurring at the
foot of the north and south slopes, namely at Vatuloaloa, Nambuonu, and
in another unnamed inland locality, are briefly mentioned on page 31.

THE SEALEVU DIVIDE.—This broad range which separates the Ndreketi and
Lambasa basins is an offshoot from the central mountains at Sealevu and
reaches the coast just east of Nanduri. Its highest part according to
the elevation given in the Admiralty chart is 1,437 feet. The road from
Sealevu to Nanduri, which crosses its broad level summit for a distance
of about three miles, does not rise over 1,100 feet. Between 800 and
1,100 feet are exposed calcareous tuffs and clays all largely made up of
palagonitic materials. The coarser might be described as sandstones. The
clays have 12 per cent. of lime and are foraminiferous and are of the
type described on page 321. The rocks displayed on the lower northern
slopes on the way to Nanduri are at first the same submarine deposits,
and afterwards decomposing basaltic andesites. It is apparent that in
the central elevated part of this range there are hills of volcanic
formation more or less completely buried beneath these deposits.


           THE DISTRICT BETWEEN NANDURI BAY AND WAILEVU RIVER

The sea-border between Nanduri Bay and Middle Point, nearly four miles
to the east, consists of a fringe of lowland margined by the
mangrove-belts and banked by a line of hills between a quarter and
two-thirds of a mile inland. These hills form a continuation of the
Nawavi coast range of mountains extending from Ravi-ravi Point to
Nanduri. They attain their greatest height in the case of Ulu-i-sori, a
cockscomb-like peak 1,141 feet above the sea. Another of these hills,
Vatu-tangiri, is capped by a remarkable obelisk-like rock. Behind this
coast range lies a hill with an elevation of nearly 1,400 feet.

The rocks exposed for the first mile or two along the coast east of
Nanduri are agglomerates and basic tuffs. The blocks of the
agglomerates, however, are made of an altered grey porphyritic rock
which has the characters of a porphyrite of a rather acid type.[63] This
composition of the agglomerate is quite exceptional and indicates the
antiquity of the volcanic rocks in this locality. Farther along the
coast the typical agglomerates occur, where the blocks, 3 to 10 inches
across, are composed of the usual semi-vitreous black basaltic rock
showing plagioclase phenocrysts. Nearer Middle Point a decaying
doleritic basalt is displayed at the surface. It is similar to the
prevailing rock of the Ndreketi plains, and is referred to the ophitic
rocks forming genus 9 of the augite-andesites.

The elevated promontory of Middle Point is a prolongation of a spur of
Ulu-i-sori. Where it is crossed by the road it is about 350 feet above
the sea. On its west slopes are exposed yellowish-white tuff-like rocks,
evidently the prevailing basic clay-tuffs which have become bleached
through the hydration accompanying the weathering process. Beneath these
deposits lies an amygdaloidal augite-andesite which is bared in places.
The rock is semi-vitreous and the amygdules it contains are often a
centimetre long. They are composed of a white mineral with fibro-radiate
structure and made up of needle-like prisms. It gives off water, but it
is not easily fused, and does not gelatinise in HCl.

From the top of the promontory the road strikes inland in an
east-south-east direction for Tambia, passing inside the coast range,
which is here 600 feet in height, and descending gradually through a
region of basaltic andesite into the valley of the Tambia river. (This
rock, which has a specific gravity of 2·84, displays more or less
parallel stout felspar-lathes, ·23 mm. in length, and has a little
interstitial glass. It belongs to genus 13 of the augite-andesites.) Low
hills shut in the little valley on all sides except where the river
breaks through the coast range. The town of Tambia is not over 100 feet
above the sea. About a mile to the north exist hot springs of
considerable extent which are described on page 32.

The road from Tambia to the Wailevu River traverses an undulating
district varying from 100 to 300 feet above the sea. A basalt containing
a little olivine, with a specific gravity of 2·91, is commonly exposed
at the surface in a disintegrating condition. Here and there occur basic
tuffs. In one locality, there is displayed a dyke-like mass in a small
stream course, 200 feet above the sea, of an altered grey and compact
andesite marked with parallel red streaks or bands. It is an aphanitic
augite-andesite; and is to be referred to genus 13 of the augite
sub-class. It displays closely crowded felspar-lathes, ·07 mm. in
length, in flow-arrangement. The bands are due to the gathering of the
residual glass in streaks parallel to the flow. Chalcedonic flints, some
of them showing the agate-structure, together with fragments of
silicified corals, are found occasionally on the surface in this
district.


                           THE LAMBASA PLAINS

These remarkable inland plains, about ten miles long and three to five
miles broad, are well described in the Admiralty chart as a low
undulating country covered with grass, screw-pines, and Casuarina trees.
They are backed by the mountains forming the central axis of the island,
whilst broken groups of hills, not usually more than 500 or 600 feet in
height and attaining in Ulu-i-Mbau an elevation of 1,160 feet, intervene
between them and the sea-border. They are traversed by the Wailevu,
Lambasa, and Ngawa rivers which after breaking through the seaward
hill-ranges pass through broad mangrove-belts to reach the coast. The
tide ascends these river-courses for several miles; and in the case of
the Lambasa river boats can follow its winding course for ten miles
penetrating into the heart of the plains. Much of this level inland
region is less than 100 feet above the sea; whilst the contour line of
300 feet by which the region is defined in the map attached to this book
fairly well indicates the higher levels.

The features which we have described in the instances of the Sarawanga
and Ndreketi plains are in the main reproduced in the Lambasa region;
but in the last-named each of the three rivers has a system of hot
springs along its course, namely (as described in Chapter III.), at Na
Kama on the Wailevu River, at Vuni-moli on the Lambasa River, and at
Mbati-ni-kama on the Ngawa River. Basaltic andesites, often exposed at
the surface in a decomposing condition, form the foundation of the
plains. They are overlain by submarine clays containing pteropod shells
and tests of foraminifera; and over these in their turn coarse
palagonitic tuffs and agglomerate-tuffs are found in places. Formations
still more recent are represented by elevated reef-rock on the seaward
side of the hills that bound the plains. Nodules of chalcedony,
silicified corals, and other siliceous rocks, together with fragments of
impure limonite, lie on the surface over much of this region.

The basaltic rocks of this region rarely show olivine, and belong as a
rule to the basaltic andesites, being referred to genera 13 and 21 of
the augite-andesites, the specific gravity being about 2·8. The
felspar-lathes, ·12 to ·14 mm. in length, are in flow-arrangement, and
the augite is at times semi-ophitic; whilst there is a little
interstitial glass. The basic rocks prevailing between Vatu-levoni and
Vandrani belong to genus 13 of the augite-hypersthene-andesites and have
a groundmass of much finer texture, the felspars only measuring ·05 mm.
Their specific gravity ranges between 2·7 and 2·75.

The overlying foraminiferous and pteropod clay rocks, the so-called
“soapstone,” are exposed over large areas of the surface. A good idea of
the important part they take in the formation of the lower plains may be
formed by visiting the hot springs of Vunimbele, close to Vuni-moli,
which issue from the side of a deep trench cut into these deposits. As
generally displayed at the surface, they have been subjected to so much
hydration in the weathering process that they appear as yellowish-white
clay-rocks deprived of their lime; and it is only now and then that the
remains of foraminifera and pteropods can be detected. They are,
however, fairly well preserved around the base of Ulu-i-mbau in the
vicinity of Koro-wiri, where they contain, besides the shells of
pteropods and foraminifera, portions of decaying coral, and extend to
200 feet and over above the sea. Here they are overlain by rather
coarser basic tuffs of mixed character, containing 5 or 6 per cent. of
carbonate of lime and some palagonite, which I followed as high as the
track lay, rather over 500 feet above the sea.[64] The reef-limestones,
already noticed as exposed in the low hills between Wailevu and Lambasa,
lie a mile or two inland and reach to 100 feet above the sea.

The fragments of siliceous rocks, which with occasional bits of impure
limonite, occur at intervals all over the surface of these plains and
largely form the gravel and pebbles in the river-beds, include nodules
of chalcedony, fragments of jasper or iron-flint, white quartz-rock
formed of chalcedonic silica, silicified corals, &c. They are especially
frequent in the vicinity of Nasawana and Koro-utari, and include fine
specimens of agates and of onyx.



                               CHAPTER X

      DESCRIPTION OF THE GEOLOGICAL AND GENERAL PHYSICAL FEATURES
                             (_continued_)


                           THE VA-LILI RANGE

THIS range extends from the Na Raro Gap before mentioned to the
Ndreke-ni-wai river. It is partly isolated on the north-east from the
Korotini Range, the extension eastward of the mountainous axis of the
island, by a depression or saddle which at its lowest part is not more
than 1,200 or 1,300 feet above the sea; but there is no real break in
the line of mountains. It is, however, convenient to make this
distinction, and I have named the dip between the two ranges, the
Waisali Saddle. The range now to be described attains its greatest
elevation in the summit of Va-lili, which is 2,930 feet above the sea.
There are two or three other peaks that exceed 2,000 feet, and much of
the range is not under 1,700 feet. My acquaintance with this range is
not extensive; but it will serve to illustrate its general geological
structure.

The summit of Va-lili is very conspicuous from most points of view. From
the north, east, and south-east, it has a remarkable broad and
square-topped profile with a little conical elevation in the centre.
From the south-west, it displays a different outline with a solitary
squarish block on the top, and this is the form most familiar to the
navigator. Unfortunately, for reasons given below, I did not quite reach
the summit, and although I was able to obtain sufficient data for
forming a general idea of the structure of this part of the range, the
structure of the actual summit has yet to be ascertained.

(1) ASCENT OF VA-LILI FROM NARENGALI.—This village, which is elevated
400 feet above the sea, lies about two miles in a direct line, N.N.E.
from the peak. In traversing the intervening country, one crosses the
Loma-loma ridge, elevated 1,000 feet, on the top of which was once
situated the village of Loma-loma visited by Horne in 1878. The rocks
exposed on the surface are scanty, a hard palagonite-tuff, which owes
its induration to a calcitic cement, occurring on the upper part of the
ridge; the original site of the village being marked by a large block of
this stone.[65] The track then descends into the valley of the Loma-loma
river, about 400 feet above the sea, in the bed of which occur blocks of
an amygdaloidal basaltic andesite, containing phenocrysts of both
rhombic and monoclinic pyroxene, and referred to genus 1 of the rhombic
pyroxene andesites. The amygdules are formed of calcite.

                 PROFILE-SKETCHES OF THE VA-LILI RANGE.

[Illustration: View from the south-east near Savarekareka.]

[Illustration: View from the south-west.]

Beyond the river the ascent of the northern slope of Va-lili begins. As
high as 1,100 feet occur basic agglomerates overlying fine and coarse
palagonite-tuffs, which are at times horizontally bedded, the finer
kinds being sometimes calcareous, and like that of the Loma-loma ridge
above mentioned. At 1,300 feet is a line of tall cliffs which extend for
some distance at intervals along the mountain-slope, and are indicated
by some fine waterfalls. My track struck these cliffs at a place named
“Nangara-ravi” (the leaning cave-rock) where they have a height of 150
feet or more. The tall cliff leans slightly forward, so that it forms a
shelter at its foot, and hence the name. It is composed of a
tuff-agglomerate, the blocks, which are formed of a semi-vitreous
basaltic andesite of the augite class, being not generally more than 3
or 4 inches across. These blocks, which are rounded on the outer exposed
side and angular on the imbedded side, are inclosed in a hard, probably
calcareous matrix. The whole face of the cliff has the appearance of
having been worn smooth by attrition, and there are not to be observed
the projecting blocks from its surface which are so characteristic of
other agglomerate-cliffs. It shows no stratification; but at its base
flush with the cliff-face are large masses of a basic massive rock. But
few portions of rock have been detached from the cliff. However, I found
in the midst of a huge fallen fragment of the agglomerate a dyke-like
mass of a basaltic andesite, which differs chiefly from the rock forming
the blocks of the agglomerate in being more crystalline. This dyke must
have been about 15 feet thick.

Having regard to these various features, I am inclined to consider that
this leaning cliff represents one side of a large fissure in the
agglomerates which was occupied by a dyke. Reference has been above made
to the fact that the agglomerates may be seen overlying the tuffs
farther down the slope, so that the conditions favourable for landslips
exist. I have shown on page 111 that the origin of the Mbenutha cliffs
where agglomerates lie on clayey tuffs may be thus attributed to a
landslip. In the case of the Nangara-ravi cliffs, the occurrence of this
fragment of a large basaltic dyke is of some importance in connection
with the origin of the basic agglomerates of this locality.

The top of the mountain-ridge is about 700 feet above Nangara-ravi, or
2,000 feet above the sea. The tuffs and agglomerates that once existed
here have been stripped off to a great extent and the deeper rocks of
the range are in part exposed. The upper part of this ridge (1,700 to
2,000 feet) is formed of a rubbly pitchstone where a basic glass has
been broken up and then consolidated, the interstices being filled up
with palagonite as described in other cases on page 313. Though
non-vesicular, it is just such a rock as one would expect to find on the
surface of a lava-flow or on the sides of a dyke.

The crest of the range is here only a narrow ridge. I followed it along
in a north-west direction, gradually ascending on the way, and in time
the rubbly pitchstone gave place to a hardened palagonitic clay rock,
which was observed as high as 2,300 feet. It apparently contains a
little lime, and probably was at one time foraminiferous; but it is now
much affected by hydration. Soon after this, we arrived at the foot of
the steep ascent leading to the summit of Va-lili. We were now rather
over 2,400 feet above the sea; but my natives refused to go on, the
heavy rain having made the slope too slippery for a safe ascent. With
much reluctance I retraced my steps; and as the bad weather continued
for several days after, I did not make another attempt. There would,
however, be no difficulty in dry weather.

(2) TRAVERSE OF THE VA-LILI RANGE FROM THE OLD SITE OF LOMA-LOMA TO
WAIWAI ON THE COAST OF SAVU-SAVU BAY.—This route, which was taken by Mr.
Horne, the botanist, in the reverse direction in 1878, is thus described
in his book, _A Year in Fiji_ (p. 19): “The path, rough and apparently
not much used, ran along streams, up steep ascents and down awkward
descents, over slippery boulders and fallen trees, up the sides and
along the crests of densely wooded mountains.”

Ascending the north slope of the range I found at the Tangi-nandreli
cave, which is 1,050 feet above the sea, a palagonite-tuff-sandstone
underlying the basic agglomerate. This tuff, which is of the type
described just below, does not effervesce with an acid, and shows no
tests of foraminifera when examined with a lens. Further up the slopes
large masses of agglomerate occur. At 1,350 feet I came upon a large
mass of a fine-grained compacted palagonite-tuff made up chiefly of
fragments of palagonitized vacuolar basic glass. Among the mineral
fragments occur plagioclase, augite, and rhombic pyroxene, and a little
fresh olivine, which is of very rare occurrence in these
palagonite-tuffs. It contains little or no lime, and shows no tests of
foraminifera in the slide. The summit of the range is here about half a
mile broad, and is relatively level. I placed its elevation at 1,760
feet, which is not far from Horne’s estimate of 1,800 feet. The southern
slope, which is the rainy side of the range, is much cut up into gorges.
In the upper 200 feet palagonite-tuffs, similar to those above referred
to, are displayed, and basic agglomerates occur lower down.... This part
of the range is remarkable through being completely covered over by
palagonite-tuffs and agglomerates. It has been pointed out above that
this is not the case with the range close to the highest peak, where the
underlying rocks are in part exposed at the crest of the range.

(3) THE EASTERN PEAK OF THE VA-LILI RANGE.—This hill, about 1,100 feet
in height, overlooks the Mbale-mbale branch of the Ndreke-ni-wai river.
At its foot near the river there is exposed at the roadside a rubbly
pitchstone formed of a basic glass, inclosing porphyritic crystals of
plagioclase, augite, and olivine, which is described on page 313. Here
also occurs an agglomerate made up of blocks of a semi-vitreous basaltic
andesite (sp. gr. 2·78), showing prismatic pyroxene in the groundmass,
and referred to genus 20 of the augite-andesites.

The upper part of the hill displays the same agglomerate, and a
tuff-agglomerate in which small fragments of the basaltic andesite are
inclosed in a matrix largely made up of fine debris of basic glass.
There protrudes through these detrital rocks at the top of the hill a
broad dyke-like mass of the same basaltic andesite that forms the
agglomerate around; and the structure of the hill is thus displayed as
that of an old volcanic neck. It has evidently an axis of massive
basaltic rocks, more or less covered over with agglomerates and tuffs.

(4) THE NAMBUNI SPUR.—This singular spur runs down to the coast between
Waiwai and Wailevu; but it is partly separated by a deep gap from the
main range. It attains a height of 550 to 600 feet, and has two little
peaks which the natives call Vatu-tolutolu and Vatu-tangitangiri. Its
position is shown in one of the profile-sketches of Va-lili, given on
page 141. The crest of the spur is formed by a dyke-like mass, 25 to 30
feet thick, which is composed of a basic agglomerate passing down into a
palagonite-tuff. The blocks of the agglomerate are composed of a
semi-vitreous basaltic andesite, showing minute felspar-lathes in
flow-arrangement in an abundant smoky glass, the fine pyroxene being not
differentiated. The tuff, into which the agglomerate passes down is
non-calcareous, and displays no organic remains. It is, however,
composed of fragments, which do not generally exceed a millimetre in
size, of palagonitised vacuolar glass, basic andesites, plagioclase,
monoclinic and rhombic pyroxene, &c.

This dyke-like mass forms the axis of the ridge and protrudes vertically
about 100 feet, the bulk of the spur being composed of a compacted
brecciated palagonite-tuff made up mainly of fragments a centimetre in
size, of a basic vacuolar glass, sometimes fibrillar, which is
extensively palagonitised.

The filling up of a fissure in a mass of tuff-breccia by
palagonite-tuffs and agglomerates probably occurred during the
submergence, the original dyke-rock having been removed by marine
erosion. After the emergence the subaerial denuding agencies reshaped
the surface, and as a result of the less yielding character of the
materials filling the fissure, they protrude as a dyke-like mass from
the crest.

In a cliff-face of the adjacent main range there are displayed an
agglomerate of basaltic andesite and a pitchstone-breccia, composed
of fragments of but little altered basic glass, the interstices
being filled up with palagonite. In the case of the Kiombo flow I
have endeavoured to explain the origin of a closely similar
pitch-stone-breccia (page 92).

(5) THE SEA-BORDER AND THE LOW-LYING DISTRICTS AT THE BASE OF THE
VA-LILI RANGE.—It may be generally remarked that palagonite-tuffs and
clays, often foraminiferous, prevail in these localities. Thus in the
sea-border between Waiwai and the mouth of the Ndreke-ni-wai basic
agglomerates are displayed where the mountains approach the coast; but
further west a broad tract of undulating land, elevated usually 100 to
300 feet, intervenes between the range and the sea-border, and here
coarse and fine palagonite-tuffs predominate.... On the north-west the
foraminiferous tuffs and clays of the Ndreketi plains approach the
Va-lili range in the vicinity of Vuinasanga, and extend for at least 200
or 300 feet up its sides.... At the east end of the range, where the
slopes descend to the plains of the Waisali valley, a little west of
Mbale-mbale, there are exposed bedded palagonite-tuffs, tilted up at an
angle of about 20° to the south-west. They contain a little lime and
display microscopic tests of foraminifera, the palagonite being minutely
vacuolar, the cavities also being filled with the altered glass. I
noticed those submarine deposits at an elevation of 100 feet, but
probably they reach much higher.

The inference to be drawn from the data above given concerning the
Va-lili range seems clearly to be this. We have here indicated the
emergence of a submarine mountain-ridge covered over with
palagonite-tuffs and agglomerates, the last being uppermost. These
coverings have been in places stripped off by the denuding agencies and
the underlying massive basic rocks exposed. These rocks, however, vary
much in texture, some being vitreous, as in the case of the pitchstones,
others hemi-crystalline as in the case of the basaltic andesites; and it
is to be gathered from this and other similar indications that different
submarine vents were formed along a fissure or fissures at the
sea-bottom. No evidence of subaerial eruptions came under my notice.
After the vents became extinct they were buried beneath the
palagonite-tuffs and agglomerates. During and after the emergence the
denuding agencies reshaped the surface of the range and left but little
of its original form.

Since it is my object to build up a theory of the origin of the
ridge-mountains as I proceed with the systematic description of the
island, it will be here convenient to follow up the preceding remarks on
the Va-lili Range by a preliminary reference to the great ridge district
lying east of it.

When a panoramic view of this region is obtained, one observes a series
of lofty ridges more or less parallel and running about N.W. and S.E.
There are the Va-lili, Narengali, and Sealevu ridge-mountains with
lesser ridges between. The intervening valleys are elevated about 400
feet above the sea, whilst the mountains rise up to over 2,000 feet. In
many localities this configuration of the surface would be attributed
mainly to subaerial denudation. In this island I will endeavour to show
that these mountain-ridges existed before the emergence. They do not owe
their form to the rivers that flow through the valleys, though no doubt
river-erosion has brought these features into greater relief.

In Vanua Levu, as there will be frequent occasion of showing, rivers
often flow in valleys that they have not made. This is especially
pointed out on page 151; and it is necessary to emphasise it here,
before proceeding farther with the description of the geological
structure of the mountain-ridges.


                           THE WAISALI SADDLE

This saddle, which connects the Va-lili and the Koro-tini ranges, has
probably a minimum elevation of not over 1,200 or 1,300 feet. To
understand this district thoroughly a regular survey is, however,
necessary. It is only at times in this densely wooded range that a view
of the surrounding country is obtained; but in spite of this drawback I
was able by a diligent use of watch, aneroid, and prismatic compass, to
obtain a fair general notion of the surface-configuration.

The track that proceeds westward from Waisali to Narengali leads also to
the villages of Na Sinu and Sealevu. About 1½ or 2 miles from Waisali,
the track branches off to the westward for Narengali and to the
northward for Na Sinu and Sealevu. After half an hour’s walk along this
last-named path, one comes to a place where at an elevation of about 900
feet it branches off to the left for Na Sinu, crossing the lowest part
of the saddle, and to the right for Sealevu across the Koro-tini Range.
It may here be remarked that since the natives are gradually abandoning
their mountain-villages and are settling at the coast, many of the
mountain-tracks used by me will before long be overgrown and forgotten.

In taking the path from Waisali to Narengali one soon enters the hilly
country where large masses of basic tuffs and basic agglomerates, the
last formed of blocks of a compact basaltic andesite, occur on the
surface up to 700 or 750 feet above the sea. The rock just named has a
specific gravity of 2·84, and since it displays rhombic pyroxene amongst
its phenocrysts, it is placed in genus 1 of the hypersthene-augite
andesites. Above this elevation, and as far as the top of the range,
1,800-1,900 feet above the sea, porphyritic basaltic andesites, having a
specific gravity of 2·8, prevail at the surface. They display small
porphyritic crystals of plagioclase, augite, and rhombic pyroxene in a
groundmass composed of small felspar-lathes, prismatic pyroxene, and
much smoky glass, and are referred to genus 5 of the same pyroxene
andesites. It is probable, judging from one of these exposures, that
such rocks are dyke-like masses: but on account of the thick soil-cap it
is not possible to obtain a good view of them.

In the stream-courses occur large blocks of altered basaltic andesites
of the propylitic type, having a specific gravity of 2·64 to 2·70, and
exhibiting abundant alteration products, such as calcite, viridite, &c.
These propylites, I presume, constitute the deeper portion of the range.
It will often be necessary to distinguish between the altered basaltic
andesites, such as are above referred to, and the relatively fresh rocks
of the same type. The former are light coloured (sp. gr. 2·6 to 2·75),
and are only exposed in gorges and stream-courses that deeply score the
mountain-slopes. The latter are blackish (sp. gr. about 2·8), and at
times penetrate the covering of tuffs and agglomerates.

Descending the opposite or north-west side of the saddle-range, one
finds the same basic andesites, both fresh and altered, down to about
1,100 feet above the sea. Then the track leads one down a precipitous
slope into the picturesque gorge traversed by the head-waters of the
Narengali River. At its lower end the gorge opens out into the broad
Narengali valley, and here the dense forest of the higher districts
gives place to the scanty vegetation of the “talasinga” region.

The rocks exposed in the sides of the gorge are basic agglomerates
overlying palagonitic tuffs of mixed composition and evidently
sedimentary. On the bottom lie huge masses, some of them 70 or 80 tons
in weight, of altered grey aphanitic or non-porphyritic
augite-andesites, penetrated in some cases by thin veins of white
quartz, and at times displaying a rudely columnar structure, the columns
being 12 to 14 inches across. Sometimes the alteration is mainly
confined to the filling of the fissures with chalcedonic quartz, minute
nests of the same material occurring in the groundmass. At other times
the small augite granules are also decomposing. The specific gravity
varies from 2·64 to 2·73; the rocks being referred to genus 16, species
A, of the augite-andesites. Occasional detached masses of a propylitic
basic andesite, displaying porphyritic plagioclase and pyroxene, also
occur in this gorge, the felspar phenocrysts being largely occupied by
calcitic and other alteration products, whilst much viridite occurs in
the groundmass. It exhibits both monoclinic and rhombic pyroxene; and on
account of the prism form of the groundmass pyroxene it is placed in the
2nd sub-order of the hypersthene-augite andesites. These altered rocks
are deep-seated intrusive masses that were originally covered over by
the basic agglomerates and palagonite-tuffs exposed in the sides of the
gorge.

Below the gorge there is an extensive exposure in the sides and bed of
the river of light-coloured calcareous tuffs which were originally
composed of palagonitic materials; but owing partly to hydration, and
partly to other secondary changes, the original structure is much
disguised.

Crossing the river in the midst of these tuffs there is a dyke, 15 feet
thick, formed of a propylitic basaltic andesite, a semi-vitreous rock in
which calcitic and zeolitic materials have been developed in quantity.
The dyke, which is not columnar, is steeply inclined at an angle of 45°
to the north-east.... Further down the river-valley as far as Narengali,
occur basic tuff-agglomerates.


                         THE TRACT OF NAKAMBUTA

This is a tract of broken country that projects from the mountainous
backbone of the island (between the Va-lili and Koro-tini ranges) into
the heart of the Ndreketi plains in the vicinity of Natua. As limited by
the 300-feet contour line, it is indicated in the map attached to this
work. Its general level varies between 300 and 600 feet in elevation;
but a number of isolated peaks are included within this area. More than
one of these hills attain a height of 1,000 feet, Nakambuta a very
conspicuous hill being as much as 1,500 feet. Basaltic andesites with
basic agglomerates and palagonite-tuffs prevail.

Towards Natua the basaltic andesites, which are often much decomposed,
are of the doleritic type referred to in the account of the Ndreketi
Plains on page 133. Inland, towards Narengali and Va-lili, these rocks
are often more or less glassy and take the form of pitchstones; whilst
the agglomerates have the same character. The first probably represent
submarine flows of basaltic lava which have spread far and wide over the
Ndreketi plains. The inland rocks are, as is pointed out below, the
products of vents that, as in the case no doubt of Mount Nakambuta, rose
out of a shallow sea. The palagonite-tuffs and clays, often
foraminiferous, which cover the Ndreketi plains, are extensively
represented in the lower levels up to 400 feet or more.

Between one and two miles to the westward of the Narengali valley, and
immediately north of Va-lili, the agglomerates, overlying
palagonite-tuffs, form lofty precipices. The agglomerates are composed
of blocks of more or less vitreous basaltic-andesites, some of them
semi-vitreous and amygdaloidal, some in the form of pitchstone, and
others again as tachylyte that fuses in the lamp-flame. The underlying
palagonite-tuffs are bedded, and are composed of fragments of basic
glass that originally inclosed porphyritic crystals of plagioclase. In
the slide it is observed that the glass and mineral fragments have often
been re-fractured as they lie in the tuff and that the former have
rounded angles and eroded edges. The interstices are filled with a more
or less palagonitised magma. Similar rocks occur in other localities,
and they will all be found described on page 334. It may, however, be
remarked here that in all cases these rocks would seem to have undergone
some crushing, the heat developed in the process being sufficient to
partially remelt the glass. A high temperature was not required to
effect this fusion, since splinters of the tachylyte occurring in the
overlying agglomerate fuse in an ordinary lamp-flame. It is pointed out
on page 341 that tuffs of this character differ in origin and in
characters from the prevailing foraminiferous palagonite-tuffs.

The road from Narengali to Natua traverses the length of this district.
At and near the mouth of the Narengali valley there are exposed basic
tuffs and agglomerates, the blocks in the last case being formed of a
semi-vitreous, vesicular or almost scoriaceous basaltic andesite. In
this neighbourhood the track passes across the top of a waterfall which
is the result of the existence of a huge dyke-like mass of a compact
basaltic andesite showing a little interstitial glass and referred to
genus 13 of the augite-andesites. It lies in a district of tuffs and
agglomerates. Farther on, about two miles north-west of Narengali, the
track crosses some rounded hills, elevated about 600 feet, on the top of
which is displayed a concretionary pitchstone, showing little nodular
concretions of the size of filberts, and having the microscopical
characters of “variolite,” as described on page 313. This is the only
locality of this rock that is known to me.

My acquaintance with the tract of Nakambuta is, however, very imperfect.
But it is apparent that in the pitchstones and in the semi-vitreous
basic rocks, sometimes vesicular and amygdaloidal, we get a nearer
approach to the products of subaerial eruptions than is to be observed
in most other portions of the island. The examination of the Nakambuta
peak by some future investigator will bring to light some interesting
facts concerning this region. It is not unlikely that during a late
stage of the emergence of this region Nakambuta and the other peaks
around protruded as active vents above the surface of a shallow sea, at
the bottom of which the products of their eruptions accumulated.


          THE VALLEY OF THE NDREKE-NI-WAI AND ITS TRIBUTARIES

Ndreke-ni-wai, which signifies “the hollow of the water,” is the name of
a broad tidal estuary, opening into Savu-savu Bay, which is formed by
the union, about half a mile above its mouth, of two rivers, the
Mbale-mbale River flowing from the north-west past a village of that
name, and the Vatu-kawa River, the largest, flowing from the eastward,
which I have also named after a village on its banks. The valleys of
these two rivers are separated by a mountainous dividing-ridge connected
by a saddle with the main range. Its highest peaks rise to 2,100 feet
above the sea, the elevation of this “divide” rapidly decreasing as it
approaches the coast, where, within a mile of the beach, it terminates
in some low hills 200 or 300 feet in height.

It may be observed here that a mouth of the river was originally
situated 700 or 800 yards to the west of its present site. This old
mouth is now represented by a lagoon communicating with the Mbale-mbale
River above, but closed by the sand-mound of the beach at its lower end,
which, however, is occasionally broken through when the rivers are in
flood. This lagoon is shown in the view facing page 153.

The valley of the Mbale-mbale River, which is much the smaller of the
two rivers, is bounded on the north by the precipitous slopes of the
Koro-tini Range, which rise to over 2,000 feet, and on the south and
west by the lofty Va-lili Range. The valley, above the village of
Mbale-mbale, is broad and low-lying; and one can ascend it to the
vicinity of Waisali, three to four miles from the river’s mouth, without
attaining an elevation of 100 feet above the sea. The main stream, which
flows down from Waisali, is joined near Mbale-mbale by a more impetuous
stream that descends the steep mountain-sides just to the east of the
Koro-tini Bluff.

The valley of the Vatu-kawa River is bounded by lofty mountain-ranges
that rise to elevations varying from 2,000 to 3,500 feet. On the south
side lies the Mariko Range, on the east lies Mount Mbatini, the most
elevated peak of the island, whilst on the north rise up the steep
slopes of the Koro-tini Range and of the mountainous “divide.” The
valley has such a gentle gradient that one can follow it inland for five
or six miles from the estuary to the vicinity of Nukumbolo without
exceeding an elevation of 100 feet above the sea. Below Na Salia the
valley is confined between the hills that approach the river; but above
that village it is very broad; and on account of its slight fall the
river here often changes its course, so that the floor of the valley is
strewn with water-worn blocks and pebbles marking the old channels.

The Vatu-kawa River, which rises on the west slopes of Mbatini, flows
with a placid current past Nukumbolo and Na Salia, until it reaches the
village of Vatu-kawa, where it is joined by its impetuous tributary, the
Wai-ni-ngio, “the river of the shark.” This affluent, after descending
the steep slopes of the Koro-tini mountains, bursts through the dividing
range that separates the Mbale-mbale and the Vatu-kawa valleys. It would
seem that the Wai-ni-ngio without any great effort on its part might
become a tributary of the Mbale-mbale River.

The great character of these two valleys, as shown above, is their
little elevation above the sea. For miles inland the level does not
attain 100 feet, and high ranges rise steeply in each case on either
side to 2,000 feet and over. Here, as in the instance of most of the
large valleys of the island, the original configuration of the surface
was not dependent on river-erosion. Rivers no doubt have done much to
carve out the lesser and to deepen and widen the greater valleys; but,
as is often remarked in this work, the main features of the surface were
in existence before the emergence of the island from the sea.

The geological formation of the slopes of these two valleys is described
in the accounts of the various ascents of the mountains bounding them.
Since foraminiferous tuffs occur high up their sides, up to elevations
of 2,000 feet and over, the valleys themselves were at one time no doubt
also covered with these submarine deposits, which, however, have been in
great part stripped off by the denuding agencies. They are still to be
found, containing large tests of foraminifera, between Mbale-mbale and
Waisali; but the basaltic andesites, originally underlying them, are
more frequently exposed. One of these rocks found a little east of
Waisali, which has an aphanitic appearance and a specific gravity of
2·82, is merely a basic glass in its early stage of crystallisation,
being made up of very minute crystallites 1/5000 of an inch in length.
On the surface in this locality there also occur basic agglomerates
containing scoriaceous rocks, the products of some of the last stages of
volcanic action in this part of the island.... In the case of the broad
part of the Vatu-kawa valley above Na Salia blocks of basic rocks
derived from the mountains around strew the bottom in great abundance.
Lower down, where the valley is confined between the hills, basic
agglomerates and coarse tuffs are displayed in the hill-sides.

Mention should be made here of the various hot springs existing in these
valleys in the low levels near the rivers and stream-courses. In the
Vatu-kawa valley they exist at Nukumbolo, and in the Mbale-mbale valley
at Natoarau, Waitunutunu and other localities. These springs are
described in Chapter III.

[Illustration: DUNIUA LAGOON, representing an old mouth of the
Ndreke-ni-wai. Behind rises the Korotini Tableland (2,000-3,000 feet).
The cliff-like declivity over the head of the lagoon is the Korotini
Bluff.]



                               CHAPTER XI

      DESCRIPTION OF THE GEOLOGICAL AND GENERAL PHYSICAL FEATURES
                             (_continued_)


                   THE KORO-TINI RANGE OR TABLE-LAND

THE level-topped range that forms the mountainous backbone of the island
for a distance of nearly 10 miles is one of the remarkable features of
Vanua Levu.[66] In the general profile of the island it is named the
Koro-tini Table-land on account of the level profile which it presents
whether viewed from the north or from the south. But this is merely its
appearance _en masse_. When it is examined in detail it is found that
although much of the range has an elevation between 2,000 and 2,400 feet
above the sea, it attains an elevation of about 3,000 feet in the case
of two gently sloping peaks. With regard also to its table-top, it is
necessary to remark that whilst in some portions of the range the summit
is broad and level, in others it is much cut up into ridges, and in
others again it presents a single narrow crest. Nor can we realise on
looking at the profile the extent to which its slopes have been carved
out by river-erosion, and we get no indication of the several lofty
spurs that descend north and south far into the plains, as in the case
of the spur west of Sueni and in that terminating in the Koro-tini
Bluff. In the profile the eye ignores the details with which the
investigator during many toilsome ascents has filled the pages of his
note-books. To this extent it is useful in that it enables him to rise a
little above the level of his facts, and permits him (to employ a
figure-of-speech) to regard the style and general character of the
edifice without being exclusively absorbed in the study of the bricks.

This range, which extends from a mile or two west of Sealevu to a couple
of miles east of Sueni, is connected on the west with the Va-lili Range
by the Waisali Saddle before described, and on the east with the
Thambeyu or Mount Thurston Range by a broken chain of mountains, of
which Koro-mbasanga is the most conspicuous. It is connected by an
elevated _col_ with Mount Mbatini and the Mariko Range to the southward.
The name of Koro-tini has been applied to this range because it is
familiar to the natives. It signifies “ten towns,” and was given to a
once populous district on the slopes of the lofty bluff overlooking on
the north the mouth of the Ndreke-ni-wai. I crossed the range in four
places, namely, between Waisali and Sealevu, between Mbale-mbale and
Vandrani, between Vatu-kawa and Vandrani, and between Nukumbolo and
Sueni.

(1) TRAVERSE OF THE KORO-TINI RANGE FROM WAISALI TO SEALEVU.—Starting
from Waisali by the Narengali track, I ascended the east slope of the
Waisali Saddle, as described on page 146, until an elevation of about
750 feet was reached, when my way lay to the northward across the
Koro-tini Range to Sealevu. At 850 feet a singular altered tuff was
displayed in position in a stream-course. It shows calcite and pyrites,
and is interesting from the fact that although it is made up largely of
basic glass the tuff does not seem to have undergone the palagonitic
change.

Afterwards, there was a fairly steep ascent to the summit of the range,
2,400 feet above the sea, which has merely a ridge-like crest. Between
an elevation of 1,400 feet and the top there are exposed at the surface
compacted coarse and fine palagonite-tuffs and agglomerate-tuffs formed
of the same materials. They contain often abundant organic remains, such
as valves of “Cardium” and “Pecten” shells, macroscopic tests of
Foraminifera, and some curious scale-like bodies, showing a concentric
structure and about an inch across, which look like fish-scales. It is
probable that these interesting rocks extend to a greater elevation than
2,400 feet, which was merely the highest level reached in the traverse,
but is not the highest point of the range.

These deposits are made up in mass of a more or less palagonitised basic
glass originally containing phenocrysts of plagioclase and pyroxene. The
palagonitic process is nearly always far advanced; but it is seen in all
its stages, the least altered materials fusing under the blow-pipe into
a black glass. The fragments are usually sub-angular in the case of the
coarse tuffs; but small rounded pebbles up to half an inch in size and
fine water-worn gravel are not infrequent. The matrix is composed of
palagonitic debris, portions of crystals of plagioclase and pyroxene,
fine gravel, occasional tests of foraminifera; and it often contains a
fair amount of carbonate of lime, in one specimen tested as much as 13
per cent. The amount of lime, however, varies, being in some places
scanty.

The term “conglomerate” could not be applied to the coarser deposits,
since the sub-angular and angular fragments always predominate. They
could scarcely be deemed “breccias” on account of the mixture with
pebbles and gravel. Their character is therefore intermediate between
the two. I have used the expression “agglomerate-tuff” because it best
describes their appearance. A specimen of such a rock presents a curious
mixture, in the well-compacted mass, of angular and sub-rounded
fragments of palagonite up to an inch in size, small pebbles and fine
gravel of the same material, and detached valves of “Cardium,” entire
and broken. One is forced to draw the inference that these materials
accumulated in shallow water. They are such as might have been produced
by the marine erosion of an emerging volcanic island endeavouring to
hold its own above the waves. But from the occasional occurrence of
blocks of a scoriaceous basaltic rock it would appear that during the
formation of the deposits there were periods of eruption.

At times massive and comparatively fresh-looking basaltic rocks are
exposed _in situ_ on the mountain sides in the midst of these submarine
deposits. A specimen obtained at 1,800 feet is a semi-ophitic
porphyritic olivine-basalt with a specific gravity of 2·86 and showing a
little interstitial glass. The mode of exposure did not admit of my
ascertaining the exact relation of these rocks to the deposits. They are
no doubt dyke-like masses representing the original fissures of eruption
of a submarine vent; and during the emergence they were covered up with
tuffs and deposits, the work of the marine erosion of the emerging land.
These, however, are points on which light will be thrown when we come to
examine other localities.

Descending the northern slopes of the range from the summit to Sealevu
the general course was N.N.E. Several valleys were crossed, of which
that occupied by the Na Sinu river was 600 feet in depth, the rivers and
streams all flowing to the north-west into the Ndreketi basin. Basic
tuffs and agglomerates were exposed at the surface all the way down to
Sealevu, 400 feet above the sea.

At the head of the Sealevu valley, about a mile or rather more above the
village, and a little east of the track followed in the descent above
described, the mountain-range terminates abruptly in lofty cliffs 400 or
500 feet in height. At their base, which is about 1,000 feet above the
sea, once stood the village of Lovutu. These cliffs are formed of basic
agglomerate-tuffs which display a horizontal arrangement, but there is
no distinct bedding. They have the castellated appearance that often
characterises horizontally bedded sedimentary formations. The inclosed
rock-fragments vary in size from 18 inches to half an inch and smaller.
The larger are angular or sub-angular, and are composed of
hemicrystalline basaltic andesites, scoriaceous and vesicular and
sometimes amygdaloidal. The smaller fragments are more or less rounded
and of the same material. The matrix is made up of fine detritus of the
large fragments and of lapilli of a vacuolar palagonitic basic glass,
whilst small crystals of calcite fill the cavities and line the
fissures. The phenocrysts of plagioclase and augite inclosed in the
altered glass also display extensive alteration, and in the first case
are largely replaced by calcite, secondary quartz, and other products.
No organic remains came under my notice; but on account of the
alteration of the tuff-matrix their preservation could hardly be
expected. Bearing in mind, however, the fossiliferous character of the
tuffs and agglomerates in the higher part of the range, it can scarcely
be doubted that the agglomerate-tuffs of the Sealevu cliffs are also
submarine.

Each traverse of the great Koro-tini Range will provide us with new
facts to aid us in framing an explanation of the origin of this long
mountain-ridge. The principal lesson to be learned from the journey
across the range from Waisali to Sealevu, and from the visit to the
cliffs, is concerned with the great extent and thickness of these
submarine basic tuffs and agglomerates. From 1,000 feet above the sea up
to the summit, 2,400 feet in height, they are almost the only rocks
exposed, excepting the occasional masses of basaltic rocks, which
probably represent dykes. Their maximum thickness must amount to some
hundreds of feet.

(2) TRAVERSE OF THE KORO-TINI RANGE FROM MBALE-MBALE TO VANDRANI.—In
this traverse the track before ascending to the summit crosses a spur of
the Koro-tini Bluff, and then descends into the valley of the Natoarau
river on the east side of it. It will therefore be convenient to
describe the bluff before giving my description of the journey across
the range.

The Koro-tini Bluff is a lofty headland (if I may so term it), lying
about four miles inland from the mouth of the Ndreke-ni-wai. It attains
an elevation of about 2,000 feet, and terminates above in a line of
precipices 300 or 400 feet in height. It represents the southern edge of
the level-topped mountain range behind, and like the Sealevu cliffs on
the north side it affords a natural section of its mass. It is shown in
the plate facing page 153, where it rises at the back of the lagoon.

Approaching the bluff from Mbale-mbale, one crosses a low-lying district
less than 100 feet above the sea before striking the spur. Here and in
the lower few hundred feet of the spur are exposed basic agglomerates,
and occasionally in the mass a semi-vitreous vesicular olivine-basalt,
almost like a pitchstone, and displaying large porphyritic crystals of
plagioclase, 5 or 6 millimetres long, the agglomerates being made up of
the same material. Higher up, at elevations between 1,000 and 1,500
feet, are exposed coarse palagonite-tuffs made up of fragments, usually
1 to 3 mm. in size, of extensively palagonitised basic vitreous rocks,
such as occur in the cliffs above. These tuffs become coarser as one
approaches the precipitous bluff, the base of which lies about 1,650
feet above the sea. Here the cliffs present a bare rocky face, some 200
feet high. The lower portion is composed of an agglomerate-tuff, and the
upper portion mainly of agglomerates. These deposits display no bedding
excepting a single plane of division inclined steeply to the north at an
angle of perhaps 40°.

The blocks in the agglomerate-tuff are either angular or sub-angular,
and are less than a foot across. They are all composed of more or less
vitreous porphyritic olivine-basalts, showing large crystals of
plagioclase a fifth of an inch (5 or 6 mm.) in length. But they vary
somewhat in character. Some of them, that are vesicular and almost
scoriaceous, may be termed from their glassy nature porphyritic
pitchstones. Others again, where the groundmass is hemi-crystalline, may
be designated porphyritic compact basalts, and are referred to genus 37
of the olivine-basalts.

The matrix of the agglomerate-tuff is made up of angular fragments, up
to 5 mm. in size, of singular vitreous and semi-vitreous
olivine-basalts, in part palagonitised. There is evidence of crushing
_in situ_ of some of the porphyritic felspar crystals; but it is not so
marked as elsewhere noticed. The palagonite is also in part
interstitial, a character that goes to support the view advanced on page
342, that the palagonite may be connected in its origin with the heat
developed during crushing, only a moderate temperature being required
for the partial fusion of the glass.

In crossing the range by this route from Mbale-mbale one first ascends,
as above observed, the spur of the Koro-tini Bluff up to a height of
1,200 feet. The track then descends into the valley-gorge of the
Natoarau river on the east, the bottom of which is 750 feet above the
sea, and from here the climb begins. One ascends the bed of the stream
course, clambering over slippery rock surfaces up to 1,200 or 1,300
feet, where the stream is left, and the mountain-slopes, often steep and
precipitous, are then followed to the summit, 2,000 feet in height.
Coarse and fine palagonite-tuffs and agglomerate-tuffs of the same
character are exposed on the surface from the commencement of the ascent
up to 1,850 feet; but they are displayed much more extensively in the
stream-course than in the soil-covered upper slopes.

The tuffs are grey except when hydrated, when they turn yellowish-brown.
Some of them contain lime, as much at times as 10 or 12 per cent.;
whilst others possess little or none. Tests of foraminifera are not
infrequently inclosed, even as high as 1,850 feet. A description of one
of these tuffs containing a few tests of Globigerina, which was obtained
at 1,200 feet, is given on page 331, under sample D. It will be there
seen that they are derived from different basic rocks, some containing
but little glass, others mainly vitreous, only the more glassy
constituents being palagonitised. The palagonite-tuff sandstones exposed
in large blocks on a bare spur at 1,850 feet contain 12 per cent. of
lime, the largest tests of foraminifera being not over half a
millimetre.[67] These tuffs occasionally show bedding. At 1,000 feet
they dip gently to the S.S.W., and at 750 feet they are inclined about
15° in the same direction. In this last locality they consist of
alternating layers, 1 to 4 inches in thickness, of fine and coarse
tuffs, the coarser looking like sandstone.... The blocks in the
agglomerate-tuff are sub-angular, and of an olivine-basalt with
hemi-crystalline groundmass,[68] their size ranging from 2 feet to an
inch. I noticed one large block of this rock, measuring 2 × 1½ × 1 feet,
imbedded alone in the tuffs at 1,200 feet. At one place a tuff
containing small fragments of basalt displayed a concretionary
structure, indicating probably the proximity of a dyke, the globular
masses being 4 feet across. A little lime occurs in the matrix of the
agglomerate-tuff.

The summit of the range, 2,000 feet in height, is “ridgy,” about half a
mile in width, and cannot therefore be described as table-topped. The
rocks exposed in blocks on the surface are composed of a semi-ophitic
olivine-basalt containing a large amount of interstitial glass which
shows the fibrous crystallites of the early stage devitrification. It is
referred to genus 33 of the olivine-basalts.

Descending the northern side of the range I followed the steep slopes
down to 1,000 feet above the sea. A rubbly doleritic olivine-basalt,
semi-ophitic, and assigned to the same genus (33), prevailed on the way;
and it is probable that a waterfall with a drop of 50 feet or more that
is situated on these slopes indicates a large intrusive mass of this
rock. During the rest of the descent to Vandrani, which lies in a valley
at the foot of the range, and is elevated about 300 feet above the sea,
basic agglomerates and palagonite-tuffs, together with deposits
intermediate in character, were exposed at the surface. At times a
semi-ophitic doleritic basalt similar to those displayed above, but
without olivine, occurred in position. The blocks in the agglomerates
are formed of a compact semi-vitreous basaltic rock, and are sometimes
vesicular. At one place the palagonite-tuffs exhibited signs of
alteration, being traversed by small fissures not over a third of an
inch broad (5 to 8 mm.), and filled with a zeolite behaving like
natrolite.

In some cliffs by the river at Vandrani are displayed fine and coarse
non-calcareous palagonite-tuffs, bedded and dipping about 15° N. by W.
They are penetrated by cracks, 5 mm. in breadth, which are filled with
chalcedony. These tuffs are evidently in part derived from acid as well
as from basic rocks, though mainly from the latter; and they show other
alteration-characters. At the mouth of the Vandrani valley there are
exposed in the river-bed coarse palagonite-tuff sandstones containing a
little lime, and probably a few tests of foraminifera.

Reference may here be made to the mountain of Ravi-koro which, when seen
from the north-east, rises up as a partially independent peak, with a
broad base and a conical truncated summit, immediately west of the track
followed in the descent from the summit of the range to Vandrani. It is
probably not much under 2,000 feet in height, and exhibits bare
precipitous cliff-faces on the north side. It would be worthy of the
attention of the future investigator.

Recurring to the principal features of the range between Mbale-mbale and
Vandrani, one may remark the extensive occurrence of basic agglomerates
and tuffs on both slopes, the prevalence of olivine-basalts, the
frequency of the semi-vitreous and vitreous or rather pitchstone
condition of these rocks, and their semi-ophitic character, especially
on the summit and north slopes. From the vesicular structure of the
rocks of the Koro-tini Bluff and from the character of its tuffs and
agglomerates, it is to be inferred that they are the direct products of
eruptions, probably in shallow seas. On the other hand, the tuffs (often
foraminiferous) as well as the agglomerate-tuffs of the north and south
slopes of the range are in part suggestive of marine erosion. Intrusive
masses of basalt are to be observed occasionally, and doubtless to this
cause may be attributed the concretionary structure of the tuffs in
places, and the alteration of these deposits in one or two localities,
where they are penetrated by cracks filled with chalcedony.

(3) TRAVERSE OF THE KORO-TINI RANGE FROM VATU-KAWA TO VANDRANI.—On
leaving Vatu-kawa[69], which is not more than 50 feet above the sea, the
ascent for the first 600 feet up the steep mountain-side lies along the
rocky bed of the Wai-ni-ngio River, which from its rapid fall has more
the character of a torrent. On its sides are exposed basic agglomerates
and agglomerate-tuffs; whilst the large boulders in its bed are composed
of a somewhat altered olivine-basalt. At 600 feet the track abandons the
stream-course for the steep mountain slopes, and thence up to 1,100 feet
similar agglomerates and tuffs prevail. At this last-named elevation
there are displayed fine and coarse indurated palagonite-tuffs, a little
altered in character and with little or no lime. A specimen of the
former, of which the materials composing it do not exceed ·2 mm. in
size, shows in the slide an occasional “Globigerina” test filled with
palagonitic debris. Such a marine deposit is evidently not of
shallow-water origin. The coarser tuff is made up of compacted
sub-angular fragments, not over 2 mm. in size; but contains no organic
remains. The prevailing rocks exposed between 1,100 and 1,900 feet, a
little below the summit of the range, are somewhat altered compacted
non-calcareous breccia-tuffs, composed of sub-angular fragments 5 or
6 mm. in size, of a more or less glassy and often vacuolar basic or
basaltic andesite, only in part palagonitised, the vacuoles as well as
the interstices between the fragments being sometimes filled with a
zeolite.[70]

The summit of the range may be described as a “ridgy” table-land. Though
about 2 miles in breadth, its level only varies between 2,000 and 2,200
feet, the inequalities being probably the effect of denudation. Here, as
in many other similar localities, on account of the dense forest it was
only possible to determine the surface-configuration by the use of
compass, watch and aneroid. The prevailing rocks displayed in this
region are grey non-calcareous basic tuffs, somewhat altered in
character, and composed of fragments usually not exceeding 1 mm. in size
of a basic glass, the palagonitic process being masked by other changes.
These tuffs often become brownish-yellow through hydration. Tests of
foraminifera are enclosed, but they are very scanty.

On the north slopes basic agglomerates and palagonite-tuffs are the
predominant rocks down to the foot of the range. A specimen of the tuffs
taken at 1,300 feet is calcareous in patches and probably contains tests
of foraminifera; but it is too much weathered to enable one to speak
with certainty on this point. The interesting feature of this slope is
the exposure at 1,600 to 1,700 feet of large blocks of a dark grey
hypersthene-augite andesite referred to the orthophyric order of those
rocks described on page 290. Lower down (1,000-1,300 feet) occasional
solitary blocks of the same rock, but somewhat altered, occur imbedded
in the palagonite-tuff. This type of rock which is characterised by the
orthophyric structure of the groundmass and by other features is rarely
represented in Vanua Levu.

Summing up the general results of this traverse we observe that here, as
in other parts of the range, basic agglomerates, breccias, and tuffs,
the last however scantily foraminiferous, occupy a great extent of the
slopes and summit. The alteration of these deposits on the southern
slopes is noteworthy. The only deeper seated massive rocks observed were
the pyroxene-andesites above alluded to.

(4) TRAVERSE OF THE KORO-TINI RANGE FROM NUKUMBOLO TO SUENI.—The hot
springs at Nukumbolo, which are described on page 24, rise up through
agglomerate-tuffs. Around the bathing pools lie large masses of altered
palagonite-tuffs which give the first indication of the region of
altered rocks that extends from Nukumbolo to the lower slopes of the
range, a distance of about three miles.

For about a mile and a half or two miles from this place the track lies
through a broken country and does not rise to a height more than 300
feet above the sea. A variety of altered rocks are here exposed in
position in the stream courses. Some of them are fine and coarse basic
tuffs showing secondary calcite, quartz and opal, as alteration
products. Others are palagonite-breccias with the vacuoles of the
altered glass filled with opal. Others again are massive basic rocks,
such as fine-textured augite-andesites, or doleritic basaltic andesites,
semi-ophitic in character, the plagioclase phenocrysts being more or
less occupied by calcitic and other products. The alteration is not
always far advanced, but it is sufficiently marked to give a common
character to the rocks of the district.

Ascending the lower slopes of the range up to 800 feet one finds the
altered rocks still exposed in the stream-courses; but the changes
exhibited are not always the same. A specimen from 500 feet looks like a
tuff, but in the slide it appears as a semi-vitreous augite-andesite,
its substance being penetrated by fine veins of chalcedonic quartz and
opal, whilst the same material is developed within the larger
plagioclase crystals. Another specimen from 800 feet, which is
apparently a tuff, contains so much lime that it effervesces freely with
an acid. It was composed originally of fragments of a hemi-crystalline
basic rock, of which the plagioclase phenocrysts have been replaced by
calcite; whilst the augite and interstitial glass is now represented by
viridite and a chloritic mineral. It is to be inferred that at some time
hot springs were very numerous in the district between Nukumbolo and the
lower slopes of the range, those at Nukumbolo, as far as I know, alone
existing in our time.

From a height of 1,100 or 1,200 feet the mountain slopes rise steeply to
the summit rather over 2,000 feet in elevation. At the foot are exposed
_in situ_ aphanitic augite-andesites,[71] which in some specimens show a
little alteration in the chalcedonic quartz filling minute cracks, and
in one case there is an irregular cavity, ¾ inch across, filled with
milk-white opal. Another rock exposed at the foot of the steep ascent is
a semi-vitreous basaltic-andesite, doleritic in texture and ophitic
structure, but apparently not much changed.[72] At 1,700 feet is
displayed a vesicular basic andesite, semi-vitreous in character, and
above this I found a porphyritic basaltic andesite.

The summit of the range is 1½ or 2 miles in breadth and is relatively
level, its undulating surface varying in elevation between 1,900 and
2,200 feet. The prevailing rocks exposed on this elevated plateau are
vitreous pitchstone-like rocks finely vesicular and scoriaceous, the
cavities being filled either with aragonite or with opal. The basic
glass, of which they are formed, shows incipient crystals, and begins to
fuse in an ordinary flame. One specimen obtained here is a doleritic
basaltic andesite, slightly ophitic and containing a fair amount of
residual glass.[73] However, the vitreous and scoriaceous character of
most of the rocks on the summit is very remarkable. (Similar rocks occur
on the top of Mount Thambeyu where the slopes of the mountain are
covered with submarine tuffs and agglomerates.) There is a precipitous
descent on the north side of the range to Sueni at its foot, massive
basaltic andesites being exposed at first, whilst basic tuffs and
agglomerates are displayed lower down.

The special features of this traverse of the range are the alteration of
the tuffs and massive rocks between Nukumbolo and the lower southern
slopes, the variation in character of the basic rocks in the upper
southern slopes, the occurrence of vitreous vesicular and scoriaceous
rocks on the summit, and the restriction of the ordinary basic
agglomerates and tuffs to the northern slopes. Any attempt on my part to
explain the structure of this part of the range from the data here given
would be futile without comparing them with those obtained from other
parts of the range. It will be subsequently pointed out that the
difficulties will be in part removed if it is assumed that the submarine
palagonite-tuffs and agglomerates, that so often cover the flanks of the
mountains to their summits, have been in this case largely stripped off
by the denuding agencies.

(5) THE SUENI VALLEY.—My acquaintance with the extreme eastern part of
the Koro-tini Range is restricted to the descent of the picturesque
valley from Sueni to Koro-utari. It is occupied by a tributary of the
Lambasa River, and is bounded on the east side by the lofty slopes of
the main range, and to the westward by a mountainous spur that projects
far into the Lambasa plains. Sueni lies by the river-side in the midst
of mountains which rise steeply on most sides to heights of 2,000 feet
and over, and often display precipitous bare faces apparently of
volcanic agglomerates. Numerous waterfalls may be observed on their
flanks, which, as in other localities, doubtless indicate the occurrence
of large intrusive dykes. Sueni is situated about 300 feet above the
sea, the descent to Koro-utari at the mouth of the valley, a distance of
3 to 4 miles in a direct line, being about 150 feet.

The river as it flows down the valley from Sueni to Koro-utari traverses
a region of basic agglomerates and agglomerate-tuffs. These deposits, as
they are displayed in the hill-slopes lying W.S.W. and at the back of
Sueni, are composed of blocks of the size of the fist of a vesicular
basaltic andesite; whilst the large masses on the surface are made of
the same, but non-vesicular, rock. The blocks in the agglomerates
between Sueni and Koro-utari range usually from a few inches to a foot
in diameter. A specimen obtained from one of them is made of a partly
vitreous basaltic andesite; whilst in another case the rock is an
altered basic andesite, the glassy groundmass being largely impregnated
with colloid silica looking like opal under the lens.[74]

Nearly a mile below Sueni, within a space of less than 60 yards, there
are exposed at the river-side in the agglomerates three vertical or
nearly vertical dykes, 4 to 6 feet in thickness. They trend roughly N.E.
and S.W., and are non-columnar, except in the case of the one farthest
up the river, which has rude, transverse joints.[75] The rocks composing
these dykes are somewhat doleritic basaltic andesites, olivine being
very rare or absent. The two highest, which are only 15 to 20 feet
apart, are made of similar rocks characterised by abundant interstitial
glass, and having a sp. gr. of 2·78. The rock of the third dyke, about
50 yards farther down the river, has but scanty glass in the groundmass,
the sp. gr. being 2·89. The differences between the two types
represented in the three dykes are mainly concerned with the degree of
crystallisation, and it is probable that though not contemporaneous they
were derived from the same fluid magma which, as we may infer from the
proximity and lie of the dykes, was situated at no great depth.[76]

GENERAL INFERENCE RESPECTING THE KORO-TINI RANGE.—If we can imagine a
line of vents, protruding in some cases above the surface of the sea,
that were ultimately worn down to a common level through marine-erosion,
and were then largely covered over with submarine tuffs and
agglomerates, we should have in our mind’s eye the first and most
important stage in the formation of this range. If we then assume that
there followed a period of emergence characterised by a renewal on a
very extensive scale of marine-erosion, during which the agglomerates
were mainly formed, and that since that period the sub-aerial denuding
agencies have been for ages in operation, we shall, I think, obtain some
idea of the history of the Koro-tini Range.



                              CHAPTER XII

      DESCRIPTION OF THE GEOLOGICAL AND GENERAL PHYSICAL FEATURES
                             (_continued_)


                        THE KORO-MBASANGA RANGE

AS is illustrated in the accompanying profile-sketch, the relatively
level-topped range of Koro-tini gives place at its eastern end to a
broken line of mountains, of which the round-topped Koro-tambu, 2,753
feet in height, and the pinnacled Koro-mbasanga,[77] 2,537 feet, are the
highest peaks. Further east lies the broad Vuinandi Gap which separates
the Koro-mbasanga and Mount Thurston, or Thambeyu, ranges. The
twin-peaks of Mount Mbatini, the highest mountain of the island, appear
in the background in the sketch, and to the left rises Thambeyu, the
second highest summit.

We enter here another complex region of mountains; and if the character
of the rocks are sometimes different we shall yet have to bear in mind
in our interpretation of its geological features the lesson derived from
the examination of the Koro-tini Range. Before and behind all our facts
of observation lie the two great periods of marine-erosion and the later
ages of sub-aerial denudation.

When approached from the north, the western part of the range has a rude
crescentic form, and looks like the remnant of a gigantic crater-cavity
about two miles across. At the back rise, as shown in the second of the
profile-sketches, the precipitous slopes of Koro-mbasanga proper; whilst
the two spurs descending from it, one on the west, the Sokena spur,
towards Koro-utari, the other on the east, at the back of Nasawana, give
the crescentic figure open to the north. The last-named village lies
nestled in this great hollow, the floor of which, though not in its
lowest part below the level of the Lambasa plains, is not over 200 feet
above the sea. However, the facts adduced in the following description
of this region do not give much support to this view of its
surface-configuration.

For the convenience of description, I will first describe the peak of
Koro-mbasanga, and then the Sokena ridge and, lastly, the Lovo valley
that cuts through the range to the eastward.

[Illustration: These three sections form a continuous profile-sketch of
the mountainous axis of Vanua Levu for a length of 15 or 16 miles and
include the Thambeyu, Koro-mbasanga, and Koro-tini ranges as viewed from
the northward near Na Kama. The eastern section is at the top and the
western section at the bottom. The summit of Thambeyu was covered with
clouds.]

[Illustration: Koro-mbasanga from the north-north-east.]

(1) KORO-MBASANGA.—The ascent of Koro-mbasanga is best made from
Nasawana, a village at its base, elevated rather over 200 feet above the
sea, and distant about a mile and a half north-east from the peak. On
the way to the foot of the mountain we traverse an undulating region of
basaltic andesite,[78] which is merely the extension to the base of the
mountains of the basaltic Lambasa plains. After commencing the ascent of
the steep slopes, we find exposed in a stream-course, 700 to 800 feet
above the sea, a sedimentary basic tuff, presenting layers of coarse and
fine materials, and partly palagonitic in composition. It is a little
calcareous, and apparently incloses tests of foraminifera. These
submarine deposits have evidently been stripped off the basaltic
low-lands beneath. It thus becomes evident that the structural features
of the Lambasa plains (basaltic rocks overlain by submarine deposits)
are preserved to the base of the range.

These submarine deposits, as exposed in the stream-course, lie beneath
agglomerates which repose horizontally upon them; and from this level up
to the bare rocky peak of the mountain, agglomerates and
agglomerate-tuffs are alone displayed either as large detached masses or
in cliff-faces.

In the lower part of the mountain the blocks, usually sub-angular, are
about a foot across; but they become smaller as one ascends towards the
summit, where they are 3 or 4 inches in diameter. At the top there are
extensive exposures in cliff-faces of the agglomerate-tuffs; and here
the finer materials of the matrix include a few rounded pebbles not
exceeding half an inch in size. This is a fact of importance in
connection with the submarine origin of these formations.

As regards their composition, the blocks of the agglomerates have not
the uniform character we would expect to find in the case of materials
directly ejected from a volcanic vent. The most frequent type of rock
represented is a grey hypersthene-augite-andesite, having a specific
gravity of 2·72-2·78. It displays small phenocrysts of plagioclase and
of rhombic and monoclinic pyroxene, but in other respects it exhibits
much variety, not only in the arrangement and average length of the
felspar-lathes (·08 to ·18 mm.) but in the form of the pyroxene of the
groundmass (either granular or prismatic) and in the amount of residual
glass, sometimes abundant, sometimes scanty. Two distinct genera (1 and
5) of the sub-class are therefore represented.

Other rocks found in these agglomerates contain no rhombic pyroxene, and
are referred to genera 13 and 16 of the augite-andesites according to
the presence or absence of plagioclase phenocrysts. In the last case we
have a dark aphanitic rock (sp. gr. 2·74), sometimes scoriaceous, where
the average length of the felspar-lathes may be as little as ·04 mm. One
of the blocks was composed of a highly scoriaceous semi-vitreous rock,
the cavities being filled with a zeolite. Another was composed of a
black porphyritic augite-andesite, showing large crystals of
plagioclase.... The matrix of the agglomerate-tuff is formed of
sub-angular and rounded fragments, up to a centimetre in size, of the
same andesites, the interstitial material being formed of fine detritus
and palagonitic debris.

Though the agglomerates of the peak of Koro-mbasanga are composed of a
variety of rocks, all the rocks are to be referred to the
pyroxene-andesites with specific gravity below 2·8 but above 2·7. They
are therefore less basic than the olivine-basalts and basaltic andesites
of the Koro-tini range, where the density is usually 2·8 and over. Their
variation, however, is more consistent with the characters of an
agglomerate formed by marine erosion. The same may be said of the
sorting of the blocks according to their size and of the occasional
occurrence in the matrix of small rounded pebbles. That these deposits
of agglomerates were formed under the sea is indicated also by their
overlying submarine sedimentary tuffs near the base of the mountain.

(2) THE SOKENA RIDGE.—To the west of Koro-mbasanga, and forming a spur
of the same range, is the flat-topped hill of Sokena, which rises about
1,100 feet above the country at its base and about 1,600 feet above the
sea. From a distance it has the appearance of being formed in its higher
portion of nearly horizontal strata dipping gently northward. In its
upper part it terminates in a line of cliffs about 200 feet in height,
and there is a similar line of cliffs lower down the slopes. These
cliffs are composed of bedded fine and coarse non-calcareous tuffs,
dipping about 10° N.N.W., in which are imbedded without any arrangement
blocks, ranging in size from 2 or 3 inches to 3 or 4 feet, of a
remarkable blackish pitchstone-like rock displaying opaque plagioclase
phenocrysts. It is referred to genus 18 (see page 289) of the
hypersthene-augite andesites, both rhombic and monoclinic pyroxene being
represented in the phenocrysts and in the groundmass where they take the
form of minute prisms (·03 mm.). There is a considerable amount of pale
brown glass. A rock very similar occurs in the Thambeyu agglomerates
(see page 178). The tuffs are formed largely of palagonitic materials,
the angular fragments in the coarser beds being ½ to 2 centimetres in
size, the palagonite being often vacuolar but much affected by
hydration.

These tuffs and agglomerates of the Sokena cliffs apparently contain no
organic remains. They appear to have accumulated under water as the
result of the eruptions of a neighbouring vent without the intermediate
agency of marine erosion.

(3) THE ASCENT OF THE LOVO VALLEY.—About two miles to the east of the
peak of Koro-mbasanga the picturesque Lovo valley cuts deeply in a
southerly direction into the mountainous backbone of the island. The
site of the old town of Lovo lies within the valley about two miles from
its mouth. “Lovo” is the Fijian word for a cannibal-oven; and I gathered
from my natives that in the old times this vale was noted for its
cannibal orgies. It is occupied by the Nasawana tributary of the Lambasa
River, and often becomes so narrow that it may be described as a gorge.
I followed the valley from its mouth, where it is elevated about 300
feet above the sea, for some miles in a southerly direction up to a
height of 1,000 feet, where the northern slope of the great
mountain-mass of Mbatini commences.

On either side of the Lovo valley rise precipitous mountain-slopes,
displaying in their cliff-faces and in the large detached rock-masses
basic agglomerates. The same formation is also usually displayed in the
sides of the river. The blocks composing the agglomerates are formed of
the usual type of hemi-crystalline or semi-vitreous blackish basaltic
andesite so characteristic of these deposits. It is generally compact,
but is at times amygdaloidal. Some distance below the old site of Lovo,
and at an elevation of about 500 feet above the sea, there is an
interesting exposure in the river-side, where the agglomerates overlie
bedded coarse calcareous basic tuffs containing large flat tests of
foraminifera with pieces of molluscan shells, and dipping about 15° S.W.
These tuffs can be traced up the valley towards Lovo.

Displayed in mass in the bed of the river in the same locality, and
beneath the submarine tuff just referred to, is a porphyritic basaltic
andesite (sp. gr. 2·79) containing but scanty interstitial glass, the
felspar-lathes being ·15 mm. in average length. It is referred to genus
1 of the augite-andesites. The same rock is exposed at intervals in the
river-bed as far as Lovo, which is about 850 feet above the sea. At one
place it exhibits a rudely columnar structure, the columns being
horizontal and 2 to 2½ feet in diameter, the trend of the dyke-like mass
being W. by S. and E. by N. Near Lovo a small dyke, 6 feet thick and
trending N.N.W. and S.S.E., pierces the agglomerate. It is composed of a
somewhat aphanitic augite-andesite closely resembling the rocks exposed
in the river-course for a mile or so above Lovo up to an elevation of
1,000 feet. In this upper part of the valley whilst agglomerates are
exposed in the cliffs and precipitous mountain-slopes on either side,
pyroxene-andesites, somewhat aphanitic in texture and with a specific
gravity of 2·68 to 2·7, are displayed in mass in the river bed. These
last-named rocks, which are closely similar to those found on the lower
slopes of Mount Mbatini (see page 173) are, as I should have also
remarked in the case of the basaltic andesite above mentioned, a little
altered, as is indicated by the existence of calcite and viridite in the
groundmass.[79]

From this instructive ascent of the Lovo valley we may learn that whilst
the mountain mass is formed, to a considerable depth, of agglomerates
with underlying submarine tuffs, the deeper seated rocks exposed in the
river-beds are massive intrusive rocks. The overlying agglomerates have
preserved the submarine tuffs from destruction, and there is no
difficulty in assuming that they also were accumulated under the sea,
but in shallow water, as evidenced by the character of the tuffs. I
found no signs of alteration in these tuffs, and except in the case of
the small dyke above noticed there is no sign of the dykes penetrating
the agglomerates. We have here a section into the heart of the
mountain-range; and assuming that the large intrusive masses of basic
andesites had penetrated these deposits, there would certainly have been
some evidence of this in the extensive exposures of agglomerates far up
the mountain-sides. As it is, however, we find such rocks only in the
deeply excavated river-bed. If we imagine a submarine volcanic mountain,
or one but slightly raised above the surface of the sea, to be subjected
during a long period of emergence to marine erosion, the “basal wreck”
of the mountain would ultimately be covered over by submarine tuffs and
agglomerates. This is the condition that seems to be presented here.

I did not make the ascent of Koro-tambu, the other principal peak of the
Koro-mbasanga Range. This round-topped mountain is well seen from the
summit of Mbatini from which it bears N. 30° W. by compass. It is
probably the peak marked 2,753 feet in the Admiralty chart, and is
connected with Mbatini by a saddle not under 1,500 feet in elevation.

Some of the most important features in the above account of this
district may here be emphasised. We have seen that in the peak of
Koro-mbasanga and in the Lovo valley agglomerates and agglomerate-tuffs,
several hundred feet thick, overlie sedimentary submarine tuffs. In the
last-named locality the deeper massive basic rocks are also exposed; and
we may infer in both instances that the agglomerate-formation is a
submarine deposit. On the other hand, in the Sokena Ridge, which is a
spur of the main range, we have apparently the accumulation of materials
on a sea-bottom, directly ejected from a vent without the intervention
of the agency of marine-erosion. In regard to this and other districts
in this part of the island it should be remembered that east and west
occur undoubted evidences of extensive submergence. It has already been
shown that submarine tuffs containing tests of foraminifera and other
organic remains occur at heights of 2,000 feet and over on the summit of
the Koro-tini Range, and it will be subsequently shown that similar
deposits are to be found on the neighbouring slopes of Thambeyu as high
as 2,100 feet.


                             MOUNT MBATINI

According to the Admiralty chart this is the highest mountain in Vanua
Levu, its elevation being 3,437 feet. It has twin peaks which lie either
N.W. and S.E. or W.N.W. and E.S.E. with each other. The northerly or
westerly peak is pointed and tooth-like. Hence probably arises its name
of Mbatini (mbati-tooth). The southerly or easterly peak is known as
Soro-levu. It has a broadly conical outline with a truncated summit. The
mountain is named Koro-mbasanga in the Admiralty chart, a name that
really belongs to a peak lying about 3 miles nearly due north (N. 5°
W.). The natives are very clear in this matter; but it must be remarked
in this connection that Koro-mbasanga, which signifies “a forked
eminence,” would be a very suitable appellation for the double-peaked
summit of Mbatini.[80] By the natives of the surrounding district the
whole mountain is known as Mbatini; but by the natives of the eastern
shores of Natewa Bay, it is usually known as Soro-levu, since the
western peak is often more or less hidden from view or is less
conspicuous. The profile of this mountain and of the neighbouring region
is shown in the accompanying profile-sketches and also in one of those
illustrating the Koro-mbasanga range on page 167.

As viewed from the top of Mariko to the southward, Mbatini presents
itself as a long mountain-ridge, trending W.N.W. and E.S.E., which is
connected on the north with Koro-tambu, the highest peak of the
Koro-mbasanga Range, by a saddle probably not over 1,500 feet above the
sea, and on the south with the mountain-ridge of Mariko by a _col_ which
appears not to be under 1,000 feet in elevation.

[Illustration: Mount Mbatini from the top of Koro-mbasanga. The distant
peak on the right is one of the summits of the Mariko ridge.]

[Illustration: View from Muanaira on the south coast of Natewa Bay.]

My ascent of this mountain was made from the north by the way of the
Lovo valley. In ascending the Lovo valley one reaches, at an elevation
of about 1,000 feet, the foot of the north slope of Mbatini. The slope
is somewhat steep up to 2,000 feet, the rocks exposed on the surface
being closely similar in the groundmass to those displayed in the upper
part of the Lovo valley. They are compact-looking blackish
augite-andesites (sp. gr. 2·7), the very small felspar-lathes of the
groundmass, which are in flow arrangement, averaging only ·05 mm. in
length. Like the rocks below, they are a little altered; and here the
interstitial glass is also scanty. But they differ in the absence of
rhombic pyroxene and are therefore referred to the augite-andesites
(genus 13).

At 2,000 feet, where one crosses the foot-track from Nukumbolo to
Korolau, the ascent of the true Mbatini ridge begins, the summit lying
nearly two miles to the south-east. Whilst following along this lofty
mountain-ridge we were for the greater part of the time in the
rain-clouds, so that very little was seen of our surroundings. The crest
is densely wooded so that our progress was very slow. The rocks are but
sparingly exposed. At the commencement of the ridge (2,100 feet) is
displayed an altered hypersthene-augite andesite, rudely columnar blocks
of which, up to 2 feet in diameter, were lying about. It belongs to
genus 1 of this sub-class (see page 286) which also includes the rocks
exposed farther along the ridge. In these rocks the felspar-lathes are
small (·05-·07 mm. long) and are not in flow arrangement. The
interstitial glass varies in amount, and the specific gravity is about
2·7.

The ascent is very gradual for the first one and a half miles, when an
elevation of 2,600 feet is attained. From here one ascends the
steep-sided peak of Mbatini, which rises some 700 or 800 feet from the
ridge. As one nears the highest point the crest becomes very narrow,
between 15 and 20 feet across; and on either side there is apparently a
drop of several hundred feet. The actual peak, which is bare and rocky,
is yet narrower; and when it is enveloped in dense mist as it was in my
instance, it is not a very secure situation for a geologist. It is
highly magnetic, as is the case with most of the other bare peaks of the
island. The rocks exposed in the upper 500 feet, that is, in the peak
proper, are highly altered semi-vitreous, but extensively weathered,
hypersthene-augite-andesites which are referred to genus 1 of that
sub-class. Much of the glassy groundmass is replaced by viridite,
silica, calcite, &c. Less altered specimens display in a brown opaque
glass small felspar-lathes averaging less than ·1 mm. in length. They
exhibit phenocrysts of rhombic pyroxene and augite, the first
prevailing.

I did not climb Soro-levu, the other of the twin-peaks. Its ascent
should be made either from Nukumbolo or from one of the villages on the
neighbouring shore of Natewa Bay. My acquaintance with Mbatini, although
very incomplete, enables me however to point out a few of its general
features. As remarked before, there is a general uniformity in the type
of its rocks. The olivine-basalts and basaltic andesites, prevailing in
the Koro-tini Range, are not here represented, nor are the dacites or
acid andesites to be found. The characteristic rocks are more or less
altered hypersthene-augite-andesites having a specific gravity in the
least altered and least vitreous condition of about 2·7; whilst the
average length of the felspar-lathes is always less than ·1 mm. The same
type prevails from the upper part of the Lovo valley to the summit of
Mbatini; but it is only in the actual peak that these rocks show much
glass in the groundmass, though extensively affected by alteration.
Neither tuffs nor agglomerates came under my notice; but they might be
expected to occur on the other slopes. I am inclined to regard this
mountain-ridge as a huge dyke-like mass or sill, representing the
remains of a volcanic vent that has been subjected at different periods
to marine-erosion and in later ages to sub-aerial denudation.


                            THE VUINANDI GAP

I have given this name to the break between the Thambeyu (Mount
Thurston) and Koro-mbasanga ranges, where the level of the mountainous
backbone of the island descends to about 1,200 feet above the sea. This
is the route taken by the track from Vuinandi on the shores of Natewa
Bay across the island to Lambasa.

At Vuinandi the mountains recede from the coast leaving a broad level
plain extending about two miles inland to the village of Tarawau without
rising over 60 feet above the sea. Basaltic rocks are exposed in the
spurs that descend from the mountains to the coast on each side of the
plain. After traversing the low-lying region that lies between Vuinandi
and the main range, one finds on ascending the eastern slopes, _en
route_ to Lambasa, basaltic andesites of the usual type prevailing up to
1,000 feet. The upper portion of the dividing range, 1,000 to 1,200
feet, is composed of a more compact basaltic andesite which is often
rubbly and in this condition is penetrated by fine cracks, 1/8 of an
inch broad, filled with chalcedony. This rock, which has a specific
gravity of 2·85, has a very fresh-looking appearance in the slide, and
the segregation of silica does not therefore appear to arise from an
alterative change. The felspar-lathes, which are in flow-arrangement,
average ·11 mm. in length, and there is a little residual glass.

The mountains rise on either side of the Vuinandi Gap to about 2,000
feet. Descending on the west side of the range one follows a
stream-course down to a level of 400 feet above the sea, agglomerates
and coarse basic tuffs being exposed on the way. The rocks forming the
agglomerates are for the most part to be referred to genus 1 of the
hypersthene-augite andesites. They are sometimes compact and sometimes
amygdaloidal, the amygdules being formed of chalcedony and other
minerals, whilst the glass of the groundmass is often altered.

The track then lay across a spur, 800 feet in height, principally
composed of a greyish porphyrite, exhibiting large opaque crystals of
plagioclase, 4 to 7 mm. long, in an almost holo-crystalline groundmass
formed of stout lamellar felspars with large augite granules. It is
described on page 268 under the porphyritic sub-genus of genus 2 of the
augite-andesites, and is an unusual type of rock for this island. After
this I descended into the picturesque gorge of the Satulaki River, which
is only elevated about 200 feet above the sea, agglomerates prevailing.
In the vicinity of Satulaki a rather compact basaltic andesite (sp. gr.
2·82) is commonly exposed in position. It is referred to genus 13 of the
augite andesites and belongs to the species with felspar-lathes less
than ·1 mm. in average length. It occurs both north and south of this
place and in the hill-spurs on either side. This is the bed-rock of the
Lambasa plains which here begin and extend to the north coast, being
usually covered with submarine tuffs and clays.


                  THE THAMBEYU OR MOUNT THURSTON RANGE

Mount Thurston is the name given in the Admiralty charts to the highest
peak (3,124 feet) of this range. There does not appear to be any general
native name. The highest peak visible from the Lambasa side is known as
“Thambeyu.” The lofty mountain-mass, as it is viewed from Vuinandi, is
known as Ulu-i-ndiri-ndiri.[81] The whole mountain-range has yet to be
properly explored. It is a much more complicated system of
mountain-ridges than is indicated in the chart, my acquaintance with it
being restricted to the Thambeyu ridge, the elevation of which is 2,600
feet above the sea. It trends N.N.W. and S.S.E.; but its relation to the
highest peak of the range could not be ascertained, as we were in the
rain-clouds during the two days we were on the mountain.

I made the ascent from the village of Numbu-ni-a-vula about three miles
to the westward, which is only 200 feet above the sea. In the
intervening low district a basaltic andesite is exposed in the
stream-courses. The structure of the ridge, as indicated by the ascent
of its western slope, is shown in the accompanying diagram. The core or
central axis is formed of massive basic rocks which protrude at the
summit and in one or two of the crests of the spurs. The flanks are
composed of submarine tuffs and clays overlaid by agglomerates of
considerable thickness. The tuffs reach to within 50 feet of the top,
whilst the agglomerates extend to within 400 feet of the summit. The
results obtained from this ascent are specially interesting, since it
afforded me the opportunity of studying in a satisfactory manner the
junction of the agglomerates with the tuffs.

There are two caves on the mountain-side which can be used for
night-shelter by those exploring the range. The lowest, 1,500 feet above
the sea, is the Taloko Cave (na-ngara-taloko). The highest is the Ndromo
Cave,[82] 2,100 feet, known to the natives as “na-ngara-vatu-ni-ndromo.”
Like most of the caves all over the island they occur at the junction of
the agglomerates and tuffs, and are to be attributed to the more rapid
weathering of the underlying tuffs.... In describing the results of my
examination of this mountain-ridge, I will deal in succession with the
tuffs, the agglomerates, the junction between these two deposits, and
the axis or core of basic rocks.

[Illustration: [Image: Ideal Section of Thombeyu]]

(1) _The submarine tuffs and tuff-clays._—As exposed in the
stream-courses near and at the foot of the mountain and as high as the
Taloko Cave, these deposits are bedded horizontally. At higher levels,
owing to insufficient exposure the bedding is not so clear. Up to 700 or
800 feet coarse palagonite-tuffs prevail; but they do not effervesce
with an acid, and apparently contain but scanty organic remains. At 950
feet coarse and fine sedimentary tuffs alternate, the last being
greenish foraminiferous tuff-clay rocks, somewhat compacted and
containing 10 per cent. of carbonate of lime. The tests of the
foraminifera, which are abundant and of the Globigerina type, are filled
with calcite. Several fragments, of a semi-vitreous basic rock, not
however exceeding ·2 mm. in size, are inclosed in the deposit; but the
mass of it is made up of yet finer materials of the same rock,
palagonitic detritus, plagioclase fragments, fine calcitic debris, tests
of foraminifera, &c. These fine tuff-clays were evidently formed in
relatively deep-water.

At the Taloko Cave (1,500 feet), where there are exposed rather coarse
tuffs containing bands about a centimetre thick of a fine clay-tuff, the
last-named effervesce freely with an acid, whilst the first contain only
a little carbonate of lime. No sections have been made of these
deposits; but when powdered and examined under the microscope they
appear to have the same general composition as the deposit described
above from an elevation of 950 feet. They are probably foraminiferous
though scantily. The tuffs found at the Ndromo Cave (2,100 feet) contain
4 per cent. of carbonate of lime and small tests of foraminifera are
visible with a lens. The mineral fragments include plagioclase and
rhombic pyroxene, and there are inclosed rounded gravel-fragments, 5 mm.
in size, of a semi-vitreous rock. Palagonitic debris make up the mass of
these tuffs. A coarse deposit from 2,500 feet is non-calcareous, but has
the same general composition.

(2) _The agglomerates._—These deposits are best represented in the upper
part of the mountain, between 1,500 and 2,200 feet above the sea. Here
they often present vertical precipices having a drop varying between 100
and 400 feet, with the submarine tuffs exposed at their base. Such
cliffs, however, display no structure. Their vertical faces are to be
attributed to joints and to the extensive “slips” that frequently occur
on these slopes, when large masses of agglomerate, undermined by the
percolation of springs through the tuffs beneath them, roll far down the
mountain-sides. The blocks of the agglomerates are fairly uniform in
size, being usually 4 or 5 inches across. They are composed of a
semi-vitreous hypersthene-augite andesite, containing both augite and
rhombic pyroxene, but of an unusual type. It is a blackish rock carrying
opaque phenocrysts of plagioclase, and is characterised by the prismatic
form of the pyroxene (monoclinic) of the groundmass. A very similar rock
from the Sokena agglomerates has been before described. It is referred
to genus 18 of the class, and the prismatic sub-order to which that
genus belongs is described on page 289.

(3) _The junction of the agglomerates and submarine tuffs._—This is well
displayed at the Taloko Cave. Here the agglomerates lie conformably on
the sedimentary tuffs; but the line of junction is sharply defined and
the only evidence of transition is afforded by the great diminution in
the size of the blocks of the agglomerates, which are 1 to 2 inches
across. Immediately beneath the agglomerate is a layer 2½ centimetres
thick of a rather coarse sedimentary palagonite-tuff having the
composition of the deposits above described, but not effervescing with
an acid, and showing no foraminiferous tests. The size of its “grain” is
about a millimetre. This passes downward rather abruptly into a
chocolate-coloured marl-like rock, a centimetre thick, which is formed
of the same materials but in a clayey condition. Beneath this is the
calcareous foraminiferous palagonite-tuff referred to in the first
paragraph.

It is apparent that for some time before the agglomerates began to
accumulate on the sea-bottom there had been a fairly uniform deposit of
submarine tuffs, evidently in rather deep water. Then followed a period
during which the finest mud was deposited which is represented by the
thin layer of chocolate-coloured clay. This was succeeded by the
deposition of coarser sedimentary tuffs forming a layer about an inch in
thickness. Then commenced the accumulation of the agglomerates, of which
the materials were at first small and afterwards larger in size.

(4) _The core or axis of volcanic rocks._—This is represented on the
summit by masses, 2 to 5 feet across, of two kinds of hypersthene-augite
andesite, which are referred to genus 1 of that sub-class. One is a
compact grey rock (sp. gr. 2·72) carrying phenocrysts both of rhombic
and monoclinic pyroxene, the former prevailing, and displaying a small
amount of interstitial glass. It is magnetic and exhibits marked
polarity, as noticed in Chapter XXVI. The other is a scoriaceous rock
containing numerous round steam-pores, ranging up to 5 millimetres in
diameter and generally filled with clear quartz-crystals and lined by
chalcedony. It contains semi-opaque glass in abundance, and is
apparently a semi-vitreous form of the rock just described. Both rocks
are to some extent altered.... On the crest of a spur, 500 feet below
the summit, is exposed in position an augite-andesite, assigned to genus
13, sub-genus 1, species B, of that sub-class. It is non-scoriaceous and
exhibits a considerable amount of greenish alteration products. (Sp. gr.
2·79.)


                            THE AVUKA RANGE

This high range, which lies immediately to the east of Lambasa, attains
its greatest elevation in Mount Avuka, which is 1,976 feet above the
sea. It represents the extension northward to the coast of the inland
Thambeyu mountains that culminate in Mount Thurston. In its upper
portion Mount Avuka presents bare precipitous faces apparently of
agglomerates and some hundreds of feet in height. My acquaintance with
this range is scanty. In a traverse from Lambasa to Ngele-mumu I crossed
it a mile or more south of Mount Avuka, where it is only 700 feet in
elevation. I also rounded the end of the range where it reaches the
coast between Lambasa and the valley of Mbuthai-sau. This last locality,
which is described on page 218, derives especial interest from the
circumstance that here the regions of basic and acid rocks meet. The
basic rocks that occupy nearly all the sea-border from Naivaka to
Lambasa here become mingled with, and finally give place to, the acid
rocks which prevail in all the region eastward as far as Undu Point.

In crossing the range on the way from Lambasa to Ngelemumu, I noticed as
high as 450 feet basic non-calcareous tuffs displaying a concretionary
arrangement suggestive of the proximity of an intrusive igneous rock.
Further up the western slope occur basic agglomerates, whilst at and
near the top (700 feet) there lie on the surface large boulders of a
dark grey hypersthene-gabbro having a specific gravity of 2·7 and
belonging to the type of plutonic rocks described on page 249. It is
very probable that this gabbro forms the axis of the range; and we have
here no doubt one of the oldest of the mountain-ridges in the island.



                              CHAPTER XIII

      DESCRIPTION OF THE GEOLOGICAL AND GENERAL PHYSICAL FEATURES
                             (_continued_)


                           THE VALANGA RANGE

THIS range, which trends N.W. and S.E. between the Mariko mountain-ridge
and the head of the valley of Na Kula, attains a height of 1,880 feet at
its N.W. and of 1,710 feet at its S.E. end. The average elevation,
however, is probably not over 1,300 or 1,400 feet. My acquaintance with
the range is only partial, but it is sufficient to bring to light some
of its leading structural features. Those who follow me will find in
these mountains a very interesting region for their geological
explorations.

(1) TRAVERSE OF THE VALANGA RANGE.—In making the journey from Valanga to
Vunimbua, I crossed the range where its elevation was about 1,300 feet.
Basic agglomerates, containing sometimes amygdaloidal blocks, are
displayed in the low district between Valanga and the foot of the range.
In the stream-course at the base of the slope the deeper seated rocks of
the range are at once exposed. Large masses, 5 or 6 feet across, of
altered grey pyroxene andesites lie in the bed of the stream. Some of
them show opaque porphyritic felspar and have the appearance of
porphyrites (sp. gr. 2·67). They belong to the type described on page
271 under genus 6 of the augite-andesites. Others are grey propylitic
varieties of a basic semi-doleritic andesite penetrated by cracks
containing calcite, and displaying in a groundmass, exhibiting much
viridite and a little pyrites, calcitic pseudomorphs of the felspar
phenocrysts and more or less parallel felspar-lathes, ·15 mm. long and
somewhat altered. Another of the deeper-seated rocks commonly exposed on
the upper west slopes of the range is a dark grey rock showing much
porphyritic pyroxene (sp. gr. 2·72). It has a micro-felsitic groundmass
and is referred to the fourth order of the hypersthene-augite-andesites
described on page 291.

About two-thirds of the way up the western slope of the range, there is
exposed a coarse palagonite-tuff, evidently an incrusting deposit. Stout
crystals of augite can be picked out of it, and it contains also lapilli
up to an inch in size of a basic vesicular semi-vitreous basalt.

Descending the eastern slopes one observes between 1,200 and 1,000 feet
large blocks of the same grey hypersthene-augite-andesite above
mentioned and of a grey granitoid rock of the gabbro type. This last is
a hypersthene-gabbro with specific gravity of 2·75, and belongs to the
group of plutonic rocks described on page 250. Its pyroxene phenocrysts
are often represented by fibrous bastite. One can scarcely doubt that
this gabbro is the plutonic equivalent of the prevailing grey
pyroxene-andesites.

Lower down the slope only small fragments of rocks were exposed,
probably derived from an agglomerate. One of the specimens here obtained
is a doleritic basaltic andesite (sp. gr. 2·77). Another is a very
interesting rock displaying large porphyritic crystals of a mineral like
bronzite in a groundmass originally to a large degree vitreous; but the
glass is now replaced by viridite and secondary crystalline silica. The
“bronzite” is the result of the conversion of associated rhombic and
monoclinic pyroxene into fibrous bastite.

From the results of the traverse across this part of the Valanga Range
it may be inferred that more or less altered grey basic andesites
passing into gabbros chiefly compose it. No doubt at one time it was
largely covered with basic tuffs and agglomerates, but these deposits
have been almost completely stripped off by the denuding agencies, and
were only noticed in one place on the western flank.

That the northern part of the range towards the Mariko ridge has a
similar structure is shown by the character of the loose blocks in the
upper course of the Vunimbua River, which takes its rise on these
slopes. Amongst those in the river above the village I noticed a
solitary block of a coarsely crystalline diorite containing prisms of
brown hornblende a centimetre in length.[83] But the rocks most
frequently represented were propylitic grey hypersthene-andesites, in
which the pyroxene is mostly changed into bastite, whilst the surface
often sparkles with pyrites (see page 297).

(2) NGONE HILL.—This is a curious conical hill, about 700 feet in
height, that rises up on the right side of the Vunimbua River about 1½
miles above the village of that name and near the foot of the range. It
evidently represents a “volcanic neck,” and doubtless this vent was the
source of the large blocks forming the basic agglomerate that occurs in
huge masses in the river-course in the vicinity of this hill. On its
lower flanks is exposed a hard compacted tuff, showing pyroxene
crystals, which is composed principally of fragments of a palagonitised
vacuolar basic glass, the minute cavities being often filled with opal.
In the upper part of the hill is displayed a massive altered
augite-andesite penetrated by fine veins of chalcedony. Numerous
irregular cavities filled with the same material occur in its dark
opaque glassy groundmass.

The blocks of the agglomerates found in the vicinity of the hill vary in
size from 4 to 18 inches. They are composed of a compact blackish
semi-vitreous basic andesite (sp. gr. 2·73) of the type characteristic
of the basic agglomerates over most of the island. The matrix of the
agglomerate is hard and somewhat altered, and is chiefly made up of
fragments, ranging up to 5 mm. in size, of a vacuolar basic glass,
sometimes but slightly changed, though usually converted into
palagonite, the vacuoles being filled with chalcedonic opal. The large
masses of coarse tuffs displayed in the bed of a stream-course close to
Ngone Hill are non-calcareous and composed of palagonitic materials.
Palagonite-tuff clays are also exposed in the river-course a little
above Vunimbua. About half-way between the village and the hill there
occurs in position at the river-side an amygdaloidal basic rock, the
amygdules being formed of chalcedonic opal.

It is apparent that this hill represents a lesser vent which probably
dates back to the period before the emergence. All the products of its
eruption are, however, more or less altered. From the absence of sorting
in the blocks of the agglomerates, and from the character of the matrix,
it may be inferred that these deposits have been accumulated directly
from the ejected materials without the intervention of the agency of
marine erosion.

(3) THE WESTERN FLANK OF THE VALANGA RANGE.—One of the boldest pieces of
coast in the island lies on the eastern side of Savu-savu Bay, between
the mouth of the Ndreke-ni-wai River and Valanga Harbour. Here a number
of lofty headlands separated by broad valleys descend with precipitous
fronts to the shore, some of them, as in the case of the Nambathi
promontory on the north side of Valanga Harbour, retaining an elevation
of 1,000 feet within a few hundred yards of the coast.

By following the coast-track from the Ndreke-ni-wai River to Valanga one
crosses some of these headlands. As far as Vatu-lele altered red tuffs,
basic agglomerates, and massive basaltic andesites are the prevailing
rocks. The red tuffs exhibit a double alteration. They were originally
composed of finely pulverised basic vacuolar glass, which subsequently
became palagonitised, and afterwards there was an extensive deposition
of chalcedonic silica and of red iron oxide. No organic remains appear
to exist; whilst the scanty calcite present is evidently an alteration
product. Where the road “tops” the headland on the north side of
Vatu-lele Bay, there is exposed a dyke-like mass of a rubbly
semi-vitreous basaltic rock penetrated in all directions by veins, 1 to
3 inches thick, of a tachylytic glass, splinters of which fuse readily
in the ordinary spirit-lamp flame. The numerous fissures were doubtless
produced during the consolidation of the rock; and subsequently they
were filled with the still fluid residual portion of the magma, which
would be composed of the most fusible constituents. This subject, which
bears on the origin of palagonite, is discussed in Chapter XXIV.

Between Vatu-lele and Urata, palagonite-tuffs and basic agglomerates are
chiefly displayed. On the north slope descending to Urata there is
exposed in the foot-path a dyke-like mass of a dark-grey
hornblende-pyroxene-andesite, an unusual type of rock which is described
on page 298. Just south of Urata I observed an agglomerate containing
large blocks, 3 or 4 feet across, of the deeper-seated altered grey
pyroxene-andesites that with the gabbros and diorites form the axis of
the range.

(4) THE VALLEY OF NA KULA.—In crossing from Sava-reka-reka to Natewa
Bay, one ascends the remarkable valley of the Kula and traverses the
ridge at its head. This ridge, which is about 700 feet in height and
forms the termination of the Valanga Range, is composed of altered grey
hornblende-pyroxene-andesites and of similar holo-crystalline rocks
representing the gabbro or plutonic type of the same. One of these rocks
is described on page 250, under the head of hornblende-gabbro. Another
is referred provisionally to the hypersthene-gabbros (page 249); but it
is extensively occupied by chlorite, viridite, and other alteration
products. Here, as with the other rocks of the Na Kula Ridge, the
plagioclase phenocrysts are opaque, the result of the numerous fine
cracks with decomposition products in the interior of the crystals....
It is thus seen that in general structure the Na Kula Ridge represents
the main axis of the Valanga Range to the north.

The valley of Na Kula is occupied by a river which does not empty
itself, as one would expect, into Savu-savu Bay, but turns off sharply
to the south at right-angles to its previous course, and after breaking
through the coast range, opens into Naindi Bay. This peculiarity has
attracted the attention of the natives. The village of Sawa-Ndrondro,
which lies about 1½ miles up this valley, is not elevated more than 50
feet above the sea. The gradient is evidently not only very slight but
is also irregular, so that in their upper course about 3 miles inland,
where the elevation is only 130 feet, the waters of the river are
partially checked in their flow and form extensive swamps where the
“vitho” or wild sugar-cane flourishes.

(5) CONCLUDING REMARKS ON THE VALANGA RANGE.—It may be inferred from the
geological structure of the range that it is one of the oldest in the
island. The agglomerates and tuffs that enter so largely into the
formation of most of the other mountain-ridges are here to a great
extent absent, except in the lower flanks; and we have exposed the axis
of the range composed of more or less altered grey pyroxene and
hornblende-pyroxene andesites passing, as appears to be the case, into
gabbros and diorites. It is true that the exposure of the gabbros is
limited and that only a single block of diorite came under my notice;
but this might be looked for where the plutonic rocks are deeply seated.
Although far overtopped by the neighbouring agglomerate mountain-ridge
of Mariko, the Valanga Range would seem to date back to a much earlier
stage in the history of the island.


                      THE MOUNTAIN-RIDGE OF MARIKO

This mountain-ridge, which trends nearly east and west and joins the
Valanga Range, rises in mass to a height of rather over 2,000 feet.
Above this elevation it terminates in several short conical peaks, of
which the highest, 2,890 feet, is named Mariko, the Drayton Peak of the
chart. One of the peaks, lying a little to the east of the summit, and
apparently between 100 and 200 feet lower, is called the Vatu-mbutho or
White Rock. In the profile of the range, as seen nearly “end-on” from
the distant south shore of Natewa Bay, it would appear to be rounded in
its upper part. Its true outline, however, when viewed in length, is, as
described above, namely, a massive ridge with various peaks.

When viewed from the top of the hills behind Valanga, this mountainous
range has a very imposing appearance. On the south side it rises
precipitously to the summit, but the northern slopes below an elevation
of 1,800 or 1,900 feet descend with a very easy gradient for 1½ or 2
miles into the valley of the river Ndreke-ni-wai. In the first case the
average angle of the slope would be from 15 to 20 degrees and in places
often more; whilst in the second case the average inclination would be
about 7 degrees. The contrast between the two sides of the range is very
striking and one ought, I think, to find a parallel in the broken-down
rim of a large crater with a gentle outer slope and a precipitous inner
face. When descending recently the outer slope of Monte Somma, the
ancient Vesuvian vent, I found reproduced some of the features of the
northern slope of Mariko. The tuffs and agglomerate-tuffs that cover
their outer flanks are in both mountains deeply scored by the gorges and
ravines worn by the torrents. After the description of the geological
structure of the Mariko Range, we shall perhaps be in a better position
to consider this question; but until a proper survey of the region has
been made it will not be possible to give a final answer. There are also
many other uncertainties which would be removed by the accurate mapping
of the district, such for instance as the mode of connection between the
Mariko and Valanga Ranges.

The highest peak of the Mariko Range is irregularly square-topped and is
only a few paces across. It has a soil-cap and supports small trees and
shrubs, whilst there is a precipitous rocky face on the east and south.
Like most of the other lofty peaks of the island it is magnetic, and as
remarked on page 368, it markedly deflects the compass-needle.

I made two ascents of this mountain from Vunimbua, one to the highest
peak (2,890 feet), and the other across the range to Nukumbolo at a
point half a mile or more to the west of the summit, where its elevation
is 2,200 feet. Basic agglomerates and agglomerate-tuffs prevail on both
the slopes up to 1,800 or 2,000 feet, the blocks being composed of a
dark semi-vitreous basic andesite referred to the hypersthene-augite
sub-class with specific gravity 2·75. It contains much glass in the
groundmass, and since the pyroxene of the groundmass is prismatic, this
rock belongs to the prismatic sub-order described on page 289. Ordinary
basic tuffs are also well represented on the north flank. On the south
or precipitous side they are usually more or less altered. Here, for
instance, they may take the form of a hard breccia-tuff containing
vesicular lapilli, up to half-an-inch in size, of a semi-vitreous basic
rock, the small steam-holes being either empty or filled with opal or
chalcedony. The matrix of the rock is made up of finer fragments of a
basic vacuolar glass, showing a few felspar microliths, but often more
or less palagonitised. Evidence of further alteration is afforded by the
small cracks and crevices filled with chalcedony.

Other altered tuff-rocks are exposed on the south slope. At an elevation
of 400-450 feet above the sea, and underlying the agglomerates and
breccia-tuffs, I found exposed in a stream-course a hard dark rock
looking like a compact andesite. Under the microscope, however, it is
shown to be an altered palagonite-tuff composed in part of angular
fragments of plagioclase and of rhombic and monoclinic pyroxene, not
exceeding ·15 mm. in size, and containing also similar-sized fragments
of a basic hemicrystalline rock. The base is made up of palagonitic
material and contains a few “Globigerina” tests sometimes displaying
calcite in their interior. Fine cracks filled with chalcedonic silica
testify to a subsequent alteration of the deposit. At 1,500 feet occurs
a hard red altered palagonite-tuff, having a similar composition and
being altered in like fashion, but not displaying tests of foraminifera
in the slide.

The foregoing remarks refer to the main undivided mass of the range,
that is, up to 2,000 feet. The highest peak of Mariko probably
represents in structure the other peaks rising to various heights on
either side of it. Here, at elevations between 2,000 feet and the
summit, a rubbly agglomerate prevails of a somewhat different character
from that occurring at lower levels. It is well exposed in some
cave-cliffs at a height of 2,500 feet and also in the rocky face of the
peak. The rock composing the blocks is a dark-grey aphanitic
augite-andesite (sp. gr. 2·65), referred to genus 20 of that sub-class
and displaying prismatic pyroxene in the groundmass. Smoky residual
glass exists usually in fair amount; whilst in the blocks of the
cave-cliffs it is so abundant that the rock may be termed semi-vitreous.
In the locality just named the blocks are scoriaceous, the steam-pores,
which are drawn out to a length of 5 or 6 mm. and more, being partially
or completely filled with calcite and occasionally with opal. At times
the steam cavities are much larger. In one of my specimens there is an
elongated cavity 5 cm. (2 inches) in length, which has a thin lining of
chalcedony, from the surface of which pyramidal crystals of calcite
project into the interior. (I found the same grey andesite exposed _in
situ_ lower down the south slope at an elevation of 1,800 feet, but
non-scoriaceous.) The matrix of the agglomerate principally consists of
fine palagonitic material with small fragments of plagioclase and
pyroxene but apparently no lime.

At heights of about 2,800 feet on the south side of the peak, and of
1,600 feet on the north flank of the range, are exposed non-calcareous
greyish tuffs remarkable for the quantity of crystals of rhombic
pyroxene, entire and in fragments, that they contain. This is a
characteristic feature of the more acid andesitic tuffs of the island,
and it is to these deposits that the Mariko tuffs in question make a
near approach. They contain at times subangular fragments of more basic
rocks; and are true tuffs in the sense that although perhaps deposited
on a sea-bottom they represent the ejected materials of a subaerial
vent.

The crest of the range, where it is crossed by the road from Vunimbua to
Nukumbolo and for 200 feet below, is formed of a decomposed rock,
perhaps a breccia. A fragment of the rock obtained from the crest is a
grey somewhat altered hypersthene-augite andesite (sp. gr. 2·75) with an
orthophyric groundmass, and referred to the order described on page 290.
This rock may be connected with the tuffs above alluded to.... Reference
may here be made to a black basaltic rock (sp. gr. 2·88) of which, at an
elevation of 2,500 feet at the foot of the peak, I found a portion of a
columnar block about 18 inches across. It may prove to be an
olivine-basalt; but no section has been made of it.

It is apparent from the foregoing description of the Mariko Range that
in general structure it does not differ materially from the other
mountain-ridges of the island, although in the types of the rocks it
presents some variety. Here also we have agglomerates prevailing on the
flanks and forming the summit. As far as the characters of the rocks can
guide us, we cannot determine whether the range has been built up by a
number of vents on a great fissure, or whether it represents the remains
of a huge crater. In this uncertainty we can only appeal to the contrast
between the gentle gradient of the north slopes and the precipitous
descent of the south slopes as favouring the last supposition. We
cannot, however, doubt that the agglomerates of the upper portion of the
range are the products of an eruptive vent or of vents that rose above
the surface of the sea, since the blocks are all of one kind of andesite
and are often scoriaceous. We can be fairly certain that at such a time
the lower slopes were in part submerged, seeing that foraminiferous
tuffs underlying the agglomerates are now exposed. But we have to
distinguish between these submarine basic tuffs of the lower slopes
which may in part be the result of marine-erosion and the grey
rhombic-pyroxene-tuffs of the upper levels which are probably derived
from subaerial eruptions.


                        THE SAVU-SAVU PENINSULA

I include in this district the promontory west of Naindi Bay and
Sava-reka-reka Bay. Although its surface is much cut up, it has, when
viewed from a distance, a fairly even profile and attains a maximum
height of rather over 800 feet. From the region east of it, it is
separated by the Naindi Gap. Here one can cross the peninsula between
the two bays above named without rising more than 50 feet above the sea.
The elevated interior is divided into two parts, which are divided by a
_col_, about 250 feet in elevation, which is ascended in crossing from
Naithekoro on the south coast to Na Kama on the north coast. Much of the
surface is clothed with the usual “talasinga” vegetation. Close to the
north shore, with which it is connected by the reef-flat, rises the
small island of Na-Wi, and off the extremity of the peninsula, which is
known as Harman’s Point, is the islet of Naviavia, formed of raised
reef-limestone as described on p. 8. The celebrated boiling springs
known as Na Kama are situated on the north coast opposite Na-Wi. It may
be remarked in passing that besides finding an exit in the springs, the
hot water oozes through the beach and below the tide-marks for several
hundred yards along the shore. These springs are described in detail on
p. 25.

This is one of the few districts of the island in which elevated
reef-masses occur at the sea-border. These old reefs, which attain a
maximum elevation of 250 feet above the sea, are principally restricted
to the neighbourhood of Naindi Bay. (They are referred to in detail in
Chapter II.) But they indicate only a part of the submergence which this
region has experienced. There is an exposure of a very interesting rock
in a stream-course that is crossed on the road from Yaroi to Naindi,
less than a mile from the first-named place, and about 30 feet above the
sea. Here we find a dark, impure “Globigerina” limestone, or, as it
might be also designated, an altered calcareous palagonitic
clay-tuff.[84] The larger fragments in it average only ·2 mm., and it
affords evidence of a period of submergence during which the hill-tops
of the Savu-savu Peninsula were below the sea-level.

We get the same indication, but in a more pronounced degree, in the
stratified sedimentary clay-tuffs which are exposed on the shore-flat of
the south side of the neighbouring Sava-reka-reka Bay. These beds, which
within a distance of fifty paces are inclined 10-15° to the south-west
and the same amount to the north-west, have apparently a quâquâversal
dip. In places they exhibit a spheroidal and concentric structure, and
are penetrated by cracks containing some calcite, but mostly filled with
a white zeolitic mineral.[85] One of these rocks is a bright green, hard
and compact deposit, containing but little lime, and evidently an
altered palagonitic clay-tuff. It contains a few minute tests of the
“Globigerina” type; and on account of the small size of its fragments of
minerals, which range from .01 to .04 mm., it may be regarded as a
relatively deep-water sediment.[86] It is interstratified with a
coarser, somewhat altered palagonite-tuff, which shows but little lime
and only a suspicion of tests of foraminifera. The size of the larger
included fragments does not exceed half a millimetre.... The low hill,
near Yaroi, on which the magistrate’s house is built, is composed of
fine and coarse tuffs, probably submarine. It is doubtful whether any
but sedimentary tuffs occur in this peninsula.

In the hills of the western part of the peninsula, that is, west of Na
Kama and Naithekoro, a particular type of basaltic andesite prevails,
characterised by rhombic pyroxene as well as augite phenocrysts, and
referred for the most part to genus 13 of the hypersthene-augite
andesites. Their specific gravity ranges from 2·76 to 2·83, and the
interstitial glass may be fair or scanty in amount. The average length
of the felspar-lathes is unusually small, ·04-·06 mm. In these respects
the basaltic andesites of the Savu-savu Peninsula differ from the
basaltic andesites found in most other parts of the island, where, as
exemplified by those of the Wainunu, Solevu, and Seatura regions, the
felspar-lathes average between ·1 and ·2 mm. in length, and there is
practically no rhombic pyroxene. A somewhat scoriaceous semi-vitreous
form of pyroxene andesite is exposed on the south slopes above
Nukumbalavu, where it is covered by basic agglomerates. The pyroxene in
the groundmass is here prismatic, and not granular, and for the most
part rhombic; and the rock is referred to the prismatic sub-order of the
hypersthene-augite andesites described on p. 287.

The basaltic andesites of the peninsula are often extensively decomposed
through the weathering process, a spheroidal structure being then
displayed. It rarely happens that the basaltic rocks of this locality
assume a propylitic character. Yet, if this change is due to
hydrothermal metamorphism, we ought to find altered rocks of this kind
in the vicinity of the boiling-springs. Such rocks did not come under my
notice at the surface; but this only indicates that if this alteration
has taken place here, it has been effected at some depth; and, indeed,
it would seem probable that the alteration known as “propylitic” is a
change produced generally in deep-seated rocks.

A semi-ophitic basaltic andesite that is exposed in the small
stream-course at the back of the springs, and not 100 yards distant,
displays no propylitic change, and is only affected by hydration. The
basaltic andesites found on the hill-slopes further inland from the
springs exhibit no change of such a nature. However, rocks of this
description occur at and near the coast about a mile to the westward.
One of them, which is light green in colour, might be taken for a
limestone, since it effervesces with an acid. When examined in the slide
it is shown to be the prevailing basaltic andesite greatly altered. The
porphyritic rhombic pyroxene is replaced by viriditic material; the
plagioclase phenocrysts are replaced by calcite, secondary silica, and
other alteration products; and the structure of the groundmass is
disguised by chalcedony, calcite, viridite, &c. Another rock from this
locality displays great alteration. The structure of the groundmass is
obscured by secondary silica, and is traversed by fine cracks passing
through the felspar phenocrysts and filled with blood-red films of
hematite.

On the hill-slopes behind Harman’s Point, at an elevation of 300 to 400
feet, blocks of a reddish, volcanic rock, greatly altered by the
deposition of silica, were displayed on the surface. The ground was here
strewn in places with beautiful pyramidal prisms of clear quartz,
ranging up to an inch in length. They contain numerous inclusions, their
faces being sometimes deeply etched or eroded. These crystals appear to
have been formed rather rapidly in some highly siliceous thermal
underground waters.

I did not ascend the hills of the portion of the peninsula lying east of
Na Kama and Naithekoro. But whilst crossing the saddle between these two
places, I perceived that the prevailing basaltic andesites extended up
the slopes to the east. The neighbourhood of Naindi Bay offers several
features of interest. The bay, which is circular in shape, is closed in
on the east and west by projecting points, where we find elevated
reef-limestone, 40 or 50 feet above the sea, displaying massive corals
and large “Tridacna” shells in their natural position, and overlaying a
cement-stone composed of blocks of volcanic rocks in a calcareous
matrix. On the beach on the west side of the bay there is exposed a
reddish-grey altered pyroxene-andesite, which, as regards the size of
the felspars of the groundmass and other characters, appears to be an
altered form of the prevailing basaltic andesites of the peninsula. In
the midst of the low passage that isolates the peninsula, which I have
termed the Naindi Gap, there is displayed a highly altered basic
andesite which contains a white, zeolitic mineral in its numerous
cracks.

The small island of Na-Wi consists of two low hills, the highest 130
feet in height, connected by a mangrove swamp and a sandy beach. There
is no trace of a crateral cavity. The prevailing rock is a porphyritic,
compact, basic andesite, differing from the other rocks of the
neighbourhood in the greater amount of glass it contains. Though it is
not easy to find a good, unweathered specimen of the rock, it would
appear that Na-Wi represents an old volcanic neck.

We may infer from the above description of this peninsula that it has a
history similar to that of most other parts of the island. There is
evidence in the upraised reefs and in the “Globigerina” clays and
limestones of considerable submergence at one period; and it is highly
probable that the prevailing basaltic andesites are the products of
submarine eruptions. In my account of the hot springs given on page 26,
reference is made to the absence of any trace of a crateral cavity in
that locality. The same is true, as far as my observation goes, of the
whole peninsula. Altered rocks do not occur in the vicinity of the
springs, but they are to be found at distances a mile and more away. It
does not seem possible to restore in imagination the original form of
this part of the island. The present contours are the results of more
than one reshaping of the surface through the agencies of marine erosion
and sub-aerial denudation.


           THE DISTRICT BETWEEN NAINDI BAY AND THE SALT LAKE

Three or four of the peaks of this hilly district rise to about 1,000
feet or rather over, the highest being that of Na Suva-suva, which
attains a height of 1,110 feet. Since my acquaintance with this region
is incomplete, I will confine my remarks to the localities actually
examined.

Through the kindness of Mr. F. Spence, I was able to make use of a track
cleared to the top of Na Suva-suva. This eminence, which forms a
conspicuous landmark for many miles, both landward and seaward, has a
rounded summit and is to all appearance an old volcanic neck. It is
composed in mass in its upper half of a heavy dark olivine-basalt (sp.
gr. 3·01), seemingly non-columnar, and referred to the highly basic
rocks forming genus 16 of the olivine-basalts. There is such a thick
soil-cap on the lower slopes that I was unable to ascertain the
character of the rocks there. It is, however, noteworthy that a very
similar olivine-basalt (sp. gr. 2·99) crops out on the coast south of
this hill and to the east of Naindi Bay. They both contain abundant
small olivine-phenocrysts and a little residual glass, the
felspar-lathes averaging ·1-·14 mm. in length. Since their localities
are rather more than a mile apart, it is not possible to say without a
further examination of the locality whether or not we have here the same
intrusion.

On the coast between Naindi Bay and Salt Lake Passage, calcareous tuffs,
probably fossiliferous, are occasionally exposed in the low spurs
descending to the sea, whilst islets of elevated reef-rock front the
beach.

The coast immediately west of the Salt Lake Passage is of exceptional
interest. Here the sea-cliffs and the shore-flat are formed of an
agglomerate tuff penetrated in all directions by veins of calcite, an
inch and under in thickness. The matrix of this deposit, which is a
little calcareous, is principally made up of fragments, ranging up to 3
or 4 millimetres in size, of vacuolar palagonite, the minute vesicles
being filled with some alteration product. It also contains large macled
augite crystals 5 or 6 mm. in size, which can be picked out in numbers
by the fingers. The blocks vary from a few inches to two feet across,
and are usually composed of an augite-andesite, containing large
porphyritic crystals of augite, and are often amygdaloidal, the
amygdules, 3 or 4 mm. in size, being formed of a zeolite. But blocks of
very different rocks also occur in this agglomerate tuff. One, about two
feet across, was composed of a coarsely crystalline diorite made up, as
described on page 251, of large crystals of hornblende, 2 to 2·5
centimetres long, and of large opaque crystals of acid labradorite.
Another was made of hornblende-hypersthene andesite belonging to the
ortho-phyric order of that sub-class (see page 299). There is a little
altered glass in the groundmass, and large secretions of brown
hornblende, more than an inch in size, are to be observed in the rock.

It is probable that this singular deposit represents a submarine
accumulation of materials ejected from some neighbouring vent. Organic
remains did not come under my notice; but apart from the palagonitic
character of the matrix and the abundance of veins of calcite, the
submarine origin is indicated by the existence of upraised reefs in the
coast districts east and west of this locality. The block of diorite
affords an important clue as to the character of the deep-seated
plutonic rocks in this part of the island. A similar diorite was found
by me amongst the blocks in the bed of the Vunimbua River; and on page
185, reference is made to the probability of such rocks forming the
nucleus of the Valanga Range.

The hills on the west side of the Salt Lake are worth further
examination. On the coast of the Natewa Bay side of this district,
in the vicinity of Vuni-tangaloa and between that place and
Vuni-sawana, there are displayed agglomerates formed of blocks of
hornblende-andesite, some of the specimens being very similar to
that obtained from the block of hornblende-andesite noticed in the
agglomerate-tuff on the neighbouring south coast.


                             THE SALT LAKE

The low isthmus, about 2½ miles in breadth, which connects the Natewa
Peninsula with the rest of the island, can be crossed without rising
more than 40 or 50 feet above the sea. From the occurrence of upraised
reefs in the islets and in the low sea-cliffs of the south coast it may
be inferred that at no distant period in the history of Vanua Levu this
isthmus was submerged.

The lake, which is oblong in form, is about four-fifths of a mile long
and about two-fifths broad. Its maximum depth according to the Admiralty
chart is 3 fathoms; but the usual depth in the centre varies, as I
found, between 2 and 2½ fathoms. It communicates with the sea on the
south coast by a long narrow passage, rather over a mile in length,
which for the greater part of its course, excepting near its seaward
mouth, is only between 25 and 30 feet broad. Mangroves flourish around
the lake and also line the passage; whilst elevated reef-rock is to be
observed on the sides of the passage. Mr. Horne was informed that corals
abound in the lake-waters; but I find no reference to this point in my
notes. Judging from the density of the effluent water, the specific
gravity of the lake-water is that of the sea. The “rise and fall,” as
noticed below, is considerably less than in the case of the tides at the
coast.

Near the centre of the lake there is a low islet, some 40 paces across,
and only raised about a foot above the level of the lake at the time of
high-water. It is chiefly made up of coral blocks; but there are a few
fragments of basaltic andesite lying about, which were probably brought
there by natives. This islet is mentioned in Mrs. Smythe’s account[87]
of the visit made by Colonel Smythe to the lake in 1860; and by reason
of its little elevation it may be accepted as a rude datum-mark of the
relative level of land and sea in this region. From this it would appear
that there has been no appreciable change of level in this region for
the last forty years.

Except on the north and north-west sides, the lake is more or less
surrounded by hills reaching up to 400 or 500 feet, the passage
representing a break in the range. On the Natewa Bay side the level of
the surface is much lower. The low strip of land that intervenes between
the north-west corner of the lake and Natewa Bay is about a mile across,
and does not attain a greater elevation than 40 or 50 feet above the
sea. On its surface, fragments of basic volcanic rocks are displayed;
but no reef debris came under my notice. At its north-east side the lake
is only separated from Natewa Bay by a neck of land 300 to 400 yards in
breadth and about 100 feet high. It was across this neck that the
natives in old times used to drag their large canoes.

Mr. Horne[88] who visited this neighbourhood in 1878, suggested that the
Salt Lake occupies a crater-cavity. The hills around are of volcanic
formation, and I am rather inclined to support this view; but certainty
is scarcely possible now, on account of the great degradation which the
surface has evidently experienced during and since the emergence; whilst
subsequent reef-growth has also to some extent masked the original form
of the district. It is noteworthy that a somewhat parallel condition of
things is presented a few miles to the west by the circular Naindi Bay
and the low passage, not more than 50 feet above the sea, that partly
isolate the Savu-savu Peninsula.

The peculiar behaviour of the tides in connection with the Salt Lake and
its passage attracted my attention during two visits to this locality.
On the first occasion I noticed that between two and three hours after
the tide at the coast had commenced to rise there was still a strong
flow through the passage from the lake, and that the current was only
reversed in the latter half of the rising tide. During my second visit
at the end of May, 1899, when I was accompanied by Mr. Smallwood, I
spent a night in observing the behaviour of the spring tides at a spot
below the narrow portion of the passage 600 or 700 yards from the
opening on the coast. Here the breadth was about 100 feet, the depth at
low-water 5 feet, and the rise of the tide 4 feet. The current ran
seaward at a velocity varying from 1,500 to 2,500 yards per hour; and it
continued to flow in this direction for 2½ hours after the tide had
begun to rise on the coast. (In the narrow part of the passage the rate
of the current would probably be not over 3 knots.) It is curious that
at the place of measurement the bottom was formed of mud into which the
pole sank six feet without striking a hard substratum. The observations
on the current were made with a vertical float immersed about 3 feet.

The point of difficulty in the behaviour of the tides is this. The water
is running rapidly out of the lake for nine hours; whilst during the
remaining three hours there is a sluggish return-flow up the passage
into the lake. A far greater quantity of water finds an exit by the
passage than is returned by the same channel; and I can only explain
this by assuming that there is an extensive percolation of water from
Natewa Bay into the lake. It is easy to show that with such a narrow
effluent, which cannot have a sectional area exceeding 180 square feet,
the level of the lake would be only lowered 2 or 3 feet, if the average
velocity during the nine hours was two nautical miles. The great bulk of
the water would thus remain unchanged. The ultimate result of such
conditions would be a lake of brine. Since, however, the sea-water of
the lake possesses the ordinary density, it is apparent for this reason
only that there is some other means of supply than by the present narrow
passage leading to the sea. The mean level of the Salt Lake is evidently
rather above that of the sea, perhaps a foot or two; and the
“rise-and-fall” is probably very small.



                              CHAPTER XIV

      DESCRIPTION OF THE GEOLOGICAL AND GENERAL PHYSICAL FEATURES
                             (_continued_)


                          THE NATEWA PENINSULA

THIS remarkable peninsula is connected with the rest of the island by
the low-lying Salt Lake district, a narrow isthmus, described in the
preceding chapter, which one can cross without rising 50 feet above the
sea. My acquaintance with this region is far from complete; but from the
following notes a fair general idea of its geological characters may be
gathered.

By referring to the map it will be seen that there are three groups of
mountains. The north-eastern culminates in Mount Freeland or Ngala,
2,740 feet; the southern is formed by the rugged Waikawa Range, 1,540
feet; whilst the Lea Range to the west attains in Ngalau-levu a height
of 1,960 feet. They are much cut up by the denuding agencies, and all
bear the stamp of an ancient land-surface. Though hot springs are not
infrequent, as at Ndreke-ni-wai, Waikatakata, Ndevo, and Navuni, no
evidence of recent volcanic action came under my notice. Submarine
deposits occur at intervals on the surface up to elevations of 1,000
feet and over; but with the exception of the comparatively recent
upheaval or emergence of some 20 or 30 feet, indicated by the raised
reef-masses and foraminiferous tuffs and clays in different parts of the
coast, there is nothing to suggest that these changes did not occur ages
since. In the frequent alteration of its andesitic rocks, and in the
occasional occurrence of porphyrites, we have sufficient indication of
the antiquity of this part of the island as far as its volcanic history
is concerned.

I will commence the description of this peninsula at its western end.

The broken elevated district that extends eastward from the Salt Lake to
Fawn Harbour on the south coast, and to the mouth of the Ndreke-ni-wai
River on the north coast, is divided into two principal masses, which
are connected by a ridge or _col_ about 400 feet above the sea, which is
situated a little east of Viene. The western portion, which may be named
the Viene sub-district, attains a maximum height of 1,000 feet. The
eastern portion reaches in the peak of Ngalau-levu, a height of 1,960
feet, and may be termed the Lea sub-district.

THE VIENE SUB-DISTRICT.—The cliffs on the north coast between Muanaira
and a little east of Viene are mainly formed of basic tuffs, often
calcareous. At a place about 1½ miles east of Viene, these tuffs as
exposed in a coast spur display large flat spiral tests of shallow-water
foraminifera 4 or 5 millimetres across. They may be described in this
locality as palagonitic calcareous tuff-sandstones, more or less
compacted, and containing fragments of palagonitised basic rocks. When
crossing the _col_ above referred to one finds similar palagonitic
calcareous sandstones and clays exposed on its slopes up to its summit
(400 feet).

On the south side, in the vicinity of Vunilangi Inlet, foraminiferous
clays and reef-limestones are displayed at the foot of the slopes; and
the coast between this place and Tathelevu to the westward is bordered
by low cliffs of reef-limestone raised 6 to 8 feet above the high-water
mark and displaying massive corals in their position of growth. Near
Tathelevu there occur raised reefs 10 to 15 feet above the sea; whilst
the hills, 250 to 300 feet in height, at the back of this place are
composed of fine and coarse tuffs and tuff-sandstones containing little
or no lime, and apparently no organic remains. They are sedimentary
tuffs of mixed composition, made up of fragments of plagioclase, rhombic
and monoclinic pyroxene, brown hornblende, portions of semi-vitreous
basic andesite, and palagonitic debris. In the lower levels they are
fine textured with a grain of ·2 to ·3 mm. In the higher part their
grain is ·5 to 1 mm., and they are more basic in character and come near
to the palagonite-tuffs. At an elevation of 200 feet they form inland
cliffs, 50 feet high, in which are imbedded blocks, 2 feet across, of a
blackish pyroxene-andesite with a specific gravity of 2·73, and
belonging to the prismatic sub-order of the orthophyric order of the
hypersthene-augite andesites. It is remarkable for the pyroxene prisms
of the groundmass, and shows a little interstitial glass. These cliffs
are well displayed behind Navelatha, about half a mile from Tathelevu.
Between this locality and the Salt Lake Passage, elevated
reef-limestones, forming low cliffs 6 to 8 feet high, together with
occasional tuff-agglomerates, occur at the coast.

THE LEA SUB-DISTRICT.—This region, which includes the mountain-range of
Ngalau-levu at the back of Lea, is limited by Fawn Harbour and Vunilangi
Inlet on the south coast, and by the Ndreke-ni-Wai River and a point
between Viene and Lea on the north coast. Its structure, as is shown
below, is very complicated, acid and basic rocks being associated in a
remarkable manner; whilst over all lie the submarine tuffs. Marine and
sub-aerial denuding agencies have shaped and re-shaped the surface to
such a degree that it is now impossible to restore it in imagination.

On the north coast of this sub-district, about two miles east of Viene,
is exposed an altered darkish porphyrite displaying large opaque
crystals of plagioclase, 5 to 7 mm. long, the rock-mass being penetrated
by fine veins of chalcedonic quartz, which also traverse the
phenocrysts. Its specific gravity is 2·6; but on account of the
imperfect development of the felspar-lathes and the amount of altered
glass in the groundmass, which also contains a little calcite, it can be
only generally referred to the augite-andesites. A greenish altered
foraminiferous tuff showing fine cracks filled with chalcedony composes
a spur in this locality. A propylitic or highly altered dolerite is
exposed half-way between Viene and Lea.

As one nears Lea from the west the lofty spurs of the mountain of
Ngalau-levu reach the coast, and basic tuffs and agglomerates prevail.
The blocks in the agglomerate are composed of a vesicular semi-vitreous
hypersthene-augite andesite, which is assigned to the second prismatic
sub-order, since it carries prismatic pyroxene in the groundmass. The
town of Lea is picturesquely situated on the coast at the foot of the
steep mountain-slopes, being closed in on the east and west by elevated
spurs descending to the sea. Fragments of jasper and chalcedony occur in
the beds of the streams that here drain the precipitous sides of the
range. Two dykes of dark basic rocks protrude through the beach in Lea
Bay. They are composed of augite-andesites referred to genus 13 of the
augite-class; but the two rocks belong to different species of that
genus. In the one the felspar-lathes are only ·04 mm. in length, and
there is a little altered glass in the groundmass, the specific gravity
being 2·63. In the other the felspar-lathes average ·2 mm. in length,
and the rock has a coarser texture, whilst the specific gravity is 2·7.
The augite granules are large (·03 mm.), and there are irregular lacunar
spaces filled with calcite and lined by a brown palagonite-like
material.

I ascended the second highest peak of the Ngalau-levu mountain, which
rises to a height of 1,680 feet behind the town of Lea, the highest
summit lying to the eastward. Ngalau-lailai, which I also ascended, is a
lesser peak, 1,400 feet in height, situated yet nearer to the town.
Basic tuffs and agglomerates similar to those exposed in the western
spur of the bay occurred all the way up to the bare rocky pinnacles
forming the summits. The blocks in the agglomerates are made up of a
semi-vitreous augite-andesite, which is sometimes scoriaceous or
amygdaloidal, and at other times pseudo-vesicular. Augite crystals, 5 or
6 millimetres in length, are inclosed in the tuffs which contain
palagonitised materials, but apparently no organic remains.

In the spur on the east side of Lea Bay occurs a light-coloured altered
hornblende andesite. The brown hornblende is mostly represented by black
pseudomorphs. Such a rock appears in strange contrast with its basic
surroundings. This is followed, as one proceeds eastward along the
coast, by basic tuffs and agglomerates. It should have been before
observed that blocks of a blackish-brown olivine basalt (sp. gr. 2·89),
referred to genus 13 of the olivine class, occur at intervals on the
coast between Viene and Ndreke; but the rock never presented itself in
position. The tiny felspar-lathes (·03 mm. long) are in flow
arrangement; but there is little or no residual glass, and the augite
granules (·01 mm.) occur in great abundance.

About two-thirds of the way between Lea and Ndreke-ni-wai there lie
close to the shore two islets, 20 to 25 feet high, of reef-limestone, in
which massive corals may be observed in their position of growth.
Further east, about half a mile west of Ndreke-ni-wai, there is exposed
at the coast a bedded light-coloured non-calcareous compacted tuff-rock,
dipping 12° to 15° to the southward. It contains pebbles and blocks of
acid and basic andesitic rocks, and may be described as an altered
hornblende-andesite tuff. Basic agglomerates occur as one approaches
Ndreke-ni-wai. This town lies at the mouth of the river of that name,
the first river that one meets on the north side of this peninsula.
There exist here between the tide-marks some hot springs, to which
reference is made on page 34.

When crossing the Natewa peninsula from Ndreke-ni-wai to the head of
Fawn Harbour, one reaches a height of 660 feet above the sea. This ridge
represents the “divide” between the Lea and Waikawa mountain-ranges. In
the lower part of the northern slopes of this ridge occur basic tuffs
and agglomerates; but between 200 and 400 feet a light-coloured acid
rock of the hornblende-andesite type prevails, both in the form of
agglomerate and of loose blocks. This rock is described under the second
order of the hornblende-hypersthene andesites on page 299.

Descending the south slope, I found at an elevation of 500 feet a single
large mass, about 4 feet across, of quite another type of volcanic rock,
which is referred to the orthophyric order of the hypersthene-augite
andesites (page 290). It is a dark grey almost holo-crystalline rock
(sp. gr. 2·69) showing porphyritic pyroxene to the eye and displaying in
its relatively scanty groundmass short stout felspars, ·05 mm. in
length. On the surface of the lower two-thirds of this southern slope
occur basic tuffs and agglomerates, basaltic blocks being found in the
streams. The tuffs are palagonitic and contain a few calcareous
particles. They apparently contain some foraminiferous shells and are
doubtless of submarine origin.

It would seem that the axis or deeper portion of this ridge is composed
of the hornblende and hypersthene-augite andesites, whilst the basic
tuffs and agglomerates form the slopes.

With reference to the south side of the Lea sub-district, it may be
observed that whilst on the north side the mountains rise up close to
the sea-border, here they are separated from the coast by a broad tract
of lowland, where bedded pteropod and foraminiferous clay-rocks are
exposed, dipping gently to the south-east. Usually between Fawn Harbour
and Vunilangi Inlet the coast is margined by low cliffs of
coral-limestone, showing the massive corals in position; but sometimes
the deposits above noticed compose the low cliffs and even the islets
close by.

THE WAIKAWA MOUNTAINS.—This range occupies nearly the whole area of the
broad and elevated promontory that is only separated from Taviuni by the
narrow straits of Somo-somo, which, however, have a minimum depth of 120
fathoms. These mountains extend to the vicinity of Mbutha Bay on one
side and to near Fawn Harbour on the other. Several of the peaks reach
to over 1,000 feet, the greatest height given in the Admiralty chart
being 1,540 feet. The whole region has a very rugged aspect, the
mountains rising up near the coast, whilst the surface is much cut up
into ridges and valleys.

A single traverse across the range was alone made, but the results
obtained are very suggestive and may doubtless be applied to much of
this rugged promontory. I crossed the mountains from Loa to Waikawa. The
summit was about a mile broad and undulating, the level varying between
900 and 1,150 feet. At one place on the top there was a deep hollow,
some 300 or 400 yards across, perhaps the remains of an old
crater-cavity; but the higher slopes were so densely wooded that it was
not possible to get a clear view of my surroundings. Basic tuffs and
agglomerates prevailed on either side from the foot to the top of the
range. Specimens obtained from between 800 and 900 feet above the sea
are characteristic palagonitic tuffs of varying degrees of coarseness
containing 5 to 10 per cent. of carbonate of lime and a few tests of
foraminifera. At 1,100 feet I obtained a specimen which on account of
the large proportion of carbonate of lime (35 per cent.) and the
abundance of foraminiferal tests may be termed an impure foraminiferal
limestone belonging to the group of these rocks described on page 319.
The tests range up to a millimetre in size, and there are also inclosed
a few large fragments, 1 to 2 centimetres in size, of shells and
crystalline limestone. The residue is made of the detritus of
semi-vitreous basic rocks, palagonitic debris, fine clayey material and
minerals (15 per cent.), the last including beside plagioclase abundant
more or less perfect pyroxene prisms, mostly of the rhombic type.

Near the summit there occurred in one place blocks of a highly basic
blackish olivine-basalt (sp. gr. 2·99), marking evidently the situation
of a dyke. This rock is referred to genus 13 of the olivine-class. It
displays abundant phenocrysts of olivine and augite with but little
plagioclase. Interstitial glass is scanty, the groundmass consisting of
stoutish felspar-lathes (·06 mm. long), abundant augite granules and
prisms, and magnetite.

The Waikawa mountains would thus seem to possess the same general
structure that characterises many of the ranges of the island. Submarine
basic tuffs and agglomerates cover their sides and their summits, the
deeper rocks forming the axis of the range being in this case not so
frequently exposed.

On the coast between Waikawa and Navuni, rather over a mile east of Fawn
Harbour, basic agglomerates, palagonitic tuff-sandstones, and calcareous
clay-rocks containing pteropod and foraminiferous tests, prevail. The
tuff-sandstones and clay-rocks are bedded, the stratification being
often well shown in the horizontal sections displayed in the shore-flat.
In one locality within an area a few hundred yards across, a
quaquaversal dip was exhibited. At Navuni, where the hills reach the
coast, the same formations occur. I ascended the stream-course there for
about a mile, basic tuffs and agglomerates being exposed in its sides,
whilst blocks of a heavy dark olivine-basalt[89] lay in the bed. The hot
springs which issue inland at the side of this stream are described on
page 35.

THE BASIN OF THE NDREKE-NI-WAI RIVER.—With the region, which is bounded
on the north by the Mount Freeland Range or the Ngala mountains and on
the south by the Waikawa Range, I have but slight acquaintance, except
in the case of the coast fronting Natewa Bay. A little way up the course
of the river Ndreke-ni-wai, which drains this area, lies the town of
Koro-ni-yasatha, where Mr. Horne, the botanist, spent some days in 1878.
Probably much of this area is not over 200 feet above the sea, and
apparently there is a good deal of talasinga country.

THE COAST BETWEEN THE MOUTH OF THE NDREKE-NI-WAI RIVER AND THE FOOT OF
THE NGALA OR MOUNT FREELAND RANGE.—Between this estuary and Valavala,
two miles to the eastward, occurs a bedded calcareous palagonitic tuff
of sedimentary origin, dipping steeply to the north. In one locality
there is a rudely columnar dyke of a porphyritic augite-andesite. Coarse
basic tuffs exposed in the cliffs and shore-flat of Ko-nandi-nandi Point
on the side of Valavala Bay display a spheroidal structure, due probably
to the vicinity of some igneous intrusion. The sea-border extending from
this bay to Natewa, and farther on to Waikatakata, near the foot of the
Ngala mountains, is in most parts a broad low strip of coast-land, where
rock-exposures are infrequent. A dark grey andesite forms the blocks of
the agglomerate in this locality. It is noticeable on account of the
prismatic pyroxene of the groundmass; and it is assigned to genus 5 of
the second (prismatic) sub-order of the hypersthene-augite andesites. A
blackish semi-vitreous pyroxene-andesite occurs in the vicinity of
Natewa.

At Waikatakata (the Fijian word for “hot water”), where an outlying spur
of the Mount Freeland or Ngala Range reaches the coast, hot springs
issue on the hill-side, as described on page 34. On the slopes around
the springs lie huge masses of an aphanitic basaltic andesite having a
specific gravity of 2·81 and referred to genus 16 of the augite
sub-class. It displays a characteristic andesitic groundmass, showing
crowded felspar-lathes in flow-arrangement with average length of
·17 mm., and containing scarcely any residual glass.

Proceeding along the coast east of Waikatakata, one enters the region of
altered pyroxene-andesites, for which Mount Freeland, or the Ngala
Range, is remarkable. There are first to be observed on the shore blocks
of a grey altered ophitic dolerite, which belongs to the non-porphyritic
division of genus 10 of the augite-andesites and is described in detail
on page 275. Afterwards the characteristic rocks of the district occur.
The lofty spurs of the Ngala Range here reach the shore; and between
them lies the coast village of Ngara-vutu, from which the ascent to the
summit is best made.


                   MOUNT FREELAND OR THE NGALA RANGE

This high range forms a conspicuous object in the profile of this part
of the island. It derives its Fijian name from the old war-town of Ngala
that was situated at an elevation of 1,500 or 1,600 feet overlooking
Ngara-vutu on the west. The main mass of the range takes a crescentic
sweep, 3 or 4 miles in extent and facing Natewa Bay. It incloses the
coast district of Tunuloa. The steep mountain-slopes here rise to 2,000
feet and over, the greatest elevation being 2,740 feet. The densely
wooded spurs of the range occupy most of the area between the town of
Natewa, Kumbulau Point, and Mbutha Bay.

The summit is a long narrow ridge covered with a dense entangled mass of
Freycinetia stems which render progress very difficult. The rocks
exposed all the way up from the stream-courses in the vicinity of
Ngara-vutu to the top, are almost all of the same type, namely altered
augite-andesites. They are dark grey in colour and effervesce slightly
with an acid, whilst occasionally they show a little pyrites.[90]
Agglomerates were not observed and tuffs only in one locality, 1,200
feet above the sea, where a highly altered tuff composed of debris of
the prevailing rocks was exposed. At an elevation of 500 feet there
occurred a grey doleritic porphyritic rock, displaying large opaque
crystals of plagioclase, and looking like a porphyrite. It exhibits a
semi-ophitic groundmass, and is probably an intrusive mass. A
description of it will be found under genus 10 of the augite-andesites
(page 274).

The general petrological characters of the principal summit of Mount
Ngala point to its high antiquity as a volcanic mountain. It probably
ranks among the oldest extinct volcanic vents in the island.

Proceeding along the coast from Ngara-vutu to Navetau, the Buli’s town
of the Tunuloa district, one observes blocks of basic rocks. Farther to
the eastward for a distance of two or three miles basic agglomerates and
basic tuff-agglomerates prevail along the sea-border, being often
extensively exposed in the cliff-faces of the hills. The rock composing
the blocks is a hemi-crystalline pyroxene-andesite remarkable for the
prismatic pyroxene in the groundmass, and referred to genus 17 of the
hypersthene-augite sub-class.

I crossed the range to Ndevo on the other side of the peninsula from a
place three or four miles west of Kumbulau Point, rising on the way to a
height of 930 feet. For the lower 300 feet on the north slope basic
agglomerates and basic tuff-agglomerates are exposed. They are made up
of the same materials as those above described on the coast. Near and at
the summit occur compacted brecciated tuffs made up of palagonitised
basic materials but containing no lime, and these are associated with
light greenish fine-textured hard tuffs of an acid character, but
without lime or organic remains. In this last case the deposit is formed
of mineral fragments (oligoclase, rhombic and monoclinic pyroxene, &c.),
the debris of a hemi-crystalline volcanic rock, and a quantity of
greenish alteration products.

On the south side in the vicinity of Ndevo the sea-border is composed of
calcareous palagonitic clays and tuffs containing pteropod shells and
large tests of foraminifera in abundance. These deposits, which are
horizontally bedded, extend a mile inland and reach to between 250 and
300 feet up the slopes. A hot spring is stated to occur between the
tide-marks near Ndevo.

At intervals all along the coast between Ndevo and Nuku-ndamu, passing
on the way the villages of Koro-i-vonu, Tuvumila and Kanakana, fine and
coarse basic tuffs, often calcareous, are to be seen exposed in the
sea-cliffs and in the low hills behind. They are bedded and dip 5° to
10° W.S.W. near Ndevo and 15° to 20° S. b W. south of Kanakana. The
lower slopes of the Ngala mountains here approach the coast, and it is
highly probable that the submarine tuffs which form the sea-border
extend a considerable distance inland and to some height above the sea.
Between Nuku-ndamu and Tukavesi, where the mountains rise close to the
beach, occur basic agglomerates and agglomerate-tuffs, derived from
basaltic andesites. A specimen of the last named represents an uncommon
type of basaltic rock (sp. gr. 2·85), which on account of the character
of the felspars and of the pyroxene of the groundmass is referred to the
prismatic sub-order of the orthophyric order of the hypersthene-augite
andesites (page 290).

It may be inferred from the foregoing remarks that whilst the main
elevated mass of the Ngala Range is composed of altered basic andesites,
the product of ancient eruptions, the basic agglomerates, tuffs, and
clays, which occur on the lower slopes and in the outlying spurs, are of
later date. These tuffs and clays are evidently of submarine origin, at
least in the lower levels; but although fossiliferous deposits were not
observed at greater elevations than 300 feet above the sea, it is
probable that future investigators will find them at much higher levels.
Their discovery, as before noticed, at an elevation of 1,100 feet in the
adjacent Waikawa mountains, renders it likely that as in the case of the
mountainous districts of the main portion of the island the whole of the
Natewa peninsula was at one time submerged, or at least all the region
excepting the summit of Mount Ngala.



                               CHAPTER XV

      DESCRIPTION OF THE GEOLOGICAL AND GENERAL PHYSICAL FEATURES
                             (_continued_)


                  THE NORTH-EAST PORTION OF THE ISLAND

THIS large area, which extends for a distance of nearly forty miles from
the eastern slopes of the Mount Thurston Range to Undu Point, forms the
region closing in Natewa Bay on the north. It would be difficult to
imagine an area of this size with a greater variety of surface or
showing such a lack of arrangement of its principal features. The hills
and mountains on the north side gather at the coast, and extensive
inland plains, raised but a few feet above the sea and strewn with
silicified corals, occupy a portion of its interior. A long valley with
a very small gradient extends nearly across its breadth; and the rivers
are for the most part tidal estuaries fed, except in one or two cases,
by insignificant streams. There is, however, a lofty range of
ridge-mountains in its broadest part attaining a height of 2,500 feet;
whilst away to the east stretches the great Undu Promontory singular for
the straightness of its form.

Volcanic rocks of acid types, such as oligoclase-trachytes,
quartz-porphyries, and white pumice-tuffs prevail in the northern part
between Undu Point and the promontory opposite Mali Island. In the
southern part, from the foot of Mount Thurston and Vuinandi to the
vicinity of Tawaki, massive rocks and tuffs and agglomerates of basic
characters predominate. Although my acquaintance with this area is
incomplete, the data below given will be sufficient to enable a general
idea to be formed of its structure. The more conspicuous features in its
geology will gradually come into prominence as the various localities
visited are described.


  THE SOUTHERN SEA-BORDER BETWEEN VUINANDI AND THE VICINITY OF TAWAKI

The basic rocks, which characterise this long extent of coast, give
place about two miles west of Tawaki to the acid rocks. I will proceed
methodically with the description from Vuinandi eastward.

(_a_) _The coast between Vuinandi and Nakarambo._—Along this coast,
spurs from the Mount Thurston Range reach the borders of Natewa Bay,
forming a succession of small bays a mile or so across, on the shores of
which are situated the villages and towns of Vuinandi, Ndaku-ndaku,
Korotasere, and Nakarambo. The basaltic rocks characteristic of the
vicinity of Vuinandi are predominant here; but agglomerates and tuffs
are of rare occurrence, the prevailing rocks exposed in the spurs being
fine textured basaltic andesites (sp. gr. 2·82). A calcareous tuff-clay
is, however, to be observed in the east point of Vuinandi Bay, where it
is apparently penetrated by a dyke of the above mentioned rock. In
Ndaku-ndaku Bay there issue from the shore-flat between the tide-marks
some hot springs which are referred to on page 34.

(_b_) _The coast between Nakarambo and Waimotu._—In this tract we meet
with rocks of a somewhat different character, though still basic in
their type. Through the agglomerates and agglomerate-tuffs of the spurs
protrude coarsely crystalline grey pyroxene-andesites.[91] They are to
some degree altered, and are characterised by their abundant phenocrysts
of plagioclase and of rhombic and monoclinic pyroxene, the groundmass
being relatively scanty. The agglomerates are formed of similar
materials. There is an interesting exposure in the point east of
Nakarambo. My specimens from this locality were unfortunately lost; but
in my notes reference is made to the coarsely crystalline grey andesite
above noticed, to the later intrusion of a dark amygdaloidal rock, and
to an altered calcareous tuff.

(_c_) _The coast between Waimotu and Natasa Bay._—A low belt of land
often forms the sea-border. Between these two localities there is a
broad estuary, the village of Vanuavou being situated on the right side
and Malati on the left side. Tuffs usually calcareous and probably
submarine are here displayed together with basaltic andesites.

(_d_) _The coast between Natasa Bay and the vicinity of Tawaki._—This
extensive stretch of sea-border, nearly 15 miles in length, is
characterised by hilly spurs and long intervening low-lying tracts. The
prevailing rocks exposed are tuff-clays, somewhat hard and altered, and
coarser basic tuffs sometimes calcareous and overlying the former. These
sedimentary deposits, which are evidently submarine, are often bedded
and have a fairly constant dip over most of this region of 10 to 15
degrees to N.W. or W.N.W. They are frequently pierced by dykes, 6 to 10
feet thick, of compact basic rocks. A specimen of one of these
dyke-rocks from between Natasa and Vatu-karoa is a doleritic basalt with
scanty olivine (sp. gr. 2·81)[92]; but the rock of the dykes east of
Vatu-karoa is less basic (sp. gr. 2·72), and may be described as an
augite-andesite[93] with a doleritic structure in part disguised by
alteration. Between Natasa and Sangani there is a remarkable exposure of
an intrusive opaque white rhyolitic rock associated with altered tuffs.
This rock, which has undergone some degree of alteration, is described
on page 311. It is the only indication of the vicinity of a region of
acid rocks that I came upon on this tract of coast. It probably, if
traced inland, would be found connected with the district of acid rocks.
In following along the coast, however, towards Undu Point, the region of
acid rocks is not reached until within 2 or 3 miles west of Tawaki.
Reference should be made here to two hot springs that, as described on
page 33, rise up on the coast at Natuvu, a little east of Lakemba.

           THE INLAND MOUNTAINOUS REGION BETWEEN THE EASTERN
      FOOT OF THE MOUNT THURSTON RANGE AND THE VICINITY OF TAWAKI

This inland region corresponds for the most part to the tract of coast
before described between Vuinandi and the vicinity of Tawaki. We have
here a very much broken area traversed in a northerly and southerly
direction by mountain-ridges and penetrated in places on the north side
by prolongations of the Wainikoro and Kalikoso plains. The two loftiest
summits, the Ndoendamu and Savu-riti peaks of the chart, rise
respectively to heights of 2,481 and 2,238 feet. The former is the Hale
Peak of the Wilkes’ Expedition; but as regards the native names there is
some confusion and I have not been able to clear it up. Several of the
lesser peaks rise to over a 1,000 feet, and four or five of them to
rather over 1,500 feet.

With the exception of the fixing of the positions of the more
conspicuous peaks, the interior of this part of the island is
unsurveyed. I was able to lay down my positions approximately; but
before a systematic examination can be made of the geology of this
region a survey is necessary. Several traverses and ascents were made by
me; but much more in the way of exploration remains to be done. I
venture, however, to think that from the data below given a fairly
accurate notion of some of the leading geological features of this
region may be formed. The districts visited will be described in their
order from east to west.

(_a_) _The mountainous district east of the mountains of Vungalei and
Nailotha, extending to the vicinity of Tawaki._—Mr. Thomson[94] in his
map of the sea-border of this district shows a continuous coast range
from near Tawaki to Natasa lying about two miles inland. This is the
appearance when the region is seen in profile from a distance; but when
viewed from an elevation in its interior its surface is seen to be for
the most part traversed by a series of mountainous and hilly ridges
trending roughly north and south, the greatest height not exceeding
1,500 or 1,600 feet.

The geological character of the sea-border of this district has already
been described on page 208. It is there shown that fine and coarse
sedimentary tuffs, sometimes calcareous, and often penetrated by basic
dykes, here prevail. Their general story is one of great denudation, and
we have the same testimony impressed on us when examining one of the
inland mountain-ridges.

I followed the crest of one of these ridges from the hamlet of Nawi, on
the headwaters of the Vui-na-savu River, in a south-easterly direction
for about 4 miles to the peak of Uthulanga, which overlooks Lakemba on
the shore of Natewa Bay. During the walk my level rose fairly gradually
from that of Nawi, less than 100 feet above the sea, until near Lakemba,
where a height of 1,400 feet was attained. Here the ridge abruptly
terminates in the round-topped peak of Uthulanga, which rises steeply
from its crest for another 150 feet or rather more, and has a
precipitous face on the southern side descending far down the
mountain-slope. This peak, which is about 1,550 feet above the sea, when
seen from the opposite shores of Natewa Bay is very conspicuous and
looks like a thumb or a nose. The first part of the Fijian name of this
peak signifies a nose.

The hamlet of Nawi, from whence the ridge begins, is built on a low
mound-like “rise,” which is composed of a dark-grey hypersthene-gabbro.
Since plutonic rocks are of very rare occurrence in this island, the
Nawi gabbro has a particular interest. It belongs to the group of
plutonic rocks described on page 250; and is to be referred to the more
basic kinds, its specific gravity being 2·84. In appearance it is like a
diallage-gabbro, and in the slide displays monoclinic and rhombic
pyroxene filling the interspaces between the large plagioclase crystals
and undergoing respectively the diallage and bronzite stages of
schillerisation. This rock forms the type of the group and need not be
referred to in more detail here. It should be added that in the bed of
the neighbouring river occur blocks of a basaltic andesite (sp. gr.
2·86) referred to genus 1 of the hypersthene-augite andesites. It is a
less crystalline form of the gabbro just mentioned.

The prevailing rock exposed for the first 2 miles on the crest of the
ridge (by starting from Nawi) was a greenish porphyrite displaying large
opaque crystals of oligoclase, but showing much alteration of the
propylitic kind, its specific gravity being 2·5, but the structure of
the groundmass is much disguised. During the next mile, basic
agglomerates and a massive hypersthene-augite andesite were exposed. The
last-named is semi-vitreous, and on account of the prismatic pyroxene of
the groundmass is placed in genus 5 of its sub-class. Its specific
gravity is 2·7. In the last and fourth mile of the ridge was exposed a
dark grey doleritic basalt (sp. gr. 2·85), which rings like clinkstone
under the hammer, and weathers in a honeycombed fashion. The
felspar-lathes average ·3 mm. in length, the residual glass being
scanty. The pyroxene phenocrysts include some of rhombic pyroxene; and
the rock is therefore referred to genus 4 of the hypersthene-augite
andesites. The actual peak of Uthulanga, as it rises abruptly about 150
feet from the end of the ridge, is of agglomerate, the blocks being
composed of a compact grey andesite.

The peculiar succession of rocks displayed in this mountain-ridge along
its length of four miles is worthy of notice. Neither vesicular nor
scoriaceous rocks came under observation; and it is to be assumed that
there existed originally a line of submarine vents, some of them
ejecting acid and others basic materials. Mainly on account of the great
marine erosion during the period of emergence, but partly also on
account of the subsequent sub-aerial denudation, a plain ridge now
represents this line of vents. Probably the peak of Uthulanga, which is
evidently an old volcanic neck, represents the last of the stages in the
volcanic history of this ridge; but a very long period must have since
elapsed. When, however, we look at the exposed gabbro at the other end
of the ridge, we have to carry the period much farther back, since here
the superjacent surface volcanic rocks have been stripped off
completely. On page 2 I have referred to an island in the Solomon Group
where we have such a chain of ancient vents of acid and basic rocks. In
that case the forms of the separate hills indicating the original vents
are still to be recognised. In this old mountain-ridge of Vanua Levu no
such outlines remain except in the instance of the terminal peak.

(_b_) _The Nailotha and Vungalei Range._—This lofty range, which towers
above all around it, attains a height of 2,481 feet in the peak of
Nailotha and of 2,238 feet in that of Ndrukau or Vungalei. The first is,
as I infer, the Ndoendamu or Hale Peak of the charts; but it is
uncertain whether the name Savu-riti should be applied to the northern
peak of the range as it is in the charts, or whether it belongs to an
independent peak farther to the east. By the natives in the vicinity the
northern peak is known as Ndrukau, and I have added the name of Vungalei
from the village at its foot. The southern peak is that of Nailotha. The
range runs roughly north and south. It is, however, obvious that the
internal topography of this part of the island is but scantily known.

Brief reference will first be made to the country bordering the range on
the north-west and west sides in the vicinity of the villages of
Vungalei and Tembe. In proceeding south from Kalikoso, which lies in the
midst of the low-lying Wainikoro plains, to Vungale one traverses this
level district, of which the elevation is never more than 150 feet above
the sea and often much less. About a mile south-east of Kalikoso the
limit of the region of quartz-porphyries and of acid tuffs is passed and
the area of basic rocks is entered, a dark semi-vitreous
pyroxene-andesite with a flinty fracture prevailing at the surface as
far as Vungalei. This rock displays a few small phenocrysts of
oligoclase and pyroxenes in a blurred glassy groundmass exhibiting the
felspar and pyroxene microliths in process of development (sp. gr.
2·46).

Vungalei itself is only elevated about 130 feet above the sea. In
proceeding from this village to Tembe, lying about two miles S.S.W., one
crosses a line of hills, about 600 feet in height, which form a spur of
the main range. Basic agglomerates similar to those found on the slopes
of Mount Vungalei, as described below, prevail in the district between
these two villages up to the top of the intervening hills. In places one
notices that they overlie the ordinary sedimentary deposit, known as
“soapstone,” a slightly calcareous clay-tuff but showing no organic
remains to the eye. The rock exposed in the stream-courses is a
semi-ophitic basaltic andesite (sp. gr. 2·74) which contains a
considerable amount of interstitial glass.[95] When proceeding S.S.W.
from Tembe on the way to Nandongo one passes on either hand, as
described on page 216, hills about 700 feet high displaying vertical
cliffs formed of agglomerates over-lying finer sedimentary beds
apparently of the “soapstone” character.

I made the ascent of Mount Vungalei from the village of that name. The
peak is also known as Ndrukau. Basic agglomerates were exposed all the
way up to an elevation of about 2,000 feet, where a bed of a somewhat
scoriaceous basaltic andesite was displayed, the upper 200 feet being
inaccessible but of the same agglomerate. At a height of 300 feet was
observed another layer of the same basaltic rock. These intercalated
beds appeared to be lava-flows rather than horizontal dykes. The
agglomerates become less coarse in the upper part of the mountain where
they take the character of agglomerate-tuffs. The blocks are composed,
like those in many other parts of the island, of a dark semi-vitreous
basaltic andesite, but often scoriaceous. But the most remarkable
features of these agglomerates are the indications of two distinct
pauses in their deposition afforded by the occurrence at elevations of
900 and 1,700 feet of a single horizontal bed, two to three feet thick,
of coarse palagonite-tuffs. Each bed is exposed at the foot of a tall
cliff of agglomerate forming a line of escarpment along the
mountain-side. The larger fragments making up these tuffs are usually
from 2 to 3 millimetres in size; but the process of palagonitisation is
not complete; and we find inclosed abundant angular pieces of a dark
fresh tachylyte-glass, finely vesicular, and fusing readily in the
ordinary spirit-lamp flame. The tuffs contain little or no lime and
seemingly no organic remains. In the lowest bed I noticed a little
water-worn gravel. Both beds pass gradually into the underlying
agglomerate; but their upper limits are well defined and the agglomerate
commences abruptly.

It is thus seen that in the formation of the agglomerates there were two
pauses when they gradually gave place to deposits of fine detritus made
up of a vesicular basic glass that has since been largely converted into
palagonite. Then followed a sudden renewal of the agglomerate-producing
process. We can scarcely doubt that the agglomerates, with their
scoriaceous blocks, and the palagonite-tuffs are in the main the direct
products of volcanic eruptions. The rival claim of marine denudation as
the agency concerned can be mostly but not altogether excluded. The
agglomerates and tuffs of Mount Vungalei cannot be distinguished from
those so often described in the case of the great inland ranges of the
other parts of the island, the submarine origin of which is frequently
demonstrated by the inclosed organic remains. It would seem that in the
instance of Vungalei these deposits took place around the shores of a
volcanic mountain that rose above the sea. On page 315 it is pointed out
that in Stromboli with its dribbling eruptions we have a good
illustration of the manner in which such deposits could be formed.

My examination of the mountain of Nailotha was restricted to the lower
slopes up to an elevation of 600 feet; but the results obtained are
very suggestive. When following up the stream-course on the way from
Tembe to the mountain one notices in its bed blocks of a
light-coloured rock like a compact quartzite. It is, however, a highly
altered oligoclase-trachyte (sp. gr. 2·53) with its structure much
disguised by secondary quartz. This is the first intimation one gets,
on leaving behind the district of basic agglomerates about Tembe, of
the otherwise unexpected character of the acid rocks displayed in the
lower part of Nailotha.[96]

A torrent here cuts deeply into the mountain-side. At the base, between
200 and 300 feet above the sea, a bluish-grey rather scoriaceous rock
with the steam-pores drawn out to a length of from 1 to 2 centimetres,
is first exposed. Its specific gravity after allowing for the cavities
is less than 2·6. It shows a groundmass partly disguised by secondary
quartz and containing numerous small vesicles lined by quartz and filled
with viridite and epidote. Where the alteration is less advanced small
parallel felspar-lathes with fine decomposing pyroxenes are shown. The
lathes give extinctions of two or three degrees and average ·04 mm. in
length. In its somewhat scoriaceous nature, in the absence of
phenocrysts, and in its less altered condition, this rock differs from
those exposed higher up the ravine; but it is evidently to be referred
to the same acid type. At a height of 300 feet a light grey highly
altered oligoclase-trachyte (sp. gr. 2·43) is exhibited. It contained
originally some phenocrysts of felspar and pyroxene, which, however,
have been more or less replaced by calcite, quartz, and other materials;
whilst the groundmass, originally hemi-crystalline but now blurred by
the deposition of silica, shows felspar-lathes in process of
development.

A little farther up the gorge there is displayed another highly altered
rock with a siliceous appearance, such as has been noticed above as
forming blocks in the stream near the foot of the mountain. It is an
oligoclase-trachyte impregnated with crystalline silica and exhibiting a
singular prismatic structure, the small columns or prisms being only 3
or 4 inches in diameter. Between 400 and 500 feet occurs a
light-coloured compact rock sparkling with pyrites and also displaying a
columnar structure, the columns being between 4 and 6 inches across. It
looks like a limestone and effervesces freely with an acid; but it is in
fact a highly altered oligoclase-trachyte (sp. gr. 2·5) of the propylite
type. It seems originally to have inclosed a few phenocrysts of
oligoclase, sanidine, and pyroxene, and here and there a stout
felspar-lathe is to be noticed in the groundmass giving straight
extinction. The whole texture of the rock, however, is more or less
impregnated with secondary quartz, calcite, chlorite, viridite, pyrites,
&c. Farther up the ravine, between 500 and 600 feet in elevation, is
displayed a remarkable quartz-porphyry which exhibits opaque porphyritic
crystals of felspar as well as rounded crystals of quartz in a grey
compact base. It has been subjected to considerable alteration and will
be found described on page 310.

At 600 feet large slabs of the ordinary sedimentary clay-tuffs,
containing but little lime and showing no organic remains to the eye,
lay about on the more level slopes, and evidently formed a surface
deposit incrusting the altered massive rocks. If my ascent had lain up
the mountain-side away from the stream-courses, these sedimentary rocks
would alone have been observed. In the gorge, however, is exposed to
view the deeper-seated rocks that make up the mountain’s mass. We have
here a thickness of about 400 feet of altered acid rocks. All are
doubtless intrusive, and were subjected to alteration after the
development of the columnar structure and before the deposition of the
overlying clay-tuffs or “soapstones,” which are no doubt of submarine
origin. These sedimentary tuffs belong probably to the period when the
submarine agglomerates and palagonite-tuffs of the neighbouring peak of
Vungalei were formed.

(_c_) _Traverse of the coast range from Nandongo to Vanuavou on the
shore of Natewa Bay._—This route was taken by Mr. Horne, the botanist,
in 1878. I approached Nandongo from Tembe to the northward. The road at
first lay between hills about 700 feet in height displaying in their
precipitous faces agglomerates overlying fine sedimentary tuffs. These
deposits in the form of slightly calcareous basic tuff-clays, the
so-called “soapstones,” are exposed in the bed of the Wainikoro River as
one nears Nandongo. This village, which is situated on the headwaters of
the Wainikoro at an elevation of about 180 feet above the sea, lies near
the foot of the range. In its vicinity there is a small thermal spring
which is referred to on page 33.

Proceeding south from Nandongo one notices in the stream-course at the
foot of the slopes the sedimentary tuff deposits above mentioned,
bedding and dipping gently to the west. Farther up the slopes, higher
than 250 feet above the sea, there are exposed the deeper-seated rocks
of the range in the shape of compact reddish rocks (sp. gr. 2·48), which
appear under the microscope to be highly altered acid andesites or
oligoclase-trachytes originally displaying flow-structure and a fair
amount of glass, but now much disguised by the formation of secondary
quartz. On this north slope of the range I also found an amygdaloidal
variety of the same altered rocks containing irregular amygdules, 5 or 6
millimetres long, of fibrous quartz or chalcedony. Blocks of basaltic
andesite were observed on the summit, which has an elevation of 950 to
1,000 feet. On the southern slopes descending towards Natewa Bay coarse
basic tuffs together with blocks of basaltic andesite are chiefly
exposed. The last-named probably represent dykes both on the south slope
and on the summit. The rocks exhibited on the portion of the coast of
Natewa Bay corresponding to this range are dark and light-coloured
sedimentary tuffs usually calcareous, with occasional basaltic andesites
indicating dykes.... From this traverse it would appear that the range
has an axis of altered acid rocks overlain by basic sedimentary tuffs
and pierced by basaltic dykes.

(_d_) _The mountainous district lying between the head waters of the
Wainikoro River and the Mount Thurston Range._—Of this region I know
very little. The highest peak according to the chart has an elevation of
about 1,600 feet. Some indication of the character of the inland rocks
may be obtained from that of those exposed on the coast between
Nakarambo and Waimotu where, as observed on page 208, grey
pyroxene-andesites, coarse in texture and almost holo-crystalline in
structure, protrude through agglomerates of the same materials. When on
the way from Ngelemumu to Wainikoro I crossed the extreme northern
prolongation of this range where the elevation above the sea is only 700
feet. Non-calcareous basic tuff-clays occur on the slopes; but the
deeper-seated rocks, judging from an exposure on the east side, are dark
grey altered pyroxene-andesites penetrated by fine cracks filled with a
mosaic of quartz and having a specific gravity of 2·7. On the summit I
found a gritty sandstone-like rock, of which my specimen has been lost.
In a stream at the foot of the east slope occur small blocks of basaltic
andesite probably derived from a dyke. The region of acid rocks, such as
quartz-porphyries, oligoclase-trachytes, &c., is not entered until about
two miles south-west of Wainikoro.


          THE COAST RANGES AND SEA-BORDER BETWEEN MBUTHAI-SAU
                       AND THAWARO OR MBEKANA BAY

We have in this region the mountains and hills at the coast and the
low-lying plains inland. This feature of the north side of Vanua Levu is
very remarkable. For some sixty miles, that is to say, for more than
half the length of the island, between the mouth of the river Ndreketi
and Thawaro or Mbekana Bay, Vanua Levu possesses this character. The
coast ranges west of Lambasa, where basic rocks evidently prevail, have
been referred to on pages 135, 136. Those east of the Lambasa mountains
as far as Thawaro Bay will be dealt with here; and instead of basic we
find acid rocks, such as quartz-porphyries akin to the rhyolites,
oligoclase-trachytes, pumice-tuffs, &c.

The sea-border is here characterised not by a continuous range running
parallel to the coast, as in the case of the district between Nanduri
and the Ndreketi River, but by a number of separate groups of hills and
lesser mountains, separated by deep gaps or valleys which are occupied
by tidal rivers and extensive mangrove swamps. The tide ascends these
rivers into the plains behind the coast ranges, so that a depression of
only 30 feet would convert these groups of hills into separate islands
and would cover much of the inland plains with the sea. The hills attain
an elevation of 1,200 or 1,300 feet a mile or two inland and descend
often as bold promontories to the coast. I will refer in order to the
different parts of this sea-border.


           (1) THE SEA-BORDER BETWEEN LAMBASA AND MBUTHAI-SAU

In the sea-border between Lambasa and Mbuthai-sau we have the junction
of the regions of basic and acid rocks, the former extending westward to
Naivaka, the latter reaching to Undu Point. In such a locality the two
types of rocks might be expected to be associated, and this is what
occurs. Acid pumice-tuffs and basic agglomerates, sometimes associated,
are here displayed. In the low hills between Lambasa and Vuni-ika Bay,
which lies west of Mbuthai-sau, I found basic agglomerates prevailing,
together with some acid pumice-tuffs. The blocks in the agglomerates are
composed of a blackish semi-vitreous pyroxene-andesite (sp. gr. 2·68),
which is characterised by prismatic pyroxene in the groundmass, and is
referred to genus 6 of the second sub-order of the hypersthene-augite
andesites described on page 287.

East of Vuni-ika on the way to Mbuthai-sau, at an elevation of about 50
feet above the sea, dark tuffs containing small fragments of
reef-limestone are exposed in a cutting. A little farther on there is a
considerable deposit of a pale grey rhyolitic pumice-tuff, a soft stone
easily worked, and indeed now quarried by the plantation authorities. It
contains no lime and in microscopical characters is scarcely
distinguishable from a sample of fine pumice debris collected by me in
the Chirica district of Lipari Island. It is made of fragments up to a
centimetre in size, of ordinary fibrillar pumice in a matrix of much
finer material of the same nature. Portions of crystals of glassy
felspar (oligoclase and sanidine) also occur in it, together with some
quartz and rhombic pyroxene.

The association in this locality of acid and basic eruptive products was
observed by Dana in 1840 in the cliffs of “Mali Point.”[97] It is not
quite clear whether Mali Island, which lies immediately adjacent to the
coast, is here alluded to, or whether it is a headland opposite to it.
Dana describes the deposits displayed in these cliffs as coarse
aggregates of fragments of pumice and decomposing trachyte, which pass
on the one side into fine clayey material, and on the other into an
agglomerate formed of angular blocks of vesicular and compact basalt
with the interstices filled with pale yellow tufaceous material.

East of the Avuka Range that limits the Lambasa plains on this side is
the picturesque valley of Mbuthai-sau. This broad valley runs in a
south-east direction into the heart of the island. So small is the
gradient that the river flowing down it can be ascended in boats for
some miles; whilst Ngele-mumu, a village situated between 5 and 6 miles
up the valley, is not much over 50 feet above the sea. This valley in
its lower part roughly divides the regions of acid and basic rocks that
lie east and west of it respectively. It has, however, been above
pointed out that the two regions overlap in the coast region on the west
side of the valley. The two types of rocks are also associated in the
coast hills immediately east of the valley, since when striking inland
from Lloyd Point to the Mbuthai-sau sugar-cane district, one leaves
behind the acid rocks at the coast and traverses a region of basic
agglomerates. With these qualifications, therefore, the above line of
demarcation holds good.


 (2) THE SEA-BORDER BETWEEN THE MBUTHAI-SAU AND THE LANGA-LANGA RIVERS

The rocks predominating in this district are white and pale-yellow
pumice-tuffs and pumice agglomerates, with quartz-porphyries and
oligoclase-trachytes as intrusive masses. The light colour of the
sea-cliffs, thus composed, makes them conspicuous from seaward. Their
appearance evidently led Mr. Horne into an error in 1878 when he viewed
this coast from his canoe during his sea-passage from Lambasa to Tutu.
“The coast from Lambasa”—he says—“is a series of bold projecting bluffs
of agglomerate interspersed with seams of coralline sandstone.”[98] The
principal feature of this coast, however, is the prevalence of
light-coloured pumice-tuffs, though it is not improbable that elevated
reef-formations may occur in some localities.

I particularly examined the coast districts in the vicinity of the
Wainikoro and Langa-Langa rivers. In a spur that descends to the right
bank of the river, a little above the mouth of the Wainikoro, is exposed
an open-textured rhyolite or quartz-porphyry containing, as described on
page 310, phenocrysts of glassy felspar (oligoclase and sanidine),
quartz, and a little hornblende. Some interesting exposures occur on the
coasts between Narikosa Point (Nari-Roso Point of the chart) and the
mouth of the Wainikoro. Here the pumice-tuffs and agglomerates are
pierced by dykes, some of them vertical and 6 feet across and composed
of a light-grey rhyolitic rock, similar except in its compact texture to
the rock above mentioned.[99] The small quartz crystals, which are 1 to
2 millimetres in size, are sometimes bipyramidal. Many of them are
rounded and have a fused-like surface. The pumice-tuff, which displays
no effervescence with an acid, is composed mainly of finely pulverised
vacuolar and fibrillar istropic colourless glass, and contains also
small fragments of glassy felspar and small rounded quartz crystals,
such as occur in the rock forming the dykes in these deposits. Imbedded
in the tuffs in places are small masses, up to 4 inches in size, of a
pretty grey vesicular rhyolite-glass exhibiting the intermediate
condition between compact obsidian and pumice which is so characteristic
of the rocks of Vulcano in the Lipari Islands. A more detailed account
of this rock is given on page 311. In the pumice-agglomerates of this
locality occur pale-yellow decayed and altered pumice blocks.

The headland, known as Narikosa Point, which lies between the mouths of
the Wainikoro and Langa-langa rivers, terminates in a rocky spur formed
by a large intrusive mass or dyke of a reddish oligoclase-trachyte
altered in character and displaying in places a rudely columnar
structure.[100] The pumice-tuffs and agglomerates are well exposed at
the coast between Narikosa Point and the Langa-langa River. In the
vicinity of Songombiau, a village here situated, the tuffs are
penetrated by dykes. One dyke at the back of the village is 4 or 5 feet
thick and has vitreous margins. It is composed of a darkish
pyroxene-andesite which in the interior of the intrusion in
hemi-crystalline and a little vesicular, but in the margins it is more
glassy and rather scoriaceous.[101] Though not a basic rock in itself,
its specific gravity in the compact state being only about 2·55, it is
relatively basic when contrasted with the rhyolitic and trachytic rocks
of the district, which have when compact a specific gravity of not over
2·3.

Pumice-tuffs and agglomerates appear in the cliff-faces of the hills on
either side of the lower course of the Langa-langa River from the
vicinity of Kalikoso to the sea. At one place about 3 miles up the river
a dyke of trachyte, much altered and showing columns 12 to 18 inches
across, was exposed at the bank.


      (3) SEA-BORDER BETWEEN THE LANGA-LANGA RIVER AND THAWARO BAY

The next part of the coast which I visited was that opposite Tutu
Island. Here a range of high hills, 1,100 or 1,200 feet in height, sends
a lofty spur to the sea, in the precipitous faces of which are exposed
breccia-tuffs and agglomerates derived from acid rocks. A specimen of
the tuffs shows, besides fragments of altered rhyolitic or trachytic
rocks, portions of decomposing pumice, the vacuoles and tubular cavities
of which are filled with alteration products. The blocks in the
agglomerates are altered oligoclase-trachytes, both compact and
vesicular. These deposits are non-calcareous and rarely display bedding;
but in one place there was a rude arrangement of the materials, the dip
being to the north-west. I crossed this coast range where it is only 600
feet above the sea, and descended into the Kalikoso plains in the
vicinity of Numbu.

In the coast district between Tutu Island and the village of Naua, 3½
miles to the east, the same altered coarse pumiceous and trachytic
tuffs, occasionally bedded and dipping W.N.W., form the low hills near
the sea. In one place, where the elevation was less than 200 feet, I
noticed on the surface small fragments of branching corals. They had
perhaps weathered out of the tuffs, and though in part silicified still
effervesced freely in an acid.

Between Naua and the town of Visongo occur low hills formed of acid
tuffs. At an elevation of 200 feet the tuffs display a remarkable
character. When examined with the lens they are seen to be crowded with
the minute tests of foraminifera of the “Globigerina” type. The deposit
forms a fairly hard light-grey clay-rock which according to the usual
acid-test has no lime. Under the microscope it is seen to be composed
nearly in equal proportion of the tests of foraminifera and of very fine
detritus apparently derived from acid rocks, the materials being
generally not over ·01 mm. in diameter, though some of the felspar
fragments measure ·05 mm. across. A secondary process of silicification,
as exhibited in the occurrence of minutely granular quartz, has affected
the matrix of the clay as well as the tests of the foraminifera. The
deposit is of deep-water origin and appears to have been formed from the
fine washings of a coast some distance away.

It is noteworthy that the surface in this locality is in places strewn
with quantities of angular fragments and round pebbles of chalcedonic
flint, together with large nodules which when broken across are found to
be occupied by radiating crystals of quartz and were doubtless formed in
the cavities of some rock. Reference has already been made to the
partially silicified coral fragments lying on the surface of the low
coast hills a few miles to the west of this locality. They indicate a
relatively recent emergence and we get the same indication from the
flints on the surface of this district. These evidences, however, relate
only to the last stage of the emergence. The testimony of the silicified
“Globigerina” clay carries us back to the earlier periods of these
changes of level, and probably dates back to a time when the greater
part of this portion of the island was submerged, with the exception of
the mountain peaks. Not the least interesting feature of the emergence
is the silicifying process that accompanied it. This is illustrated on a
much greater scale in the neighbouring inland plains of Kalikoso and
Wainikoro which are described in Chapter XXV.

Proceeding eastward from Visongo to Namukalau one traverses a coast
district not elevated more than 300 feet above the sea. Here there are
displayed whitish and pale yellow compacted tuffs differing in aspect
from those prevailing to the westward and often bedded, the dip being
about 20° N.W. or N.N.W. They show no lime and apparently inclose no
organic remains. Where the upper surface of a bed is bared, it shows
regular shrinkage-lines inclosing hexagonal spaces 2 to 3 inches across;
but there is no corresponding columnar structure in the bed-mass. The
rock is very light in weight and homogeneous in texture and looks a
little like China-clay. In a section its structure appears obscure; but
it seems to be formed of the finest detritus, derived from some acid
partly devitrified glass, the pumiceous structure being in places
faintly indicated; but the whole mass appears to have been subjected to
a process of alteration perhaps similar to the ultimate palagonitic
change in basic rocks. In the slide a few small felspar fragments, about
·1 mm. in size, are displayed.

Just east of Namukalau is the mouth of the Vui-na-savu River, the Na
Savu River of the chart. This is a tidal river, and is navigable for
boats for several miles. In the lower part of its course agglomerates
and tuffs prevail, probably in part at least derived from acid rocks.
Near Rauriko, which lies about 5 miles up the river, a bare bluff,
overlooking the valley on the east, is formed of a decomposed trachytic
rock remarkable for the fact that it displays, as described on page 370,
faint magnetic polarity. Above Vitina, a mile or two farther up, I found
a similar rock but amygdaloidal in character. On the head-waters of the
river is situated the hamlet of Nawi, where, as mentioned on page 211, a
plutonic rock of the gabbro type occurs. Tuffs and agglomerates appear
to prevail in the low coast tract between the Vui-na-Savu river and
Thawaro Bay.

                  *       *       *       *       *

In drawing some general inferences respecting the acid formations,
mostly fragmental in character, that are displayed in the sea-border
between Lambasa and Thawaro Bay, it is necessary to distinguish between
the deposits west and east of the Langa-langa River. Between Lambasa and
the river just named the tuffs may usually be regarded as the products
of sub-aerial eruptions. Some of the specimens might have been obtained,
as far as their characters go, from the pumice-districts in the island
of Lipari. Their limeless condition and the apparent absence of organic
remains are noteworthy features, though of course the products of
sub-aerial eruptions may be deposited under the sea. It is, however,
remarkable that no compact obsidian came under my notice. The fragments
of rhyolitic glass, intermediate in structure between compact obsidian
and pumice, that were imbedded in the pumice-tuffs in one locality, were
probably ejected from some sub-aerial vent.

In the region between the Langa-langa River and Thawaro Bay acid tuffs
and agglomerates prevail; but they have all been subjected to alteration
by the deposition of secondary silica, and the pumice-structure when
present is largely disguised. They have evidently, in part at least,
been derived from compact rhyolitic and trachytic rocks, and are
probably in some measure the products of marine erosion. Although
neither lime nor organic remains were detected, the presence of the
altered “Globigerina” clay near Visongo is very suggestive and indicates
a considerable submergence of this region at some distant period.

Much, however, remains to be done in the examination of the peaks of the
coast ranges of this part of Vanua Levu; and it is likely that some
interesting results will be obtained from such an exploration.



                              CHAPTER XVI

      DESCRIPTION OF THE GEOLOGICAL AND GENERAL PHYSICAL FEATURES
                             (_continued_)

           THE NORTH-EAST PORTION OF THE ISLAND (_continued_)


                   THE WAINIKORO AND KALIKOSO PLAINS

THESE extensive inland plains occupy a considerable area in this part of
the island. I estimate that there is an area of about 20 square miles
that does not exceed an elevation of 100 feet above the sea and is often
much less. Of the two villages situated in the midst of these plains,
about 5 miles inland, Kalikoso is about 30 feet and Wainikoro is
scarcely 20 feet above the sea; whilst much of the surrounding district
is similarly low. Taking the 300-feet contour-line as a guide, this
low-lying region, as shown in the map, would be much larger. The plains
are backed on the south by the high mountain-range of Vungalei and
Nailotha. On the north the coast-ranges, which attain a height of 1,100
or 1,200 feet, separate them from the sea-border; whilst on the west
they are shut off from the Mbuthai-sau valley by a line of hills, of
which the minimum elevation is about 700 feet. This region is occupied
by the scanty open vegetation characteristic of the “talasinga” or
“sun-burnt” districts. The tall “Ngasau” reed is common; the Pandanus
trees are frequent; and amongst the bushes and small trees are
represented “Dodonæa viscosa,” “Morinda citrifolia,” and a species of
“Hibbertia.”

Different rivers and streams, rising in the mountains on the south side
of the plains, traverse this area; and after breaking through the
coast-ranges reach the sea. They acquire an exaggerated size from the
circumstance that they are in great part tidal estuaries. The tide
ascends them for several miles reaching behind the coast-ranges into the
plains; whilst the dense mangroves, which line their lower courses
amongst the hills, extend beyond into the low-lying level districts
farther inland. Boats can follow up the winding course of the Wainikoro
River until they reach the village of that name, which lies about five
and a half miles in a direct line from the coast and nearly in the
centre of this part of the island. The mangroves extend up to the
village. The Langa-langa River, which is of much smaller size, is
similarly navigable for three or four miles. The mangroves that line its
banks spread out in broad tracts when in ascending the river we emerge
from the hills into the plains. Above the influence of the tide it
dwindles into a stream ridiculously small. The same remarks apply to the
river two miles to the eastward. The Vui-na-savu River near the eastern
margin of this low-lying region can be ascended by boats into the heart
of the island.

From the foregoing remarks it will be expected that some portion of this
low-lying inland region will be occupied by swamps. This is in truth the
case. About one and a half miles north-east from the village of Kalikoso
is a small fresh-water lake, about 100 yards across, which lies in a
slight depression in the plains and is surrounded by a wide margin of
swampy ground, occupied by reeds and sedges, in which one sinks
knee-deep when approaching the banks. The level of the surface of the
lake is not over 25 feet above the sea, and only a foot or so below that
of the plains around. Since the depth, as I ascertained it, is 15 to 18
feet, it follows that the bottom of the lake is only a few feet above
the high-tide level. The mangroves extend to within a mile of its
border; and it is possible that though lying about five miles inland, it
may be indirectly affected by unusually high tides. It would be
interesting to determine whether the water is ever brackish.

This small lake is, or was, regarded with superstitious awe by the
natives on account of the floating islands that it contains. Different
legends are connected with it, and the Fijians have given it the name of
“Vaka-lalatha,” in allusion to the drifting of the islands from one side
of the lake to the other, the small trees growing upon them acting “in
the manner of sails.” Mr. Horne, who was in this locality in 1878,
refers to the lake in his book, _A Year in Fiji_ (p. 24); but does not
appear to have seen it. Mr. Thomson[102] visited it in 1880, and
describes it as containing a single floating island, a quarter to a half
an acre in extent. The island was in existence, the natives said, in the
days of their great-grandfathers, a statement indicating that the people
of the district had no reason to doubt its antiquity. A chief, who
formed one of the party, dived under the island.

When I visited this lake in 1899 there were three floating islands,
named by the natives “Wanga levu” and “Wanga lailai,” that is, “Large
canoe” and “Small canoe.” The largest was 90 or 100 feet long, whilst
the two smaller were about 50 feet long, the breadth being less than
half the length. They are composed of a dense growth of reeds and sedges
supporting small living trees 10 to 17 feet in height, swamp ferns, and
other smaller vegetation, the whole forming fairly solid
standing-ground, and doubtless possessing considerable thickness.

The origin of these floating islands is probably to be found in the
circumstance that the dense mass of swamp-vegetation lies on a rocky
substratum, and that during some unusually heavy flood large portions of
the swampy soil-cap became detached and floated up. A floating island
occurs near the sea in the Lauthala district on the Lower Rewa in Viti
Levu. The floating island in Derwentwater in the north of England is
said to be “a blister of sublacustrine turf.” Those in Russia and
Hungary, according to Mr. Hanusz, appear to be felted masses of
tree-trunks, branches, and marsh-plants thinly covered with soil.[103]

I will preface my remarks on the geology of the Wainikoro and Kalikoso
plains by observing that this region displays three conspicuous
features. In the first place, it is a region of acid rocks, mostly
altered tuffs, derived from quartz-porphyries, the alteration consisting
in the deposition of quartz often of the chalcedonic type. In the second
place, a silicifying process has been in operation here on an extensive
scale, as evidenced by the abundance of silicified corals lying on the
surface, especially in the district around the lake and by the abundance
of concretions of chalcedony, of fragments of chalcedonic flints, and of
portions of white chalcedonic quartz-rock that in places strew the
ground. In the third place, earthy limonite, or bog iron ore, has been
produced in quantity, particularly around the lake; and in places small
round concretions of impure carbonate of iron cover the soil. Before
referring more in detail to the different parts of this region, it may
be remarked that the silicified corals, flints, iron-ore deposits, &c.,
of this and other parts of the island are dealt with in Chapter XXV.

This region of acid rocks extends about two miles to the west and
south-west of the village of Wainikoro. Here prevail white acid tuffs
often decomposing and altered by the formation of secondary quartz. They
are derived from the degradation of quartz-porphyries or rhyolitic
rocks. The surface is irregular but the elevation is small, the 100-feet
level lying about one mile and the 200-feet level about two miles west
of Wainikoro.

The villages of Wainikoro and Kalikoso are from two to three miles
apart, the intervening district not attaining a greater elevation than
70 feet above the sea. Decomposing altered white acid tuffs here occur
with occasional large blocks, a couple of tons in weight, of apparently
a quartz-porphyry or trachyte with its structure disguised by
silicification. Fragments of siliceous concretions, as chalcedony,
chalcedonic flints, &c., lie on the surface, the soil being friable and
of a deep ochreous red; whilst in places the ground is covered with
round marble-sized concretions composed of a mixture of carbonate of
iron and limonite which I have described on page 356. A few hundred
yards to the north-west of Kalikoso, where there is a little rise, a
decomposed quartz-porphyry is exposed displaying rounded crystals of
quartz fractured in position, the matrix of the rock being impregnated
with chalcedony. In one part of this mound there is a friable white
rock, composed of a crumbling mass of small irregular quartz-grains,
rarely showing crystal-faces. It appears to be disintegrated
quartz-felsite.

In taking the track from Kalikoso to the village of Vungalei, which is
distant about 2½ miles to the south-east, one traverses after the first
mile, where the acid rocks terminate, a rather more elevated region
which rises to a maximum height of 180 feet above the sea. Though the
acid rocks give place to a semi-vitreous pyroxene-andesite as described
on page 212, small fragments of chalcedonic flints are frequent on the
surface over both areas. The district that intervenes between Kalikoso
and the landing-place on the Langa-langa River, about two miles in
length, is not more than 30 feet above the sea, and is crossed by small
sluggish streams, in the beds of which occur numerous fragments of
silicified corals, up to 3 or 4 inches in size, belonging to the Porites
and Astræan type. In these stream-beds also occur bits of mamillated
chalcedony and of onyx, flattish agates, 3 or 4 inches across, and
pebbles of the compact pyroxene-andesite above alluded to, the last
probably derived from the south side of the plains.

The plains extend in a north-east direction as far as Numbu, which lies
between 2½ and 3 miles north-east of Kalikoso. The country between these
two places is usually elevated between 60 and 100 feet and is never more
than 130 feet above the sea. It is known as the Kuru-kuru district. The
surface is strewn with the fragments of flints and of a white
chalcedonic quartz-rock; whilst the soil is friable and deep-red in
colour, limonite in abundant fragments occurring on the ground. Here and
there one passes slabs of a hard white silicified tuff, small portions
of which are frequent on the surface.

Silicified corals and earthy limonite are to be found in abundance
scattered over the surface of the plains immediately surrounding the
small lake of Vakalalatha. We also find lying on the ground in this
district bits of chalcedony and onyx, portions of chalcedonic flints,
and nodules of the size of the fist, which when fractured either
disclose the regular layers of the agate or radiating crystals of
quartz. The silicified corals occur usually in fragments not over 4
inches across, and include portions of branching corals of the Madrepore
habit, bits of massive corals of the Astræan type, small rounded nobs of
“Porites” just as one commonly observes on a reef-flat, &c. They have an
ancient weathered look, and in some cases it is evident from the
existence of boring-shells in the fractured end of a branching coral
that it once lay as a dead fragment on the surface of a reef-flat.

In Chapter XXV. the characters of the silicified corals of the island
are discussed in detail; and I have there advanced the view that the
extensive silicification of the Kalikoso and Wainikoro plains took place
during the consolidation of the calcareous muds of a reef-flat whilst
the land was emerging. I have already alluded on page 222 to an area of
silicification on the neighbouring sea-border between Visongo and Tutu.
There can be no doubt that during the last stage of the emergence the
present situation of the fresh-water lake near Kalikoso was occupied by
the sea, which also extended far over the surrounding plains. The
process of silicification would of necessity be restricted to the period
that has since elapsed; and the discussion is therefore confined to the
nature of the conditions under which this change was induced. As a
factor in the process we cannot disregard the acid character of the
rocks of the district.


                          THE UNDU PROMONTORY

The north-east portion of the island terminates in a long narrow
promontory which I have named after Undu Point at its extremity.
Commencing at Thawaro Bay and near Tawaki it runs in a straight line for
a distance of between 13 and 14 miles, its breadth varying between 1½
and 2½ miles. Its greatest elevation of nearly 1,600 feet is attained at
its western end; and it diminishes irregularly in height as one proceeds
towards Undu Point, where a height of 400 feet is maintained about a
mile from the cape. It is bordered by reefs sometimes a mile in breadth,
and the reefs prolong the promontory for another three miles beyond Undu
Point. As indicated by the 100-fathom line, the submarine contour
corresponds to that of the land, and the extent of marine erosion during
the existing relations of land and sea is evidently displayed in the
breadth of the reefs. I found no sign of upraised reefs; and although
diligent inquiries were made nothing could be learned of any hot
springs.

It will be seen from the following remarks that pumice-tuffs,
quartz-porphyries, and oligoclase-trachytes, are the prevailing rocks.
On the north side they may be regarded as continuous with the acid rocks
of the region extending from near Lambasa to Thawaro Bay. On the south
side they commence in the vicinity of Tawaki.


     (1) THE DISTRICT EXTENDING TWO AND A HALF MILES WEST OF TAWAKI

When proceeding eastward along the north coast of Natewa Bay one enters
the region of acid rocks between 2 and 2½ miles west of Tawaki. Here the
country is much broken, picturesque hills with bare precipitous faces
rising up near the coast, one of which named Natoto has a rudely conical
and truncated form. Grey oligoclase-trachytes having a specific gravity
of 2·4 and possessing the characters described on page 308, prevail in
the district extending west of Tawaki. Sometimes they occur in mass; but
they often form agglomerates. A singular pitchstone-agglomerate occurs
at the coast at the foot of Natoto. The pitchstone, which has a specific
gravity of 2·48, is a semi-vitreous form of a hypersthene-augite
andesite. It shows abundant small pyroxene prisms in its glassy
groundmass and is referred to the prismatic sub-order (5) described on
page 289.


                       (2) NAITHOMBOTHOMBO RANGE

A high range of hills, forming the backbone of this part of the island,
extends eastward for about five miles from Thawaro and Tawaki. It is
named “Naithombothombo” in the Admiralty chart. From it rise two
conspicuous peaks, Thawaro Peak (1,573 feet) at its western end, where
it overlooks the village of that name, and Mount Thuku (1,288 feet) near
its eastern end. West of Thawaro Peak this range is connected with the
hills beyond by a saddle 600 feet in height, which is ascended when
crossing the promontory from Thawaro Bay to Tawaki.

(_a_) _Thawaro Peak._—This represents an old “volcanic neck” of
agglomerate rising out of the tuffs that are exposed on its slopes to an
elevation of about 600 feet. As viewed from the saddle above mentioned,
the upper part of the hill presents bare precipitous sides, several
hundred feet in height, of agglomerate, the blocks of which are composed
of a compact hypersthene-augite andesite (sp. gr. 2·48). It displays a
few small phenocrysts of medium andesine and of rhombic and monoclinic
pyroxene; and is referred to the prismatic sub-order (5) described on
page 289, characterised by prismatic pyroxene in the groundmass.

(_b_) _South coast between Tawaki and the foot of Mount Thuku._—The tall
cliffs that rise to a height of from 200 to 300 feet behind Tawaki are
composed of white tuffs and agglomerate-tuffs derived from the acid
rocks of the district. Eastward from Tawaki to the base of Mount Thuku
the coast scenery is particularly fine. A little inland a line of hills,
named “Na Kula,” rises precipitously to a height of 700 or 800 feet, and
in the vertical sides are displayed tuffs and agglomerates probably of
the character of those above noticed. Light-coloured tuffs are sometimes
exposed at the coast in which are inclosed fragments varying in size of
a pitchstone[104] (sp. gr. 2·36) approaching in structure a trachytic
glass. At one place the tuffs were evidently sedimentary and bedded, the
dip being about 15° N.W.

The massive rock most frequently exposed at the coast and on the
hill-slopes between Tawaki and Mount Thuku is a quartz porphyry[105]
displaying abundant porphyritic crystals of quartz and felspar in a
groundmass originally semi-vitreous but now obscurely felsitic in
character. The shore-flat for more than half a mile west of Mount Thuku
is strewn with great numbers of detached columnar blocks, 12 to 15
inches across, of a slightly vesicular oligoclase-trachyte of the type
described on page 308.

(_c_) _North coast between Thawaro Bay and the foot of Mount
Thuku._—Coarse and fine tuffs prevail at the coast and on the
neighbouring hill-slopes; and in the bare rocky faces of the hills
inland they are also to be observed. The streams have worn deep gorges
into their mass. Towards Thuku they are acid and pumiceous, and are
evidently the products of eruption. Towards Thawaro, they are more basic
and darker, and are in part at least to be attributed to marine
degradation.

(_d_) _Mount Thuku._—My ascent of this hill, which is 1,288 feet in
height, was made from the north coast. I found it to be composed from
the foot to the summit of white pumiceous tuffs without any evident
arrangement. It has a narrow top and shows no sign of a crateral
cavity. The hills east and west, as viewed from its summit, are
ridge-shaped and display nothing in their configuration at all
suggestive of craters. The pumice-tuffs of Mount Thuku are
non-calcareous, and exhibit greyish pumiceous lapilli in an abundant
white matrix formed of fine pumice-debris. Under the microscope it
shows the characteristic vacuolar and fibrillar structure; but the
material has not the fresh appearance of ordinary pumice and the
minute cavities are often filled with alteration products. The two
rocky points on the north coast opposite the hill are formed in one
case of a somewhat altered oligoclase-trachyte and in the other of a
quartz-porphyry. Both no doubt represent intrusive masses, the almost
horizontal columns, 12 to 15 inches in diameter, of the former
indicating a nearly vertical dyke.


              (3) THE UNDU PROMONTORY EAST OF MOUNT THUKU

East of Mount Thuku the hilly backbone of the promontory is of much less
elevation. About three miles to the eastward the highest hill is 630
feet, and thence to within a mile or two of Undu Point the hills retain
a height of 400 to 500 feet.

(_a_) _The north coast between Mount Thuku and the coast village of
Nuku-ndamu._—On this stretch of coast, about five miles in length, the
shore-cliffs are composed of white and pale-yellow, coarse and fine
stratified pumice-tuffs, the beds being either horizontal or with a
gentle dip northward. They are as a rule non-calcareous, and contain
some quartz grains and small bits, 1 to 3 millimetres in size, of bottle
green compact obsidian, much as one finds in Lipari pumice-tuffs. In
general character, both naked-eye and microscopic, they correspond to
the Mount Thuku pumice-tuffs above described. Large blocks of basic
rocks are occasionally to be observed on this coast, sometimes probably
indicating dykes, but in one place near Mount Thuku forming an
agglomerate. The rock is a dark-grey augite-andesite with a specific
gravity of 2·77. It is compact and has a hemi-crystalline
groundmass.[106]

(_b_) _The south coast between Mount Thuku and Moala, a distance of
about five miles._—Pumice-tuffs and agglomerates are displayed at the
coast, the former often bedded and in one place having a dip of 35 or 40
degrees to the north. A quartz-porphyry, somewhat banded and a little
altered, and displaying rounded quartz crystals 3 or 4 millimetres in
size, is the prevailing massive rock exposed on the hill-slopes and
occasionally at the coast. It is well exhibited about a mile east of
Mount Thuku. Blocks of a grey oligoclase-trachyte also occur. These
rocks are described on pages 308, 309.

(_c_) _The terminal portion of the promontory from Nuku-ndamu and Moala
to Undu Point._—The same pumiceous tuffs, usually non-calcareous, form
the shore-cliffs on the south coast from Moala to Mr. Bulling’s station
at Ndothiu, which lies about 2 miles from the point. On the
corresponding part of the north coast between Nuku-ndamu and Ndothiu
these tuffs are often calcareous; and near the first-named place they
contain sub-angular bits of coral of the size of a walnut. On the beach
at Vunikondi in this locality they are overlain by nearly horizontal
beds of basic lava, the upper surface of which when exposed displays the
smooth, “ropy” crust of a lava flow. The rock is a dark slightly
vesicular augite-andesite, hemi-crystalline in structure, and containing
a fair amount of residual glass.[107] Since the underlying tuffs were
evidently deposited on a sea-bottom, it follows that this is a submarine
flow. I intended to revisit this locality, but was prevented. A detailed
examination of it would be worth undertaking.

From Ndothiu to Undu Point, about 2 miles distant, the interior of the
promontory has an undulating surface, the elevation being usually 200 or
300 feet and rising to 400 feet. On the coasts are exposed bedded
pumiceous tuffs, steeply inclined and usually calcareous. As displayed
in the hill-slopes of the interior they are horizontally stratified and
as a rule non-calcareous. These deposits sometimes exhibit a spheroidal
arrangement indicative of the proximity of a dyke. In one or two places
at the coasts occur basic agglomerates, formed of the same
augite-andesite lava-rock of which the Vunikondi beds are composed, but
scoriaceous and containing more glass in the groundmass. In the
hand-specimen beside me, the steam cavities are of all sizes, from that
of a pin-prick to a third of an inch (8 mm.) and are generally
elongated.

A careful search of the tuff-deposits in this part of the Undu
promontory ought to result in the discovery of remains, both of plants
and of marine mollusks. Mr. Chalmers informed me that fossilised
tree-trunks occur on the coast near Vunikondi; but I was unfortunately
not able to discover them.

                  *       *       *       *       *

BRIEF SUMMARY OF THE GENERAL CHARACTERS OF THE UNDU PROMONTORY FROM
THAWARO AND TAWAKI TO UNDU POINT.—One suggestive negative feature of
this region is to be found in the absence, as far as I could ascertain,
of any trace of a crater. Here also, as in the area of acid rocks
extending westward along the north side of the island to near Lambasa,
hot springs are not to be found. The prevailing rocks are in the first
place the pumice-tuffs, which not only as a rule form the coast-cliffs,
but occur inland as high as the summit of Mount Thuku almost 1,300 feet
above the sea. They were probably in great part ejected from sub-aërial
vents, though no doubt, as in the vicinity of Undu Point, they were
often deposited under the sea. The acid and basic tuffs in the vicinity
of Tawaki and Thawaro are, as I imagine, largely derived from marine
degradation. Next to the pumice-tuffs, massive quartz-porphyries and
oligoclase-trachytes are the characteristic rocks. They are probably in
most cases intrusive, and present themselves sometimes as vertical
columnar dykes, evidently of considerable dimensions.

The basic rocks, however, are not unrepresented. They occur in one or
two places as agglomerates, as in the vicinity of Mount Thuku on the
north coast and near Undu Point; whilst they form “flows” overlying the
pumiceous tuffs at Vunikondi. Occasional blocks lying on the surface on
the north coast are indications of dykes. The basic rocks, nevertheless,
take a very secondary part in the composition of the Undu Promontory.
They are in most cases to be referred to the augite-andesites with a
specific gravity 2·6 to 2·77; but some, as in the case of those forming
the agglomerates of Thawaro Peak, are hypersthene-augite andesites with
specific gravity of about 2·5. Olivine-basalts are not represented.

The vents, from which the materials forming this promontory were
ejected, were arranged in a single straight line for a length of 14
miles. Along this linear fissure, which was probably submarine, vents
were at different times formed; and though owing to sub-aerial and
marine degradation the present surface has been since shaped and
reshaped, their situation may still be recognised in the “necks” of tuff
and agglomerate that form the peaks, and in the large dykes or sills of
quartz-porphyry and oligoclase-trachyte.



                              CHAPTER XVII

                    THE VOLCANIC ROCKS OF VANUA LEVU


THE varied character of the volcanic rocks in my collection is brought
out in the following Table, where I have grouped about 400
rock-sections, excluding those of the tuffs and finer detrital deposits.
The small proportion of plutonic rocks should be noted.

 Olivine-basalts                             23 per cent.
 Augite-andesites                            40  "   "
 Hypersthene-augite-andesites                17  "   "
 Acid andesites, including hornblende
    and quartz-andesites, &c.                12  "   "
 Oligoclase-trachytes, quartz-porphyries
    or rhyolites                              6  "   "
 Hypersthene-gabbros and diorites             2  "   "
                                           ----
                                            100

In order to avoid the necessity of frequently describing rocks of the
same type it has been found requisite to devise a method of
classification. In carrying out this somewhat laborious task I have
often been surprised at the readiness with which rocks, the relations of
which had been previously very difficult to ascertain, fell into their
place in the scheme. Although many of my uncertainties have been thus
removed, a large number of doubtful points remain. I venture to think,
however, that others may be able to employ and also to extend the method
of classification here employed.

The general plan followed has been worked out in detail for the
olivine-basalts and the pyroxene-andesites of the more basic type;
whilst lack of materials has prevented its further elaboration in the
case of the acid andesites, oligoclase-trachytes, quartz-porphyries, &c.
The treatment of all the classes has been uniform, the scheme being the
same whether applied to a basalt or to a dacite.

In describing the general method I will take the Augite class. As will
be seen in the Synopsis that follows, this class is first divided
according to the absence or presence of a groundmass into two
sub-classes, the first referring to the plutonic rocks, which, however,
are not represented in my collection, the second comprising the
augite-andesites which make up 40 per cent. of the total. These
andesites are again divided into four orders, according as the
groundmass presents parallel or non-parallel felspar-lathes, or short
and stout felspars (orthophyric), or displays a felsitic character. The
last two orders are practically unrepresented here, though many examples
of them are found amongst the more acid andesites. Each order is then
split up into three sub-orders depending on the nature of the pyroxene
of the groundmass, whether granular, prismatic, or ophitic.

Each sub-order is broken up into two sections, one displaying
plagioclase-phenocrysts, the other without them, or possessing very few
of them. The first section is divided into two genera, according to the
character of the plagioclase-phenocrysts, whether glassy or opaque, the
second genus often comprising rocks allied to the porphyrites. The
second or aphanitic section is subdivided into two genera according to
the character of the pyroxene-phenocrysts; in the one case they are
macroporphyritic; in the other they are either small or absent. The
genera are split into four species according to the length of the
felspar-lathes, a method which readily separates out the doleritic
rocks. In cases where the materials are abundant, the genera have been
first divided into porphyritic and non-porphyritic sub-genera, based on
the macroporphyritic or the micro-porphyritic character of the
plagioclase-phenocrysts, when present. The species can be also split
into sub-species, according to the degree of basicity of the rocks, as
indicated by the specific gravity.

This method is fully worked out in the later pages and need not be
further described here. With abundant material from different regions it
appears to me that a ready mode is here afforded of assigning to a rock
its place in the scheme. In this way it would be possible to follow the
systematist in his method of comparing plants and animals from different
localities. To facilitate this end, I have suggested in the synopsis the
employment of abbreviations, so that the description of the critical
characters of a rock can be condensed into a formula capable of easy
interpretation.

As an example of the use of these abbreviations I will take the instance
of a common form of augite-andesite which is represented by the
formula:—“_Plag, aug, matr, flu, gran, non-phen, parv, ·1-·2 mm._” This
is the formula for an aphanitic augite-andesite, and it signifies that
it is a rock of the plagioclase-augite class possessing a groundmass
showing parallel felspar-lathes (between ·1 and ·2 mm, in average
length) and granular pyroxene, and displaying no phenocrysts of
plagioclase or only a very few of small size, whilst pyroxene
phenocrysts if present are micro-porphyritic.

As another example the following formula for a type of porphyrite found
in this island may be given:—_Plag, hypersth-aug, matr, orth, prism,
phen, opac_. This is an andesite in which rhombic and monoclinic
pyroxene are associated both in the phenocrysts and in the groundmass
where the pyroxene is prismatic and not granular. The plagioclase
phenocrysts are opaque and the felspars of the groundmass are of the
orthophyric type.

There are one or two points that require further reference. In the first
place the early employment in the scheme of the characters of the
felspars of the groundmass for distinguishing the orders scarcely seems
needed in the cases where they take the lathe-form; but the importance
of its early use is shown in the acid andesites where it is certainly of
prime importance in an early stage of the classification to distinguish
the rocks by the character of the felspars of the groundmass, whether
lathe-like, orthophyric, or felsitic. It may also be objected that the
two orders obtained by dividing the lathe group into “parallel” and
“non-parallel” divisions are not equivalents of the two other orders,
the orthophyric and the felsitic. The distinction, however, between the
flow or non-flow arrangement, though in practice not always readily
established, is a far-reaching one. On _à priori_ grounds the first
division might be expected to have no plutonic equivalent; whilst in the
second division, it is easy to trace the gradations through the
doleritic stage, where the felspar-lathes are very large, to the
granitoid condition. Then, again, the ophitic habit is as a general rule
confined to rocks with a doleritic or semi-doleritic groundmass, where
the felspar-lathes are coarse and form a mesh-work. Two quite distinct
lines of development unite in the felspar-lathes and begin to diverge
with the difference in their arrangement in the groundmass.

The nature of the difference between the flow and non-flow arrangement
of the felspar-lathes is well brought out in some dykes of basalt and
augite-andesite that I examined in this island and also in the Valle del
Bove on the Etna slopes. In the outer vitreous portion the
felspar-lathes, which are fairly well represented, are all about the
same length and are more or less parallel with the sides of the dyke. In
the central more crystalline portion two sets of lathes can be
distinguished, one (_A_) corresponding in the average length and in the
flow-arrangement with the lathes of the borders, the other (_B_) being
about half the length and forming a plexus between the larger parallel
lathes. Those of the _A_ set, which are those that usually catch the eye
in a section, are contemporaneous in their origin with those in the
margins of the dyke; whilst those of the _B_ set have been subsequently
formed. In the preliminary “stiffening” of the first stage of
consolidation, the whole mass of the dyke would be affected. To this
stage the lathes of the _A_ set belong; whilst to the later stage of
consolidation which would proceed much more slowly in the interior than
at the margins of the dyke, the lathes of the _B_ set are to be
referred. This distinction so plainly illustrated in a dyke must be
postulated for all intrusive masses; but I have not yet found it
possible to make much use of it. Much ground will have first to be
cleared before it can be safely employed, since it is apparent, for
instance, that there are often all gradations in a slide between a lathe
and a phenocryst, and that the term “phenocryst” is applied to crystals
having very different histories.

[Illustration: Diagram illustrating the two sets of felspar-lathes in a
dyke: _A_, long and parallel, dispersed through the mass. _B_, short and
non-parallel and found only in the centre.]

With regard to the ophitic habit of some of the basaltic rocks the
following conclusions may be drawn:

(_a_) Typical ophitic “plates” are not very common in the slides. More
frequently the habit of the pyroxene is semi-ophitic.

(_b_) This character is as a general rule associated with the plexus or
non-fluidal arrangement of the felspar lathes.

(_c_) The felspar-lathes are nearly always large, frequently averaging
more than ·2 mm. in length. This coarse doleritic groundmass is almost
diagnostic of an ophitic rock.

                  *       *       *       *       *

It is not always possible to allow for the influence of locality in
drawing up such a classification as this, since it is well known that in
each volcanic region the rocks have a particular facies recognisable in
hand-specimens as well as in the slide, though it is not easy to express
such a distinction in a definition. Perhaps this is represented in
“adaptation” as we find it in the organic world, and the question arises
as to the value of such characters for critical purposes. Regional
variation plays such an important part that it cannot be ignored in
rock-classification.


                                SYNOPSIS

               I. CLASS.—OLIVINE ROCKS (_Plag., oliv._).

 +----------------+------------------+---------------------------------+
 | SUB-CLASS.     |     DIVISION.    |            ORDER.[108]          |
 +----------------+------------------+---------------------------------+
 |    1.          |                  |                                 |
 | No groundmass  |                  |                                 |
 |  (_non-matr._) |                  |                                 |
 | LIMBURGITES,   |                  |(Not represented in collection.) |
 | PERIDOTITES,   |                  |                                 |
 | OLIVINE        |                  |                                 |
 | GABBROS,       |                  |                                 |
 |     etc.       |                  |  {              1.              |
 |                |                  |  { Felspar lathes not           |
 |                |{        1.       |  { in flow-arrangement          |
 |                |{Olivine abundant |  { (_non-flu._).                |
 |                |{    (_cop._)     |  {              2.              |
 |                |{                 |  { Felspar lathes               |
 |     2.         |{                 |  { in flow-arrangement          |
 | Groundmass     |{                 |  { (_flu._).                    |
 | (_matr._)      |{                 |                                 |
 | OLIVINE-BASALTS|{                 |  {              3.              |
 |                |{                 |  { Felspar lathes not           |
 |                |{        2.       |  { in flow-arrangement          |
 |                |{Olivine          |  { (_non-flu._).                |
 |                |{relatively       |  {           4.                 |
 |                |{scanty           |  { Felspar lathes               |
 |                |{(_pauc._).       |  { in flow-arrangement          |
 |                |                  |  { (_flu._).                    |
 +----------------+------------------+---------------------------------+

                II. CLASS.—AUGITE ROCKS (_Plag., aug._).

 +----------------------+----------------------------------------------+
 |       SUB-CLASS.     |                     ORDER.                   |
 +----------------------+----------------------------------------------+
 |                      |                                              |
 |        1.            |                                              |
 | No groundmass        |  Not represented in collection.              |
 | (_non-matr._)        |                                              |
 | GABBROS, in part.    |                                              |
 |                      |{                          1.                 |
 |                      |{ Felspar lathes not in flow-arrangement      |
 |                      |{ (_non-flu._).                               |
 |        2.            |{                                             |
 | Groundmass (_matr._) |{                          2.                 |
 | BASALTIC ANDESITES;  |{ Felspar lathes in flow-arrangement          |
 |  AUGITE-ANDESITES;   |{ (_flu._).                                   |
 | PYROXENE-ANDESITES   |{                          3.                 |
 |     in part.         |{ Felspars short and broad (orthophyric,      |
 |                      |{ _orth._).[109]                              |
 |                      |{                          4.                 |
 |                      |{ Groundmass felsitic (_fels._)[109]          |
 +----------------------+----------------------------------------------+

    III. CLASS.—HYPERSTHENE-AUGITE ROCKS (_Plag., hypersth.-aug._).

 +----------------------------+---------------------------------------+
 |        SUB-CLASS.          |            ORDER.                     |
 +----------------------------+---------------------------------------+
 |            1.              |                                       |
 |No groundmass (_non-matr._) |Represented by the Hypersthene-gabbros |
 |    GABBROS, in part.       |  in collection.                       |
 |                            |                                       |
 |                            |{             1.                       |
 |                            |{Felspar lathes not in flow-arrangement|
 |                            |{  (_non-flu._).                       |
 |                            |{                                      |
 |                            |{             2.                       |
 |            2.              |{Felspar lathes in flow-arrangement    |
 |    Groundmass (_matr._)    |{  (_flu._).                           |
 |PYROXENE-ANDESITES, in part.|{                                      |
 |                            |{             3.                       |
 |                            |{Felspars of the groundmass short and  |
 |                            |{  broad (orthophyric, _orth._).       |
 |                            |{                                      |
 |                            |{             4.                       |
 |                            |{Groundmass granular or presenting a   |
 |                            |{  mosaic (felsitic, _fels._).         |
 +----------------------------+---------------------------------------+

       IV. CLASS.—HYPERSTHENE ROCKS.
        V. CLASS.—HORNBLENDE-HYPERSTHENE ROCKS.
       VI. CLASS.—QUARTZ-HORNBLENDE-HYPERSTHENE ROCKS OR DACITES.

These last three classes are merely provisional. They include the Acid
Andesites of Vanua Levu, which are all characterised by the prevalence
of rhombic pyroxene amongst the phenocrysts and by its predominance or
rather by its usually exclusive occurrence in the groundmass. All the
classes are capable of being split up into two sub-classes and four
orders as in the case of the third class. The characters of these rocks
are given in Chapter XXI.

                     VII. CLASS.—HORNBLENDE ROCKS.

In this class are included those rocks where hornblende is the only
ferro-magnesian mineral. It is only represented by two diorites
described on page 251.

 VIII. CLASS.—OLIGOCLASE-TRACHYTES.           }
   IX. CLASS.—QUARTZ-PORPHYRIES AND RHYOLITES } Described in Chapter XXI.


                  FIRST ORDER OF THE OLIVINE BASALTS.

      CHARACTERS.—Olivine in abundance. Felspars of the groundmass
                        not in flow-arrangement.

                FORMULA.—_Oliv., matr., cop., non-flu._

 +------------+---------------+------------+------------+--------------+
 | SUB-ORDER. |   SECTION.    |   GENUS.   |  SPECIES.  |              |
 +------------+---------------+------------+------------+              |
 |            | Presence or   |            |            |  Number of   |
 |  Pyroxene  | absence of    | Character  |   Length   |  specimens.  |
 |   of the   | plagioclase   |   of the   | of felspar |              |
 | groundmass.| phenocrysts.  |phenocrysts.|  lathes.   |              |
 +------------+---------------+------------+------------+--------------+
 |            |               |{    1.     |{·02-·1 mm. |      1       |
 |            |               |{  Glassy   |{ ·1-·2 "   |      2       |
 |            |{     1.       |{plagioclase|{ ·2-·3 "   |              |
 |            |{ Plagioclase  |{phenocrysts|{ ·3-·5 "   |              |
 |            |{ phenocrysts  |{ (_vitr._) |            |              |
 |            |{  (_phen._)   |{           |            |              |
 |            |{              |{    2.     |{·02-·1 mm. |              |
 |            |{              |{  Opaque   |{ ·1-·2 "   |       1      |
 |            |{              |{plagioclase|{ ·2-·3 "   |       2      |
 |            |{              |{phenocrysts|{ ·3-·5 "   |              |
 |            |{              |{ (_opac._) |            |              |
 |    1.      |{              |            |            |              |
 | Granular   |{              |{    3.     |            |              |
 | (_gran._)  |{              |{  Large    |{·02-·1 mm. |              |
 |            |{              |{phenocrysts|{ ·1-·2 "   |              |
 |            |{              |{of olivine |{ ·2-·3 "   |              |
 |            |{    2.        |{    and    |{ ·3-·5 "   |       1      |
 |            |{    No        |{ pyroxene, |            |              |
 |            |{ plagioclase  |{ over 2 mm.|            |              |
 |            |{ phenocrysts  |{ (_magn._) |            |              |
 |            |               |{           |            |              |
 |            |{ (_non-phen._)|{    4.     |            |              |
 |            |               |{  Small    |{·02-·1 mm. |              |
 |            |               |{phenocrysts|{ ·1-·2 "   |              |
 |            |               |{of olivine |{ ·2-·3 "   |              |
 |            |               |{ and       |{ ·3-·5 "   |       1      |
 |            |               |{  pyroxene |            |              |
 |            |               |{under 2 mm.|            |              |
 |            |               |{ Pyroxene  |            |              |
 |            |               |{scanty and |            |              |
 |            |               |{  often    |            |              |
 |            |               |{  absent   |            |              |
 |            |               |{ (_parv._) |            |              |
 |            |               |            |            |              |
 |            |               |{5. _Vitr._ |}           |              |
 |            |{  3. _Phen._  |{           |}           |              |
 |    2.      |{              |{6. _Opac._ |}           |              |
 | Prismatic  |{              |            |}           |{ Not         |
 | (_prism._) |{              |{7. _Magn._ |}           |{ represented.|
 |            |{4. _Non-phen._|{           |}           |              |
 |            |               |{8. _Parv._ |}           |              |
 |            |               |            |            |              |
 |            |               |{9. _Vitr._ |}           |              |
 |            |{ 5. _Phen._   |{           |}           |{ Not         |
 |    3.      |{              |{10. _Opac._|}           |{ represented.|
 |  Ophitic   |{              |            |            |              |
 |  (_oph._)  |{              |{11. _Magn._|            |{ Not         |
 |            |{6. _Non-phen._|{           |            |{ represented.|
 |            |               |{12. _Parv._| ·1-·2 mm.  |      1       |
 +------------+---------------+------------+------------+--------------+


                  SECOND ORDER OF THE OLIVINE-BASALTS.

      CHARACTERS.—Olivine in abundance. Felspars of the groundmass
                          in flow-arrangement.

                  FORMULA.—_Oliv., matr., cop., flu._

 +------------+---------------+------------+------------+--------------+
 | SUB-ORDER. |    SECTION.   |   GENUS.   |  SPECIES.  |              |
 +------------+---------------+------------+------------+              |
 |            |  Presence or  |            |            |   Number of  |
 |  Pyroxene  |   absence of  | Character  |   Length   |   specimens. |
 |   of the   |  plagioclase  |   of the   | of felspar |              |
 | groundmass.|  phenocrysts. |phenocrysts.|  lathes.   |              |
 +------------+---------------+------------+------------+--------------+
 |            |               |{   13.     |{·02-·1 mm. |       3      |
 |            |               |{  Glassy   |{ ·1-·2  "  |       2      |
 |            |               |{plagioclase|{ ·2-·3  "  |       1      |
 |            |{      7.      |{phenocrysts|{ ·3-·5  "  |              |
 |            |{  Plagioclase |{ (_vitr._) |            |              |
 |            |{  phenocrysts |{           |            |              |
 |            |{   (_phen._)  |{   14.     |{·02-·1 mm. |              |
 |            |{              |{  Opaque   |{ ·1-·2  "  |       3      |
 |            |{              |{plagioclase|{ ·2-·3  "  |              |
 |            |{              |{phenocrysts|{ ·3-·5  "  |              |
 |            |{              |{ (_opac._) |            |              |
 |     4.     |{              |            |            |              |
 |  Granular  |{              |{   15.     |            |              |
 |  (_gran._) |{              |{  Large    |{·02-·1 mm. |       2      |
 |            |{              |{phenocrysts|{ ·1-·2  "  |              |
 |            |{              |{of olivine |{ ·2-·3  "  |              |
 |            |{              |{   and     |{ ·3-·5  "  |              |
 |            |{              |{ pyroxene, |            |              |
 |            |{              |{ over 2 mm.|            |              |
 |            |               |{ (_magn._) |            |              |
 |            |{      8.      |{           |            |              |
 |            |{      No      |{   16.     |}           |              |
 |            |{  plagioclase |{  Small    |}           |              |
 |            |{  phenocrysts |{phenocrysts|}           |              |
 |            |{ (_non-phen._)|{of olivine |}·02-·1 mm. |       1      |
 |            |               |{   and     |} ·1-·2  "  |       5      |
 |            |               |{ pyroxene, |} ·2-·3  "  |       2      |
 |            |               |{under 2 mm.|} ·3-·5  "  |       1      |
 |            |               |{ Pyroxene  |}           |              |
 |            |               |{ scanty and|}           |              |
 |            |               |{   often   |}           |              |
 |            |               |{  absent   |            |              |
 |            |               |{ (_parv._) |            |              |
 |            |               |            |            |              |
 |            |               |{17. _Vitr._|}           |              |
 |            |{  9. _Phen._  |{           |}           |              |
 |     5.     |{              |{18. _Opac._|}           |{ Not         |
 |  Prismatic |{              |            |}           |{ represented.|
 | (_prism_.) |{              |{19. _Magn._|}           |              |
 |            |{ 10.          |{           |}           |              |
 |            |{ _Non-phen._  |{20. _Parv._|}           |              |
 |            |               |            |            |              |
 |            |               |{21. _Vitr._|}           |              |
 |            |{ 11. _Phen._  |{           |}           |              |
 |     6.     |{              |{22. _Opac._|}           |{ Not         |
 |  Ophitic   |{              |            |}           |{ represented.|
 |  (_oph._)  |{              |{23. _Magn._|}           |              |
 |            |{ 12.          |{           |}           |              |
 |            |{ _Non-phen._  |{24. _Parv._|}           |              |
 +------------+---------------+------------+------------+--------------+


                  THIRD ORDER OF THE OLIVINE-BASALTS.

       CHARACTERS.—Olivine scanty. Felspars of the groundmass not
                          in flow-arrangement.

                FORMULA.—_Oliv., matr., pauc., non-flu._

 +------------+---------------+------------+------------+--------------+
 | SUB-ORDER. |    SECTION.   |   GENUS.   |  SPECIES.  |              |
 +------------+---------------+------------+------------+              |
 |            |  Presence or  |            |   Length   |   Number of  |
 |  Pyroxene  |   absence of  | Character  | of felspar |   specimens. |
 |   of the   |  plagioclase  |   of the   | lathes and |              |
 | groundmass.|  phenocrysts. |phenocrysts.|  prisms.   |              |
 +------------+---------------+------------+------------+--------------+
 |            |               |{    25.    |{·02-·1 mm. |       1      |
 |            |               |{  Glassy   |{ ·1-·2  "  |      13      |
 |            |{      13.     |{plagioclase|{ ·2-·3  "  |       5      |
 |            |{  Plagioclase |{phenocrysts|{ ·3-·5  "  |              |
 |            |{  phenocrysts |{ (_vitr._) |            |              |
 |            |{   (_phen._)  |{           |            |              |
 |            |{              |{    26.    |{·02-·1 mm. |              |
 |            |{              |{  Opaque   |{ ·1-·2  "  |       4      |
 |            |{              |{plagioclase|{ ·2-·3  "  |              |
 |            |{              |{phenocrysts|{ ·3-·5  "  |              |
 |    7.      |{              |{ (_opac._) |            |              |
 | Granular   |{              |            |            |              |
 | (_gran._)  |{              |{    27.    |            |              |
 |            |{              |{   Large   |{·02-·1 mm. |}             |
 |            |{              |{phenocrysts|{ ·1-·2  "  |} Not         |
 |            |{              |{of olivine |{ ·2-·3  "  |} represented.|
 |            |{              |{    and    |{ ·3-·5  "  |}             |
 |            |{      14.     |{ pyroxene, |            |              |
 |            |{      No      |{ over 2 mm.|            |              |
 |            |{  plagioclase |{ (_magn._) |            |              |
 |            |{  phenocrysts |{           |            |              |
 |            |{ (_non-phen._)|{    28.    |            |              |
 |            |               |{   Small   |            |              |
 |            |               |{phenocrysts|{·02-·1 mm. |}             |
 |            |               |{of olivine |{ ·1-·2  "  |} Not         |
 |            |               |{    and    |{ ·2-·3  "  |} represented.|
 |            |               |{ sometimes |{ ·3-·5  "  |}             |
 |            |               |{ pyroxene, |            |              |
 |            |               |{under 2 mm.|            |              |
 |            |               |{ (_parv._) |            |              |
 |            |               |            |            |              |
 |            |               |{29. _Vitr._|}           |              |
 |            |{ 15. _Phen._  |{           |}           |              |
 |     8.     |{              |{30. _Opac._|}           |{ Not         |
 |  Prismatic |{              |            |}           |{ represented.|
 | (_prism_.) |{              |{31. _Magn._|}           |              |
 |            |{ 16.          |{           |}           |              |
 |            |{ _Non-phen._  |{32. _Parv._|}           |              |
 |            |               |            |            |              |
 |            |               |            |{·02-·1 mm. |              |
 |            |               |{33. _Vitr._|{ ·1-·2  "  |       3      |
 |            |{ 17. _Phen._  |{           |{ ·2-·3  "  |       1      |
 |     9.     |{              |{           |{ ·3-·5  "  |       2      |
 |   Ophitic  |{              |{           |            |{ Not         |
 |  (_oph_.)  |{              |{34. _Opac._|            |{ represented.|
 |            |{              |            |            |              |
 |            |{              |{35. _Magn._|}           |{ Not         |
 |            |{ 18.          |{           |}           |{ represented.|
 |            |{ _Non-phen._  |{36. _Parv._|}           |              |
 +------------+---------------+------------+------------+--------------+


                  FOURTH ORDER OF THE OLIVINE-BASALTS.

      CHARACTERS.—Olivine scanty. Felspar-lathes of the groundmass
                          in flow-arrangement.

                  FORMULA.—_Oliv., matr., pauc., flu._

 +------------+---------------+------------+------------+--------------+
 | SUB-ORDER. |    SECTION.   |   GENUS.   |  SPECIES.  |              |
 +------------+---------------+------------+------------+              |
 |            |  Presence or  |            |            |              |
 |  Pyroxene  |   absence of  | Character  |   Length   |   Number of  |
 |   of the   |  plagioclase  |   of the   | of felspar |   specimens. |
 | groundmass.|  phenocrysts. |phenocrysts.|  lathes.   |              |
 +------------+---------------+------------+------------+--------------+
 |            |               |{    37.    |{·02-·1 mm. |      2       |
 |            |               |{  Glassy   |{ ·1-·2  "  |     12       |
 |            |               |{plagioclase|{ ·2-·3  "  |      6       |
 |            |{      19.     |{phenocrysts|{ ·3-·5  "  |              |
 |            |{  Plagioclase |{ (_vitr._) |            |              |
 |            |{  phenocrysts |{           |            |              |
 |            |{   (_phen._)  |{    38.    |{·02-·1 mm. |              |
 |            |{              |{  Opaque   |{ ·1-·2  "  |      2       |
 |            |{              |{plagioclase|{ ·2-·3  "  |              |
 |            |{              |{phenocrysts|{ ·3-·5  "  |              |
 |     10.    |{              |{ (_opac._) |            |              |
 |  Granular  |{              |            |            |              |
 |  (_gran._) |{              |{    39.    |            |              |
 |            |{              |{   Large   |{·02-·1 mm. |}             |
 |            |{              |{phenocrysts|{ ·1-·2  "  |} Not         |
 |            |{              |{of olivine |{ ·2-·3  "  |} represented.|
 |            |{              |{   and     |{ ·3-·5  "  |}             |
 |            |{              |{ pyroxene, |            |              |
 |            |{      20.     |{over 2 mm. |            |              |
 |            |{      No      |{ (_magn._) |            |              |
 |            |{  plagioclase |{           |            |              |
 |            |{  phenocrysts |{    40.    |            |              |
 |            |{ (_non-phen._)|{   Small   |            |              |
 |            |               |{phenocrysts|{·02-·1 mm. |              |
 |            |               |{of olivine |{ ·1-·2  "  |              |
 |            |               |{ and       |{ ·2-·3  "  |      2       |
 |            |               |{  sometimes|{ ·3-·5  "  |      1       |
 |            |               |{ pyroxene, |            |              |
 |            |               |{under 2 mm.|            |              |
 |            |               |{ (_parv._) |            |              |
 |            |               |            |            |              |
 |            |               |{41. _Vitr._|}           |              |
 |            |{ 21. _Phen._  |{           |}           |{ Not         |
 |     11.    |{              |{42. _Opac._|}           |{ represented.|
 |  Prismatic |{              |            |            |              |
 | (_prism._) |{              |{43. _Magn._|            |{ Not         |
 |            |{ 22.          |{           |            |{ represented.|
 |            |{ _Non-phen._  |{44. _Parv._| ·02-·1 mm. |      1       |
 |            |               |            |            |              |
 |     12.    |               |            |            |              |
 |   Ophitic  |               |            |            |{ Not         |
 |  (_oph._)  |               |            |            |{ represented.|
 +------------+---------------+------------+------------+--------------+


                  FIRST ORDER OF THE AUGITE-ANDESITES.

     CHARACTERS.—Felspar lathes or prisms of the groundmass not in
                           flow-arrangement.

                    FORMULA.—_Aug., matr., non-flu._

 +------------+---------------+------------+------------+--------------+
 | SUB-ORDER. |   SECTION.    |   GENUS.   |  SPECIES.  |              |
 +------------+---------------+------------+------------+              |
 |            |  Presence or  |            | Length of  |   Number of  |
 |  Pyroxene  |   absence of  | Character  |  felspar   |   specimens. |
 |   of the   |  plagioclase  |   of the   | lathes and |              |
 | groundmass.|  phenocrysts. |phenocrysts.|  prisms.   |              |
 +------------+---------------+------------+------------+--------------+
 |            |               |{     1.    |{·02-·1 mm. |       9      |
 |            |               |{   Glassy  |{ ·1-·2  "  |       5      |
 |            |               |{plagioclase|{ ·2-·3  "  |       4      |
 |            |{       1.     |{phenocrysts|{ ·3-·5  "  |              |
 |            |{  Plagioclase |{ (_vitr._) |            |              |
 |            |{  phenocrysts |{           |            |              |
 |            |{   (_phen._)  |{     2.    |{·02-·1 mm. |       6      |
 |            |{              |{   Opaque  |{ ·1-·2  "  |       4      |
 |            |{              |{plagioclase|{ ·2-·3  "  |       3      |
 |            |{              |{phenocrysts|{ ·3-·5  "  |       3      |
 |      1.    |{              |{ (_opac._) |            |              |
 |  Granular  |{              |            |            |              |
 |  (_gran._) |{              |{     3.    |            |              |
 |            |{              |{   Large   |}           |              |
 |            |{              |{phenocrysts|}           |{ Not         |
 |            |{              |{of augite, |}           |{ represented.|
 |            |{       2.     |{over 2 mm. |}           |              |
 |            |{       No     |{ (_magn._) |}           |              |
 |            |{  plagioclase |{           |            |              |
 |            |{  phenocrysts |{     4.    |            |              |
 |            |{ (_non-phen._)|{   Small   |            |              |
 |            |               |{phenocrysts|{·02-·1 mm. |              |
 |            |               |{of augite, |{ ·1-·2  "  |       3      |
 |            |               |{   under   |{ ·2-·3  "  |       1      |
 |            |               |{   2 mm.,  |{ ·3-·5  "  |       2      |
 |            |               |{  or none  |            |              |
 |            |               |{ (_parv._) |            |              |
 |            |               |            |{·02-·1 mm. |       1      |
 |            |               |{5. _Vitr._ |{ ·1-·2  "  |       2      |
 |            |{ 3. _Phen._   |{           |            |              |
 |      2.    |{              |{6. _Opac._ |  ·1-·2  "  |       2      |
 |  Prismatic |{              |            |            |              |
 | (_prism._) |{              |{7. _Magn._ |}           |{ Not         |
 |            |{ 4.           |{           |}           |{ represented.|
 |            |{ _Non-phen._  |{8. _Parv._ |}           |              |
 |            |               |            |            |              |
 |            |               |            |{·02-·1 mm. |       1      |
 |            |               |            |{ ·1-·2  "  |       3      |
 |            |               |{9. _Vitr._ |{ ·2-·3  "  |       4      |
 |            |{ 5. _Phen._   |{           |{ ·3-·5  "  |       7      |
 |            |{              |{           |            |              |
 |      3.    |{              |{           |{ ·2-·3 mm. |       1      |
 |   Ophitic  |{              |{10. _Opac._|{ ·3-·5  "  |       1      |
 |  (_oph._)  |{              |            |            |              |
 |            |{              |{11. _Magn._|            |              |
 |            |{ 6.           |{           |            |              |
 |            |{ _Non-phen._  |{12. _Parv._|{ ·2-·3 mm. |       3      |
 |            |               |            |{ ·3-·5  "  |       3      |
 +------------+---------------+------------+------------+--------------+


                 SECOND ORDER OF THE AUGITE-ANDESITES.

   CHARACTERS.—Felspar-lathes of the groundmass in flow-arrangement.

                      FORMULA.—_Aug., matr., flu._

 +------------+---------------+-------------+-----------+--------------+
 | SUB-ORDER. |   SECTION.    |   GENUS.    |  SPECIES. |              |
 +------------+---------------+-------------+-----------+              |
 |            |  Presence or  |             | Length of |   Number of  |
 |  Pyroxene  |  absence of   |  Character  |  felspar  |   Specimens. |
 |   of the   |  plagioclase  |   of the    | lathes and|              |
 | groundmass.| phenocrysts.  |phenocrysts. |  prisms.  |              |
 +----------------------------+-------------+-----------+--------------+
 |            |               |{     13.    |{·02-·1 mm.|      10      |
 |            |               |{  Glassy    |{ ·1-·2  " |      10      |
 |            |               |{ plagioclase|{ ·2-·3  " |       4      |
 |            |{      7.      |{ phenocrysts|{ ·3-·5  " |       1      |
 |            |{ Plagioclase  |{ (_vitr._)  |           |              |
 |            |{ phenocrysts  |{            |           |              |
 |            |{  (_phen._)   |{     14.    |}          |              |
 |            |{              |{   Opaque   |}          |              |
 |            |{              |{ phenocrysts|}·02-·1 mm.|       2      |
 |      4.    |{              |{      of    |}          |              |
 |  Granular  |{              |{ plagioclase|}          |              |
 |(_gran._)   |               |{  (_opac._) |           |              |
 |            |{              |             |           |              |
 |            |{              |{    15.     |}          |              |
 |            |{              |{   Large    |}          |{ Not         |
 |            |{       8.     |{phenocrysts |}          |{ represented.|
 |            |{ No (or very  |{ of augite, |}          |              |
 |            |{few and small)|{ over 2 mm. |}          |              |
 |            |{ plagioclase  |{  (_magn._) |           |              |
 |            |{ phenocrysts  |{            |           |              |
 |            |{(_non-phen._)  |{    16.    |{·02-·1 mm.|      12      |
 |            |               |{Small augite|{ ·1-·2  " |      14      |
 |            |               |{phenocrysts,|{ ·2-·3  " |       4      |
 |            |               |{ under 2 mm.|{ ·3-·5  " |       3      |
 |            |               |{  or none   |           |              |
 |            |               |{ (_parv._)  |           |              |
 |            |               |             |           |              |
 |            |               |{ 17. _Vitr._|  ·1-·2 mm.|       1      |
 |            |{ 9. _Phen._   |{            |           |{ Not         |
 |            |{              |{ 18. _Opac._|           |{ represented.|
 |      5.    |{              |             |           |              |
 |  Prismatic |{              |             |           |{ Not         |
 | (_prism._) |{              |{ 19. _Magn._|           |{ represented.|
 |            |{ 10.          |{            |{·02-·1 mm.|       4      |
 |            |{ _Non-phen._  |{ 20. _Parv._|{ ·1-·2 "|       4      |
 |            |               |             |           |              |
 |            |               |{ 21. _Vitr._|  ·1-·2 mm.|       3      |
 |            |{ 11. _Phen._  |{            |           |{ Not         |
 |      6.    |{              |{ 22. _Opac._|           |{ represented.|
 |   Ophitic  |{              |             |           |              |
 |  (_oph._)  |{              |{ 23. _Magn._|}          |              |
 |            |{ 12.          |{            |}          |{ Not         |
 |            |{ _Non-phen._  |{ 24. _Parv._|}          |{ represented.|
 +------------+---------------+-------------+-----------+--------------+


            FIRST ORDER OF THE HYPERSTHENE-AUGITE ANDESITES.

          CHARACTERS.—Felspar-lathes not in flow-arrangement.

               FORMULA.—_Hypersth.-aug., matr., non-flu._

 +------------+---------------+-------------+-----------+--------------+
 | SUB-ORDER. |   SECTION.    |    GENUS.   |  SPECIES. |              |
 +------------+---------------+-------------+-----------+              |
 |            |  Presence or  |             |           |   Number of  |
 |  Pyroxene  |  absence of   |  Character  |   Length  |   specimens. |
 |   of the   |  plagioclase  |   of the    | of felspar|              |
 | groundmass.|  phenocrysts. | phenocrysts.|  lathes.  |              |
 +------------+---------------+-------------+-----------+--------------+
 |            |               |{     1.     |}          |              |
 |            |               |{   Glassy   |}·02-·1 mm.|      12      |
 |            |               |{plagioclase |} ·1-·2  " |       4      |
 |            |               |{phenocrysts |}          |              |
 |            |{      1.      |{ (_vitr._)  |           |              |
 |            |{ Plagioclase  |{            |           |              |
 |            |{ phenocrysts  |{     2.     |}          |              |
 |            |{  (_phen._)   |{   Opaque   |}          |{ Not         |
 |            |{              |{phenocrysts |}          |{ represented.|
 |            |{              |{     of     |}          |              |
 |      1.    |{              |{plagioclase |}          |              |
 |   Granular |               |{  (_opac._) |           |              |
 | (_gran._)   |{              |            |           |              |
 |            |{              |{     3.     |}          |              |
 |            |{              |{   Large    |}          |              |
 |            |{              |{phenocrysts |}          |{ Not         |
 |            |{              |{     of     |}          |{ represented.|
 |            |{              |{ pyroxene,  |}          |              |
 |            |{      2.      |{ over 2 mm. |}          |              |
 |            |{No plagioclase|{ (_magn._)  |           |              |
 |            |{ phenocrysts  |{            |           |              |
 |            |{(_non phen._) |{     4.     |           |              |
 |            |               |{   Small    |           |              |
 |            |               |{  pyroxene  |  ·3-·5 mm.|       1      |
 |            |               |{phenocrysts,|           |              |
 |            |               |{under 2 mm.,|           |              |
 |            |               |{  or none   |           |              |
 |            |               |{ (_parv._)  |           |              |
 |            |               |             |{·02-·1 mm.|       2      |
 |            |               |{5. _Vitr._  |{ ·1-·2  " |       2      |
 |            |{ 3. _Phen._   |{            |           |              |
 |      2.    |{              |{            |{·02-·1 mm.|       1      |
 |  Prismatic |{              |{6. _Opac._  |{ ·1-·2  " |       1      |
 | (_prism._) |{              |             |           |              |
 |            |{              |{7. _Magn._  |}          |              |
 |            |{4. _Non phen._|{            |}          |{ Not         |
 |            |               |{8. _Parv._  |}          |{ represented.|
 |            |               |             |           |              |
 |            |               |{  9 }       |           |              |
 |      3.    |               |{ 10 }       |           |{ Not         |
 |   Ophitic  |               |{ 11 }       |           |{ represented.|
 |  (_oph._)  |               |{ 12 }       |           |              |
 +------------+---------------+-------------+-----------+--------------+


           SECOND ORDER OF THE HYPERSTHENE-AUGITE ANDESITES.

            CHARACTERS.—Felspar-lathes in flow-arrangement.

                 FORMULA.—_Hypersth.-aug., matr., flu._

 +------------+---------------+-------------+-----------+--------------+
 | SUB-ORDER. |   SECTION.    |   GENUS.    |  SPECIES. |              |
 +------------+---------------+-------------+-----------+              |
 |            |  Presence or  |             |           |   Number of  |
 |  Pyroxene  |  absence of   |  Character  |   Length  |   specimens. |
 |  of the    |  plagioclase  |   of the    | of felspar|              |
 | groundmass.| phenocrysts.  | phenocrysts.|  lathes.  |              |
 +------------+---------------+-------------+-----------+--------------+
 |            |               |{    13.     |}          |              |
 |            |               |{   Glassy   |}·02-·1 mm.|       9      |
 |            |               |{plagioclase |} ·1-·2  " |       2      |
 |            |{      7.      |{phenocrysts |}          |              |
 |            |{ Plagioclase  |{ (_vitr._)  |}          |              |
 |            |{ phenocrysts  |{            |           |              |
 |            |{  (_phen._)   |{    14.     |}          |              |
 |            |{              |{   Opaque   |}          |{ Not         |
 |            |{              |{phenocrysts |}          |{ represented.|
 |            |{              |{     of     |}          |              |
 |      4.    |{              |{ plagioclase|}          |              |
 |  Granular  |{              |{ (_opac._)  |           |              |
 | (_gran._)  |{              |             |           |              |
 |            |{              |{    15.     |}          |              |
 |            |{              |{   Large    |}          |              |
 |            |{              |{phenocrysts |}          |{ Not         |
 |            |{              |{of pyroxene,|}          |{ represented.|
 |            |{              |{ over 2 mm. |}          |              |
 |            |{      8.      |{ (_magn._)  |}          |              |
 |            |{No plagioclase|{            |           |              |
 |            |{ phenocrysts  |{    16.     |}          |              |
 |            |{(_non-phen._) |{   Small    |}          |              |
 |            |               |{ pyroxene   |}          |{ Not         |
 |            |               |{phenocrysts,|}          |{ represented.|
 |            |               |{under 2 mm.,|}          |              |
 |            |               |{  or none   |}          |              |
 |            |               |{  (_parv._) |           |              |
 |            |               |             |           |              |
 |            |               |{ 17. _Vitr._| ·02-·1 mm.|       2      |
 |            |{ 9. _Phen._   |{            |           |              |
 |      5.    |{              |{ 18. _Opac._| ·02-·1 mm.|       4      |
 |  Prismatic |{              |             |           |              |
 | (_prism._) |{              |{ 19. _Magn._|           |{ Not         |
 |            |{ 10.          |{            |           |{ represented.|
 |            |{ _Non-phen._  |{ 20. _Parv._|{·02-·1 mm.|       4      |
 |            |               |             |{ ·1-·2  " |       2      |
 |            |               |             |           |              |
 |      6.    |               |{ 21 }       |           |              |
 |   Ophitic  |               |{ 22 }       |           |{ Not         |
 |  (_oph._)  |               |{ 23 }       |           |{ represented.|
 |            |               |{ 24 }       |           |              |
 +------------+---------------+-------------+-----------+--------------+


            THIRD ORDER OF THE HYPERSTHENE-AUGITE ANDESITES.

         CHARACTERS.—Felspars of the groundmass short and stout
                             (orthophyric).

                FORMULA.—_Hypersth.-aug., matr., orth._

REMARKS.—The same classification is to be employed here as in the case
of the two previous orders, but as the rocks in my collection that
belong to this order are not numerous (nine sections), it will be
sufficient to refer to the general remarks on the order on p. 290.


           FOURTH ORDER OF THE HYPERSTHENE-AUGITE ANDESITES.

          CHARACTERS.—Groundmass presenting a rudely granular
             appearance or blurred mosaic (microfelsitic).

                FORMULA.—_Hypersth.-aug., matr., fels._

REMARKS.—My sliced specimens (five) are too few for the elaboration of
this order to which the classification employed for the other orders is
scarcely applicable. This is due to the partial decomposition or
imperfect development of the pyroxene of the groundmass. The general
characters of the order are given on p. 291.


                          THE PLUTONIC ROCKS.

These rocks are very infrequent and are for the most part
hypersthene-gabbros or norites, with a few representatives of
diorites[110] without pyroxene. True plutonic rocks did not come under
my observation in the western half of the island (west of Lambasa and
Savu-savu), those of Mount Thoka-singa in the Ndrandramea district
making the nearest approach (see p. 302). The localities in which they
were found are below enumerated:—

 +----+--------------------+-----------------+------------------+------+
 |    |                    |                 |      Mode of     |      |
 |No. |  Nature of rock.   |    Locality.    |    occurrence.   |Page. |
 +----+--------------------+-----------------+------------------+------+
 | 1  | Hypersthene-gabbro | Avuka Range     | Probably forms   |      |
 |    |                    | between Lambasa | the axis of the  |      |
 |    |                    | and Mbuthai-sau | range.           | 180  |
 |    |                    |                 |                  |      |
 | 2  |       Ditto        | Nawi, at the    | Deep-seated.     | 211  |
 |    |                    | head of the     |                  |      |
 |    |                    | Vui-na-savu     |                  |      |
 |    |                    | River           |                  |      |
 |    |                    |                 |                  |      |
 | 3  |       Ditto        | Valanga Range   | Probably forms   |      |
 |    |                    | between         | the axis of the  |      |
 |    |                    | Savu-savu and   | range.           | 182  |
 |    |                    | Natewa Bays     |                  |      |
 |    |                    |                 |                  |      |
 | 4  |       Ditto        | Ridge at the    |      Ditto.      | 184  |
 |    |                    | head of Na Kula |                  |      |
 |    |                    | valley between  |                  |      |
 |    |                    | Savu-savu and   |                  |      |
 |    |                    | Natewa Bays.    |                  |      |
 |    |                    |                 |                  |      |
 | 5  | Hornblende-gabbro. |      Ditto      |      Ditto.      | 184  |
 |    |                    |                 |                  |      |
 | 6  | Diorite            | Vunimbua River, | Loose blocks in  |      |
 |    |                    | on south side of| river-bed.       | 182  |
 |    |                    | the Mariko Range|                  |      |
 |    |                    |                 |                  |      |
 | 7  |       Ditto        | Coast cliffs    | Large block in   |      |
 |    |                    | west of the Salt| agglomerate-tuff.| 193  |
 |    |                    | Lake Passage    |                  |      |
 +----+--------------------+-----------------+------------------+------+

THE HYPERSTHENE-GABBROS.—These rocks also contain monoclinic pyroxene,
and are the plutonic equivalents of the hypersthene-augite-andesites
which as a rule prevail in the localities where these rocks occur. They
are usually dark grey or steel-grey in colour with a specific gravity
ranging from 2·7 to 2·84 and have a granitoid aspect. The following
characters are common to all the specimens.

They display a mixture of plagioclase and pyroxene, the last filling up
the spaces between the felspars and apparently of later formation. The
plagioclase crystals, which are 1 to 2 mm. in size, are opaque; and
since they are traversed by numerous fine fissures filled with dust-like
decomposition products, their appearance is often semi-saussuritic. They
are much cross-macled, are at times zoned, and give lamellar extinctions
of andesine-labradorite (20°-30°).... The pyroxene includes both the
rhombic and monoclinic forms, the last with extinction angles of over
30°. They may be associated or may occur as separate crystals, the
rhombic prevailing in the less basic and the monoclinic in the more
basic rocks. The rhombic pyroxene is usually more or less converted into
bastite which by further change passes into a chloritic material; whilst
the augite sometimes undergoes the diallagic change resulting from
schillerisation.

Some special features are presented by rocks from different localities.
That from Nawi is most basic and looks like a diallage-gabbro. That from
the Valanga Range (sp. gr. 2·75) contains some quartz, apparently
secondary and filling up the interspaces. The rock from the Na Kula
Ridge shows traces of a groundmass; but it comes near the plutonic type.

A HORNBLENDE-GABBRO.—This granitoid rock, which is from the Na Kula
Ridge, has a specific gravity of 2·72. Hand-specimens display large
porphyritic crystals of hornblende (7 mm. long) in a base of opaque
felspar and smaller hornblende. In the slide we observe besides the
large crystals of plagioclase and hornblende a little pyroxene; but the
mass of the rock consists of greenish-brown hornblende, plagioclase, and
some secondary quartz, forming a coarse mosaic with a “grain” of about a
millimetre. The hornblende is displayed in regular hexagonal sections,
markedly pleochroic, and gives extinctions up to 12°. It shows no dark
resorption borders; and the larger porphyritic crystals have the same
characters. Almost all the plagioclase of the rock is traversed by
numerous fine fissures, and often acquires a semi-saussuritic appearance
from the presence of dust-like decomposition products. The lamellar
extinctions indicate andesine-labradorite. The quartz occurs mostly in
nests. The pyroxene is formed of large grains of both the monoclinic and
rhombic types.

THE DIORITES.—The rock forming blocks in the Vunimbua River has a
specific gravity of 2·78 to 2·8. It is a pretty rock showing long black
blades of hornblende, 10 mm. in length, in an opaque felspar base. In
the slide the hornblende, which is dark brown and markedly pleochroic,
shows six-sided sections with characteristic prismatic cleavage lines,
the longitudinal sections giving extinctions up to 15°. The borders in
some cases display traces of resorption. The felspar (plagioclase) is in
the form usually of broad regular crystals, 3 to 4 mm. in size, and
giving extinctions of andesine-labradorite (28°); they are “clouded”
through the presence of fine alteration products associated with
numerous fissures. The relation between the hornblende and the
plagioclase is not constant. This appears to be partially due to the
occurrence of traces of a groundmass.

The diorite forming blocks in the agglomerate of the coast cliffs, west
of the Salt Lake Passage, is a remarkable rock showing large blackish
hornblende crystals, in the shape of blades 25 mm. long and 3 or 4 mm.
broad, set in a base of opaque plagioclase felspar which surrounds the
hornblende. The last-named is deep-brown, very pleochroic, yields
extinctions up to 22°, and displays but little evidence of resorption.
The plagioclase is irregular in shape and exhibits broad lamellæ giving
extinctions of acid labradorite (28°-30°). It is traversed by numerous
fine fissures filled with decomposition products and contains abundant
dust-like materials. (Spec. grav. 2·8).



                             CHAPTER XVIII

             THE VOLCANIC ROCKS OF VANUA LEVU (_continued_)


                             OLIVINE CLASS

                              SUB-CLASS II

               THE OLIVINE BASALTS (_Plag, oliv, matr._)

THIS sub-class includes the plagioclase-olivine-basalts. Although these
rocks are not the most numerous of the basic rocks, they are well
represented in the island, being in great part confined to the western
half, and being especially characteristic of the districts of Wainunu
and Solevu and of the mountains of Seatura and Naivaka. It will be seen
from the Synopsis that this sub-class is split up into two divisions,
according to the relative abundance of the olivine. Many of the rocks
are grey basalts with the olivine more or less hematised; but the
majority are blackish with the olivine usually more or less
serpentinised. In the typical blackish rocks there is a little dark
opaque interstitial glass. In the grey basalts the groundmass is as a
rule holo-crystalline. The specific gravity ranges generally from 2·8 to
3.

It will be noticed in the scheme that the “prismatic” sub-orders, where
the pyroxene of the groundmass is for the most part in prisms, are
scarcely represented. The “ophitic” sub-orders are poorly represented,
since they only include about 10 per cent. of the total. The ophitic
olivine-basalts are indeed mostly confined to the division where the
olivine is scanty, and the felspars of the groundmass are for the most
part not parallel, the plexus arrangement, as will be often pointed out,
being almost essential for the development of the ophitic structure.
With the basaltic andesites, which cannot always be sharply separated
from the basalts with scanty olivine, the proportion of ophitic rocks is
much higher, probably about 20 per cent. The pyroxene in the
olivine-basalts is nearly always augite, intergrowths with rhombic
pyroxene being only occasionally observed.


                   I. DIVISION OF THE OLIVINE-BASALTS

CHARACTERS.—Abundant olivine.

FORMULA.—_Oliv, matr, cop._

These rocks are characterised by abundant olivine usually as phenocrysts
but sometimes represented in the groundmass. When a basalt presents much
of this mineral in an ordinary hand-specimen and displays at least five
or six phenocrysts in a slide, it is placed in this division.
Olivine-basalts of this character are well exhibited in the hills around
Solevu Bay and in the neighbouring Seatovo Range. They are also fairly
represented on the northern slopes of Mount Seatura, on the coast
between the Wainunu River and Nandi Bay, and on the Wainunu basaltic
tableland. In the Ulu-i-ndali Range, which lies east of the Wainunu
estuary, they are especially frequent. Whilst confined mostly to the
portion of the island west of Savu-savu Bay, they occur sporadically in
other localities to the eastward, as in Na Suva-suva Hill and in some
parts of the Natewa Peninsula. The grey basalts, which form one-third of
the total, are chiefly characteristic of the hill of Ulu-i-ndali, of the
Solevu district, and of the northern slopes of Seatura. Whilst the
blackish basalts usually compose the flows, the grey basalts form dykes
and volcanic necks.

Two-thirds of these olivine-basalts belong to the order presenting
flow-structure and almost all (28 out of 29) are included in the
sub-order exhibiting granular augite in the groundmass. The ophitic
structure is displayed in only one case; and the prismatic form of the
augites is never a predominant feature.


                    1. GENUS OF THE OLIVINE-BASALTS

FORMULA.—_Oliv, matr, cop, non-flu, gran, phen, vitr._

CHARACTERS.—Abundant olivine. Felspar-lathes of the groundmass not in
flow arrangement. Pyroxene of the groundmass granular. Phenocrysts of
glassy plagioclase.

DESCRIPTION.—Dark-brown or blackish rocks. Sp. gr. 2·88 to 2·93.
Phenocrysts of pyroxene occur in fair quantity in addition to those of
the olivine and plagioclase. The groundmass displays a plexus of
felspars and augite-granules with much magnetite in grains and
irregular patches. The interstitial glass is scanty or almost absent.
The olivine phenocrysts, of which the larger are 3 to 4 mm. in size,
are as a rule hematised at the borders and in the fissures, and are
sometimes partially serpentinised. In some cases small crystals of
olivine are enclosed in the pyroxene-phenocrysts. The plagioclase
phenocrysts do not usually exceed 2 mm. in size. They give lamellar
extinctions of 15°-28°, and are often cross-macled. They generally
contain magma-inclusions, which may be arranged in zones, and they
sometimes inclose small pyroxene crystals. Their borders are often
eroded. The pyroxene-phenocrysts, which frequently are 3 to 4 mm. in
size, give extinctions of 30° and over, and may be described as
composed of brown augite. It is only at times that intergrowths of
rhombic pyroxene occur. They are often twinned and are sometimes
eroded and may contain magma and other inclusions. The felspars of the
groundmass, which for the most part form a plexus, are small and
stout, their average length varying from ·08 to ·13 mm., whilst they
frequently display lamellar twinning and give extinctions of about 15°
(oligoclase-andesine). The pyroxene-granules of the groundmass, which
are of brown augite, vary in average size from ·02 to ·04 mm.

Two of the four species, where the felspar-lathes are less than ·1 mm.
and between ·1 and ·2 mm. in length, are represented in this collection.


                    2. GENUS OF THE OLIVINE-BASALTS

FORMULA.—_Oliv, matr, cop, non-flu, gran, phen, opac._

CHARACTERS.—Abundant olivine. Felspar-lathes of the groundmass not in
flow-arrangement. Pyroxene of the groundmass granular. Phenocrysts of
plagioclase opaque white.

DESCRIPTION.—Grey compact-looking rocks; sp. gr. 2·83 to 2·9.
Interstitial glass scanty. The olivine phenocrysts, which range up to
5 mm. in size, are more or less hematised; and in extreme cases of this
alteration, where schiller-planes are formed, the hand-specimen appears
to carry brown mica. There are sometimes small grains of olivine
(·1 mm.) in the groundmass. The plagioclase-phenocrysts, varying in size
from 2 to 4 mm., owe their opacity partly to their composite character,
when they present an aggregate of smaller clear crystals, and partly to
multiple macling. They give extinctions of acid labradorite (25°-32°).
Pyroxene-phenocrysts, when present, are scanty, pale brown, not over
2 mm. in size, and give the large extinctions (+30°) of augite. In their
absence small augites (·2 mm.) are frequent. Fine granules (·01-·03 mm.)
of similar augite, together with magnetite, abound in the groundmass.
The felspars of the groundmass are fairly stout and lathe-like and show
at times a few twin-lamellæ which give extinctions of 13° to 20° (medium
andesine).

Species represented:

 (_a_) felspar lathes ·1-·2 mm.
 (_b_) felspar lathes  ·2-·3 mm.


                    3. GENUS OF THE OLIVINE-BASALTS

FORMULA.—_Oliv, matr, cop, non-flu, gran, non-phen, magn._

CHARACTERS.—Abundant olivine. Felspar-lathes of the groundmass not in
flow-arrangement. Pyroxene of the groundmass granular. No plagioclase
phenocrysts. Large phenocrysts of olivine and pyroxene over 2 mm. in
size.

DESCRIPTION.—A remarkable blackish-grey rock. Although somewhat
scoriaceous, it has a sp. gr. of 2·91. It displays large phenocrysts of
olivine and pyroxene, 4 to 8 mm. in size, in a coarse-textured
groundmass of stout felspars, augite granules, magnetite and a little
glass. The olivine is extensively hematised. The pyroxene phenocrysts
are of brownish-yellow augite with regular outlines and giving angles of
extinction up to 40°. The broad lamellar felspars of the groundmass,
which are on the average ·3 to ·4 mm. long, give extinctions indicating
both acid labradorite (23° to 28°) and medium andesine (16° or 17°). The
abundant augite-grains average ·05 mm. in size; but the prism form
occurs at times.

The species with felspar-lathes ·3 to ·5 mm. long is alone represented.


                    4. GENUS OF THE OLIVINE-BASALTS

FORMULA.—_Oliv, matr, cop, non-flu, gran, non-phen, parv._

CHARACTERS.—Abundant olivine. In the groundmass the felspar-lathes are
not in flow-arrangement and the pyroxene is granular. No plagioclase
phenocrysts. Small phenocrysts of olivine and occasionally pyroxene
under 2 mm.

DESCRIPTION.—A grey coarse-grained rock. Sp. gr. 2·9. It displays
abundant small phenocrysts of olivine, all less than a millimetre in
size (·2-·8 mm.), which are hematised in the fissures and at the
borders. The felspar-lathes, which display a few twin-lamellæ giving
extinctions of 16°-18° (medium andesine), vary greatly in size. The
smaller are ·1 to ·3 mm. and the larger ·3 to ·5 mm. long; but the two
are connected by felspars of intermediate length. The abundant augite
granules average ·037 mm. in breadth. Pyroxene phenocrysts are not
represented in the slide. From its coarsely crystalline texture this
rock merits the field-name of a grey doleritic basalt; and except in the
arrangement of the felspar-lathes it does not differ from the grey
doleritic basalts of genus 16.


                    12. GENUS OF THE OLIVINE-BASALTS

FORMULA.—_Oliv, matr, cop, non-flu, oph, non-phen, parv._

CHARACTERS.—Olivine abundant. Felspars of the groundmass not in
flow-arrangement. Pyroxene of the groundmass ophitic or semi-ophitic. No
plagioclase phenocrysts. Small phenocrysts (under 2 mm.) of olivine and
occasionally a few of pyroxene.

DESCRIPTION.—A dark greenish-brown rock, with sp. gr. 2·91, showing
abundant micro-porphyritic olivine in a groundmass consisting of ophitic
pale-brown augite inclosing the felspar-lathes, together with small
olivines, patches of magnetite, and a little altered interstitial glass.
The olivine-phenocrysts are about ·5 mm. in average size and are more or
less hematised. The felspar-lathes, which average ·15 mm. in length,
often show twin-lamellæ that give extinctions of 23°-30°
(andesine-labradorite).

The species with felspar-lathes ·1-·2 mm. long is alone represented.


                    13. GENUS OF THE OLIVINE-BASALTS

FORMULA.—_Oliv, matr, cop, flu, gran, phen, vitr._

CHARACTERS.—Olivine abundant. Felspars of the groundmass in
flow-arrangement. Pyroxene of the groundmass granular. Phenocrysts of
glassy plagioclase.

DESCRIPTION.—Dark grey or dark brown rocks with sp. gr. 2·88 to 2·99.
Phenocrysts of olivine, pyroxene, and plagioclase occur in a groundmass
showing partially parallel felspar-lathes, abundant pyroxene grains, and
fine magnetite, residual glass being scanty or absent. The
olivine-phenocrysts do not exceed 3 or 4 mm. in size and in some rocks
are less than 1 mm. They are usually more or less serpentinised and
hematised at the borders and in the cracks; but sometimes they are
almost fresh and present regular outlines. In some rocks the olivine
also occurs as grains (·3 mm.) in the groundmass. When the phenocrysts
of olivine have blackish borders they are surrounded by a halo, as
though the crystal had attracted the magnetite from the groundmass
immediately around. The plagioclase phenocrysts vary from 1 to 3 mm. in
size. They often contain abundant magma-inclusions and give lamellar
extinctions of 15° to 25° (basic andesine). In some rocks they are
rudely parallel. The pyroxene phenocrysts, which are of pale
brownish-yellow augite giving extinctions of over 30°, do not usually
exceed 3 mm. They present regular octagonal cross-sections and sometimes
display lamellar twinning. Occasionally there is a suspicion of
intergrowth with rhombic pyroxene. The felspar-lathes, which according
to the species vary much in length, at times show a few lamellæ. The
augite grains of the groundmass are abundant and are as a rule about
·02 mm. in size; but in some rocks they are larger and in others
smaller.

This genus may be divided into two sub-genera, the porphyritic sub-genus
where the felspar phenocrysts are larger than 3 mm., and the
non-porphyritic where they are smaller. All four species, as indicated
by the length of the felspar-lathes, are represented.


                    14. GENUS OF THE OLIVINE-BASALTS

FORMULA.—_Oliv, matr, cop, flu, gran, phen, opac._

CHARACTERS.—Olivine abundant. Felspars of the groundmass in
flow-arrangement. Pyroxene of the groundmass granular. Opaque
plagioclase phenocrysts.

DESCRIPTION.—Dark grey rocks, with sp. gr. 2·9 to 2·93, showing
phenocrysts of olivine and pyroxene with opaque whitish phenocrysts of
plagioclase in a groundmass of felspar lathes, pyroxene grains, and
magnetite, with occasional fine olivine. The olivine phenocrysts, which
are sometimes 5 or 6 mm. in size, are often deeply eroded. They are at
times so extensively hematised along the schiller-planes that they
appear like brown mica. The plagioclase phenocrysts owe their opacity in
part to their consisting of an aggregate of lesser crystals which are
clear and glassy and give lamellar extinctions of 20° to 30° (andesine
labradorite). They do not usually exceed 3 mm. and are sometimes scanty.
The pyroxene phenocrysts, which are at times infrequent, may be 5 mm. in
size. They are of pale yellowish-brown augite, giving extinctions of
40°. The felspar-lathes are rarely lamellar; but in one such case the
angle of extinction was 17° (medium andesine). The grains of augite in
the groundmass average ·02 mm. in diameter.

The only species represented is that with the felspar-lathes ·1 to
·2 mm. long.


                    15. GENUS OF THE OLIVINE-BASALTS

FORMULA.—_Oliv, matr, cop, flu, gran, non-phen, magn._

CHARACTERS.—Abundant olivine. In the groundmass the felspars are in
flow-arrangement and the pyroxene is granular. No plagioclase
phenocrysts, but large phenocrysts, over 2 mm., of olivine and pyroxene.

DESCRIPTION.—This genus includes the most basic rocks represented in my
collection. They are compact heavy blackish rocks with sp. gr. 3 to 3·1,
and display large porphyritic crystals of olivine and pyroxene often 3
or 4 mm. in size. The olivine phenocrysts may be fairly fresh with clean
outlines, or they may be deeply eroded and stained by iron oxide, or
they may be passing into serpentine. The pyroxene phenocrysts may be
either eroded or possess regular outlines. They are of pale brown augite
and give extinctions over 30°. The pyroxene granules, which average
·02 mm. and are very abundant, are of the same augite. The
felspar-lathes are relatively scanty. In the two rocks here included
they average in length ·06 mm. and ·08 mm. When lamellæ can be
recognised they give extinctions of 30°-40° (labradorite). The untwinned
lathes give extinctions of 20°-28° (labradorite).

The only species represented is that with felspar-lathes less than
·1 mm. in length.


                    16. GENUS OF THE OLIVINE-BASALTS

FORMULA.—_Oliv, matr, cop, flu, gran, non-phen, parv._

CHARACTERS.—Abundant olivine. In the groundmass the felspar-lathes are
in flow arrangement and the pyroxene is granular. There are no
plagioclase phenocrysts; but there are numerous small phenocrysts, under
2 mm. in size, of olivine and occasionally a few of pyroxene.

DESCRIPTION.—As a highly basic genus this ranks next to the preceding
one, the specific gravity of the rocks ranging from 2·91 to 3·01. All
the four species indicated by the varying length of the felspar-lathes
are represented in my collection. The rocks of the first two, with the
felspars averaging less than ·1 mm. and between ·1 and ·2 mm., are
compact aphanitic basalts, only displaying an occasional small
phenocryst of augite and blackish-grey or bluish-black in colour. Those
of the last two species, with the average length of the felspar-lathes
·2-·3 mm. and ·3-·5 mm. respectively, are lightish-grey coarse-textured
rocks of the doleritic type. In all the rocks no pyroxene phenocrysts
are displayed in the slide; and the olivine phenocrysts, which are very
numerous, do not usually exceed 1 mm., though occasionally the average
size is 1·3 mm., and not infrequently it is only ·5 mm. In some cases
where the larger olivine phenocrysts lie athwart the current of the
felspar-lathes, the smaller (·5 mm.) lie with their long axes parallel
to the flow. The olivine is either fresh, or it may be beginning to
serpentinise in the cracks, or it may be in part hematised. The crystals
may have regular outlines, or they may be rounded and sometimes deeply
eroded. The pyroxene granules of the groundmass are of pale-brown
augite, and average ·01 to ·03 mm. in size. Occasionally a few prism
forms occur, giving extinctions of 30° to 40°. In the case of the more
compact rocks, with the felspar-lathes averaging less than ·2 mm. in
length, lamellar twinning is but scantily to be noticed in the lathes,
which give extinctions measured from the long axis of 20° to 25° and by
the twin lamellæ of 30° to 35°, indicative of acid labradorite in both
instances. With the coarser doleritic grey basalts, where the
felspar-lathes are stouter and have an average length exceeding ·2 mm.,
lamellar twinning is more frequent; the extinctions afforded by the
lamellæ range between 15° and 25° (medium andesine). Residual glass is
scanty in these rocks, and in the grey dolerites it is often difficult
to recognise any.


                   2. DIVISION OF THE OLIVINE-BASALTS

CHARACTERS.—Scanty olivine.

FORMULA.—_Oliv, matr, pauc._


                    25. GENUS OF THE OLIVINE-BASALTS

FORMULA.—_Oliv, matr, pauc, non-flu, gran, phen, vitr._

CHARACTERS.—Olivine scanty. Felspars (lathes and prisms) of the
groundmass not in flow-arrangement. Pyroxene of the groundmass granular.
Glassy plagioclase phenocrysts.

DESCRIPTION.—About two-thirds of these rocks have a common facies, being
closely similar in appearance, brownish-black in colour, and with spec.
grav. usually between 2·84 and 2·92. They belong to the species with the
felspar-lathes ·1 to ·2 mm. in length. They are essentially the rocks of
the old submarine basaltic flows; and they are often columnar, the
columns being 2 to 4 feet across. My remarks will mainly apply to this
predominant group.

To the eye they are somewhat compact and show scattered porphyritic
crystals of plagioclase. In the slide they display numerous phenocrysts
of plagioclase, with a few of olivine and pyroxene, in a groundmass
formed of stout lamellar felspar-lathes and small prisms forming a
plexus with granular augite in the meshes. There is a good deal of
magnetite and generally scanty residual glass. The plagioclase
phenocrysts are usually 2 to 3 mm. in size, but they may be smaller (1
to 2 mm.) or larger (3 to 5 mm. or more) when the rock has a porphyritic
appearance.[111] They are often cross-macled and at times show zoning.
In many slides two kinds are distinguished by the extinctions which
indicate in one case medium andesine (15° to 22°) and in the other acid
labradorite (27° to 32°). They contain inclusions of the magma and are
often eroded. The pyroxene phenocrysts are of pale brown augite, scanty
and small, and give extinctions of 30° to 40°. They are sometimes
twinned and may be eroded and contain inclusions of the magma. The
olivine phenocrysts, which do not usually exceed 2 or 3 mm., are mostly
rounded, but sometimes have the regular outlines, and are in various
stages of serpentinisation. The felspars of the groundmass, which
average ·17 mm. in length, are mostly stout and lamellar; but they
exhibit all transitions from the lathe-shape with one or two lamellæ to
broad multi-lamellar prisms where the breadth is half the length. They
give lamellar extinctions averaging 15° to 18° (andesine). The augite
granules of the groundmass are pale brown and average ·02 to ·03 mm. in
size. In a few cases they are larger (·05 mm.) which is an indication of
an approach to the ophitic type. In most slides occur a few small
augites of prism-form, two or three times the size of the granules,
which give extinctions of over 30°. Where the phenocrysts of augite are
very scanty or absent, there exist large grains (·1 mm.) of an
intermediate size. The magnetite is often abundant, occurring in
crystals, rods, and irregular masses, the last associated often with the
interstitial glass which is present in small quantities in most rocks,
being greenish or brownish and showing fibrous devitrification.

In some localities semi-vitreous rocks referable to this genus are
frequent. This is especially the case in the Naivaka peninsula, where
the rocks show a fair amount of glass in the groundmass, the porphyritic
augite being well developed, whilst the pyroxene of the groundmass is
only in part differentiated. Three of the four species are here
represented. Those with large felspar-lathes (·2-·3 mm. long) and coarse
augite granules (·05) approach the semi-ophitic rocks included in genus
33.


                    26. GENUS OF THE OLIVINE-BASALTS

FORMULA.—_Oliv, matr, pauc, non-flu, gran, phen, opac._

CHARACTERS.—Olivine scanty. In the groundmass the felspar-lathes and
prisms are not in flow-arrangement and the pyroxene is granular. The
plagioclase phenocrysts are opaque.

DESCRIPTION.—Dark grey porphyritic rocks, which, from the opacity of the
felspar phenocrysts, look like porphyrites. They are not very frequent
and occur mostly on the northern slopes of Mount Seatura. Two different
types occur in my collection which may be regarded as sub-genera. In the
most basic kind, where the sp. gr. is 2·86 to 2·89, the plagioclase
phenocrysts, 2 to 3 mm. in size, owe their opacity chiefly to their
aggregate structure. They give lamellar extinctions (15°-30°) of
andesine labradorite. Porphyritic olivine is scanty and more or less
hematised; but a fair amount of olivine grains, less than ·1 mm., occur
in the groundmass. Pyroxene phenocrysts are scanty, but microporphyritic
pale brown augite (·1 mm.) is frequent. In the groundmass are found
stoutish felspar-lathes, averaging ·2 mm. long, together with an
abundance of fine augite granules (·01-·02 mm.) and fine magnetite, the
residual glass being scanty.... In the other type the sp. gr. is 2·75;
and the plagioclase phenocrysts 4 to 6 mm. in size give extinctions of
andesine and acid labradorite (10°-30°). There is an approach to the
orthophyric structure in the groundmass, as is indicated by the number
of short broad felspars, averaging ·2 mm. in length and giving lamellar
extinctions of acid and basic andesine. The granular augite of the
groundmass is coarse (·04 mm.), and occasional prism-forms give
extinctions of 40°.


                    33. GENUS OF THE OLIVINE-BASALTS

FORMULA.—_Oliv, matr, pauc, non-flu, oph, phen, vitr._

CHARACTERS.—Olivine scanty. In the groundmass the felspar-lathes are not
in flow-arrangement and the pyroxene is ophitic or semi-ophitic.
Plagioclase phenocrysts glassy.

DESCRIPTION.—These brownish-black rocks are all of the semi-ophitic
type. Although no ophitic “plates” occur in the slide, the augites of
the groundmass have no longer the granular form, but are large,
·08-·1 mm. in size, and tend to invest the felspar-lathes. The specific
gravity ranges from 2·78 to 2·86. As in other of the ophitic and
semi-ophitic rocks of this collection (genera 9, 10, 12, of the
augite-andesites), the large size of the felspar-lathes (·2-·3 mm. long)
of the groundmass gives a doleritic texture in the slide. In most of the
other characters these rocks approach those of genus 25 which possess
felspar-lathes more than ·2 mm. in length. But they are more often
semi-vitreous, and display a considerable amount of dark smoky glass
showing numerous magnetite rods and skeletal crystals with fibrous
devitrification. The plagioclase phenocrysts, which vary much in size in
different rocks (in some 2 or 3 mm., in others 4 or 5 mm.), give
extinctions of andesine labradorite (20°-35°). They are often eroded and
contain numerous large black inclusions of the magma. The pyroxene
phenocrysts, which are of pale-brown augite, often have an
aggregate-structure, having been formed _in situ_. Others again consist
of single crystals and have been much affected by the magma. The olivine
phenocrysts, which are at times deeply eroded, are generally small and
in part serpentinised.


                    37. GENUS OF THE OLIVINE-BASALTS

FORMULA.—_Oliv, matr, pauc, flu, gran, phen, vitr._

CHARACTERS.—Olivine scanty. In the groundmass the felspar-lathes are in
flow-arrangement and the pyroxene is granular. Glassy plagioclase
phenocrysts.

DESCRIPTION.—Brownish-black rocks which cannot be distinguished, except
in the flow-arrangement of the felspars of the groundmass, from those
described under genus 25. Like them they enter into the formation of the
basaltic plains of Sarawanga and Mbua and elsewhere. Most of the rocks
of this genus group themselves into one type where the felspar-lathes
average in length ·15-·21 mm. The sp. gr. is usually between 2·87 and
2·91. Though rarely porphyritic, such rocks display to the eye a few
small scattered glassy phenocrysts of plagioclase and an occasional
grain of olivine. It is to this type of the genus that the following
description applies.

In the thin sections they display small plagioclase phenocrysts, with a
few of olivine and occasionally of pyroxene, in a groundmass where the
flow-arrangement of the felspar-lathes is well marked, the rest of the
groundmass being made up of granular augite with magnetite and generally
a little residual glass.... The plagioclase phenocrysts do not usually
exceed 2 mm. in size and contain magma inclusions. Two kinds are often
indicated in the same slide by the extinctions, namely, one of medium
andesine (17°-22°), and the other of acid labradorite (28°-33°).... The
pyroxene phenocrysts are of pale brown augite; but they are small (less
than 2 mm.), scanty, and often absent when their place is taken by
microporphyritic augite, ·2 mm. in size.... The olivine phenocrysts are
generally small. Though sometimes showing the long hexagonal sections,
they are often rounded and more or less serpentinised.... The felspars
of the groundmass present more typical lathes than are to be observed in
the non-parallel felspars of the rocks of genus 25. The twin-lamellæ,
when present, are fewer; but give similar extinctions (15°-21°) of
medium andesine.... The augite granules are, as a rule, ·02 or ·03 mm.
in diameter; but occasional more prismatic forms occur, two or three
times the length, which give extinctions of over 30°. The magnetite is
abundant, and the scanty interstitial glass is green or brown and
displays fibrous devitrification.

The following three species, as indicated by the length of the
felspar-lathes, are represented in my collection:

 (_a_) ·02-·1 mm.
 (_b_) ·1-·2 mm.
 (_c_) ·2-·3 mm.


                    38. GENUS OF THE OLIVINE-BASALTS

FORMULA.—_Oliv, matr, pauc, flu, gran, phen, opac._

CHARACTERS.—Olivine scanty. In the groundmass the felspar-lathes are in
flow-arrangement and the pyroxene is granular. Opaque plagioclase
phenocrysts.

DESCRIPTION.—Grey rocks, sp. gr. 2·78 to 2·83, showing small opaque
porphyritic crystals of plagioclase with a few phenocrysts of olivine
and pyroxene in a groundmass of parallel felspar-lathes, augite
granules, and magnetite, with very scanty, if any, interstitial
glass.... The plagioclase phenocrysts, 2 to 3 mm. in size, are often
aggregates of smaller crystals. They contain colourless granular
inclusions and are sometimes zoned, giving extinctions of medium
andesite (15°-18°), and of andesine labradorite (25°-29°).... The
pyroxene phenocrysts are pale-brown, scanty, 2 to 3 mm. in size, often
twinned and give the large extinctions of augite.... The olivine
phenocrysts, which do not exceed 2 or 3 mm., are deeply eroded by the
magma and are hematised and schillerised. Small grains also occur in the
groundmass.... The felspar-lathes, which in the species here represented
average ·15 mm. in length, are stout and give lamellar extinctions of
andesine (18°-22°).... The augite granules are pale-brown and usually
·02-·03 mm. in diameter.

The only species represented in my collection is that with the
felspar-lathes ·1-·2 mm. long.


                    40. GENUS OF THE OLIVINE-BASALTS

FORMULA.—_Oliv, matr, pauc, flu, gran, non-phen, parv._

CHARACTERS.—Olivine scanty. In the groundmass the felspar-lathes are in
flow-arrangement and the pyroxene is granular. No plagioclase
phenocrysts; but there are a few small phenocrysts of olivine and
sometimes of pyroxene under 2 mm. in size.

DESCRIPTION.—Compact-looking non-porphyritic blackish-brown rocks, sp.
gr. about 2·9. Occasionally a little vesicular. For the most part
dyke-rocks.

In the slide are displayed a few small phenocrysts of olivine and
pyroxene in a groundmass formed of more or less parallel felspar-lathes,
augite granules, magnetite, sometimes in rods, and a little greenish
devitrified residual glass.... The pyroxene phenocrysts are of
pale-brown augite and are generally less than a millimetre in size. They
may be single crystals or they may be formed of an aggregate of a few
smaller crystals.... The olivine phenocrysts rarely exceed 2 mm. in size
and are in part serpentinised.... The augite granules vary usually from
·01 to ·03 mm. in diameter.... The felspar-lathes of the rocks in this
collection are large, often exceeding ·2 mm. in length, giving the rock
a doleritic texture in the slide. In a single slide they may range from
·1 to ·6 mm. When lamellar they give extinctions of 15° to 25° (basic
andesine).

Two species are represented in this collection:--

 (_a_) with felspar-lathes ·2-·3 mm.
 (_b_) with felspar-lathes ·3-·5 mm.


                    44. GENUS OF THE OLIVINE-BASALTS

FORMULA.—_Oliv, matr, pauc, flu, prism, non-phen, parv._

CHARACTERS.—Olivine scanty. In the groundmass the felspar-lathes are in
flow-arrangement and the pyroxene is in great part prismatic. There are
no phenocrysts of plagioclase; but there are a few small phenocrysts of
olivine and sometimes of pyroxene, less than 2 mm. in size.

DESCRIPTION.—A dark grey compactish rock; sp. gr. 2·9; showing a little
macroscopic olivine; forming a dyke in the tuffs on the summit of the
hill of Vatui (p. 54).

In the section it exhibits a few phenocrysts of olivine (more or less
serpentinised) and of augite in a groundmass formed of stout augite
prisms and small augite granules with felspar-lathes, magnetite, and a
little devitrified yellowish interstitial glass. The augite prisms and
the felspar-lathes are in flow-arrangement.... The pyroxene phenocrysts,
which are pale brown and give extinctions of over 30° from the single
cleavage-lines, may be aggregates of five or six smaller crystals or
single crystals presenting sometimes lamellar twinning. The first are
doubtless formed _in situ_. The second though showing regular outlines
may have a nucleus giving a different extinction and possessing eroded
margins. The stout augite prisms of the groundmass, which are
occasionally twinned, have an average length of ·2 to ·3 mm. and give
angles of extinction with the long axis of 30° to 40°. The
felspar-lathes average only ·07 mm. in length and afford extinctions,
when untwinned, of 18° to 24° (acid labradorite).

The only species represented is that where the average length of the
felspar-lathes is between ·02 and ·1 mm.



                              CHAPTER XIX

             THE VOLCANIC ROCKS OF VANUA LEVU (_continued_)


                              AUGITE CLASS

                              SUB-CLASS II

           AUGITE-ANDESITES INCLUDING THE BASALTIC ANDESITES
                          (_Plag, aug, matr._)

THIS sub-class, which comprises 40 per cent. of the volcanic rocks, is
characterised by the absence of olivine on the one hand, and by the
rarity or absence of rhombic pyroxene on the other. On the basic side it
shades into the olivine-basalts through the basaltic andesites, and on
the acid side by intermediate stages into the hypersthene augite
andesites; and for these reasons it is not always possible to draw a
sharp line of distinction. In cases where a hand-specimen displays no
macroscopic olivine and where a solitary small phenocryst of olivine is
alone observed in the slide, it should be referred to this sub-class;
and here also all doubtful specimens as regards the occurrence of
olivine should be placed. When the question of the occurrence of rhombic
pyroxene arises, it should be remembered that the great prevalence of
monoclinic pyroxene amongst the phenocrysts and the practical absence of
rhombic pyroxene from the groundmass are essential characteristics of
this sub-class. Rhombic pyroxene is only indicated at times by
intergrowths in the phenocrysts.

The basaltic andesites enter into the formation of old “flows,” as in
the Mbua and Ndama plains. The other rocks enter principally into the
composition of dykes, necks, and agglomerates.


                I. GRANULAR SUB-ORDER (AUGITE-ANDESITES)

FORMULA.—_Aug, matr, non-flu, gran._


                    1. GENUS OF THE AUGITE-ANDESITES

FORMULA.—_Aug, matr, non-flu, gran, phen, vitr._

CHARACTERS.—In the groundmass the felspar-lathes and prisms are not in
flow-arrangement and the augite is granular. Phenocrysts of glassy
plagioclase.

DESCRIPTION.—These rocks frequently form dykes; and it is probable that
most of the instances where the nature of the exposure could not be
ascertained also fall into this category. They are dark-brown or
blackish, and their sp. gr. ranges, except in the semi-vitreous rocks,
from 2·7 to 2·83. They are sometimes vesicular, and rocks with abundant
interstitial glass are common. They admit of grouping into two
sub-genera according to the size of the plagioclase phenocrysts:

 (_a_) Porphyritic, where the average size is 3 mm. or over.
 (_b_) Non-porphyritic where the average size is less than 3 mm.

Nearly all the rocks in my collection belong to the second group.

In the sections they display phenocrysts of plagioclase and occasionally
of pyroxene in a groundmass formed of a plexus of felspar-lathes, augite
granules, magnetite, and usually a fair amount of smoky more or less
opaque interstitial glass.... The felspar phenocrysts, which are
sometimes abundant, give lamellar extinctions of andesine labradorite
(15° to 30°). They are frequently small (1 to 2 mm.) and contain often
many magma-inclusions. Whilst the corroded aspect of some indicate that
they belong to an earlier period, the aggregate character and regular
outlines of others suggest that they have been produced in position....
Pyroxene phenocrysts are absent in half the rocks. When present they are
generally small and of a pale augite which gives extinctions of 30°.
Their size does not usually exceed 2 mm.; and they may consist of single
crystals (sometimes twinned) or of an aggregate of smaller crystals. At
times there is a suspicion of intergrowth with rhombic pyroxene; but no
phenocrysts formed alone of that mineral occur.... The augite granules
of the groundmass as a rule vary from ·02 to ·04 mm. in diameter.
Occasional prism-forms giving large extinctions occur.... The
felspar-lathes vary much in length in different rocks. In some they
average as little as ·05 mm., and in others as much as ·2 mm.; but the
doleritic type with yet longer lathes is not represented in the
collection except among the altered rocks. Most of the lathes show a
single median twin-line, and when broader they display twin-lamellæ. The
angles of extinction indicate acid and basic andesine.

Three out of the eight species distinguished by the length of the
felspar-lathes occur in my collection, that with the longest lathes
(·3-·5 mm.) being not represented.


                    2. GENUS OF THE AUGITE-ANDESITES

FORMULA.—_Aug, matr, non-flu, gran, phen, opac._

CHARACTERS.—In the groundmass the felspar-lathes and prisms are not in
flow-arrangement and the augite is granular. Opaque plagioclase
phenocrysts.

DESCRIPTION.—This genus may be divided into two groups according to the
size of the plagioclase phenocrysts, the first “porphyritic,” where they
average 3 mm. and over, the second “non-porphyritic,” where they are
smaller than 3 mm., usually not over 2 mm. The former would include some
of the “porphyrites,” and to this only one of the rocks sliced is to be
referred. All the rest belong to the non-porphyritic type; and several
of them are rocks that have undergone the propylitic change, as
indicated by the formation of pyrites, chlorite, calcite, and other
alteration-products.

(_a_) PORPHYRITIC SUB-GENUS.—A greyish rock, with sp. gr. 2·78, showing
abundant porphyritic opaque plagioclase (4 to 7 mm.), from the vicinity
of Satulaki. These phenocrysts are often aggregates of lesser crystals,
or they may be extensively cross-macled. They are traversed by numerous
fine cracks and show much dust-like included material. They are in part
corroded by the magma and give evidence of fracture in their present
position, the re-union being sometimes effected by the growth of new
substance. Their lamellar extinctions (10° to 20°) are those of
andesine. The groundmass displays a plexus of stout felspar-lathes,
averaging ·1 mm. long, with the meshes occupied by coarse augite
granules, ·03 to ·05 mm., with little, if any, interstitial glass. The
felspars are often lamellar and give extinctions like those of the
phenocrysts.

(_b_) NON-PORPHYRITIC SUB-GENUS.—Reference will first be made to some of
the propylitic rocks of the dykes of the Ndriti Basin which belong to
this group (see p. 70). They are greenish or greyish, with sp. gr. 2·76
to 2·8, and often sparkle with pyrites and contain secondary calcite,
sometimes to such an extent that they might be taken at first sight for
impure limestones.

The small opaque plagioclase phenocrysts (under 2 mm.), that they
contain, are more evident in the slide than in the hand-specimen, and
scarcely give a macroscopic character to the rock. They give extinctions
(10° to 30°) ranging from those of acid andesine to acid labradorite,
and are traversed by numerous cracks occupied by calcitic and other
alteration products. The few pyroxene phenocrysts that once existed are
now entirely represented by chloritic pseudomorphs. The groundmass
displays a doleritic texture, exhibiting a plexus of long
felspar-lathes, ·2 to ·4 mm. in average length, which often present a
false resemblance to a flow-arrangement from their aggregation into
bundles. They are often clouded by secondary products, but occasionally
give lamellar extinctions (20° to 30°) indicating andesine labradorite.
The rest of the groundmass is greatly altered, the granular augite and
the interstitial glass, which originally existed in fair amount, being
replaced by calcite, chlorite, pyrites, and occasionally epidote, so
that the rock mass appears largely impregnated with alteration products.
In addition there is much secondary magnetite, and in some cases there
are a few minute cavities filled with chalcedonic silica and zeolites.

Reference may here be made to a singular rock from Ruku-ruku Bay, which
resembles the Ndriti rocks in its propylitic alteration, but the
felspar-lathes of the groundmass, ·21 mm. in length, give the small
extinctions of oligoclase. Spec. grav. 2·61.

Most of the prevailing rocks of Mount Freeland belong to this sub-genus.
They are dark grey and show small opaque plagioclase phenocrysts 1 or
2 mm. in size. They usually, however, are more or less altered, the
change being often of the propylitic type, calcite, chloritic material,
viridite, and occasionally pyrites occurring as alteration products. The
specific gravity of the altered rocks is 2·61-2·69; that of the least
affected is about 2·76. They all, however, belong to the same genus,
displaying small phenocrysts of plagioclase and augite in a groundmass
composed of minute stoutish felspar-lathes (·03-·06 mm.), augite
granules, magnetite, and a little residual glass. The plagioclase
phenocrysts owe their opacity partly to the numerous fine cracks
traversing them and partly to the alteration products. The pyroxene
phenocrysts, which are mostly of pale yellow augite, display at times
intergrowths of rhombic pyroxene.


                    4. GENUS OF THE AUGITE-ANDESITES

FORMULA.—_Aug, matr, non-flu, gran, non-phen, parv._

CHARACTERS.—In the groundmass the felspar-lathes and prisms are not in
flow-arrangement and the augite is granular. There are no plagioclase
phenocrysts, and those of augite when present are small (under 2 mm.).

DESCRIPTION.—Two groups of these rocks occur in my collection. In the
one there are vesicular and scoriaceous rocks forming dykes near
Nukunase and near the village of Ndriti. They display a plexus of
felspar-lathes with abundant smoky more or less devitrified glass, the
augite granules not being always differentiated. The felspar-lathes vary
from ·1 to ·2 mm. in average length, and when lamellar give extinctions
of basic andesine (25°). There are no pyroxene phenocrysts, and the
augite granules when present average ·02 mm. in size.

In the other group are included some propylites from the dykes of the
Ndriti Basin. They are greyish or greenish rocks, have a sp. gr. of 2·72
to 2·76, sparkle often with pyrites, and contain so much secondary
calcite that they effervesce freely with an acid. Except in the rarity
or absence of plagioclase phenocrysts, they come near to the propylitic
rocks described under genus 2. They usually display a doleritic
groundmass exhibiting long felspar-lathes, ·2 to ·33 mm. in length,
which present the same pseudo-flow arrangement from their being gathered
into bundles. The alteration corresponds precisely to that previously
described, chlorite, epidote, pyrites, &c., occurring in quantity as
secondary products.


          II PRISMATIC SUB-ORDER OF THE AUGITE-ANDESITES WHERE
             THE FELSPAR-LATHES ARE NOT IN FLOW-ARRANGEMENT

FORMULA.—_Aug, matr, non-flu, prism._

The augite-andesites, which display in the groundmass a plexus of
felspar-lathes and much prismatic pyroxene, are not frequent in my
collection. About half of the specimens belong to agglomerates, whilst
the rest are of the massive type, none apparently being obtained from
dykes. They admit of the same classification as that generally adopted
for the “granular” sub-orders; and it must be not forgotten that
granular pyroxene also occurs but is not predominant.

5. GENUS OF THE AUGITE-ANDESITES

FORMULA.—_Aug, matr, non-flu, prism, phen, vitr._

CHARACTERS.—In the groundmass the felspar-lathes and prisms are not in
flow-arrangement and the augite is for the most part prismatic.
Plagioclase phenocrysts glassy.

DESCRIPTION.—Except as regards the prismatic pyroxene these rocks do not
differ much from the “granular” augite-andesites. Those before me show
phenocrysts of plagioclase and sometimes of augite in a groundmass
displaying a mesh-work of felspar-lathes, prismatic pyroxene, and much
interstitial glass.... The plagioclase phenocrysts, 1 to 3 mm. in size,
show abundant magma-inclusions arranged either zone-wise or parallel to
the twinning-planes. They are often eroded.... The phenocrysts of
augite, which give extinctions of over 30°, are often rounded and
display glass and other inclusions.... The prismatic pyroxenes of the
groundmass vary in average length from ·03 to ·08 mm. They have the
peculiar pale muddy brown hue characteristic of the prismatic augite in
these rocks, and give oblique extinctions up to 30° and over. They may
be short and broad or long and slender, and when there is much glass in
the rock they resemble the felspar-lathes in their forked and imperfect
extremities. Granular pyroxene occurs, but is subordinate.... The
felspar-lathes, ·1 mm. long, are rarely lamellar, and give extinctions
measured from their long axis of 20° (basic andesine).


                    6. GENUS OF THE AUGITE-ANDESITES

FORMULA.—_Aug, matr, non-flu, prism, phen, opac._

CHARACTERS.—In the groundmass the felspar-lathes are not in
flow-arrangement and the augite is for the most part prismatic. The
plagioclase phenocrysts are opaque.

DESCRIPTION.—Light and dark grey rocks displaying abundant opaque
plagioclase phenocrysts not exceeding 2·5 mm. They are somewhat altered,
one of the specimens having a sp. gr. of 2·68.

In the section they exhibit phenocrysts of plagioclase, and occasionally
of augite in a groundmass of felspar-lathes, pyroxene prisms and
granules (the former predominating), with a fair amount of altered
interstitial glass.... The plagioclase phenocrysts owe their opacity to
the great number of fine and sometimes parallel cracks filled with
alteration products, that traverse them. Although much of their original
material has often disappeared, they still display the lamellar twinning
of medium andesine (15° to 20°).... The phenocrysts of pale yellowish
augite, which give the large extinctions of that mineral, exhibit but
little alteration, although lying in the same slide with those of the
plagioclase.... The pyroxene prisms of the groundmass are of the same
yellowish augite. They are broad with rounded extremities and are
arranged in a loose plexus.... The felspars of the groundmass, which
average ·1 mm. in length and give extinctions of medium andesine, are
either lathe-shaped or short and broad when they display lamellæ. The
last-named approach the orthophyric type, and such rocks come near the
porphyrites; but I do not feel justified in placing them in a separate
orthophyric order.


         III OPHITIC SUB-ORDER OF THE AUGITE-ANDESITES WITH THE
          FELSPARS OF THE GROUNDMASS NOT IN FLOW-ARRANGEMENT.

FORMULA.—_Aug, matr, non-flu, oph._

These rocks form generally ancient flows. They are for the most part
semi-ophitic, large ophitic “plates” being uncommon.


                    9. GENUS OF THE AUGITE-ANDESITES

FORMULA.—_Aug, matr, non-flu, oph, phen, vitr._

CHARACTERS.—The felspar-lathes and prisms of the groundmass are not in
flow-arrangement. The augite of the groundmass is ophitic or
semi-ophitic. Glassy plagioclase phenocrysts.

This genus may be divided into two sub-genera,

 (_a_) Porphyritic, where the average size of the plagioclase
    phenocrysts
 is 3 mm. and over.

 (_b_) Non-porphyritic, where the size is less than 3 mm.


                        A. PORPHYRITIC SUB-GENUS

DESCRIPTION.—Coarse-looking brownish-black porphyritic rocks displaying
large plagioclase crystals that often show a play of colours. Their sp.
gr. is about 2·8. None of the rocks in my collection are vesicular. On
account of the considerable porphyritic development of the plagioclase,
the groundmass is relatively diminished, the large phenocrysts occupying
about a third of the mass. They form ancient basaltic flows more
especially in the vicinity of the isolated hills and mountains of acid
andesite, as around Vatu Kaisia; whilst they may enter into the
formation of the low basaltic plains as in the region west and south of
the Ndreketi River. They are, however, limited in their extent and
occurrence. From the large amount of glass in the groundmass, they may
be inferred to belong to flows formed under different conditions from
those under which the great basaltic plateaux were formed, where the
rock contains but scanty interstitial glass.

In the slide they show the large plagioclase phenocrysts together with a
few small plates of ophitic augite in a groundmass displaying in an
abundant smoky glass a loose plexus of long stout lathe-like plagioclase
prisms partly wrapped around by lesser augites.... The plagioclase
phenocrysts, which attain a size of 4 to 6 mm., give lamellar
extinctions of basic andesine (20°-27°) and show concentric zone-lines
with transmitted light. They often polarise in brilliant colours and are
extensively cross-macled. They contain usually abundant inclusions of
the magma sometimes arranged zone-wise, and are frequently eroded....
Non-ophitic pyroxene phenocrysts are uncommon. In the slide occur one or
two small “plates,” 1 to 2 mm. in size, of ophitic pale-brown augite,
and a number of lesser augites, ·2 to ·3 mm. in size, which in part wrap
around the felspar-lathes and by their aggregation form imperfect
ophitic “plates.”... The long stout felspar-lathes, which are on the
average ·3 to ·45 mm. in length, give lamellar extinctions of 15° to 20°
(medium andesine).... The copious smoky glass is rendered partially
opaque by the abundant development of rods and skeletal crystals of
magnetite, and shows the fibrous devitrification arising from the
formation of incipient microliths. In some rocks there appear in the
smoky glass brownish-yellow patches of the residual magma which under
the microscope cannot be distinguished from palagonite.

All but one of the specimens belong to the species where the
felspar-lathes average over ·3 mm. in length.


                      B. NON-PORPHYRITIC SUB-GENUS

DESCRIPTION.—Blackish-brown semi-ophitic rocks, sp. gr. 2·74-2·77,
frequently of doleritic texture and showing a few small macroscopic
plagioclase phenocrysts. They are sometimes vesicular, and form old
flows in a few localities, as in the vicinity of Natua in the eastern
part of the Ndreketi plains and in the coast district between Lekutu and
Wailea Bay.

They display in the slide small plagioclase phenocrysts, often abundant,
in a groundmass exhibiting a loose plexus of large lathe-shaped felspar
prisms, together with occasional small ophitic “plates” of augite and
numerous smaller semi-ophitic augites, whilst there is much interstitial
smoky glass.... The plagioclase phenocrysts are as a rule 1 to 2 mm. in
size and do not exceed 3 mm. They often contain abundant inclusions of
the magma sometimes arranged schiller-fashion, and are frequently
eroded. Their lamellar extinctions (15°-30°)indicate andesine
labradorite.... The stout plagioclase lathes, which in most of my
specimens range between ·2 and ·3 mm., and contain at times
magma-inclusions, give the rocks their doleritic texture.... The
occasional small ophitic “plates” of pale augite are not over 1 mm. in
size and give extinctions of +30° from the single cleavage-lines. The
lesser augites, ·2 mm. in size, are several times larger than typical
granular pyroxenes (·02-·03 mm.), and adapt their form to the
interspaces of the felspar-lathes which they in part invest.... The
copious interstitial glass is generally smoky and sometimes quite opaque
through the deposition of magnetite. It is never clear and isotropic,
but displays fibrous devitrification and is usually a little altered.

These rocks come near to those of the previous sub-genus in several
respects, but they differ conspicuously in their non-porphyritic
character, in being sometimes vesicular, and in their general
appearance. All the four species indicated by the length of the
felspar-lathes are here represented; but the doleritic types with the
lathes exceeding ·2 mm. are the most frequent.


                   10. GENUS OF THE AUGITE-ANDESITES

FORMULA.—_Aug, matr, non-flu, oph, phen, opac._

CHARACTERS.—In the groundmass the felspar-lathes and prisms are not in
flow-arrangement and the augite is ophitic or semi-ophitic. The
plagioclase phenocrysts are opaque.

This genus may be divided into two sub-genera, porphyritic and
non-porphyritic, according to the average size of the plagioclase
phenocrysts, whether above or below 3 mm.

A. PORPHYRITIC SUB-GENUS.—This is represented by a light grey
porphyritic rock, with sp. gr. 2·75, from the lower part of Mount
Freeland. It comes near to the porphyrites, and displays large opaque
white phenocrysts of plagioclase 5 or 6 mm. long. It is a somewhat
altered rock.

In the thin section it displays the plagioclase phenocrysts in a ground
mass of doleritic and semi-ophitic texture showing a plexus of long
felspar-lathes partly invested by small augites with a fair amount of
altered greenish opaque interstitial glass.... The plagioclase
phenocrysts, which give extinctions of 11° to 15° (acid andesine), are
traversed by a network of fine cracks and contain a quantity of
colourless dust-like inclusions and alteration-products. They are long
and rectilinear in outline and are not much affected by the magma....
The felspar-lathes, which average ·26 mm. in length, are occasionally
lamellar when the angle of extinction indicates acid andesine. Like the
phenocrysts they contain many dust-like inclusions.... The augite may at
times form an aggregate phenocryst of small size; but usually it occurs
as semi-ophitic masses ·1 mm. in diameter.

B. NON-PORPHYRITIC SUB-GENUS.—The specimen representing this group is a
coarse-grained greyish altered rock, sp. gr. 2·81, found in blocks near
Waikatakata (p. 204), showing small somewhat opaque plagioclase
phenocrysts. In the section these phenocrysts are displayed in numbers
together with a few of augite. The groundmass, doleritic in texture in
this species, displays a plexus of long stout felspar-lathes with
numerous semi-ophitic lesser augites, chloritic pseudomorphs after
pyroxene, and scanty interstitial altered glass.... The plagioclase
phenocrysts, 2 to 3 mm. in size, give extinctions (22° to 28°) of basic
andesine. They are traversed by many fine cracks and contain an
abundance of colourless dust-like inclusions, apparently altered magma
products.... The small pyroxene phenocrysts consist of aggregates of
smaller crystals of pale augite. The lesser augites, ·1 mm. in size,
partly invest the felspars.... The broad felspar-lathes, which average
·25 mm. in length and give lamellar extinctions of medium andesine
(15°), contain inclusions similar to those of the phenocrysts.... The
scanty interstitial glass is converted into viriditic and chloritic
materials. Secondary calcite also occurs here and in the chloritic
pseudomorphs.

The only species represented is that with felspar-lathes ·2 to ·3 mm. in
length.


                   12. GENUS OF THE AUGITE-ANDESITES

FORMULA.—_Aug, matr, non-flu, oph, non-phen, parv._

CHARACTERS.—In the groundmass the felspar lathes and prisms are not in
flow-arrangement and the augite is ophitic or semi-ophitic. No
plagioclase phenocrysts. Augite phenocrysts when present less than 2 mm.
in size.

DESCRIPTION.—These rocks come near to the non-porphyritic group of genus
9; but differ in the absence or rarity of plagioclase phenocrysts, in
their more frequently vesicular and scoriaceous character, and in the
fresher condition of the rock. Their sp. gr. is about 2·77. They present
themselves usually as blackish-brown doleritic rocks and form ancient
flows in the coast-plains, sometimes exhibiting a columnar structure as
in the crossing of the Ndreketi above Mbatiri. They are, however, of
limited occurrence and are mostly represented in the Ndreketi plains and
in the district between the Lekutu River and Wailea Bay.

Typically they display in thin sections no phenocrysts either of
plagioclase or of pyroxene, and exhibit a plexus of usually long stout
felspar-lathes partly invested by the lesser augites in a copious
smoky glass.... The felspar-lathes, ·25 to ·4 mm. in length, give
lamellar extinctions of 10° to 20° (andesine) and contain a few
magma-inclusions.... The semi-ophitic augites, ·1 to ·2 mm. in size,
are sometimes twinned.... The smoky glass polarises feebly and
displays dark feathery aggregates of microliths. Within it are
brownish-yellow semi-opaque patches of residual glass, which polarise
faintly and behave like palagonite.

The two species with felspar-lathes ·2 to ·3 mm. and ·3 to ·5 mm. are
represented in my collection.


             4. GRANULAR SUB-ORDER OF THE AUGITE-ANDESITES

FORMULA.—_Aug, matr, flu, gran._


                               13. GENUS

FORMULA.—_Aug, matr, flu, gran, phen, vitr._

CHARACTERS.—In the groundmass the felspar-lathes or prisms are in
flow-arrangement and the augite is granular. The plagioclase phenocrysts
are glassy.

DESCRIPTION.—This genus readily splits up into two sub-genera, the
non-porphyritic, where the plagioclase phenocrysts are less than 3 mm.
in size, and the porphyritic where they are larger.

1. NON-PORPHYRITIC SUB-GENUS.—Dark-brown or blackish rocks displaying
small plagioclase phenocrysts, usually only 1 or 2 mm. in size. Three of
the four species defined by the length of the felspar-lathes are
represented in my collection.

SPECIES A.—Felspar-lathes .02-.1 mm. in average length. This may again
be sub-divided according to the degree of basicity of the rocks:—

(_a_) _Sub-species of greater basicity._—Sp. gr. 2·76 to 2·82.... Such
rocks are represented in dykes and in the prevailing basic agglomerates.
They are at times scoriaceous. The small plagioclase phenocrysts, which
are fairly numerous, give lamellar extinctions of andesine labradorite
(20 to 30°). Two kinds occur which may or may not be represented in the
same slide. In the one the crystal is much corroded and contains
abundant magma-inclusions. It belongs in such a case to an earlier
period. In the other the outlines are clean and regular, and the crystal
is often cross-macled to such an extent that it may be inferred from its
unbroken condition to have been formed _in situ_. Augite phenocrysts
when present are small and scanty, pale-yellow, idiomorphic, and giving
extinctions of +30°. The felspar-lathes, which average ·06-·08 mm. in
length, give extinctions indicating andesine labradorite. The augite
granules are small (·01-·02 mm.). Interstitial glass, generally scanty,
is sometimes abundant when it is smoky, showing fibrous devitrification,
with irregular “lacunæ” filled with a brownish yellow opaque glass like
palagonite.

(_b_) _Sub-species of lesser basicity._—Sp. gr. 2·65-2·70.... The
remarks on the plagioclase phenocrysts of the more basic sub-species
here apply, except that the lamellar extinctions indicate medium
andesine (12°-20°). The characters of the augite phenocrysts and
granules are in both groups the same; but in this case there is more
frequently a suspicion of intergrowth with rhombic pyroxene. The
felspar-lathes are very small, ·04 or ·05 mm, and give simple
extinctions of acid andesine (5-10°). Interstitial glass exists in
moderate amount.

SPECIES B.—Felspar-lathes ·1-·2 mm. in average length.

Blackish or dark-grey rather compact rocks, sp. gr. 2·75-2·79, that
cannot be readily divided into groups according to their basicity. They
form dykes and volcanic “necks” and are sometimes scoriaceous. The small
plagioclase phenocrysts, which are most evident in the slide, present
the two kinds above described under Species A. They give lamellar
extinctions varying from those of medium andesine to acid labradorite
(15-30°). The augite phenocrysts, which are small and scanty,
occasionally show intergrowths of rhombic pyroxene. The augite granules
are generally ·02 to ·03 mm. in size, and here and there a prism form
gives extinctions of +25°. The felspar-lathes which average ·11 to
·15 mm. long, are often rather stout, showing a few lamellæ that give
extinctions of medium and basic andesine. Interstitial glass occurs in
fair amount.

SPECIES C.—Felspar-lathes ·2-·3 mm. in average length.

Blackish rocks with sp. gr. 2·75-2·84. The description of Species B
applies here. The plagioclase phenocrysts are for the most part
microporphyritic. The size of the augite granules is as above given.

2. PORPHYRITIC SUB-GENUS.—This group of rocks is mostly confined to the
slopes and vicinity of Mount Seatura in the western part of the island,
being prevalent in the Mbua and Ndama plains, and occurring also as
dyke-rocks in the Nandi Gorge leading into the Ndriti Basin, and at and
near the coasts of Wainunu Bay between the Tongalevu and Wainunu rivers.
They come near in appearance to the porphyritic forms of the blackish
olivine-basalts belonging to genera 13, 25, and 37 of the olivine rocks;
but they differ in the absence of that mineral, in their lower density,
and in other characters. They are the type to which the term
“porphyritic basaltic andesite” is most frequently applied in the text
when the ophitic structure is not displayed.

They are blackish rocks having a specific gravity of 2·71 to 2·81 and
exhibiting large porphyritic crystals of plagioclase, but they vary in
their minute structure on account of the different size of the felspars
of the groundmass. Those forming dykes in the Nandi Gorge are often more
or less propylitic in character. The felspar-lathes, which have an
average length of ·2 to ·3 mm., sometimes show a few lamellæ giving
extinctions of medium andesine (12°-20°). The plagioclase phenocrysts of
the same andesine are 3 to 5 mm. in size. They are eroded and contain
abundant magma inclusions. There are a few small phenocrysts of pale
brown augite. The augite granules are ·03 or ·04 mm. in diameter, and
there is a little dark opaque residual glass.

The rocks of the Mbua and Ndama plains have a specific gravity of 2·81.
The plagioclase phenocrysts, which yield extinctions of basic andesine
(21-27°), are sometimes a centimetre in length. They are traversed by
cracks filled with dark altered glass or occupied by brownish films. The
felspar-lathes, which average ·11 mm. in length, are often stout and
lamellar and give extinctions like the phenocrysts. Augite phenocrysts
are either absent or scanty; whilst the granules average ·02-·03 mm. in
size. There is usually a little interstitial glass.

A rock, almost holocrystalline and 2·74 in density, which was obtained
from the Tongalevu district in Wainunu Bay, approaches the orthophyric
type in the character of the ground mass. The felspars are short
(·06 mm.) and stout, and yield lamellar extinctions of oligoclase
(5°-10°). The plagioclase phenocrysts are of basic andesine. Amongst the
granules (·025 mm.) of pale brown augite occur prismatic forms giving
oblique extinctions of +30°.


                   14. GENUS OF THE AUGITE-ANDESITES

FORMULA.—_Aug, matr, flu, gran, phen, opac_.

CHARACTERS.—In the groundmass the felspar-lathes are in flow-arrangement
and the augite is granular. The plagioclase phenocrysts are opaque.

DESCRIPTION.—Only two rocks are referred to this genus. One which is
dark grey with a specific gravity of 2·72 is exposed in the gorge of the
Mbutu-mbutu River below the falls of Na Savu. Flow-arrangement is
displayed both by the felspar phenocrysts and lathes. The phenocrysts, 2
to 3 mm. in size, owe their opacity to the abundance of inclusions of
brown glass. They are corroded and give extinctions of acid labradorite
(26-32°). Pyroxene phenocrysts are scanty and small. The groundmass has
a characteristic “pilotaxitic” appearance, the densely packed
felspar-lathes averaging only ·05 mm. in length, whilst the pyroxene
granules are ·01 mm. in size. Residual glass scanty.

The other rock is from the range behind Sueni. It shows large
porphyritic crystals (5 or 6 mm.) of medium andesine which contain
magma-inclusions in abundance. The average length of the felspar-lathe
is ·08 mm. and the size of the augite granules is ·02 mm. There is but
little glass. The rock is somewhat altered.


                   16. GENUS OF THE AUGITE-ANDESITES

FORMULA.—_Aug, matr, flu, gran, non-phen, parv_.

CHARACTERS.—In the groundmass the felspar-lathes are in flow-arrangement
and the augite is granular. Plagioclase phenocrysts are absent, or if
present very scanty and not usually over 1 mm. in size. When present the
augite phenocrysts are under 2 mm.

DESCRIPTION.—This is a very extensive genus, admitting considerable
variation and including most of the aphanitic augite-andesites, where
the felspar-lathes are as a rule very small (under ·1 mm. in length), as
well as some of the doleritic types where they are very large (·2 to
·4 mm. long). In assigning a rock a place in this genus some regard must
be paid to its macroscopic aspect as well as to the presence or absence
of plagioclase phenocrysts. In many cases two or three small phenocrysts
may be observed in a slide, under a millimetre in size; but they do not
give a character to the naked-eye appearance of the rock, and such rocks
cannot be distinguished from others that do not display them.

These rocks range in specific gravity from 2·55 to 2·85. This large
range is in the main concerned with different degrees of basicity
depending on the character of the plagioclase, the relative abundance of
the augite granules, &c.; but it is also connected with the amount of
interstitial glass. The variety of plagioclase ranges between oligoclase
and andesine labradorite. The fluidal structure is nearly always
well-marked, and the closely packed felspar-lathes have often the
peculiar “felted” appearance of many andesites. A little interstitial
glass is present in most rocks.

Many, perhaps nearly all, of the rocks belong either to dykes or to
larger intrusive masses. All the four species indicated by the length of
the felspar-lathes are represented in my collection, especially the two
with smallest felspars. They may again be split up into two sub-species
according to the degree of basicity of the rocks.

SPECIES A.—Felspar-lathes between ·02 mm. and ·1 mm. in average length.

(1) _Most basic sub-species_.... Sp. gr. 2·75-2·85. Dark-brown and
dark-grey compact aphanitic rocks showing no plagioclase phenocrysts to
the eye. When a few of these phenocrysts are present in a slide they are
not usually much over 1 mm. in size, and give extinctions of andesine
labradorite (20° to 30°). Augite phenocrysts are often absent, and when
present are not over 1 mm. in size and are as a rule scanty,
occasionally affording a suspicion of inter-growths with rhombic
pyroxene. The felspar-lathes which display marked flow-structure, vary
in average length in different rocks from ·05-·08 mm. Lamellar twinning
is rare, the extinctions being those of oligoclase andesine (10° to
20°). The usual extinction, as measured from the long axis of the lathe,
is 10° to 15° (medium andesine). The augite granules are small
(·01-·02 mm.) and abundant. There is generally a little interstitial
glass with small magnetite.

(2) _Least basic sub-species_.... Sp. gr. 2·55-2·75. Dark compact
aphanitic rocks especially characteristic of the Ndrawa district. When
plagioclase phenocrysts are present, they are very scanty and not
generally over a millimetre in size, possessing rectangular clean
outlines and showing but few inclusions. They may display carlsbad
twinning and zoning, or albite twinning, when they give extinctions of
oligoclase andesine (10°-15°). Pyroxene phenocrysts are either absent,
or scanty and small, being usually of pale yellow augite with occasional
indications of intergrowth with rhombic pyroxene. The felspar-lathes as
a rule average ·07 or ·08 mm. and present a dense fluidal arrangement.
They rarely display lamellar twinning and give extinctions measured from
the long axis of oligoclase and oligoclase andesine (2°-8°). The
pyroxene granules are very small, averaging ·01 mm. and less. There is
also fine magnetite. A little interstitial glass is usually present.
When abundant it is not generally smoky but shows clear fibrous
devitrification.... One of the specimens, which is semi-vitreous,
exhibits tube-like steam-pores drawn out to a length of 1-1½
centimetres. The felspar microliths are only ·02 mm. in length. The
copious glass has the character above described.

SPECIES B.—Felspar-lathes ·1-·2 mm. in average length.

This species may also be sub-divided into two sub-species (more basic
and less basic). Since, however, all but one of the fifteen rocks
belonging to the species are of the more basic kind my remarks will
mainly apply to them. They are dark-brown or dark-grey compact aphanitic
rocks, occasionally banded or streaky, in appearance, and ranging in
specific gravity from 2·75 to 2·84. They occur in several districts, but
are especially characteristic of the Ndrawa district. The plagioclase
phenocrysts, if present, are very scanty and small (1 or 1½ mm.). They
contain inclusions of the magma and give lamellar extinctions of
andesine labradorite (20°-30°). Pyroxene phenocrysts do not generally
occur. When present, they are small and of pale yellow augite yielding
large extinctions. Occasionally micro-porphyritic augite is well
represented. The felspar-lathes, which exhibit a well-marked
flow-arrangement, are generally ·13 to ·15 mm. long. Lamellar twinning
is uncommon, the extinctions measured from the long axis indicating
basic andesine (10°-20°). The augite granules are abundant and small
(·01-·02 mm.). Occasional prism-forms yield large extinctions. Magnetite
is abundant, its grains corresponding in size to the augite granules.
There is as a rule a little residual glass, which shows fibrous
devitrification and is not smoky. The banded appearance of some of the
rocks arises from the glass collecting in streaks rudely parallel and
running in the direction of the “flow” of the felspar-lathes.

The only specimen in my collection of “sliced” rocks belonging to the
less basic sub-species is an altered bluish-grey rock (sp. gr. 2·7) from
the range between the Mbuthai-sau valley and the Wainikoro plains. Its
long parallel untwinned felspar-lathes give the nearly straight
extinctions of oligoclase. Fine cracks in the rock are filled with
crystalline silica.

 SPECIES C, felspar-lathes ·2-·3 mm. long.
 SPECIES D, felspar-lathes ·3-·5 mm. long.

The rocks of these species in the collection are for the most part
dyke-rocks of the more basic kind. They are blackish or dark-brown,
almost doleritic in texture, and range in specific gravity from 2·77 to
2·87. At times they are vesicular or scoriaceous, as in the specimens
from an agglomerate at Undu Point and from a flow or dyke at Vunikondi.
The most typical of these rocks are those of some of the dykes of the
Ndriti basin, which, however, display propylitic alteration in a varying
degree. They would be described as semi-doleritic basalts without
olivine or as non-porphyritic basaltic andesites. Plagioclase
phenocrysts are typically absent, or they are scanty and not over 1 mm.
in size. Augite phenocrysts are usually scanty and small. The
felspar-lathes, which are more or less in flow-arrangement, are rather
stout, and range in average length in different rocks from ·23 to
·35 mm. They often show a few twin-lamellæ which yield extinctions of
medium and basic andesine (15-28°). The augite granules are large
(·03 mm.) in the Ndriti rocks. Magnetites, usually corresponding in size
to the augite granules, are abundant. Interstitial glass occurs often in
fair quantity and is dark and semi-opaque.

At times there can be recognised a later generation of minute felspar
microliths between the much larger lathes. They display a plexus rather
than a flow-arrangement. Whilst the larger parallel lathes of the
Vunikondi rock, above referred to, average ·23 mm. long, the felspar
microliths of the interspaces average only ·03 mm. The significance of
these two crops of felspars in the groundmass is discussed on page 237.

The only rock of the less basic sub-species in my collection is from a
dyke near Vatua-karoa. It shows secondary calcite and viridite and other
evidences of the propylitic change. The felspar-lathes, which average
·3 mm. in length, give extinctions of oligoclase (0-5°). The specific
gravity is 2·72.


            5. SUB-ORDER, PRISMATIC, OF THE AUGITE-ANDESITES

(Felspar-lathes in flow-arrangement. _Aug, matr, flu, prism_.)

This sub-order includes dark-brown or blackish semi-vitreous rocks, all
but one of which belong to the genus below described. Since the
exception (which belongs to genus 17 of the synopsis) differs only in
the presence of plagioclase phenocrysts, its separate description is not
needed.


                   20. GENUS OF THE AUGITE-ANDESITES

FORMULA.—_Aug, matr, flu, prism, non-phen, parv._

CHARACTERS.—In the groundmass the felspar-lathes are in
flow-arrangement, and the augite is prismatic. Plagioclase phenocrysts
are absent or are very small and scanty, and pyroxene phenocrysts when
present do not exceed 2 mm. in size.

DESCRIPTION.—These dark semi-vitreous rocks occur in agglomerates and as
“necks” and dykes, and are at times scoriaceous. They are usually
compact and aphanitic, showing few if any plagioclase phenocrysts and
having a semi-conchoidal fracture. The specific gravity varies, being
generally 2·6-2·65, but according to the degree of basicity and amount
of glass it may be as low as 2·5 or as high as 2·77. In the less glassy
condition, as in the case of a rock from the ridge east of Na Raro, the
felspar-lathes are relatively scanty and the groundmass is mainly formed
of augite prisms in flow-arrangement. The lathes are generally small,
less than ·1 mm., and rarely over ·2 mm. Their extinctions are those of
oligoclase and acid andesine. The pyroxene prisms, which give the large
extinctions of augite, have a pale muddy-brown hue and are as a rule
·03-·07 mm. long. Granular pyroxene, if present, is subordinate in
amount. The glass, which is always in good quantity and is sometimes
abundant, displays fibrous devitrification. In a rock from the vicinity
of Narengali a variolitic structure is exhibited in the form of
sheaf-like aggregates of fibre-like felspars and skeleton prisms of
pyroxene.


             6. SUB-ORDER, OPHITIC, OF THE AUGITE-ANDESITES

(Felspar-lathes in flow-arrangement. _Aug, matr, flu, oph._)

21. GENUS OF THE AUGITE-ANDESITES

FORMULA.—_Aug, matr, flu, oph, phen, vitr._

CHARACTERS.—In the groundmass the felspar-lathes are in flow-arrangement
and the augite is ophitic or semi-ophitic. Glassy plagioclase
phenocrysts.

DESCRIPTION.—Dark rocks, sp. gr. 2·76-2·8, forming ancient flows and
displaying at times a columnar structure as at Yanutha Point (page 123).
The ophitic character is only in part developed, which may be connected
with the flow-arrangement of the felspars. These rocks come near to the
blackish ophitic basalts with scanty olivine (genus 33 of the olivine
sub-class).

They all belong to the non-porphyritic sub-genus where the plagioclase
phenocrysts are less than 3 mm. in size. These phenocrysts, which often
contain abundant magma-inclusions, give extinctions of andesine
labradorite (20°-30°). The augite phenocrysts are small and composite in
character as often happens with these ophitic rocks. They sometimes
invest the smallest felspar phenocrysts, and occasionally display
intergrowths of rhombic pyroxene. The felspar lathes are ·1 to ·14 mm.
in length, and give extinctions of medium and basic andesine. The augite
granules are large (·02-·06 mm.), and tend to wrap around the lathes.
Interstitial glass exists in fair amount.



                               CHAPTER XX

             THE VOLCANIC ROCKS OF VANUA LEVU (_continued_)


                        HYPERSTHENE-AUGITE CLASS

              II. SUB-CLASS. HYPERSTHENE-AUGITE-ANDESITES

FORMULA.—_Plag, hypersth-aug, matr._

CHARACTERS.—The pyroxene phenocrysts usually are represented by separate
crystals of the monoclinic as well as the rhombic type, and the two
forms are often associated in the same crystal. The monoclinic form
prevails in the groundmass in most cases.

REMARKS.—It is not possible to draw a sharp line between the augite and
the hypersthene-augite-andesites; but where two or three phenocrysts of
the rhombic type occur in a slide the rock may be placed in this
division. Between this variety and that where rhombic pyroxene prevails,
both among the phenocrysts and in the groundmass, numerous intermediate
kinds exist. These rocks mostly occur in agglomerates and form small and
large dykes or sills, but rarely are found in flows. They are
distributed over most of the island except in the western portion (the
basaltic districts of Wainunu, Seatura, and Solevu), but reappear again
in the Mbua peninsula in places, as at Mount Koroma.

The pale yellow rhombic pyroxene is uniform in its optical behaviour.
The prisms are noticeably pleochroic, being nearly colourless when lying
across the long axis of the lower nicol and pale yellow when parallel
with it. The intergrowths with monoclinic pyroxene often take the form
of lamellar bands, whilst in some cases a nucleus of the one (usually
rhombic) is invested by a growth of the other.


       1 SUB-ORDER, GRANULAR, OF THE HYPERSTHENE-AUGITE-ANDESITES

              (_Felspar-lathes not in flow-arrangement._)

FORMULA.—_Hypersth-aug, matr, non-flu, gran._

 1 genus        (_Vitr._) }
 2   "          (_Opac._) } See Synopsis, p. 247.
 3   "          (_Magn._) }
 4   "          (_Parv._) }

Nearly all of the rocks of this sub-order that are represented in my
collection belong to the genus (1) with phenocrysts of glassy
plagioclase. They vary considerably in appearance and in colour (black
to grey), and occur under very different conditions, as in “necks,” old
flows, large intrusive masses, dykes, agglomerates, &c. Their specific
gravity has rather a wide range according to the degree of basicity. In
the heavier rocks where the rhombic pyroxene is scanty, it is usually
2·7 to 2·8. In the others, where rhombic pyroxene is more predominant
and where the felspar is less basic, it is 2·6 to 2·7.

In the slide small phenocrysts of plagioclase and pyroxene occur in a
groundmass of felspar-lathes and pyroxene granules, whilst there is as a
rule a fair amount of residual glass. The plagioclase phenocrysts, which
give extinctions in different rocks of acid and basic andesine and
contain abundant magma-inclusions, are generally one to two mm. in size.
The pyroxene phenocrysts are small, and may be represented by separate
crystals of the monoclinic and rhombic kinds, or by crystals displaying
intergrowths of the two sorts. The pyroxene granules vary much in size
and are evidently in great part of augite. In most of the rocks the
felspar-lathes are less than ·1 mm. in length. In those where the length
is ·1 to ·2 mm. they are sometimes stout and display a few lamellæ,
yielding extinctions corresponding to those of the phenocrysts.

A singular dark grey almost holocrystalline doleritic rock (sp. gr.
2·85) is exposed in the Thulanga Ridge (p. 211). It shows no plagioclase
phenocrysts, but those of pyroxene are numerous, which, however, do not
exceed 2 mm. in size, so that the rock would be referred to genus 4 of
this sub-order. It appears to be a doleritic form of the plutonic rock
found at Nawi in this neighbourhood (p. 211). The pyroxene phenocrysts
are mostly of brownish-yellow augite, but rhombic pyroxene, either as
separate crystals or as intergrowths, is not uncommon. The plagioclase
lathes are long and fairly stout, giving at times lamellar extinctions
of 20°. Their average length is ·3 mm., and it is to their large size
that the doleritic texture is due. The pyroxenes of the groundmass are
similarly coarse (·2 mm.), and include both monoclinic and rhombic
forms, the latter infrequent. There is a slight tendency to semi-ophitic
behaviour in places; but generally these pyroxenes are irregular in
shape or rudely prismatic.


      2. SUB-ORDER, PRISMATIC, OF THE HYPERSTHENE-AUGITE-ANDESITES

              (_Felspar-lathes not in flow-arrangement._)

FORMULA.—_Hypersth-aug, matr, non-flu, prism._

 5 genus    (_Vitr._) }
 6   "      (_Opac._) } See Synopsis.
 7   "      (_Magn._) }
 8   "      (_Parv._) }

This sub-order includes rocks varying much in appearance, but all alike
in the presence of prismatic pyroxene in the groundmass and in the
absence of flow-arrangement of the felspar-lathes. They belong to the
two genera, with glassy and opaque plagioclase phenocrysts. These
crystals are not usually over 2 mm. in size and are of medium andesine
(15°-20°). The pyroxene phenocrysts are small and may be entirely of
monoclinic or of rhombic pyroxene, or the two may be associated either
as lamellar intergrowths, or by displaying an eroded nucleus (generally
rhombic) around which a regular crystal of monoclinic pyroxene has
grown. The felspar-lathes are in some cases less than ·1 mm. long, and
in others ·1-·2 mm. The pyroxene prisms of the groundmass average
·01-·04 mm. in length, and give both straight and oblique extinctions,
the last prevailing. The specific gravity ranges from 2·55 to 2·75
according to the degree of basicity and amount of interstitial glass,
which is usually in fair quantity.


      4. SUB-ORDER, GRANULAR, OF THE HYPERSTHENE-AUGITE-ANDESITES

                (_Felspar-lathes in flow-arrangement._)


                               13. GENUS

FORMULA.—_Hypersth-aug, matr, flu, gran, phen, vitr._

CHARACTERS.—Glassy plagioclase phenocrysts.

DESCRIPTION.—This is a group of rocks that comes near the basaltic
andesites represented in genera 1 and 13 of the augite sub-class; and to
the more basic kinds the terms of basaltic andesite is equally
applicable. These rocks, however, differ in the prevalence of rhombic
pyroxene, which occurs as phenocrysts, but always accompanied by
monoclinic pyroxene, whether as separate crystals or as inter-growths.
Such rocks are intermediate between those of the augite and rhombic
pyroxene classes. They are particularly characteristic of the Savu-savu
peninsula, but they are also found in other scattered localities.
Sometimes they appear to form ancient flows, and at other times
intrusive masses and dykes; but they are rarely scoriaceous.

Almost all the rocks in my collection referred to this genus belong to
the species where the felspar-lathes of the groundmass are ·02-·1 mm.
long. They are generally blackish or dark-brown, and the specific
gravity ranges usually from 2·72 to 2·83. They display in the slide a
fair number of plagioclase and pyroxene phenocrysts in a groundmass of
felspar-lathes, pyroxene granules, and magnetite, the interstitial glass
being scanty or moderate in amount.... The plagioclase phenocrysts are
rarely as large as 3 mm., so that most of the rocks belong to the
non-porphyritic group of the genus. These phenocrysts, which are often
zoned, give extinctions of basic andesine (15°-25°). They contain
magma-inclusions, sometimes in abundance, which are arranged in
zone-lines.... The pyroxene phenocrysts are small, the two kinds being
always represented in the same slide. In some cases separate crystals
occur, and in others the two are associated as intergrowths, but in most
cases separate and compound crystals occur in the same section. Not
infrequently the phenocryst is an aggregate of lesser crystals of the
two pyroxenes. The monoclinic form is a brownish yellow augite with
large extinctions and often twinned. The felspar-lathes of the
groundmass, which usually average ·05 or ·06 mm. in length, are either
narrow and untwinned, or they may be stouter and display simple and at
times lamellar twinning, giving extinctions of medium andesine.... The
granules of pyroxene are generally ·01-·02 mm. in size; but occasional
prism-forms occur which give sometimes the straight extinction of
rhombic pyroxene and at other times the large oblique extinctions of the
augite type.


      5. SUB-ORDER, PRISMATIC, OF THE HYPERSTHENE-AUGITE-ANDESITES

                 (_Felspar-lathes in flow-arrangement_)

FORMULA._—Hypersth-aug, matr, flu, prism._

 17 genus     (_Vitr._) }
 18   "       (_Opac._) } See Synopsis.
 19   "       (_Magn._) }
 20   "       (_Parv._) }

The rocks of this sub-order that are represented in my collection admit
easily of a general description, and since the diagnoses of the genera
are given in the Synopsis, there will be no need to separately describe
each genus.

Almost without exception these rocks form a constituent of agglomerates
in various parts of the island; and they occur in this condition in some
of the highest mountains, as Mariko, Thambeyu, and Koro-mbasanga. The
exception refers to a low mound-like hill, apparently a “volcanic neck,”
that rises from the basaltic plains west of Mbua (see page 58).

In about half of the specimens the rocks are referred to section 10,
where the plagioclase phenocrysts are either small and very scanty or
are absent altogether. In a fair number these phenocrysts are opaque
(genus 18); whilst in a few they are glassy (genus 17). The rocks are
typically blackish or dark grey, and often have a pitchstone-like
appearance, the groundmass being frequently semi-vitreous in character.
Vesicular and scoriaceous rocks occur at times.

In all cases the felspar-lathes and pyroxene prisms are more or less in
flow-arrangement; whilst pyroxene granules, if present, are subordinate.
The felspar-lathes, which are either simple or once-twinned and give
extinctions of acid and medium andesine, are usually small, and average
in different sections ·05-·08 mm. in length. The pyroxene prisms are
pale brown and are ·03 or ·04 mm. long. Most of them give oblique
extinctions of over 25; but in the same slide some give straight
extinctions; the proportion varies in different rocks. The pyroxene
phenocrysts in all the specimens are small (not over 2 mm.), and are
rhombic and monoclinic. In most sections the two forms are represented
by separate crystals and are also associated in the same phenocryst.
Those of rhombic pyroxene have often dark borders. There is a
considerable amount of a pale brown glass in the groundmass, more or
less devitrified. The specific gravity varies considerably, but is as a
rule between 2·55 and 2·75, the more basic rocks containing augite in
preponderance and basic andesine, whilst the less basic possess a large
proportion of rhombic pyroxene and display oligoclase-andesine.
Sometimes, as in the case of a rock composing an agglomerate east of
Nanduri, where the porphyritic plagioclase is opaque and there is some
degree of alteration, the rock looks very like a porphyrite.


     THIRD ORDER, ORTHOPHYRIC, OF THE HYPERSTHENE-AUGITE-ANDESITES

FORMULA.—_Hypersth-aug, matr, orth._

CHARACTERS.—Felspars of the groundmass short and broad.

Since the material is insufficient for the separate description of each
genus, a general account of the order is alone given. These rocks are
often represented in agglomerates or they occur as large blocks, either
lying on the surface or imbedded in tuffs. Many of them are somewhat
altered.

They are for the most part dark grey dull-looking rocks, with a specific
gravity of 2·7 to 2·8, showing macroscopic plagioclase together with
conspicuous pyroxene phenocrysts. The plagioclase phenocrysts are
usually small (1 to 3 mm.), and give extinctions of medium andesine
(15°-20°) and in some rocks of acid labradorite (30°). They are as a
rule corroded and are penetrated by numerous fissures, whilst they
contain a considerable amount of altered magma-inclusions with sometimes
other alteration products. The pyroxene phenocrysts are from 2 to 4 mm.
in size. Brownish-yellow augite, giving extinctions of over 30°, and
pale-yellow rhombic pyroxene of the type before described occur
generally in the same slide, and are frequently associated as
intergrowths in the same crystal. They may have regular outlines or dark
eroded borders, and at times they exhibit abundant dark opaque
inclusions. The broad felspars of the groundmass are sometimes
rectangular and give lamellar extinctions of medium and acid andesine
(12°-17°). They vary in length in different rocks from ·05 to ·2 mm. and
more. The pyroxene of the groundmass is generally granular and coarse
(·02-·05 mm.). As indicated by the extinctions of occasional prism-forms
it is composed of both augite and rhombic pyroxene, the former
prevailing. The prismatic sub-order is also represented, and here the
pyroxene of the groundmass is in great part prismatic, the length of the
prisms not often exceeding ·05 mm., whilst both the monoclinic and
rhombic kinds occur. Interstitial glass varies in amount, sometimes
absent, sometimes scanty and viriditic, at other times abundant and
opaque. Magnetite abounds in the groundmass much of it often being of
secondary origin.


      FOURTH ORDER, FELSITIC, OF THE HYPERSTHENE-AUGITE ANDESITES

FORMULA.—_Hypersth-aug, matr, fels._

CHARACTERS.—The groundmass presents a rudely granular appearance or a
blurred mosaic.

This order is capable of subdivision, as in the other orders of the
hypersthene-augite-andesites (see Synopsis, page 247); but since it is
only represented by six of my rock-sections, I will confine the
description to the general characters.

These rocks are dark-grey, sometimes granitoid in appearance, with
specific gravity 2·65 to 2·75. They usually show some alteration,
arising from secondary changes within the rock-mass; and probably the
felsitic or semi-mosaic appearance of the scanty groundmass is the
result of such a secondary change. Such rocks in some respects approach
the type of the gabbros. They are frequent on the north coast of Natewa
Bay in the vicinity of Waimotu and also occur in the Valanga Ridge. They
generally present themselves as deeper-seated massive rocks exposed by
the stripping off of the superficial deposits.

There are as a rule more or less conspicuous phenocrysts, up to 3 mm. in
size, of plagioclase and pyroxene, in a relatively scanty micro-felsitic
groundmass, displaying a blurred mosaic, in which a few stout
felspar-lathes can still be recognised, and composed apparently of
felspar and crystalline silica. The “grain” of the mosaic may range in
different rocks from ·005 mm. to ·02 mm. The pyroxene of the groundmass
is largely decomposed, and the scanty residual glass is represented by
viriditic materials.

The plagioclase phenocrysts, which give extinctions of medium and basic
andesine (15°-25°), are often semi-opaque and corroded. They are
traversed by numerous cracks and often contain many whitish alteration
products, though the lamellar structure is usually well preserved. The
pyroxene phenocrysts are composed of brownish-yellow augite (ext. + 30°)
and pale rhombic pyroxene of the type described before, either as
separate crystals or associated as intergrowths. The rhombic pyroxene
crystals are often sub-rounded with dark borders; and as a rule the
pyroxene phenocrysts are much fresher than the plagioclase phenocrysts.
As far as can be ascertained, most of the original pyroxene of the
groundmass was monoclinic with a little rhombic.



                              CHAPTER XXI

             THE VOLCANIC ROCKS OF VANUA LEVU (_continued_)


                             ACID ANDESITES

      _Previous observations on the Hornblende-Andesites of Fiji_

THESE rocks were first described by Wichmann[112] from specimens
obtained by Kleinschmidt from the mountain of Mbuke Levu in Kandavu.
These Kandavu rocks had a microfelsitic base, the porphyritic brown
hornblende having usually black borders in which a change into epidote
was observed. Rhombic pyroxene was only noted as an occasional
constituent of a rock from Ono. Renard[113] described these rocks from
the vicinity of Ngaloa Harbour in Kandavu and remarked that bronzite was
of more common occurrence than the monoclinic pyroxene. In the
groundmass were numerous felspar and augite microliths, whilst there was
a porphyritic development of plagioclase, hornblende, biotite, and
pyroxene. The hornblende phenocrysts played an important part in the
rock-composition, being surrounded by a black zone of magnetite or
bordered by a bacillary aggregate of small pyroxene prisms, parallel and
colourless or greenish, with extinction-angles of 40°. There was often
also a development of biotite in the heart of the mineral, the whole
hornblende section being sometimes thus transformed.

Mr. Eakle[114] has more recently described the hornblende-andesites from
Mbuke Levu in Kandavu. As the result of his examination of a collection
of volcanic rocks made by the Agassiz expedition in Viti Levu, Kandavu,
Mbenga, Totoya, Malolo, Yasawa group, and in several other small
islands, he inferred that the hornblende-andesites are much more limited
in their occurrence in Fiji than the augite-andesites; whilst
hypersthene-andesite was only represented in the collection from
Vomo-lailai near Waia in the Yasawas. The specimens from Waia had a
microfelsitic base with pseudomorphs of hornblende and some augite. Mr.
E. C. Andrews[115] in his account of his collection of volcanic rocks,
made mostly in the Lau Group and Taviuni, makes no special reference to
hornblende-andesites, the andesites being mainly augitic, rhombic
pyroxene also occurring as a common porphyritic constituent.

It may be inferred from the above and from my own observations in Vanua
Levu below given that hornblende-andesites have a relatively limited
distribution in Fiji. They are not generally distributed as in the case
of the augitic and basaltic andesites; but are confined to certain
localities in Viti Levu,[116] Vanua Levu, Ovalau, Kandavu, Ono, Malolo,
Yasawa Islands, etc.

_The occurrence of quartz-andesites or dacites in Fiji._—In connection
with the existence of these rocks in Vanua Levu, it is noteworthy that
except in Mr. Eakle’s paper there is no reference in any of these
writings to the occurrence of quartz-andesites in Fiji. Wichmann
expressly states that the rocks he examined were free from quartz, and
that up to his time (1882) no quartz-bearing younger eruptive rocks were
known from the South Seas. Mr. Eakle in 1899 described a
holo-crystalline andesite with a felsitic aspect from Malolo and another
similar looking rock from Vatu Mbulo, in the same sub-group of the
Fijian Islands, showing quartz both in the phenocrysts and in the
microcrystalline groundmass, concerning which he observed that it was
perhaps more of a dacite than an andesite. Dacites were found by me in
1884 in the island of Fauro in the Solomon group,[117] and it is
probable that they are of more frequent occurrence in the Pacific than
has been generally supposed. As shown immediately below, they are
represented in Vanua Levu; and the extent of their distribution in the
island depends on the limits we assign to the definition of the term
“dacite.”

If we restrict the term to a hornblende-andesite carrying porphyritic
quartz and displaying a microfelsitic groundmass, such rocks, though
they form some of the highest peaks in the Ndrandramea district, namely
Ngaingai and Wawa-levu, would not be very frequent in Vanua Levu. If,
however, a microfelsitic groundmass is alone necessary to constitute a
“dacite,” the great majority of the acid andesites of the island would
fall under this designation.[118] This has long been a controverted
point in petrology. If I adopted the last procedure, my general
classification for the andesites would fall into confusion and many
rocks without any quartz would be included in the dacites.

In the Synopsis it will be seen that my classification of the
andesites is as far as concerns the great groups based on
the mineral and not on the structural characters. There
are three sub-classes closely allied to each other, the
hypersthene-andesites, the hornblende-hypersthene-andesites, and the
quartz-hornblende-hypersthene-andesites or dacites, which cannot be
distinguished at their boundaries by their petrological characters or by
their different modes of occurrence. These groups of rocks which include
all the acid andesites of the island will now be dealt with.


                    THE ACID-ANDESITES OF VANUA LEVU

                (_Comprising the hypersthene-andesites;
                   hornblende-hypersthene-andesites;
        and quartz-hornblende-hypersthene-andesites or dacites_)

These rocks compose in mass numerous isolated hills that rise up
abruptly in the interior of the central portion of the island. Such
hills, or mountains, as they might be often termed, usually attain a
height of from 700 to 1200 feet above the surrounding country, and
possess precipitous slopes and frequently perpendicular cliff-faces. In
the geological description of the island, I have referred in detail to
these mountains, when speaking of Na Raro, Vatu Kaisia, Ndrandramea,
Ngaingai, etc.; and illustrations of some of them are included in this
work. It may, however, be here remarked that they are as a rule rudely
conical with rounded or peaked summits. The ground-plan is generally
elliptical in outline; and in consequence the profile often varies from
different points of view, so that as in the case of Na Raro, it is that
of a sharp conical peak when the mountain is viewed “end-on,” or of a
broad truncated mass when seen from the side. A similar change of form
is to be noticed in the illustrations of Ndrandramea. No traces of
crateral cavities came under my notice. The rocks are neither vesicular
nor scoriaceous, and are usually massive; but exhibit at times a rudely
columnar structure.

Each hill or mountain has its peculiar variety of these rocks. This is
well shown in the Ndrandramea district. Thus the rocks of Ngaingai and
of Wawa-levu in carrying porphyritic quartz differ from those of all the
other hills around. Those of Soloa Levu are distinguished by the
orthophyric groundmass and by the absence of hornblende. Those of Mount
Ndrandramea again have no porphyritic quartz, but little hornblende, and
possess a micro-felsitic groundmass. The rocks of Na Raro and Vatu
Kaisia differ as regards specific gravity, the “grain” of the felsitic
groundmass, the presence of phenocrysts of rhombic pyroxene, etc. The
characters of these rocks from various localities are contrasted in the
table given on a later page, whilst the different sub-classes to which
they belong are described in detail below.


                    SUB-CLASS HYPERSTHENE-ANDESITES

These are dark and light grey rocks, sometimes granitoid in appearance.
They pass on the one hand into the hypersthene-augite-andesites before
described and on the other into the hornblende hypersthene-andesites to
be subsequently dealt with. From the former they are distinguished by
the great predominance of rhombic pyroxene both as phenocrysts and in
the groundmass; whilst from the latter they are separated by the absence
of brown hornblende or its pseudomorphs. These rocks are found in the
Ndrandramea, Valanga, and Vunimbua districts. They may form isolated
dome-shaped hills as in that of Soloa Levu, or they may constitute the
deeper-seated rocks of the region from which these hills arise,
as in the Ndrandramea district. In their general mode of
occurrence, however, they cannot be treated apart from the allied
hornblende-hypersthene-andesites and the dacites.

This sub-class may be divided like the hypersthene-augite-andesites into
four orders according to the character of the groundmass; and these are
enumerated in the Synopsis. Only the orthophyric and felsitic orders are
represented in my collection. Of the former the most typical rocks are
those composing the hill of Soloa Levu which is described on page 103.

These Soloa Levu rocks are lightish grey and granitoid in aspect, with
specific gravity of 2·54-2·62, and displaying abundant porphyritic
crystals of pyroxene, 2-3 mm. in size. In the slides they show a large
number of plagioclase phenocrysts together with those of pyroxene in a
relatively scanty groundmass, for the most part orthophyric in texture
and without residual glass.... The plagioclase phenocrysts, which are
not usually over 2 or 3 mm. in size, are often tabular and show distinct
zone-lines. Though they are traversed by minute cracks and have
frequently a semi-saussuritic appearance arising partly from
change-products and partly from the abundance of colourless inclusions,
they yield clear lamellar extinctions of medium and basic andesine
(15°-25°).... The pyroxene phenocrysts, which are not much altered, are
in most cases long pale-yellow rhombic prisms with rounded ends,
behaving optically as described on page 285; but intergrowths with
monoclinic pyroxene may occur and even separate crystals of augite....
The scanty groundmass, though in the main formed of short and broad
felspars, ·12 mm. long, of the orthophyric type, displays in places a
rude mosaic, apparently of quartz and felspar. It also shows abundant
small pyroxenes in the form of small prisms (·05 mm. in length), giving
extinctions nearly always straight but occasionally oblique (30°-35°).

As examples of the felsitic order of these rocks, most of which are
altered like the propylites, I will first take the case of those
deep-seated rocks that are exposed in the river-bed above Nambuna in the
Ndrandramea district. In the least altered state they are dark grey and
mottled, and have a specific gravity of 2·66-2·69. In section they
display tabular zoned plagioclase phenocrysts, usually more or less
occupied by alteration products, but at times giving lamellar
extinctions of basic andesine (20°-25°). The rhombic pyroxene is more or
less replaced by chloritoid pseudomorphs; whilst the “grain” of the
mosaic is often coarse (·03 mm.), and much of it is evidently quartz.
The more advanced stages of alteration of these rocks are described in
the account of the district given on page 106.... Similar rocks, showing
pyrites, occur amongst the blocks of Vunimbua River; but here the
rhombic pyroxene is mostly converted into bastite, and the groundmass is
in part trachytic as well as felsitic in texture. The specific gravity
is 2·7.

(In the last survey of my collection I have found a solitary specimen
from an agglomerate in the Mbua-Lekutu “divide,” which must be referred
to the order with felspar-lathes in flow-arrangement. It is a pale grey
rock showing abundant macroscopic pyroxene prisms, 2 mm. long, mostly
rhombic but showing also intergrowths with monoclinic pyroxene. The
felspar-lathes do not average more than ·1 mm., and there is a quantity
of small prisms of rhombic pyroxene in the groundmass, which also
contains a little residual glass.)


               SUB-CLASS HORNBLENDE-HYPERSTHENE-ANDESITES

This is an extensive group which includes the rocks forming several of
the hills in the Ndrandramea district as well as the isolated peaks of
Na Raro, Vatu Kaisia, etc. It passes on the one side into the
Hypersthene-Andesites before described and on the other into the
Hornblende-Hypersthene-Quartz-Andesites, the Dacites of this island.

Of the four orders established in the Synopsis (page 236) according to
the general method there adopted, the first, where the groundmass
exhibits felspar-lathes not in flow-arrangement, is not represented in
my collection.


          SECOND ORDER OF THE HORNBLENDE-HYPERSTHENE-ANDESITES

                 (_Felspar-lathes in flow-arrangement_)

This order is only represented by three rocks, all of which belong to
the prismatic sub-order where the pyroxene of the groundmass is
prismatic and not granular.

Two of these rocks are very similar in appearance and character, though
coming from different localities on the opposite sides of Savu-savu Bay,
one from the agglomerate of Vatu-ndamu in the Kumbulau peninsula (page
91), the other from an intrusive mass in the vicinity of Urata (page
184). They are dark grey, with specific gravity 2·6 to 2·7, and display
macroscopic crystals of hornblende and pyroxene. In the slide they
exhibit in addition numerous phenocrysts of plagioclase, 1 to 2 mm. in
size, in a groundmass showing small felspar-lathes (less than ·1 mm. in
length) in partial flow-arrangement and numerous pyroxene prisms
(·05 mm. long) giving straight extinctions, together with a little
residual glass.... The plagioclase phenocrysts, which give extinctions
of medium and basic andesine, are often tabular and display zone-lines.
They contain abundant pale inclusions arranged zone-wise.... The
hornblende phenocrysts are dark brown, markedly pleochroic, and give
extinctions up to 12 degrees. They have dark resorption borders and are
sometimes deeply corroded. They show in various stages the remarkable
conversion at the borders into fine pyroxene, which is described on page
306.... The pyroxene phenocrysts are more numerous in the Urata rock.
They are for the most part of the pale yellow feebly pleochroic rhombic
type that prevails in the island. A few phenocrysts of pale augite (ext.
35°) may occur in the same slide; whilst the two pyroxenes may be
associated as intergrowths.

A crypto-crystalline variety of these rocks, where the felspar-lathes
and rhombic pyroxene prisms of the groundmass are only in part
differentiated, is found on the hills of Ndreke-ni-wai on the shores of
Natewa Bay (page 201). It is a pale-grey open-textured rock, displaying
numerous small macroscopic crystals of hornblende.


          THIRD ORDER OF THE HORNBLENDE-HYPERSTHENE-ANDESITES

 (_Felspars of the groundmass, short and broad, of the orthophyric type_)

These rocks occur generally as agglomerates and are more particularly
characteristic of the district between the Mariko Range and the Salt
Lake. They belong for the most part to the prismatic sub-order of the
group and to the section with plagioclase phenocrysts, and fall
naturally into two divisions corresponding to the two genera with
_glassy_ and _opaque_ phenocrysts. The last named would be regarded by
some as porphyrites. The specific gravity of the specimens ranges from
2·52 to 2·7.

The plagioclase phenocrysts, 1 to 2 mm. in size, give extinctions
indicating in some rocks oligoclase-andesine (10°-15°) and in others
basic andesine (15°-25°). Their opacity in the porphyrites is sometimes
due to multiple macling, but more usually it arises from the numerous
fine cracks filled with decomposition products that traverse them. The
phenocrysts of dark brown hornblende are generally abundant and give
extinctions of 15 degrees. They as a rule have dark resorption borders
in which the process of conversion into fine pyroxene is in active
operation. The pyroxene phenocrysts are scanty and in most cases
rhombic; but intergrowths with augite and separate crystals of the
last-named may occur. In the altered rocks or porphyrites they are
largely replaced by bastite and viridite. The felspars of the groundmass
are broad and often rectangular and may give lamellar extinctions of
oligoclase-andesine. The pyroxene in the groundmass of the porphyrites
is often partly decomposed. It is as a rule prismatic. A little
interstitial glass, altered in the porphyrites, is generally present.


          FOURTH ORDER OF THE HORNBLENDE-HYPERSTHENE-ANDESITES

    (_Groundmass felsitic, displaying a granular mosaic structure_)

These are light and dark grey rocks showing usually macroscopic pyroxene
and hornblende. They vary considerably in appearance from the
open-textured rock to that with a granitoid coarsely crystalline aspect.
They generally carry brown hornblende phenocrysts, but frequently these
are represented by pseudomorphs; and they all have a felsitic
groundmass. They are only separated by the absence of porphyritic quartz
from the dacites of Vanua Levu, which are treated in the next sub-class.
They present all stages from the crypto-crystalline to the
holo-crystalline condition, but all show a groundmass which may be
scanty in the more coarsely crystalline rocks.

These rocks are characteristic of some of the hills of the Ndrandramea
district and of the isolated peaks of Vatu Kaisia and Na Raro. They
include a large proportion of the acid andesites of the island, and all
belong to the prismatic sub-order with prismatic pyroxene in the
groundmass, and to the section with plagioclase phenocrysts. Their
specific gravity ranges usually from 2·55 in the more acid and less
crystalline types to 2·74 in the most crystalline and basic kinds.

In the typical slides they display phenocrysts of plagioclase, pyroxene,
and brown hornblende in a microfelsitic groundmass formed evidently of
felspar and quartz together with much prismatic pyroxene. They may be
conveniently divided into three species according to the size of the
“grain” of the groundmass.

Of the first species, where the “grain” is less than ·01 mm., the rocks
of Mount Ndrandramea are typical. They have a crypto-crystalline
groundmass where the felsitic structure is in process of development and
where the pyroxene prisms or microliths are very minute. The plagioclase
phenocrysts (1 mm. in size) give extinctions of acid and medium andesine
(10°-20°), and are tabular, zoned, and contain abundant pale inclusions.
The hornblende phenocrysts, except in the case of the rocks at the foot
of the hill, are represented by pseudomorphs in various stages of
dispersion, so that this character is likely to be overlooked. The
pyroxene phenocrysts are in most cases of the pale yellow feebly
pleochroic rhombic type, but they may present intergrowths of both the
monoclinic and rhombic forms. These are light grey rocks with a specific
gravity increasing as one descends from the summit, where it is about
2·5, to the base where it is 2·7, a change corresponding with increase
of the ferro-magnesian minerals and with the more crystalline structure
of the groundmass.

The second species, where the “grain” of the mosaic is between ·01 and
·02 mm., is represented by the Vatu Kaisia rock and by that exposed in
the opposite side of the gorge. They are granitoid in appearance and
have a specific gravity of 2·68 to 2·71. Large porphyritic crystals of
pyroxene and hornblende, the last sometimes 7 mm. in length, occur in a
dark grey base. In the slide these phenocrysts together with those of
plagioclase are displayed in a somewhat scanty holo-crystalline
groundmass, where the “grain” of the mosaic averages ·012 mm. The
plagioclase phenocrysts are zoned, and give in different crystals
extinctions in some cases of oligoclase-andesine (10°-12°) and in others
of andesine-labradorite (25°-30°). The hornblende phenocrysts in their
pseudomorphism illustrate the various stages of the process of
conversion into fine granular and prismatic pyroxene. The least altered
crystals have dark resorption borders and are at times deeply corroded.
The pyroxene phenocrysts are for the most part rhombic; but intergrowths
with the monoclinic form occur. The pyroxene of the groundmass consists
for the most part of small rhombic prisms averaging ·05 mm. in length.

The third species, where the groundmass may be described as a coarse
mosaic with a “grain” between ·02 and ·03 mm., is represented by the
rocks of the peak of Na Raro and of Mount Thokasinga in the Ndrandramea
district.

The Na Raro rocks are light grey with a specific gravity in the
unweathered state of about 2·6, and display macroscopic crystals of
glassy plagioclase and hornblende. In the slide they exhibit tabular
phenocrysts of the plagioclase, together with dark pseudomorphs after
hornblende, in a coarsely felsitic groundmass (grain ·022 mm.) where a
little very fine prismatic pyroxene, apparently rhombic, occurs. There
is also a little altered interstitial glass. The plagioclase phenocrysts
are zoned and give extinctions of 15° to 25° (acid and basic andesine),
whilst they often show magma inclusions. An interesting feature of these
rocks is concerned with the rarity or absence of phenocrysts of
pyroxene. They are to be seen, however, in the process of being built up
within the substance of the hornblende pseudomorphs, which consist
entirely of minute prisms and granules of pyroxene and of fine
magnetite. The process seems to consist in the formation of parallel
layers of rhombic and monoclinic pyroxene.

The Thoka-singa rocks are more basic (spec. grav. 2·72 to 2·74), and in
the scanty holo-crystalline groundmass approach the plutonic type. They
are dark grey granitoid rocks displaying abundant macroscopic pyroxene
crystals 2 to 3 mm. long. The original hornblende phenocrysts are only
represented by traces of pseudomorphs of fine pyroxene and magnetite,
the process of dispersion, described on page 307, being almost
completed. The pyroxene phenocrysts are mostly rhombic; but intergrowths
with the monoclinic form occur. The “grain” of the mosaic of the
groundmass is coarse (·023 mm.), and there is a fair amount of prismatic
with a little granular pyroxene, the prisms, ·04 mm. long, giving
usually straight extinctions, whilst the granules are apparently
monoclinic.


      SUB-CLASS QUARTZ-HORNBLENDE-HYPERSTHENE-ANDESITES OR DACITES

These rocks are infrequent. They compose in mass the adjacent mountains
of Ngaingai and Wawa Levu in the Ndrandramea district, and appear also
on the lower slopes of the neighbouring mountain of Navuningumu. They
differ chiefly from the hornblende-hypersthene-andesites in the presence
of porphyritic quartz, which, however, is not as a rule abundant. In
their general origin and affinities and in their mode of occurrence they
cannot be separated from the two sub-classes of hypersthene-andesites
and hornblende-hypersthene-andesites before described. They all belong
to the felsitic order of the sub-class, and all are referred to the
sub-order with prismatic pyroxene and to the section with plagioclase
phenocrysts.

They are light grey rocks, with a specific gravity of 2·57 to 2·61,
showing usually dark pseudomorphs after hornblende and a little
porphyritic quartz. In the slide they display these pseudomorphs and
quartz crystals, associated with abundant plagioclase phenocrysts, in a
felsitic groundmass, evidently a mixture of felspar and quartz, with
fine pyroxene, mostly prismatic and rhombic. Pyroxene phenocrysts are
absent or rare; but they may be seen in process of formation in the
substance of the hornblende-pseudomorphs. It is only at times, as in the
instance of the Navuningumu rock, that the brown hornblende phenocrysts
are in part unchanged and that complete pyroxene phenocrysts occur. In
such cases the last may be entirely rhombic, or may exhibit at times
intergrowths with the monoclinic form.

The plagioclase phenocrysts, 2 mm. in size, are often tabular and zoned
and give two sets of extinctions, indicating acid and basic andesine.
Their abundant inclusions are arranged in zones. The quartz-crystals, 1
to 2 mm. in size, present hexagonal sections with rounded angles. They
are sometimes traversed by cracks occupied by iron oxide films. The
pseudomorphs after hornblende, which consist of fine pyroxene mixed with
magnetite, exhibit often the building up in their interior of pyroxene
phenocrysts, apparently rhombic, by long parallel rows of stout prisms.
In other cases the pseudomorphs display the different stages of
dispersion. The fine pyroxene of the groundmass consists mostly of
rhombic prisms (·02-·06 mm. long) with some granules. The “grain” of the
groundmass is usually between ·01 and ·02 mm. There is little or no
residual glass.

The rocks of Wawa Levu and Ngaingai are closely similar, but they differ
in the size of the prismatic pyroxene of the groundmass, which is
coarser in the first-named mountain (·055 mm. long) than it is in the
second (·025 mm. long). In both the “grain” of the mosaic is about the
same (·014 mm.).


        TABULAR COMPARISON OF THE ACID ANDESITES OF VANUA LEVU.

NOTE.—The figures in the columns headed “Felsitic” and “Rhombic” refer
to the size in millimetres of the felsitic “grain” and the larger
pyroxene prisms.

 Column headings:

 P: Plagioclase.
 Q: Quartz.
 H: Hornblende or its pseudomorph.
 R: Rhombic Pyroxene.
 M: Monoclinic Pyroxene.
 O: Orthophyric.
 Rh: Rhombic.
 Mo: Monoclinic.

 +------------+----+---------------------+---------------------------+
 |            |    |     Phenocrysts.    |        Groundmass.        |
 |            |    +-+-+--------+-+------+-------------+-------------+
 | Locality.  | Sp.| | |        | |      |  Structure. |  Pyroxene.  |
 |            | Gr.|P|Q|   H    |R|  M   +-----------+-+------+------+
 |            |    | | |        | |      | Felsitic. |O|  Rh  |  Mo  |
 +------------+----+-+-+--------+-+------+-----------+-+------+------+
 |Ngaingai,   |    | | |        | |      |           | |      |      |
 | summit     |2·57|+|+|   +    | |      |  + ·014   | | + ·03|      |
 |            |    | | |pseudom.| |      |           | |      |      |
 |            |    | | |        | |      |           | |      |      |
 |Ngaingai,   |    | | |        | |      |           | |      |      |
 | lower part |2·57|+|+|   +    | |      |  + ·013   | | + ·02|      |
 |            |    | | |pseudom.| |      |           | |      |      |
 |            |    | | |        | |      |           | |      |      |
 |Wawa-levu,  |    | | |        | |      |           | |      |      |
 | 1,500 feet |2·61|+|+|   +    | |      |  + ·015   | | + ·05|      |
 |            |    | | |pseudom.| |      |           | |      |      |
 |            |    | | |        | |      |           | |      |      |
 |Wawa-levu,  |    | | |        | |      |           | |      |      |
 | 1,500 feet |2·61|+|+|   +    | |      |  + ·014   | | + ·06|      |
 |            |    | | |pseudom.| |      |           | |      |      |
 |            |    | | |        | |      |           | |      |      |
 |Wawa-levu,  |    | | |        | |      |           | |      |      |
 | west side  |2·57|+|+|   +    |+|      |  + ·022   | | + ·04|      |
 | of base    |    | | |pseudom.| |      |           | |      |      |
 | but not on |    | | |in great| |      |           | |      |      |
 | mountain   |    | | |  part  | |      |           | |      |      |
 |            |    | | |        | |      |           | |      |      |
 |Ndrandramea,|    | | |        | |      |           | |      |      |
 | summit     |2·44|+| |   +    |+|  +   |  + ·007   | | + ·01|      |
 |            |    | | |  only  | |scanty|   hemi-   | |      |      |
 |            |    | | | traces | |      |crystalline| |      |      |
 |            |    | | |   of   | |      |           | |      |      |
 |            |    | | |pseudom.| |      |           | |      |      |
 |            |    | | |        | |      |           | |      |      |
 |Ndrandramea,|    | | |        | |      |           | |      |      |
 | 1,600 feet |2·58|+| |   +    |+|  +   |  + ·008   | | + ·01|      |
 |            |    | | |  only  | |scanty|   hemi-   | |      |      |
 |            |    | | | traces | |      |crystalline| |      |      |
 |            |    | | |   of   | |      |           | |      |      |
 |            |    | | |pseudom.| |      |           | |      |      |
 |            |    | | |        | |      |           | |      |      |
 |Ndrandramea,|    | | |        | |      |           | |      |      |
 | 1,200 feet |2·68|+| |   ?    |+|  ?   |  + ·008   | | + ·01|      |
 |            |    | | |        | |      |   hemi-   | |      |      |
 |            |    | | |        | |      |crystalline| |      |      |
 |            |    | | |        | |      |           | |      |      |
 |Ndrandramea,|    | | |        | |      |           | |      |      |
 | saddle at  |2·71|+| |   +    |+|  ?   |  + ·009   | | + ·01|      |
 | base       |    | | |        | |      |           | |      |      |
 |            |    | | |        | |      |           | |      |      |
 |Kalakala,   |2·61|+| |   +    |+|      |  + ·02    | | + ·01|      |
 | near base  |    | | |pseudom.| |      |           | |      |      |
 |            |    | | |in part | |      |           | |      |      |
 |            |    | | |        | |      |           | |      |      |
 |Navuningumu{|2·65|+| |   +    |+|  +   |  + ·02    | | + ·03|  ?   |
 | near base {|    | | |in part | |scanty|           | |      |      |
 |           {|    | | |pseudom.| |      |           | |      |      |
 |           {|2·48|+|+|   +    |+|      |  + ·018   | |  ?   |      |
 |            |    | | |        | |      |           | |      |      |
 |Thokasinga  |2·74|+| |   +    |+|  +   |  + ·024   | | + ·03|  +   |
 |            |    | | |pseudom.| |      |           | |      |      |
 |            |    | | |        | |      |           | |      |      |
 |Thokasinga  |2·72|+| |        |+|  +   |  + ·023   | | + ·04|  +   |
 |            |    | | |        | |      |           | |      |      |
 |Na Raro,    |2·58|+| |   +    | |      |  + ·02    | | + ·01|      |
 | upper part |    | | |pseudom.| |      |           | |scanty|      |
 |            |    | | |        | |      |           | |      |      |
 |Na Raro     |2·58|+| |   +    | |      |  + ·022   | | + ·01|      |
 |            |    | | |pseudom.| |      |           | |scanty|      |
 |            |    | | |        | |      |           | |      |      |
 |Vatu Kaisia |2·71|+| |   +    |+|      |  + ·013   | | + ·05|      |
 |            |    | | |in part | |      |           | |      |      |
 |            |    | | |pseudom.| |      |           | |      |      |
 |            |    | | |        | |      |           | |      |      |
 |Vatu Kaisia,|2·68|+| |   +    |+|      |  + ·01    | | + ·05|      |
 | opposite to|    | | |in part | |      |           | |      |      |
 |            |    | | |pseudom.| |      |           | |      |      |
 |            |    | | |        | |      |           | |      |      |
 |Soloa-levu, |2·54|+| |        |+|  +   |           |+| + ·06|      |
 | upper part |    | | |        | |      |           | |      |      |
 |            |    | | |        | |      |           | |      |      |
 |Soloa-levu, |2·57|+| |        |+|  +   |           |+| + ·05|      |
 | upper part |    | | |        | |      |           | |      |      |
 |            |    | | |        | |      |           | |      |      |
 |Soloa-levu, |2·62|+| |        |+|  +   |           |+| + ·06|  +   |
 | near base  |    | | |        | |      |           | |      |scanty|
 |            |    | | |        | |      |           | |      |      |
 |Soloa-levu, |2·61|+| |        |+|  +   |           |+| + ·06|  +   |
 | near base  |    | | |        | |      |           | |      |scanty|
 +------------+----+-+-+--------+-+------+-----------+-+------+------+

_Note on the Rhombic Pyroxene of the three foregoing sub-classes of the
Acid Andesites._—The term “hypersthene” has been here used as a
convenient expression equivalent to “rhombic pyroxene.” The mineral is
always a little pleochroic and is never colourless, and it is only in
very rare cases that the term “enstatite” could be used. As a matter of
fact there is practically only one form of rhombic pyroxene represented
in my collections whether in acid or basic andesites or in
hemi-crystalline and plutonic rocks. In the acid andesites it occurs not
only as phenocrysts but also as minute prisms forming a constituent of
the groundmass.

This mineral, when composing the phenocrysts, presents itself usually as
single untwinned prisms which exhibit the typical octagonal
cross-sections with much reduced prism-faces. The prismatic sections
give straight extinctions; whilst with the cross-sections we obtain
straight extinctions parallel with the pinakoid faces. The colour in
transmitted light is pale brownish yellow. The pleochroism, though
usually feeble, is quite distinct, the colour being pale yellow when the
prism lies parallel with the long axis of the lower nicol, and almost
white when it lies across. Not infrequently these phenocrysts behave
abnormally and give small oblique extinctions. This is often the case
when monoclinic pyroxene occurs in the same section. The association of
the two pyroxenes in one crystal can in some cases be clearly
recognised. At one time a plate of pyroxene exhibits itself as a coarse
aggregate of the two pyroxenes. At other times the two occur as parallel
intergrowths, as in the accompanying figure. But it is rarely that such
intergrowths are so typically displayed, the reason of which has been
supplied by Zirkel in his _Lehrbuch der Petrographie_; 2nd edit.: I.
271.

_Note on the “magmatic paramorphism”[119] of the hornblende
phenocrysts._—Reference has before been made in the general description
of these rocks to the dark alteration margins of the hornblende
phenocrysts. The dark borders display the “bacillary” structure noticed
by Renard in the case of some hornblende andesites from Kandavu in the
same group of islands,[120] being composed of minute granules and
parallel prisms of pyroxene and also of magnetite grains. With the
Kandavu rocks Renard observed that the tiny pyroxenes were colourless or
greenish and had an extinction angle of 40°. In the Vanua Levu rocks,
however, there is a mixture in these dark margins of both monoclinic and
rhombic pyroxene; and the process may be observed in all stages as it
advances into the interior of the crystal, until a dark pseudomorph or
paramorph of pyroxene and magnetite results. When the magnetite
prevails, the pseudomorph may ultimately form a black patch in which the
process is obscured. But when, as is generally the case, the pyroxenes
are more frequent, it occurs as a dark grey mass.

Finally follows the dispersion of the pseudomorph, which first becomes a
loosely arranged aggregate of the two pyroxenes and magnetite, and then
breaks down, and at length is only represented by small pale patches of
its original constituents. These patches are easily recognisable, and in
not a few rock-sections offer the only indication of the previously
existing hornblende phenocryst. There can be little doubt that this is
the source of much at least of the often abundant pyroxenes of the
groundmass, which are usually most frequent in the vicinity of the
patches.

In the earliest stage when the dark border alone exists, it is not easy
to distinguish the one pyroxene from the other, the granules and prisms
being colourless and very minute, less than ·01 in size. But in a far
advanced stage of the paramorphism the granules and prisms become
sometimes much larger, the first attaining a breadth of ·04 and ·05 mm.
and the last a length of ·15 mm. Finally the interior of the paramorph
is seen to be more or less completely composed of very pale brown augite
and pale yellow rhombic pyroxene in coarse grains and prisms, the first
distinguished by its oblique extinction of 30° to 35°, the last
recognised by its straight extinction and feeble though distinct
pleochroism.

Although as a rule the paramorph becomes dispersed and its pyroxene
constituents are added to the groundmass, it sometimes exhibits a change
of another character. In this case the outer portion is alone dispersed,
whilst the growth of a single large crystal of pyroxene proceeds within
the mass. In a later stage, when the dispersion of the outer part is
complete, we have a fresh-looking pyroxene phenocryst with unformed
edges, on the borders of which little granules and prisms of pyroxene
may be seen arranging themselves, as if the crystal-building was still
in progress, or rather as if it had been interrupted and left unfinished
by the too rapid dispersion of the outer portions of the paramorph.

It will be gathered from the above that the source of the pyroxene of
the groundmass is to be found in the magmatic paramorphism of the
porphyritic hornblende. The hornblende is dark-brown, markedly
pleochroic, and extinctions up to 15° are given in prismatic sections.
It is well known that the conversion of a hornblende crystal into an
aggregate of pyroxene prisms and magnetite was long since experimentally
effected by Doelter and Hussak by immersing the hornblende in molten
basalt, andesite, &c.[121] I would imagine that the transformation of
the hornblende and the dispersion of the paramorph occurred under two
conditions; in the first case whilst the “flow” was still in motion when
the resulting pyroxene would be mixed up in the magma; in the second
case after movement had ceased, but before consolidation of the
groundmass, when a paramorph or pseudomorph would be formed.


                          OLIGOCLASE-TRACHYTES

The term “trachyte” is here applied in a general sense to a group of
light-grey intrusive acid rocks, having a specific gravity when compact
of 2·4 to 2·45 and showing phenocrysts of glassy felspar, but not of
quartz. These rocks, which are especially characteristic of the
districts around Tawaki and Mount Thuku and of the Wainikoro sea-border,
are often open-textured and sometimes a little vesicular, whilst several
of them exhibit some degree of alteration in the groundmass. In all
cases they appear to be intrusions rather than surface-flows; and at
times they display a columnar structure.[122]

The difference between the oligoclase-trachytes in various localities
appears to be mainly concerned with the varying degrees of
crystallisation. There are two principal varieties. In the most
crystalline type there are small phenocrysts of glassy felspar and a few
of pale augite, the angle of extinction of the last being over 30
degrees. The felspar phenocrysts, which contain but few inclusions and
have sharp rectilinear outlines, in most cases show zoning and give
lamellar extinction of 5° to 12° indicating oligoclase; but some of them
have the tabular untwinned or simple twinned form of sanidine. The
groundmass is in the main composed of minute felspar-lathes, less than
·1 mm. in length, arranged in a dense plexus, and giving nearly straight
extinctions. But it also contains a number of scattered larger
felspar-lathes averaging ·2 mm. in length and giving extinctions of 5°
when simple, and of 8° to 10° when lamellar. There is also some small
prismatic augite in the groundmass but often decomposing. The original
interstitial glass is represented by numerous reddish-brown patches of
devitrified glass.

In the second type of these trachytes, the rock is more open in texture
and is at times vesicular, the specific gravity being usually less than
2·4. The general characters are much the same, but sanidine is better
represented among the phenocrysts, and the groundmass is more blurred;
but when the felspar-lathes are distinct they give an extinction either
nearly straight or from 4° to 8°, according as they are simple or
display lamellæ. The augite of the groundmass is scanty and more or less
decomposed; whilst the interstitial glass when unaltered is in fair
quantity and nearly isotropic.

The alteration observed in several of these oligoclase-trachytes is
restricted chiefly to the interstitial glass in which secondary quartz
and at times calcite and viridite are developed. Scarcely any of them
are quite free from these changes.[123]

The pitchstone or vitreous form of these trachytes is displayed in the
blocks of an agglomerate-tuff between Tawaki and Mount Thuku. It has a
specific gravity of 2·36, is dark-brown, and has a conchoidal fracture.
Phenocrysts of felspar, mostly oligoclase, with extinction-angles of 5°
to 11°, and often penetrated by the magma, are inclosed in a
semi-isotropic groundmass showing incipient development of felspar and
other darker microliths. There are also a few small phenocrysts of pale
augite.


                 QUARTZ PORPHYRIES AND RHYOLITIC ROCKS

Wichmann when he wrote in 1882 that no quartz-bearing younger eruptive
rocks had hitherto been observed either in Fiji or in the South Sea
Islands generally, had apparently overlooked Dana’s observations in the
Fijian group. The American geologist[124] refers to a rock found on the
north-east shores of Vanua Levu which exhibited in a greenish base
thickly disseminated crystals of quartz (bipyramidal dodecahedrons, 1/8
of an inch in diameter) and glassy felspar, together with a few sphene
crystals.

Quartz porphyries, akin to the rhyolites, are especially characteristic
of the north-east part of the island, to which in fact they are entirely
confined. They perhaps are best represented in the vicinity of Mount
Thuku and in the neighbourhood of the mouth of the Wainikoro River. None
of my specimens have the fresh appearance of the Lipari rhyolites and
all are more or less altered. Their specific gravity does not exceed
2·4, and they are for the most part intrusive in character.

A rock frequently exposed between Tawaki and Mount Thuku[125] contains
abundant phenocrysts of glassy felspar (oligoclase and sanidine) and
quartz in a greenish opaque groundmass having a blurred microfelsitic
structure. There appears to have been a secondary devitrification of the
groundmass since consolidation. The porphyritic quartz crystals are
rounded and about 2 mm. in diameter.

Another rock displayed in the coast-cliffs on the north side of Natewa
Bay, a mile east of Mount Thuku, has a somewhat banded appearance. It
shows crystals of quartz, more or less rounded and 3 to 4 mm. in
diameter, together with phenocrysts of glassy felspar (oligoclase with
lamellar extinction of 5° and sanidine). The groundmass displays traces
of spherulites and is in places semi-isotropic; but for the most part it
is microfelsitic.

The type of rock found in the Wainikoro district and in the adjacent
sea-border, where it may be observed forming dykes in the pumice-tuffs,
is light-grey and loose-textured with a specific gravity of 2·1. It
exhibits small phenocrysts of quartz and of glassy felspar (oligoclase
5° to 12°, and sanidine), with, in one locality only, a scanty amount of
dark green hornblende yielding extinctions up to 20°. The quartz
crystals, which are 1 to 2 mm. in size, are sometimes bipyramidal; but
are often rounded and have fused-like outer surfaces. The groundmass is
semi-isotropic with a blurred aspect, and shows traces of spherulites
and numerous crystallites, with occasional felspar-lathes giving a
nearly straight extinction.

An extensively altered quartz porphyry of a different type is associated
with other altered rocks at the base of Mount Nailotha. It has a
specific gravity of 2·54, and displays large opaque crystals of
plagioclase, 2 to 5 mm. in size with small quartz crystals 1 to 2 mm.
across, in a grey compact matrix. The first-named shows the felspar to
be mostly replaced by alteration products; but occasionally a lamellar
extinction of 6° or 7° can be observed. The quartz crystals are rounded
and penetrated by the magma, and contain numerous strings of
fluid-cavities. The groundmass was originally spherulitic; but this
structure is more or less disguised by the development of a mosaic of
chalcedonic quartz. It shows some micro-porphyritic patches of viridite
and calcite. A singular altered white rhyolitic rock is exposed on the
north coast of Natewa Bay between Natasa and Sangani, where it is
associated with altered tuffs. It is compact with a conchoidal fracture
and has a specific gravity of 2·48. The hand-specimen has a banded
appearance. Under the microscope it appears as a rhyolitic glass for the
most part devitrified and rendered opaque by the formation of secondary
silica. Much of it presents a microfelsitic structure, the bands
appearing as semi-opaque streaks.

Glassy forms of the quartz porphyries or intrusive rhyolitic rocks are
extensively represented in the pumice tuffs of the Undu Promontory and
of the coasts between the Langa-langa river and Lambasa. These tuffs
will be found described on page 336. Fragments of a grey rhyolitic glass
looking like perlite are inclosed in the pumice-tuffs near the mouth of
the Wainikoro River. Under the microscope it is displayed as a
colourless glass inclosing phenocrysts of sanidine, oligoclase (ext.
4°), and quartz, the last with rounded outlines and a fused-like outer
surface. The glass shows in places perlitic cracks; but it is mainly
characterised by a vacuolar structure, the minute cavities being
lengthened out in the direction of the flow and displaying eddy-currents
around the phenocrysts. The elongated steam-cavities sometimes contain
water, but are usually more or less filled with granular materials.



                              CHAPTER XXII

                Basic Glasses and Volcanic Agglomerates


                    BASIC PITCHSTONE AND BASIC GLASS

IT is not possible to draw a sharp distinction between the pitchstone
and the purely vitreous condition of these glasses. The following
remarks will therefore apply to both.

Regarded as components of the pitchstone-tuffs and palagonite-tuffs
these rocks have a very extensive distribution in the island; but in the
massive state they are hardly ever to be found, whilst in the form of
agglomerates they are only frequent in certain localities, as in the
cliffs of the Korotini Bluff, in the vicinity of Mbale-mbale, on the
slopes of Soloa Levu, and in the dividing ridge between the Mbua and
Lekutu plains. On rare occasions they are to be found in a rubbly
condition, as in the upper part of a basaltic flow described on p. 92,
or they may form veins in a more crystalline basaltic rock as at
Vatulele Bay. Their specific gravity ranges from 2·61 to 2·77, and they
fuse readily before the blow-pipe, the melting beginning in the ordinary
flame. Since they are not dissolved under any condition in HCl, they
would be referred to the old hyalomelane group of basic glasses.

One of the most interesting of these rocks occurs on the slopes of Soloa
Levu. As displayed on the south-west slope, it presents itself as a
brownish-black rock with a specific gravity of 2·61 and exhibiting large
porphyritic crystals (6 to 8 mm.) of plagioclase. It is generally
compact, but it is in places a little vesicular, the minute cavities
being often filled with a zeolite. The mode of occurrence of this
pitchstone-porphyry is described on p. 104. In the slide the plagioclase
phenocrysts give lamellar extinctions (21°-27°) of andesine labradorite,
and have regular outlines, with but few inclusions of the glassy magma.
There are also a few small phenocrysts of augite with dark rounded
borders and showing in some cases lamellar twinning. The groundmass is a
brown rather turbid glass in which dark points of devitrification occur.
It is traversed by cracks that also penetrate the felspar phenocrysts.
These cracks are filled with a feebly refractive material like
palagonite; and there are traces of the early stage of the palagonitic
change in one or two places. This is of importance, because on the
north-west side of the hill occurs the same rock, in which the basic
glass has been converted into a reddish-brown almost opaque palagonite;
but in this case the porphyritic crystals of plagioclase are more
affected by the magma, being rounded and extensively penetrated
schiller-fashion by this material; whilst the augite phenocrysts are
somewhat similarly affected. The altered glass is also vacuolar, the
cavities being filled with a zeolite. There is an indication of some
degree of crushing in the fracture of some of the felspar phenocrysts
_in situ_. There appears to be a connection, as shown on p. 342, between
the crushing of a basic glass and the formation of palagonite. It is
noteworthy that with this change the specific gravity drops from 2·61 in
the comparatively fresh rock to 2·14 in the palagonitised hydrated
condition.

As another example of these basic pitchstones I will take that forming
an agglomerate near Mbale-mbale. It has a specific gravity of 2·77 and
displays phenocrysts of plagioclase, olivine, and augite. The
first-named, which give the lamellar extinction of acid labradorite,
(22°-28°), are fresh-looking and only affected to a small extent by the
magma. Those of olivine and augite are in much the same condition. The
glass of the groundmass is rather turbid and displays numerous dark
patches of incipient crystallisation, which in some cases prove to be
composed of brush-like crystallites around a clear H-shaped nucleus, and
in other cases have a more prismatic form.

A vitreous rock having some of the characters of a variolite is found
near Narengali (see page 150). It, however, has the low specific gravity
of 2·43 and is not readily fusible with the blow-pipe. It displays an
imperfect spheroidal structure on a small scale, being made up of
nodules, the largest having the size of a filbert. In the slide it
appears as a grey glass made up of sheaf-like aggregates of fibre-like
crystallites, apparently of felspar, with minute skeleton prisms of
pyroxene in parallel arrangement, and is traversed by perlitic cracks.


                       THE VOLCANIC AGGLOMERATES

In this place my remarks will be chiefly confined to a summary of some
of the leading features of these formations. The agglomerates, which
pass by all gradations through the tuff-agglomerates into the submarine
tuffs, rank amongst the most prevalent and the most conspicuous of the
rocks exposed at the surface in this island. Their lithological
characters vary according to the type of the massive rocks of the
district. Thus in the Ndrandramea district the blocks are composed of
the prevailing acid andesites. In the Koro-mbasanga district they are
formed of hypersthene-augite-andesites. In the Korotini and Va-lili
ranges they are composed of olivine basalts and basaltic andesites. The
agglomerates derived from basaltic rocks and basic andesites are by far
the most frequent, and it is to them that the following general
observations apply.

The basic agglomerates and tuff-agglomerates are found almost everywhere
and at all elevations up to 2,500 feet above the sea and over. They
compose the inland cliffs and the long lines of precipitous declivities
that give character to the valleys and gorges of the mountainous
interior. The blocks are often scoriaceous and semi-vitreous, but the
characters of the rocks will be found described on page 316. They are
generally sub-angular and vary in size from a few inches to one or two
feet; and, though sometimes heaped together in confusion, they will
generally be found in the case of any extensive exposure to be rudely
sorted according to size, or to present a rude horizontal arrangement.

The matrix varies much in amount, being sometimes barely appreciable and
at other times so abundant that the deposit may be termed a
tuff-agglomerate. Typically it has the character of the palagonite-tuffs
of mixed composition described on page 326, being made up of fragments
of palagonitised vacuolar basic glass, portions of crystals of
plagioclase and augite, with the debris of the basic semi-vitreous and
hemi-crystalline rocks forming the blocks. When it is scanty it contains
neither carbonate of lime nor organic remains; but in the
tuff-agglomerates it may be calcareous and may inclose tests of
foraminifera and molluscan shells.

From the circumstance that the basic agglomerates overlie submarine
sedimentary tuffs and clays almost everywhere, their submarine origin
could alone be safely postulated. There are one or two localities that
throw especial light on the conditions under which these accumulations
occurred. They are dealt with at some length in the general description
of each district and only a brief reference can be made to some of their
indications here.

The testimony supplied by the interesting exposures on the slopes of
Mount Thambeyu (page 178) goes to show that after the deposition of the
foraminiferous tuffs and clays the stage of the agglomerates was ushered
in gradually. The tuffs increased in coarseness, and afterwards they
were covered up with an agglomerate formed of blocks at first only one
or two inches in size, but afterwards of larger dimensions.... Curious
evidence is afforded by the agglomerates of Mount Vungalei (page 213),
where two beds of palagonite-tuff, at elevations of 900 and 1,700 feet,
mark two pauses in the accumulation of the agglomerates. In each case
the pause was introduced by the gradual decrease of the agglomerates
which gave place by gradation to the tuffs. In each case also the pause
was followed by a sudden renewal of the deposition of agglomerates.

With reference to the maximum thickness of these deposits, it would
appear that on the slopes of the Korotini Range this amounts to some
hundreds of feet, if we also include the tuff-agglomerates. Their origin
is to be attributed partly to eruptions and partly to marine erosion.
The two agencies although often associated were in their turns
predominant in their different phases, and it is not too much to suppose
that the agglomerates without arrangement, with scanty matrix, and
composed of scoriaceous blocks, belong more to an eruptive period, and
that those with abundant tufaceous matrix and sorted blocks are mainly
the product of marine erosion. In either case the deposition was
submarine.

But the history of these agglomerates and of their associated
foraminiferous tuffs and clays must of necessity be a complicated one,
since they indicate a minimum emergence of 2,500 feet. Their
accumulation first began when a number of vents, in linear arrangement,
were striving to raise their heads above the surface of the sea. It was
continued after the waves had ultimately worn the volcanic islets down
to below the sea-level, and the shoals became covered over with
submarine deposits. Again and again no doubt this struggle between the
eruptive agencies and the waves was renewed, until at length the great
emergence began, and probably from that date the agency of marine
erosion was predominant.

When on the island of Stromboli I had presented for my observation at
least two modes of agglomerate-building under the sea. There was the
ordinary work of the marine erosion of the lava-cliffs, of which the
beach represents but a small part of the result; and there were the
dribbling eruptions of the crater, from which at intervals of only a few
minutes masses of semi-molten lava bounded down the steep slopes into
the sea.

_Note on the general characters of the rocks of the basic
agglomerates._—In appearance the basic rocks forming the blocks are
often very similar. They are usually compact blackish with a
semi-vitreous aspect and display some plagioclase phenocrysts. But to
enumerate the types to which they belong would be to go over much of the
ground traversed in the classification of the basic rocks, whether
olivine basalts, basaltic andesites, ordinary augite-andesites, or
hypersthene-augite-andesites. The groundmass as a rule contains much
smoky glass, but the hemi-crystalline portions of it vary considerably
in character. Whilst fine granular augite prevails, semi-ophitic coarser
augites are not uncommon, and prismatic pyroxene, sometimes of the
rhombic form, is represented in the groundmass of the rocks composing
the agglomerates of Mount Thambeyu and of the Sokena Cliffs. In some
localities, as on the south-west slopes of the Korotini Range, rocks of
the basic pitchstone kind are predominant.



                             CHAPTER XXIII

         CALCAREOUS FORMATIONS, VOLCANIC MUDS, PALAGONITE-TUFFS


THE classification that is adopted in my work on the geology of the
Solomon group with respect to the calcareous formations and volcanic
muds of those islands is only in part applicable to the calcareous rocks
and volcanic deposits of Vanua Levu. Deposits strictly comparable with
those of the Solomon Islands here exist, and have in some places an
extensive distribution; but many others cannot be referred to that
classification. In addition to the calcareous oozes and volcanic muds,
such as are now forming off these reef-bound coasts, the result partly
of marine erosion and partly of sub-aerial denudation, there are many
kinds of submarine deposits in Vanua Levu that have been largely formed
from the materials ejected by volcanic vents. Basic glasses, for
instance, often finely vesicular and usually converted into palagonite,
enter largely into the composition of submarine deposits that frequently
form the surface from the sea-borders to the summits of the
mountain-ranges; and it is by the degradation of a land-surface formed
of such materials that the volcanic muds comparable to those of the
Solomon Islands are mainly produced. It is therefore apparent that we
have to distinguish here between the deposits of sedimentary and
eruptive origin, a distinction, however, which is not always easy to
make, since they are in both cases submarine, and doubtless were often
in process of forming together. The deposits most prevalent in the
island are the submarine tuffs partly sedimentary and partly eruptive in
their origin and the overlying volcanic agglomerates. The first are
usually palagonitic and calcareous and often contain organic remains,
being usually associated with volcanic muds and clays mainly the product
of marine erosion.

In connection with the employment of the terms “upraised” and “elevated”
in the case of the Vanua Levu deposits I will take this opportunity to
remark that I do not thereby commit myself to the view that there has
been an actual upheaval of this region. This is a matter, however, that
will be found discussed in Chapter XXVII.


                     THE UPRAISED CORAL LIMESTONES

These reef-limestones are scantily represented in the island, though one
can scarcely doubt that they were once far more extensive, having been
largely stripped off by the denuding agencies. They are mostly found on
the south coast between Naindi Bay and Fawn Harbour, and rarely extend
to heights greater than 20 or 30 feet above the sea, usually composing
the sea cliffs and not occurring as a rule inland. Massive corals are
often to be seen imbedded in their position of growth, as described in
Chapter II.; and as far as the absence of signs of disturbance is
concerned, these ancient reefs might owe their present situation, either
to the withdrawal of the sea or to the upheaval of the land. Such
reef-limestones exist over much of the Pacific, and they belong to the
usual type of these rocks.


                  SHELLY AND FORAMINIFERAL LIMESTONES

These rocks are composed partly of reef-debris, partly of volcanic
detritus, and partly of the tests of foraminifera (usually bottom
forms), fragments of lamellibranchiate and gasteropod shells, together
with those of pteropods, and other organic remains. Occasionally
separate valves of the genera “Cardium” and “Ostraea” are inclosed in
the limestone. These rocks have been evidently formed in rather shallow
water. In places they overlie palagonite-tuffs and clays, also
foraminiferal. Similar limestones are doubtless forming at the present
time off the coast.

They are usually hard in texture and greyish or pale yellow in colour.
They contain between 25 and 45 per cent. of carbonate of lime; whilst
the residue consists of fragments of minerals (10 to 15 per cent.),
including plagioclase, monoclinic and rhombic pyroxene, and occasionally
brown hornblende, with siliceous casts of foraminifera (4 to 20 per
cent.), mostly formed of chalcedonic silica but sometimes black and
glauconitic; the remainder (30 to 40 per cent.) being composed of
rounded and sub-angular portions of palagonite and semi-vitreous basic
rocks, of which the larger vary from ½ to 1 millimetre in diameter. In
some cases the carbonate of lime of the inclosed organic remains has
been mainly replaced by more or less crystalline silica. In others a
recrystallisation of the calcitic material is in progress, as described
on page 131; and the matrix presents in places a mosaic of calcite.

These rocks have therefore been subject to some degree of alteration,
the causes probably lying within the mass. They are, however, far from
frequent. They are best represented in the upper valley of the Sarawanga
River in the vicinity of Tembe-ni-ndio, where they reach to a height of
about 250 feet above the sea. The greatest elevation at which I found
them was in the mountainous interior of the Waikawa Promontory, where
they occur at a height of 1,100 feet above the sea.

As samples, the results of the examination of two rocks from the
Tembe-ni-ndio district are here appended:—

                                   A.



          Carbonate of lime                                46 per cent.
         {Fragments of palagonite and of semi-vitreous
         {basic rocks                                      30  "   "
 Residue {Minerals                                         12  "   "
         {Secondary silica replacing the carbonate of
         {lime in the organic remains                      12  "   "
                                                          ---
                                                          100

It displays to the naked eye fragments of shells including pteropods,
and numerous tests of foraminifera 2 or 3 millimetres in diameter. In
the section it displays in addition coral debris and a considerable
quantity of rounded and sub-angular pieces of palagonite and
semi-vitreous basic rocks, usually less than a millimetre across, with
smaller fragments of minerals (pyroxene and plagioclase), and much
calcitic material in the matrix.

                                   B.


 Carbonate of lime                                 25 per cent.
 Debris of palagonite and of basic volcanic rocks  45  "   "
 Minerals                                          10  "   "
 Siliceous casts of foraminifera                   20  "   "
                                                  ---
                                                  100

The organic remains mainly consist of tests of foraminifera, many of
which occur in the residue as colourless siliceous casts. The fragments,
whether of minerals or of volcanic rocks, are usually less than half a
millimetre across. The foraminifera are mostly small and of the
“Globigerina” type; but there is a cast of a tube of some
boring-mollusc, and fragments of shells also occur.


                          PTEROPOD-OOZE ROCKS

These rocks are bluish-grey when not exposed; but through the hydration
accompanying exposure they become much lighter in colour. They are
crowded with pteropod shells, and contain also small gasteropod and
lamellibranchiate shells together with tests of foraminifera both
microscopic and macroscopic. They yield between 30 and 40 per cent. of
carbonate of lime, the residue being made up of disintegrated
palagonitic debris and fine clayey material derived from the same
source, together with a fair amount of mineral fragments (10 per cent.)
which include plagioclase, pyroxene, and brown hornblende, and measure
in the case of the larger fragments between ·1 and ·4 mm. in diameter.
Such rocks are somewhat friable and correspond with the pteropod-ooze
rocks of the Solomon Islands; but they are not very frequent, being best
represented on the flanks of the basaltic table-land between the Wainunu
and Yanawai rivers, as in the vicinity of the Nandua tea-plantation
where they extend up to 500 feet above the sea. In this particular
locality (see page 345) they overlie horizontally-stratified tuffs and
clays, composed of the debris of a basic glass usually vacuolar but now
for the most part converted into palagonite, and showing a few small
tests of foraminifera.

These deposits are always either surface or incrusting formations.
The circumstance of their passing down into characteristic
palagonite-formations is repeated in the case of the Tembe-ni-ndio
limestones, as observed on page 131; and there is no doubt that much
of their non-calcareous material is derived from the disintegration
of palagonite.

  _Sample of pteropod-ooze rock from below the Nandua tea-plantation._

          Carbonate of lime                          38 per cent.
         {Palagonitic debris and clayey material     51  "   "
 Residue {Minerals                                    8  "   "
         {Casts of foraminifera                       3  "   "
                                                    ---
                                                    100

Fine clayey material makes up the greater part (72 per cent.) of the
residue. It presents the microscopic characters of material derived from
the degradation of palagonite. Amongst the mineral fragments, of which
the larger are ·2 to ·3 mm. in size, occur brown hornblende, pyroxene,
felspar, magnetite, &c. The casts of foraminifera are usually
glauconitic, but a few are of crystalline silica. A number of curious
little pellets of palagonite, oblong or oval in form, occur in the
residue. Their size is ·3 to ·6 mm., and they apparently represent the
minute amygdules of palagonite that occupy the vacuoles in an altered
basic glass.


                   FORAMINIFEROUS VOLCANIC MUD-ROCKS

These deposits, which represent the “volcanic muds” forming around the
coasts of volcanic islands, are more or less consolidated clay-rocks.
They contain in varying numbers the tests of foraminifera with
occasionally pteropod-shells. The former are usually minute and of the
“Globigerina” type; but in some rocks larger bottom-forms prevail. The
original colour of these deposits is bluish-grey, but as generally
displayed they are pale-brown and considerably affected by hydration and
are known as “soapstones” in the group. The ultimate effect of exposure
is the production of a whitish or yellowish soapy rock that has lost all
the carbonate of lime and all the organic remains and breaks down easily
in the fingers. Such a crumbling material, when examined with a high
power, cannot be distinguished from the products of the final
disintegration of palagonite as described on page 348.

These deposits are frequently displayed in the lower regions up to
elevations of 300 feet above the sea, incrusting the basaltic plains of
Lekutu, Sarawanga, Ndreketi, and Lambasa, that occupy such a large area
of the north side of the island, and exhibiting there in the prevailing
horizontality of their beds the same indication that is presented by the
vertical position of the columns of the under-lying basalt, namely, the
comparative absence of disturbance during the emergence. At the foot of
the mountains these deposits are interstratified with and finally
overlaid by coarse palagonite-tuffs, also containing marine remains; and
these are in their turn covered over by the agglomerates that often
enter so largely into the composition of the mountainous backbone of the
island. Such beds form apparently the lowest of a series which begins
with the foraminiferous clay and ends with the agglomerate. But in some
places, as has been noticed in the cases of the pteropod-ooze rock of
Nandua, and of the shelly and foraminiferous limestone of Tembe-ni-ndio,
beds composed largely of characteristic palagonite lie beneath.

In the elevated interior of the island these volcanic mud-rocks are
usually concealed by the tuffs and agglomerates. Occasionally, however,
they are to be seen exposed by landslips high up the flanks of the
mountains, as on the slopes of Thambeyu 1,000 feet above the sea. An
interesting exposure of them is displayed in the heart of the island in
the face of the Mbenutha cliffs where the elevation is about 1,100 feet.
Here they are overlaid by a thick bed of agglomerate; and tuff-beds
largely made up of vacuolar basic glass debris and showing a few
foraminifera are interstratified with them; but they exhibit signs of
considerable disturbance (see page 109).

These deposits contain between 5 and 25 per cent. of carbonate of lime,
and as a rule about 90 per cent. of the residue consists of fine clayey
material derived from the final degradation of palagonite and of basic
rocks. The mineral fragments (plagioclase, augite, rhombic pyroxene, and
occasionally hornblende) vary much in amount, their average proportion
being 13 or 14 per cent. of the mass. Their size is usually less than
·2 mm. and does not exceed ·4 mm. Casts of foraminifera are nearly
always present in the residue and form generally 3 or 4 per cent. of the
whole deposit. Sometimes they are black and glauconitic; but more
frequently they are white and composed of chalcedonic silica. Such casts
represent on a small scale the results of the same silicifying operation
to which the flints and silicified corals that occur so frequently on
the surface in some localities owe their origin (see Chapter XXV.).

With regard to the age of these volcanic mud-rocks of Vanua Levu, it is
most likely that as in the case of similar deposits in Viti Levu, which
were examined by Mr. H. B. Brady, they are of post-tertiary origin.
Samples of the Suva “soapstone” containing 5 or 6 per cent. of lime and
displaying shells of foraminifera, pteropods, and other molluscs, were
obtained from different heights up to 100 feet above the sea. Since 87
out of the 92 species of foraminifera represented in the deposits are
known to be living now in the Pacific, Mr. Brady had no hesitation in
assigning the beds to the post-tertiary epoch.[126]

                   SAMPLES OF THE VOLCANIC MUD-ROCKS

    A. _From districts west of Ndranimako, 100 feet above the sea._

          Carbonate of lime                  20 per cent.
         {Fine debris of palagonite and
         {  semi-vitreous basic rocks        62  "   "
 Residue {Minerals                           14  "   "
         {Casts of foraminifera               4  "   "
                                            ---
                                            100

The organic remains consist mainly of tests of minute foraminifera of
the “Globigerina” type, casts of which, both glauconitic and
chalcedonic, occur in the residue. About 88 per cent. of the residue
consist of fine clayey materials less than ·25 mm. in size. The mineral
fragments, which average about ·1 mm. in diameter, are mostly of felspar
with a little pyroxene and brown hornblende.

        B. _From the Mbenutha Cliffs, 1,100 feet above the sea._

            Carbonate of lime                      15 per cent.
          { Fine material mainly derived from
          {   the degradation of palagonite        60  "   "
 Residue  { Minerals                               23  "   "
          { Casts of foraminifera                   2  "   "
                                                  ---
                                                  100

This rock is somewhat hard, so that the proportion of fine clayey
material, which is however large, cannot be accurately determined. It
shows in places dark streaks composed of an abundance of minute and
often perfect tabular crystals of zoned plagioclase and prisms of
rhombic pyroxene, the size in neither case exceeding half a millimetre,
both of them being derived from the acid andesites of the neighbourhood.
In the slide it displays minute tests of foraminifera of the
“Globigerina” type in a matrix formed mainly of palagonitic debris,
fragments of minerals and semi-vitreous basic rocks. The larger
fragments of the minerals and of the volcanic rocks do not exceed
·15 mm.; but most of the material is very fine. The tests of the
foraminifera are sometimes filled with the matrix, but often they are
entirely of calcite and exhibit in polarised light a dark cross.

   C. _From between Natua and Mbatiri, about 290 feet above the sea._

           Carbonate of lime                                25 per cent.
         { Fine material derived from the degradation of
         { palagonite and of semi-vitreous basic rocks      62  "    "
 Residue { Minerals                                          2  "    "
         { Casts of foraminifera                            11  "    "
                                                           ---
                                                           100

This is a relatively deep-water deposit, the foraminifera being minute
and of the “Globigerina” type. About 90 per cent. of the residue
consists of fine clayey material, with which the calcite is so
intimately mixed that each particle is highly refractive and effervesces
freely in an acid. The mineral fragments (pyroxene and felspar) are very
scanty, the largest being less than ·25 mm. The white casts of
foraminifera, composed of chalcedonic silica, form a conspicuous
elements in the residue.

       D. _From the vicinity of Mbatiri, 100 feet above the sea._

 Carbonate of lime                                         4 per cent.
 Fine material derived from the degradation of palagonite 90  "   "
 Minerals                                                  2  "   "
 Casts of foraminifera                                     4  "   "
                                                         ---
                                                         100

About 94 per cent. of the residue consists of fine clayey material. The
fragments of minerals are very scanty and are all less than ·2 mm. in
size. The casts of foraminifera are white and of chalcedonic silica.
From the fineness of the materials and the small size and pelagic
character of the foraminifera, this deposit may be regarded as formed in
relatively deep water.

E. _From the eastern flank of the Wainunu table-land, 200 feet above the
sea._—This is a shallow-water deposit and contains, besides small
gasteropod shells, large flat tests of foraminifera 5 or 6 mm. in
diameter. It possesses 24 per cent. of carbonate of lime, 62 per cent.
of palagonitic debris, &c., and 14 per cent. of minerals.


                       ALTERED VOLCANIC MUD-ROCKS

This group includes compact hard foraminiferous usually dark-brown
rocks, which exhibit evidence of alteration in their induration, in the
presence of pyrites, and in the chalcedonic quartz filling fine cracks
in the rock-mass. Occasionally special types of alteration occur, one of
which will be referred to in the description of some of the rocks given
below. The proportion of carbonate of lime is generally small; but
sometimes it amounts to 10 per cent. or more.

They admit of being examined in thin sections; and their true nature is
at times so much disguised that I have taken them at first for aphanitic
basic andesites. In the slide they display a few scattered tests of
foraminifera of pelagic habit in a matrix composed of the fine debris of
palagonite and of basic rocks, together with fragments of plagioclase
and pyroxene. Most of the material is very fine, and the size of the
largest mineral fragments does not exceed ·2 mm.

Such rocks, however, are not very frequent. They may be displayed on the
flanks of mountain-ranges buried beneath basic tuffs and agglomerates,
as in the case of Mount Mariko (page 187) and of the mountain-slope
behind Mbale-mbale (page 158); or they may be found at much lower levels
as at Savarekareka Bay (page 190), where, however, they assume sometimes
a peculiar character. Though in the last-named locality the alteration
is possibly connected with thermal metamorphism, it is probable that in
most instances it is a normal interstitial change occurring in beds of
some antiquity which are covered over by a considerable thickness of
later deposits. In places, where these rocks have been subjected to much
hydration in the weathering process, they become red in colour, as is
found on the flanks of Mount Mariko.


               SAMPLES OF THE ALTERED VOLCANIC MUD-ROCKS

A. From between 400 and 500 feet above the sea on the south slope of the
Mariko Range.... The characters and mode of occurrence of this rock are
described on page 187.

B. From an elevation of 1,100 feet on the south slope of the Korotini
Range.... The description of the locality will be found on page 160.

This rock is hard and compact and looks like an altered basic rock
showing a few minute specks of pyrites. It is composed of fine
palagonitic debris, and small fragments of semi-vitreous basic rocks and
of crystals of pyroxene and felspar, none of the fragments exceeding
·2 mm. in diameter. Tests of minute foraminifera, filled with matrix and
of the “Globigerina” type, occur very scantily. There is little or no
carbonate of lime; but secondary silica, both colloid and crystalline,
is present as an alteration product.

C. From the vicinity of Yaroi, 30 feet above the sea.... The locality is
described on page 189.

This is a dark grey hard compact rock, containing probably between 10
and 15 per cent. of carbonate of lime, and looking like an altered
limestone. In the section it displays minute tests of foraminifera of
the “Globigerina” type in a matrix composed of fine disintegrated
palagonitic material, impregnated with calcite and containing also
fragments of minerals (augite and felspar), none of which exceed ·2 mm.
in diameter. There are also a few similar-sized fragments of
semi-vitreous basic rocks. Some fine cracks in the rock-mass are filled
with a quartz mosaic. The tests of the foraminifera remain calcitic; but
their cavities are filled either with the matrix or with calcite or with
a colourless fibro-radiate mineral polarising in blackish-blue hues.

D. From the south shore of Savarekareka Bay.... The locality is
described on page 190.

This is a bright green hard compact rock with flinty fracture and not
effervescing with an acid. In the slide it shows a few casts of
foraminifera of the “Globigerina” type in a matrix composed mainly of
fine debris (·01-·04 mm.) of felspar and pyroxene with much greenish
opaque amorphous alteration products. The abundance of pyroxene is
remarkable. The material of the tests of the foraminifera is altogether
replaced by a greenish yellow mineral, occurring in grains and radiating
prisms, apparently epidote.

E. From an elevation of 950 feet on Mount Thambeyu.... The locality is
described on page 177.

A hard dark grey rock containing 10 or 15 per cent. of carbonate of lime
and showing fine specks of pyrites. In the slide are displayed numerous
tests of foraminifera of varying size up to ·5 mm.; scattered patches of
pyrites; fragments of a semi-vitreous basic rock, not exceeding ·15 mm.
in diameter, and of plagioclase and pyroxene; in a matrix of the finest
debris of the same materials impregnated with granules of calcite. The
tests of the foraminifera are filled either with calcite, showing a
black cross in polarised light, or with a zeolite, or with pyrites, or
with the matrix.


               SUBMARINE BASIC TUFFS OF MIXED COMPOSITION

These tuffs, which are composed not only of palagonitic materials but
also of the fine detritus of usually semi-vitreous basic rocks, rank
first in frequency amongst the volcanic sedimentary deposits of the
island. In their character they pass on the one hand into the
foraminiferous volcanic mud-rocks or clay rocks and on the other into
the tuff-breccias and tuff-agglomerates. We have here a series beginning
with the agglomerate and ending with the clay that represents in a
general sense the successive stages of the degradation of the same
materials.

These tuffs occur at all elevations from the sea-border, where they may
form the shore-cliffs, to the upper slopes and summits of the
mountain-ranges, where they are found at elevations between 2,000 and
2,500 feet above the sea. In the interior of the island they are
generally to be observed underlying the basic agglomerates. Wherever an
extensive exposure of the agglomerates exists in the mountainous
districts, these tuffs are as a rule to be found at the base of the
cliffs. The precipitous bluff of agglomerate, that so often gives a
character to the mountainous interior, and the line of cliff of the same
formation that runs along the slopes, represent the work of landslips,
as is shown by the huge masses of agglomerate lying on the ground below.
These “slips” are not uncommon, and are due to the undermining influence
of the springs that percolate through the tuffs and clays underlying the
agglomerates.

When the tuffs are well displayed they as a rule show stratification.
The bedding may be indicated either by distinct parting-lines or by
alternating bands of varying degrees of coarseness. That these deposits,
when occurring in mass in the upland regions, are often horizontal or
but slightly inclined, is evidenced by the Nganga-turuturu Cliffs, 1,200
feet above the sea, which are described in Chapter VIII., in the line of
cliffs behind Sealevu (Chapter XI.), and high up the slopes of Mount
Thambeyu (Chapter XII.) as high as 1,500 feet. This is also shown in the
circumstance that the line of junction with the overlying agglomerate,
except in rare cases, as in that of the Mbenutha Cliffs, is generally
horizontal. It is, however, not uncommon to find the beds exposed on the
mountain-flanks dipping away from the axis of the range at a small
angle, as on the slopes behind Mbale-mbale and in the Sokena Cliffs. In
the lower regions, where these deposits are associated with the volcanic
mud-rocks on the basaltic plains, they are but slightly inclined. On the
other hand, as in the Kumbulau district, the sea-cliffs for some miles
may be composed of tuffs more or less steeply tilted.

These tuffs are generally more or less compacted and have a greyish
colour; but as usually exposed in a weathered condition they are often
pale brown or yellowish and are more friable. They may be uniform in
structure, or they may display thin seams of a marl-like clay, or they
may contain numerous lapilli of vesicular basic glass extensively
palagonitised. Not uncommonly they contain larger fragments of basic
rocks, and when these are at all frequent the terms “agglomerate-tuff”
or “tuff-agglomerate” have been employed according to the preponderance
of either material.

Many of these tuffs show no effervescence with an acid; and this is
especially the case with specimens at all weathered. On the other hand
there are just as many that contain a little carbonate of lime, not
usually more than 3 or 4 per cent., but sometimes amounting to 12 or 13
per cent. It often happens in the case of a series of tuffs, apparently
non-calcareous, that an occasional thin band of a fine clay-like rock
contains a good percentage of lime. It is pointed out below, however,
that the absence of effervescence does not necessarily imply the absence
of foraminiferous tests.

Tests of foraminifera, often macroscopic bottom forms, together with
shells of small gasteropods, are displayed at times; but they are as a
rule in such cases not frequent. I found foraminiferous tuffs at
considerable heights in some localities, as for instance between 2,000
and 2,500 feet on the slopes of Mount Thambeyu (page 178), at an
elevation of 1,850 feet on the south slope of the Korotini Range above
Vatu-kawa (page 158), and between 2,000 and 2,400 feet on the summit of
the range between Waisali and Sealevu (page 154). In the last-named
locality, where the tuffs are coarse and often of the nature of
agglomerate-tuffs, they are highly fossiliferous; but such a character
is exceptional.

The submarine origin of the tuffs can often be demonstrated in the
absence of evidence of organic remains, as by their interstratification
with foraminiferous clay rocks, such as we find at an elevation of 1,000
to 1,100 feet on the top of the “divide” between the Ndreketi and
Lambasa basins. A single seam of marl-like rock displaying only a
solitary test of a foraminifer in the slide may throw light on the
origin of the coarser tuffs with which it is associated. The use of the
microscope is essential in the case of some of the harder tuffs, where
there has been a little alteration. Here casts of foraminifera may be
observed, although no carbonate of lime is indicated by an acid. In some
localities where no organic remains are evident in the tuff, fine
waterworn gravel is to be noticed.

These deposits are composed as a rule of sub-angular fragments of
semi-vitreous basic or basaltic rocks and of palagonite, together with
fragments of plagioclase and pyroxene, the interspaces being filled with
fine debris of the same materials. The relative proportion, however, of
the three principal constituents varies considerably, the palagonite,
for instance, being sometimes scanty and sometimes abundant. The size of
the larger fragments in a tuff of the most common kind is about a
millimetre; but deposits rather finer and rather coarser are also
frequent. In the very coarse tuffs and in the breccia-tuffs, where the
larger materials are mostly of palagonite, the larger fragments may be a
centimetre in size and even more, the interspaces being filled with fine
debris of the same character cemented together often by carbonate of
lime.

The fragments of semi-vitreous basic rocks forming a regular constituent
of these tuffs are usually dark and opaque and display a few plagioclase
lathes. They correspond with the type of the semi-vitreous basalt or
basaltic andesite, of which the blocks of the overlying agglomerates are
as a rule composed and are doubtless derived from the same source.
Fragments of unaltered basic glass are rarely to be observed in these
tuffs. It is as a rule all converted into palagonite. This material
presents itself in various stages of secondary alteration, from the
compact greenish or yellowish waxy mass to a white friable pulverulent
substance, which represents the last stage of degradation. These changes
will be found described on page 348. It can, however, be stated here
that they are mainly concerned with hydration. In the case of the
lapilli of finely vesicular basic glass, that is, of basic pumice, which
are inclosed in some of the tuffs, all stages of the secondary
alteration of palagonite are often exhibited, and the last stage of the
change is merely indicated by a white powdery patch containing a few
minute siliceous amygdules. The puzzling little white patches so common
in basic tuffs merely represent lapilli of basic pumice that have been
palagonitised, and then bleached and disintegrated by hydration.

The minerals are more or less abundant and may constitute a third of the
whole mass. They include plagioclase, augite, rhombic pyroxene, and
magnetite, olivine being rare and scanty. Entire crystals of any size
are infrequent. However, crystals of augite, 5 or 6 mm. in length, are
found in the tuffs at Naivaka and of the coast cliffs near the Salt Lake
Passage.... It may be observed that zeolitic minerals which are
frequently developed in the tuffs consisting almost entirely of
palagonite are not typical of the tuffs of mixed composition.

There is nothing suggestive of recent eruptions in any of these
formations. They were formed ages since on the sea-floor at varying
depths around volcanic vents. Sometimes a cone was able to rear itself
above the level of the sea; but in most cases it rapidly succumbed to
breaker-action. Three agencies, concerned with sub-aerial eruptions,
submarine eruptions, and marine denudation, have co-operated in the
production of these deposits, but their parts in the process have varied
greatly. The last is indicated when the tuff is formed of a variety of
basic rocks with but little palagonite. The tuffs containing much
palagonite representing an original vacuolar basic glass are regarded as
mainly the products of submarine-eruptions. In those cases where lapilli
of altered basic pumice occur in the deposit, sub-aerial eruptions are
directly indicated. When an extensive exposure of these tuffs occurs, as
in the case of the Nganga-turuturu Cliffs and in that of the section
displayed near the hill of Korolevu (page 48), all three agencies are
often illustrated.


       SAMPLES OF THE SUBMARINE BASIC TUFFS OF MIXED COMPOSITION

A. As examples of the non-fossiliferous tuffs where the palagonite
constituents do not predominate, I will take those exposed in beds in
the coast cliffs and in the low hills in the vicinity of Na Tokalau in
the Kumbulau peninsula (page 90). They are grey in colour and have the
texture of a sandstone, being more or less compacted and showing no
effervescence with an acid. They are composed of fragments of basic
rocks and of minerals, varying in size from ·5 to 1 mm., in a scanty
matrix made up of fine detritus of the same materials and of palagonitic
debris. The fragments of volcanic rocks are rounded and sub-angular, and
formed mainly of a basaltic rock, with a black opaque groundmass showing
some small plagioclase crystals and in places more or less
palagonitised. There are also portions of a hemi-crystalline basic rock
showing small augite crystals. The minerals entire and in fragments,
which make up quite a third of the mass, are mostly of plagioclase, but
monoclinic and rhombic pyroxene are also well represented. It is evident
that through the alteration of the palagonitic constituents, which were
probably more frequent when the tuffs were deposited, the structure of
the matrix is somewhat disguised.

B. A yellowish grey tuff composing the cliffs on the north coast of
Naivaka affords a good example of a tuff where the palagonitic materials
predominate. It is somewhat fine-textured and displays a tendency to
lamination. The powdered material effervesces slightly with an acid. In
the slide it is exhibited as composed mainly of palagonite and of
fragments of minerals, the latter making up about one-third of the whole
and ranging in size usually between ·2 and ·5 mm. A number of more or
less parallel fine cracks, filled with calcite and traversing also the
inclosed crystals of plagioclase, together with small fragments of basic
rocks are displayed in the section. There are a few fragments of
semi-vitreous basic rocks, as just indicated, but the palagonite is the
principal constituent. It shows numerous minute amygdules occupying the
original vacuoles of the basic glass; and in its substance occur
irregular patches formed of a colourless semi-isotropic mineral which is
either zeolitic or a form of opal. Plagioclase and augite compose the
mineral fragments, the former prevailing. Although these tuffs are
derived from a vent that was probably the last in eruption in this
island, they display considerable alteration which is mainly connected
with the secondary changes affecting the palagonite since the deposition
of these materials.

C. As an example of the banded tuffs composed of coarse and fine
materials I will take a compact grey rock forming one of the horizontal
beds in the natural section exposed near the hill of Korolevu (p. 49).
It is also an example of those tuffs which whilst not effervescing with
an acid display a few casts of foraminifera in the slide. The
alternating bands which are about a centimetre in thickness pass
gradually into each other. The bands of finer materials are made up of
sub-angular fragments, ·1 to ·2 mm. in size, of the dark opaque
groundmass of a semi-vitreous basic rock and of a grey hemi-crystalline
groundmass of an augite-andesite, together with palagonite more or less
decomposed, and fragments of plagioclase and augite, whilst the
interspaces are filled with the finer debris. The layers of coarse
materials have much the same composition, the fragments varying usually
between ·5 and 1 mm. in diameter, with occasional larger pieces of
palagonite 2 to 5 mm. in size representing original lapilli of a
vacuolar basic glass. The tests of the foraminifera, which occur
scantily in the layers of fine material, are all minute. They are filled
with palagonitic material.

D. The tuffs prevailing on the higher flanks of the mountainous backbone
of the island are well represented by those exposed at an elevation of
1,200 feet above the sea on the south slope of the Korotini Range behind
Mbale-mbale. It is a somewhat coarse-grained rather hard grey rock
effervescing feebly with an acid. It is composed of sub-angular or
partly rounded fragments, 1 to 4 mm. in size, of various basic rocks,
and of rather smaller fragments of plagioclase and augite, the
interstices being filled up with fine debris of the same materials, in
which a few minute tests of foraminifera of the “Globigerina” type may
be observed. The basic rocks of which the fragments are formed comprise
the following: (_a_) a grey aphanitic augite-andesite, with but little
interstitial glass, presenting a parallel arrangement of the minute
felspar-lathes which have an average length of ·05 mm.; (_b_) a grey
augite-andesite of coarser texture but in other respects similar, the
felspar-lathes being about ·1 mm. in length, whilst there is a little
microporphyritic plagioclase; (_c_) a semi-vitreous basaltic rock
showing small porphyritic crystals of plagioclase and augite in a
groundmass usually black and opaque, but sometimes smoky and displaying
felspar microliths; (_d_) a vacuolar basic glass more or less
palagonitised.

E. As a specimen of the calcareous tuffs those exposed on the south
slope of the Korotini Range at an elevation of 1,850 feet may be given.
They contain about 11 per cent. of carbonate of lime and inclose a few
tests of foraminifera 1 to 2 mm. in diameter. The other constituents are
fragments of semi-vitreous basic rocks and of palagonite, together with
fragments of plagioclase and pyroxene crystals and of an amorphous
siliceous mineral which behaves optically like chalcedonic silica. When
the rock is gently rubbed down, minute fragments of this white mineral
can be picked out. They have a wrinkled surface and an irregular form
and are not affected by acids. In polarised light they display a rude
mosaic or an imperfect radiate structure.

F. As specimens of the fossiliferous agglomerate-tuffs composed mainly
of palagonite, those exposed on the high mountain slopes of the Korotini
Range at heights of 2,000 feet may be here cited. They are described on
p. 154.

_Note._—The examples of mixed tuffs above given represent only some of
the principal types of these deposits. Short descriptions of others will
be found in the detailed account of the geology of the island.


                ALTERED BASIC TUFFS OF MIXED COMPOSITION

These form a group of hard compact rocks, the fragmental character of
which is not always apparent in hand-specimens, microscopical
examination of thin sections being usually required for the
determination of their true nature. They are commonly exposed on the
southern flanks of the Korotini Range at the back of Vatu-kawa and
Nukumbolo. They are composed of compacted fragments, varying in
different localities from 1 to 5 or 6 mm. in size, of a variety of
semi-vitreous basic rocks, the matrix being scanty but often containing
zeolites and secondary silica, whilst occasionally secondary calcite is
developed. They contain no organic remains, and palagonite when present
is usually scanty, whilst viridite and similar materials represent the
decomposition of the pyroxene.... A hard breccia-tuff found on the
flanks of Mariko and in one or two other localities contains vesicular
fragments, where the steam-holes are filled with opal or chalcedony, and
the cracks traversing the matrix are also filled with this mineral.

The alteration of these tuffs has evidently arisen from a variety of
causes. In some cases the change appears to be purely interstitial. In
other times it has arisen from contact-metamorphism, or from
hydro-thermal agencies, as in the case of the altered tuffs near the hot
springs at Nukumbolo (p. 161).


                      SUBMARINE BASIC PUMICE TUFFS

These deposits, which, however, are not of frequent occurrence, are
interstratified with volcanic mud-rocks in certain localities, as at the
Mbenutha Cliffs (p. 110), and in the vicinity of the hill of Korolevu
(p. 47). They indicate periods of volcanic activity during the
deposition of the foraminiferous muds, with which they are associated,
when the fine materials ejected from sub-aerial vents fell into the seas
around.

Such tuffs are more or less compact and usually fine in texture. When
the glass fragments are but slightly altered, the tuff-rock is dark
grey; but when the palagonitic change is well advanced, it becomes pale
and yellowish. They are made up chiefly of small fragments of a
bottle-green basic glass, which are as a rule vacuolar and sometimes
fibrillar; but it never happens that the pumiceous character is as
pronounced as in acid pumice; and in some cases the vacuoles or
steam-pores are to be observed only in the minority of the fragments
displayed in a slide. The size of the glass fragments is as a rule
small, in some tuffs averaging only ·1 mm. and in others ·5 mm.; but
occasionally they may be 1 or 2 mm. in diameter.

Fragments of minerals (plagioclase and pyroxene) corresponding in size
to the glass fragments are as a rule well represented, forming a fourth
or a third of the mass. A little fine detritus of a semi-vitreous basic
rock also occurs. Tested in an acid several of the tuffs either do not
effervesce or give an indication of a small percentage of carbonate of
lime; whilst others effervesce freely. They usually display a few minute
tests of foraminifera of the “Globigerina” type, the cavities of which
are filled with fine palagonitic debris.

The palagonitic alteration is to be noticed in all cases; but it varies
considerably in its extent. In the dark grey tuffs it affects the
margins only of a few of the glass-fragments. In the paler tuffs it has
extended more into their substance, although the alteration is never
more than partial. The pale greyish material filling up the interspaces
is composed of disintegrated palagonite. The steam-pores or vacuoles are
sometimes empty; and at other times, especially where the palagonitic
change has begun, they are filled with a granular alteration product.
The glass fuses readily; but is not affected by acids. It is clear and
isotropic, showing however a few scattered microliths. These tuffs
correspond with the hyalomelan-tuff from the island of Munia in this
group as described by Wichmann; but in that instance no mention is made
of inclosed tests of foraminifera.


                  “CRUSH-TUFFS” FORMED OF BASIC GLASS

This is a remarkable group of compacted tuff-like rocks which as
hand-specimens would be generally regarded as pitchstone-tuffs. Their
detrital origin is, however, often very doubtful. They are composed of
fragments of basic glass, carrying plagioclase phenocrysts, with the
interspaces occupied by palagonite and by the finer debris of the glass
and felspar. The larger glass fragments, which vary in different rocks
from 1 or 2 to 4 or 5 mm. in size, have been crushed _in situ_, the
broken portions often remaining more or less in position. These
fragments are invested by palagonite and have eroded borders, as shown
in the figure on page 342. The glass is bottle-green, non-vacuolar,
fuses readily, and only at times displays incipient crystallisation. The
explanation of the origin of these rocks is attempted in Chapter XXIV.
They contain neither carbonate of lime nor organic remains. The most
typical example is present in a bed underlying a pitchstone-agglomerate
near Narengali (see page 149). It is not uncommon to find evidence of
crushing in the glassy matrix of a pitchstone-agglomerate or of rubbly
pitchstone, as in the Va Lili Ridge (142), the Korotini Bluff (157), and
Mount Soloa Levu (313); and here also palagonite has been produced
around the crushed fragments.


                    COARSE ZEOLITIC PALAGONITE-TUFFS

These deposits represent coarse kinds of the submarine tuffs of basic
glass, in which the palagonitic change is far advanced, and where
zeolites and at times secondary calcite have been produced in abundance
as a result of the alteration. They present themselves in the mass as
mottled grey rocks which when examined in thin sections are seen to be
composed in great part of fragments of more or less palagonitised
vacuolar basic glass, whilst zeolites are extensively developed in
numerous irregular cavities and in the interspaces. Although displaying
no organic remains, their submarine character is indicated as at Nandua
by the circumstance of their occurring as horizontal beds overlaid by
pteropod-ooze deposits, or as at Tembe-ni-ndio by their forming part of
a series of horizontal beds with a shelly limestone and a foraminiferous
palagonite clay overlying them.

The fragments of bottle green basic glass vary usually between 1 and 4
millimetres. They were originally vacuolar and at times fibrillar from
the lengthening out of the minute steam-pores; but through the
palagonitic change these characters have been often disguised, and it is
only at times that the unaltered glass is observed. Plagioclase and
sometimes augite and occasionally olivine formed phenocrysts in the
original glass. The zeolites, which include chabazite and natrolite, may
be so extensively developed that they make up a fourth or a fifth of the
rock mass. One may observe them in cavities where the walls are lined by
fibrous natrolite with the cube-like crystals of chabazite occupying the
interior. The calcite is usually subordinate to the zeolites, but
sometimes the tuff contains as much as 10 per cent. of this mineral,
which is evidently of secondary origin.... The history of these tuffs in
the district of Nandua and Ulu-i-ndali is no doubt applicable to these
deposits in other localities. They are the products of submarine
eruptions which shattered into fragments the extensive palagonite crusts
of flows of basaltic lava. In Chapter XXIV. I have attempted to show how
palagonite is formed on a large scale in the case of such submarine
displays of volcanic activity.


           CHOCOLATE-COLOURED FORAMINIFEROUS PALAGONITE-MARLS

We have here hard, somewhat calcareous, clay-rocks which consist in
great part (nine-tenths) of fine palagonite debris with some fragments
of minerals and a little fine detritus of semi-vitreous basic rocks.
Some hand-specimens would be taken for pure palagonite; but the
fragmental nature appears at once in the slide. This is especially the
case with a rock exposed in a stream-course near Rewa on the shores of
Savu-savu Bay (see page 95). The materials composing them are
exceedingly fine, the largest fragments not usually exceeding ·2 mm. As
a rule they contain a little carbonate of lime and sometimes as much as
10 per cent., whilst a few tests of minute foraminifera are to be
noticed in the slide. These deposits are horizontally bedded, and
underlie a pteropod-ooze rock at Nandua and a shelly impure limestone at
Tembe-ni-ndio. They are not very frequent, and sometimes approach in
characters the volcanic-mud rocks, which, however, are much more mixed
in composition. I regard them in the main as sedimentary deposits
derived from the disintegration of the palagonitised vitreous surface of
a submarine basaltic flow. They pass downward at Nandua, as described on
page 345, into a rock of pure palagonite; and they are only to be found
in localities where basaltic plains or plateaux are covered over with
submarine deposits.


                           ACID PUMICE TUFFS

The general characters of these deposits are described on pages 10, 218,
220, 222, 223, 231, 233, &c. Such tuffs are restricted to the north-east
part of the island east of Lambasa and Tawaki, and are well displayed in
the coast cliffs. They are pale yellow or whitish, and are usually
non-calcareous. They are composed of the debris of a vacuolar and
fibrillar isotropic glass, nearly colourless and in some localities
altered. Small crystals of quartz and of glassy felspar with bits of
obsidian (up to 3 mm.) and lapilli of rhyolitic glass are inclosed in
them. In places inclosed pieces of coral and coral rock indicate
submarine deposition.



                              CHAPTER XXIV

                               PALAGONITE


FROM the sea-border to the mountain-top in almost every part of the
island, palagonite occurs in a fragmental condition. It is only where
tuffs are not found, as in the mountainous mass of Seatura, or where
these deposits are formed of acid rocks as in the north-east portion of
the island, that palagonite has not been observed. Perhaps, it is not
too much to say that the later if not all the stages in the history of
Vanua Levu are bound up with the history of this material. In this place
I will only deal with certain features in the problem connected with the
origin of palagonite which seem to receive further elucidation from my
observations in this island. The literature is already extensive, and
those interested in the matter will find in Zirkel’s _Petrographie_ and
in the _Challenger_ Report on Deep-Sea Deposits by Murray and Renard a
good introduction to the subject.

In Vanua Levu we are confronted with the same difficulty that has
perplexed geologists in various parts of the world. If we expected to
find in this island the source of the enormous quantities of the basic
glass that are represented by the palagonite of the tuffs, we should
look in vain. Basic or basaltic glass usually occurs in agglomerates in
the form of tachylytic pitchstones, as described on page 312, and is
also found at times in basic pumiceous tuffs, as described on page 333;
but it is far from frequent. Palagonite-rock, that is to say, a basaltic
glass converted in mass into this substance, never came under my notice.

In order to clear the ground for the discussion of my own observations,
I will quote from the report on deep-sea deposits above named. Fragments
of basic glass undergoing the palagonite change are found everywhere in
these deposits and especially in the red-clay areas. The hydro-chemical
modifications determining the decomposition of these fragments into
palagonite, and at the same time the formation of zeolites, have
likewise resulted in the complete transformation of these lapilli into
ferruginous argillaceous matter (p. 309). The authors, however, of this
report do not attribute the frequent occurrence of fragments of basic
glass on the bottom of the ocean to the buoyant powers of basic pumice.
Unfortunately, the problem does not permit of such a simple solution.
Basic volcanic glass, writes Prof. Renard, though known only from a few
geological formations and from a few eruptions of recent volcanoes at
the surface of the continents, appears in abundance and in most typical
form among the products of submarine eruptions, as if the deep oceans
had been in some way specially favourable to the development of this
lithological type (p. 299).

The palagonite-tuffs of this island are described in detail in Chapter
XXIII., and a few general remarks are alone needed here. This altered
glass enters into the composition, to a greater or less extent and in
varying stages of disintegration, of nearly all the submarine basic
tuffs and clays. In the volcanic muds, however, and in the tuffs of
mixed character, which are the prevailing deposits, it is associated
with other components. Here the question of the origin of palagonite
within the deposit does not as a rule arise, since there is nothing to
indicate that this material was not derived from rocks previously
palagonitised, and the point of main interest is connected with the last
stages in the degradation of this substance. There are not a few cases,
however, where, unless we assume that the lapilli of vesicular basic
glass were ejected in the palagonitic condition from a volcanic vent, we
must apparently regard the alteration as having occurred in the tuff.
But even this will prove to be by no means a necessary consequence if it
can be shown, as I have attempted to do below, that the palagonitic
condition exists potentially in a particular type of basic glass and
that the effect of hydration is not so much to produce but to make
evident a condition that was previously latent.

It will be therefore of interest to determine whether palagonite occurs
in this island independently of the tuff-deposits, and under such
circumstances that it may be regarded as having been produced within the
rock-mass. An example is afforded in the case of a basaltic flow near
Soni-soni Island, which is fully described on page 92. Whilst the lower
part of this flow is composed of a hemicrystalline basalt with scanty
olivine, the upper portion is made of a basaltic glass which has been
broken up or crushed “in situ,” the spaces between the fragments being
filled with palagonite. It would seem from the peculiar erosion of the
glass fragments that after the crushing a liquid magma occupied the
interspaces, and afterwards solidified and underwent the palagonitic
change.

[Illustration: MAGMA-LAKELET, ·25 mm. in size, magnified 290 diameters,
from a basalt at Navingiri. The groundmass, which is a smoky devitrified
glass containing abundant felspar-lathes, is coloured black. The
magma-lakelet is pale yellow in the slide and displays concentric lines
of congelation. It behaves like palagonite.]

In this connection it is noteworthy that in the sections of the lower
hemicrystalline portion of the flow there are shown in the groundmass
collections of a palagonitic material forming, as I have termed them,
“magma lakelets” of microscopic dimensions (·25 mm. in average size).
These “lakelets” are irregular in form, and are not uncommon amongst a
certain type of basaltic rocks. One of them is figured above; and it may
be added that they are best examined when displayed in a groundmass
containing much smoky, partly devitrified, glass. They are usually more
or less opaque and reddish-brown or yellowish in colour, whilst they
have often a marked zoned structure, the concentric bands conforming to
the irregular contours of the lakelet. In the least affected stage the
zones show fibrous devitrification across their breadth, but as the
palagonitic change progresses the material becomes opaque. In the
secondary changes, such as those associated with the early alteration of
the propylites, these “magma lakelets” are the first affected. They then
present alternating layers of calcite and viridite and are often
bordered by magnetite.

If these “lakelets” were to be described as collections of residual
glass, the description would be insufficient, since they may occur in
the midst of a smoky, partially devitrified, glass. During the last
stage in the consolidation of the basaltic mass, the magma-residuum that
still retains its fluidity collects here and there in the crevices of
the groundmass, and forms little pools of usually microscopic dimensions
into which the felspar-lathes often protrude from the sides. These
little pools or lakelets represent that portion of the yet fluid magma
that during the last stage of consolidation is imprisoned in the
stiffening mass—like the whey in a cheese—whilst the greater part of it
has been squeezed into the cracks of the cooling mass, as occurs in a
dyke-like intrusion below described, or has been extruded on its
surface, as in the case of the basaltic flow above referred to.

As a suggestive instance of the formation of palagonite “in situ,” I
will now refer to a basic tuff-agglomerate on the plateau of Na Savu
(see p. 81) which is penetrated by veins, a few inches thick, apparently
composed of a finely brecciated pitchstone-tuff. In the section the
material forming the veins is seen to be composed of fragments of basic
glass (carrying porphyritic plagioclase and augite) which have been
crushed in position, the interspaces being filled up with the finer
debris of the glass and of the minerals together with palagonitic
material. The glass fragments, which have lost their sharp edges and
angles, are often palagonitised at the borders, and we thus get a patch
of isotropic brown glass with a yellowish margin formed of a feebly
refractive turbid substance. Where this border is not so evident, it is
noticed that the edge of the glass is peculiarly eroded. The indication
appears to be that the fissures in this agglomerate were filled with a
basic magma that after its solidification into a glass was subjected to
a crushing process, and that during this process a partial remelting of
the glass took place which resulted in the molecular change
characteristic of palagonite. Since the unaltered glass-fragments fuse
in the ordinary flame, it would seem that the heat developed during the
crushing might be sufficient to partially remelt the glass without
affecting the rock penetrated by the veins.... It is of importance to
note that in the palagonite-tuffs of the Canary Islands the change is
often most complete along fissures, which thus present the appearance of
being occupied by veins of pitchstone.[127]

In this connection allusion may be made to a dyke-like mass of a rubbly
semi-vitreous basaltic rock exposed at Vatu-lele Bay, described on page
184. It is penetrated in all directions by veins, 1 to 3 inches thick,
of a tachylytic glass which begins to fuse in the ordinary flame. The
glass is traversed by cracks which sometimes contain palagonite. The
basalt, penetrated by the veins, has a smoky groundmass displaying
imperfect felspar-lathes in a feebly refractive glassy base and
containing a few small “magma-lakelets” that cannot be distinguished
from palagonite.[128]

Near the mouth of the Narengali valley (see page 149) I found what
appears to be a palagonite-tuff overlain by agglomerates formed of
tachylytic pitchstone and of semi-vitreous amygdaloidal basalts. The
tuff consists of fragments of a brown basic glass, the larger 1 to 2
millimetres in size, carrying porphyritic plagioclase, and fractured in
position, the interspaces being filled with palagonite. The glass
fragments possess the eroded margins indicated in the accompanying
figure. It may be remarked that this type of tuff differs from that of
the prevailing palagonite-tuffs in being rarely vacuolar, in the absence
of marine organic remains, and in its homogeneous composition. It is
described on page 334 under the head of “crush-tuffs.” Whether it is
derived from the destruction of a mass of basic glass that had
previously undergone crushing and partial palagonisation I cannot say;
but its characters point in the direction of this conclusion.[129]

In the foregoing pages it has been attempted to show that
palagonitisation has taken place in the veins of basaltic glass
traversing in one case a basic tuff agglomerate and in another case an
intrusive basaltic mass, and that it has also occurred in the upper
vitreous portion of a basaltic flow and in the materials now composing a
so-called “crush-tuff.” In order to explain this group of facts I
venture to propose this theory.

In certain types of basaltic lava,[130] when cooling and consolidation
take place under peculiar conditions, such as we would expect to find in
submarine eruptions, there is a residuum of the magma with relatively
low fusibility that remains fluid after general solidification of the
mass is well advanced. As the rock continues to consolidate, portions of
this magma residuum become imprisoned in the mass, like whey in a
cheese, giving rise to the “magma lakelets” above described; whilst
other portions, during the contraction and fissuring accompanying the
cooling process, are squeezed out into the cracks thus formed, or are
intruded on the surface of the consolidating mass, as in the case of a
submarine lava-flow. This solidified magma-residuum differs from the
ordinary basic glass not only in its lower degree of fusibility but in
its mineral composition and in its molecular condition. It probably in
the first place does not differ much in appearance from the typical
glass, but it is an unstable substance and is capable under certain
hydro-chemical conditions of developing the characters of palagonite.

[Illustration: Showing fragments of glass with eroded borders and of
plagioclase with more even edges in a matrix of palagonite traversed by
cracks. The length of the largest fragment is half a millimetre. The
glass has been evidently fractured in position and this is true of one
of the felspar fragments. It is also apparent that whatever its cause
the erosion of the margins of the glass has been produced since the
fracture.]

In those cases where the occurrence of palagonite is associated with
evidence of crushing, the process appears to be in a sense reversed,
since partial palagonitisation of an ordinary basic glass takes place as
a result of the elevation of temperature due to the crushing. The heat
thus developed is sufficient to partly fuse the glass; but since it is
not great, it only affects the most fusible constituents, and the
remelted material corresponds therefore to the magma-residuum of the
consolidating mass, which is referred to in the previous paragraph. It
has the same unstable characters and the same tendency to assume the
palagonitic condition.

This theory centres around the relatively low fusibility of the
magma-residuum that gives rise to palagonite. This degree of fusibility
has yet to be ascertained, since according to the views here advanced it
may even be much lower than that of tachylyte. It is, however,
noteworthy that the melting-point of tachylyte is far below that of the
more crystalline basaltic rocks, since it can be readily determined, as
I have done in the instance of a dyke-like mass penetrated by
tachylyte-veins before referred to, that the veins are composed of a
much more fusible material than the rock-mass. From a very crude
experiment I would infer that the melting-point of ordinary tachylyte is
not much above that of lead (335° C). The fusion-point of an ordinary
hemicrystalline basalt, according to the well-known experiments on the
lavas of Vesuvius and Etna, would probably be over 1,000 degrees C.

Two interesting experiments, the one artificial, the other natural, may
be here cited in connection with this view. Bunsen[131] more than half a
century ago, as a result of some experiments in which he produced
palagonite, arrived at the conclusion that the tuffs formed of this
material are submarine deposits derived from the breaking up of
previously formed palagonite-masses. Having obtained this substance by
placing powdered basalt in an excess of melted potash-hydrate
(Kalihydrat) and then adding water to the silicate of potash thus
formed, he concluded that palagonite results from the reaction between
glowing augitic-lavas and rocks rich in lime and other alkalies.
Although Zirkel quotes in this connection the example where this
material has been produced in the Cape de Verde Islands by basic lava
flowing over limestone, he rejects Bunsen’s explanation as inapplicable
to extensive palagonite districts, such as occur in Iceland, though
allowing that it would account for the local production of this
substance.

I venture to think, however, that in these two experiments the general
principles involved in the production of palagonite are partly
illustrated. We may accept the results of an experiment without
acquiescing in its interpretation. As I take it, it is in the partial
wet fusion of the powdered basalt that the secret of this successful
production of palagonite lies. In both these experiments some of the
conditions of a submarine flow have been reproduced.

Whilst Rosenbusch established the true character of palagonite as the
product of a peculiar alteration of a basic glass, Renard pointed out
the conditions under which it was most typically and in greatest
abundance formed. But Bunsen was happy in his suggestion that
palagonite-tuffs are submarine deposits derived from the breaking up of
previously formed palagonite masses. The question thus resolves itself
into one concerning the conditions of submarine eruptions and the
behaviour during consolidation of a submarine basaltic flow. In the
nature of things the field of investigation is mainly restricted to the
examination of ancient submarine basaltic flows that have been raised
above the sea.

A remarkable series of beds exposed in a stream-course below the Nandua
tea-estate may be here described in connection with the question of the
origin of palagonite formations. As observed on page 86, this locality
lies on the flanks of a basaltic plateau, which are incrusted with
recent submarine deposits. A pteropod-ooze, containing also the tests of
large and small foraminifera and the shells of small bivalves, is
displayed on the sides of the stream-course for the first 150 feet of
the descent. Below this, as shown in the diagram, is a declivity with a
drop of 60 or 70 feet where there is a waterfall. Horizontal beds of the
pteropod-ooze rock are exposed in the upper-third of this declivity; but
below, they pass into a chocolate-coloured marl-like deposit also
horizontally bedded, and sometimes having a banded appearance from the
alternation of layers of different degrees of fineness. This rock
contains 5 or 6 per cent. of carbonate of lime and incloses a few
scattered tests of minute foraminifera of the “Globigerina” and
“Nodosaria” types. In the slide the rock appears to be of massive
palagonite inclosing a few felspar-lathes ·1 to ·3 mm. in length, and
exhibiting a zeolite and calcite in the crevices and cracks. But it was
not until I had discovered the tests of the foraminifera and had
observed some fragments of larger crystals of plagioclase and a little
detritus of a semi-vitreous basaltic rock that its clastic character was
disclosed. The palagonite change has here to a great extent disguised
the character of the deposit.

This palagonite-marl formation is 20 or 30 feet in thickness. It passed
downward into a reddish-brown rubbly unstratified rock which falls to
pieces in one’s hands, breaking up into little cube-like masses an inch
or two across. These masses display in their interior a radiate
prismatic structure; but after drying they crumble into small fragments
exhibiting the same minute prismatic structure, the miniature prisms
being about a millimetre in diameter. Fine cracks, filled with calcite
and a zeolite, traverse this rock in all directions, and no doubt this
peculiar structure arises from shrinkage.

My idea that I was dealing with a clay-rock affected by the proximity of
an igneous intrusion was dispelled when the powdered material presented
itself as pure palagonite with scarcely any mineral fragments. Unlike
the marl above, it does not effervesce with an acid; and appears as a
mass of compacted minute fragments of basic glass converted into
palagonite, which is seemingly non-vacuolar, and containing about 15 per
cent. of water.

In connection with the diagram it should be remarked that I did not find
the palagonite-rock actually passing down into the basalt which,
however, is exposed in the river-bed below. The whole district is
characterised by columnar basalt, and the series of deposits here
described have been formed on the flank of the great basaltic table-land
of Wainunu. It is noteworthy that in the uppermost deposits of the
pteropod-ooze palagonite forms a noticeable proportion (10-20 per cent.)
of the residue; and perhaps most of the fine clayey material is thus
derived. As noted on page 321, minute pellets of pure palagonite are not
infrequent in the residue. Probably about 90 per cent. of the underlying
marl consists of palagonite. In the lowest palagonite-rock the
proportion would be quite 98 per cent.

[Illustration: Diagram showing the succession of deposits below the
Nandua tea-estate. The total thickness is about 250 feet. The figures
refer to the proportion of palagonite.]

Whilst it is apparent that we have represented in this series the
covering of a submarine basaltic flow with submarine deposits, it is
also evident that the mode of junction between the flow and the
overlying deposits is of an unexpected nature. Before drawing any
inferences, it is necessary to point out that when we begin on _à
priori_ grounds to frame our notions as to the course of events on the
surface of a submarine basaltic flow, we are entering a little known
region of inquiry. I would, however, suggest in the light of the theory
before advanced, the following explanation of the appearances presented
by this series.

During the consolidation of the flow much of the magma-residuum that
still retained its fluidity was extruded on the surface, where after
solidification it became palagonitised. According to my view this would
be the typical behaviour of submarine basaltic flows; but, owing to the
unstable and perishable nature of the palagonitic crust of the flow, it
would be rarely preserved in upheaved volcanic regions. There would
probably be, as in the case of the Nandua series, no sharp line to be
drawn between the palagonite-crust and the deposits subsequently
covering it, deposits indeed that would derive no inconsiderable
proportion of their materials from the disintegration of the crust
itself. During and after the emergence of such a district of submarine
eruptions the unstable palagonitic crust would be further subjected to
the hydration resulting from weathering and similar agencies; and as a
result of its final degradation there would often alone remain a bed of
reddish argillaceous material.

In concluding these remarks on palagonite the following summary of the
principal points here dwelt upon may be added:

(_a_) The basic glass, that undergoes the palagonitic change, is the
vitreous form of the magma-residuum that in a particular type of basalt
and under certain conditions remains fluid after the mass of the rock
has solidified. During the last stage of the consolidation it is in part
imprisoned in the “magma-lakelets” of the groundmass; whilst the rest of
it is squeezed into cracks and fissures, or extruded on the surface of
the flow.

(_b_) This glass differs from ordinary basic glass in its molecular
condition, its mineral composition, its low degree of fusibility, and in
its unstable character.

(_c_) The formation of palagonite in connection with the crushing of a
basic glass is to be explained by the hypothesis that the heat developed
during the crushing is sufficient to partially re-fuse the glass, the
material thus produced corresponding to the magma-residuum of low degree
of fusibility, which is above referred to.

(_d_) In submarine eruptions are to be found the conditions favouring
the production of palagonite on a large scale. In the case of such
basaltic flows it is probable that their upper portions are formed
entirely of palagonite arising from the alteration of a vitreous
magma-residuum extruded on the surface in the manner above described.
Such a crust, as a result of shrinkage and other processes, would
probably present itself to the geologist as a somewhat friable material,
passing gradually into the overlying submarine deposits.

                  *       *       *       *       *


     _Note on the type of basalt found associated with palagonite._

The type is characterised, it would appear, rather by its structural
features than by its mineral composition. It is the basalt of ophitic or
semi-ophitic habit that would seem to be usually associated with
palagonite; and since this habit is as a rule to be found where the
groundmass displays large felspar-lathes in plexus arrangement, coarse
augites, and at least a fair amount of smoky glass, it follows that a
hemi-crystalline, ophitic or semi-ophitic, doleritic basalt is the type
to be associated with palagonite.

This is the type of rock that forms the lower part of the basaltic flow
near Kiombo, the upper part of which is largely palagonitic. To this
structural type also belong most of the basalts in my collection where
palagonite exists in the form of “magma-lakelets” in the groundmass.
These “lakelets” are almost diagnostic of this type of basalt. Here also
belongs the famous globular basalt of Acicastello on the coast of
Sicily.[132] In such rocks the felspar-lathes form a mesh-work and vary
usually in average length between ·1 and ·3 mm. The augites of the
groundmass, typically semi-ophitic, range up to ·1 mm. in size. They are
always large, that is, over ·03 mm., and this coarseness is another
important indication.


   NOTE ON THE CHANGES PRODUCED THROUGH THE HYDRATION OF PALAGONITE.

Most of that which is detailed below is not according to my views
palagonitisation, but the effect of hydration in the disintegration of
this material. The initial molecular condition and the other characters
which represent potentially the palagonitic change are not connected
with hydration; but are concerned with the causes before explained that
led to the formation of a basic glass of such an unstable constitution.
Indeed, there is good reason to believe that the changes to be now
described may be observed under the ordinary influences of weathering in
a wet region.

The early stages of alteration are well displayed in some of the tuffs
formed mainly of basic vacuolar glass, the submarine character of which
is often indicated by a few tests of foraminifera. Whilst the glass
retains its original bottle-green colour, it loses the clean sharp
conchoidal edges and displays rough and uneven or granular borders. With
a high power the surface of the fragment is seen to be minutely pitted
or pock-marked in places, the shallow circular pits, less than ·01 mm.
in diameter, being sometimes arranged in a row like a number of
overlapping rain-prints. This process proceeds until all the surface is
affected, and from this cause there is often an appearance of polygonal
markings. The pock-marking, however, continues; and as the pits encroach
more and more on each other an irregularly wrinkled rough surface
results. Up to this time the glass has retained much of its original
colour; but its clearness is replaced by turbidity, and collections of
very minute rounded, rod-shaped, and irregular granules, composed of a
colourless feebly polarising material, are displayed here and there in
its substance, whilst some of the previously empty vacuoles are now
filled with water.

In the next stage the hydration of the iron-oxides begins, and the glass
becomes opaque and yellowish or reddish-brown, and has a more granular
appearance, polarising feebly. Cracks now traverse the substance, and
penetrate into the vacuoles, which, as they become filled with the
alteration products, whether palagonitic, zeolitic, or siliceous, become
ruptured and curiously distorted. The hydration and consequent
disintegration continue until the deep stain of the iron-oxide is
removed, and a semi-pulverulent whitish material remains. This is the
history of the little bleached powdery patches so common in basic tuffs,
each representing originally a lapillus of basic pumice. This powder
when examined with the microscope is shown to be made up of fine
granular and tubercular materials which lose much of their distinctness
when mounted in Canada balsam. It is not affected by boiling in HCl, and
contains usually an abundance of minute siliceous oval amygdules that
have been freed in the last stage of the disintegration of the
palagonite.

Such is the story of the degradation of the palagonite daily in
operation in the basic tuffs of this island. From this source is
doubtless derived much of the finest constituents of the submarine clays
so common over Vanua Levu.


_Supplementary note on the occurrence of palagonite in the glassy matrix
of pitchstone-agglomerates and in rubbly pitchstones._—In my last
revision of the proofs I find that I have not laid sufficient stress on
the production of palagonite under these conditions. The evidence of
crushing is often very evident, and especial references to this point
will be found in the index under “Pitchstone,” and on page 334 under
“Crush-tuffs.”



                              CHAPTER XXV

                      SILICIFIED CORALS AND FLINTS


SILICIFIED corals, together with siliceous minerals (quartz, chalcedony,
jasper, &c.) and siliceous concretions are evidently widely distributed
in these islands. Kleinschmidt in his journal refers to large blocks of
flint on the island of Ono, from which the natives used to obtain their
musket-flints,[133] and he collected from this island as well as from
Viti Levu, Ovalau, &c., numerous specimens of these and other siliceous
minerals and rocks, such as hornstone, chalcedony and jasper, which were
examined by Wichmann and described in his paper.[134] Mr. Andrews
observed silicified corals on the summits and higher slopes of Vanua
Mbalavu.[135] The Fijian name for flints, “ngiwa” (thunderbolt) or
“vatu-ngiwa” (stone-thunderbolt), affords a good instance of that
curious superstition connected with the origin of these stones, which
came also under my notice in the Solomon Islands,[136] and in fact is
widely spread.

In Vanua Levu these siliceous rocks and minerals are in places abundant.
They are especially frequent on the surface of the extensive low plains
on the north side of the island which constitute the basins of the
Sarawanga, Ndreketi, Wailevu, and Lambasa rivers; but it is in the
low-lying district of Kalikoso, in the north-eastern part, that they
exist in the greatest quantity. They do not occur usually at greater
elevations than 300 feet, and are found as a rule at much lower levels.

It must be understood that reference is not here made to quartz-veins,
such as are found in certain localities and of which mention is made on
pages 106, 116. It is not with the ordinary products of contact or
general metamorphism that we have here to deal; but with the remarkable
surface-collections of silicified corals, nodules and flints of
chalcedony, fragments of white quartz-rock, bits of jasper, and certain
curious siliceous concretions, that occur often in association with
fragments of limonite in these low-lying regions. All the siliceous
materials above named have, as the microscope indicates, a common
character, chalcedonic silica in a greater or less degree being the
basis of all of them, whether coral, flint, white quartz-rock, or
jasper. It soon became apparent whilst examining these districts that
one general condition prevailed whilst this extensive deposition of
silica and the formation of the beds of limonite were in progress. It
cannot, however, be pretended that these processes are actually in
operation on the plains now. Except in the case of the limonite in a few
localities the processes have been suspended; but they were in active
operation not long ago: and an examination of the general characters of
the districts will probably disclose some of the conditions under which
these products have been formed.

On the surface of the Kalikoso plains, where these materials are most
abundant, we find silicified corals associated with fragments and
nodules of chalcedony, flints, white quartz-rock, limonite, concretions
of carbonate of iron, &c., in the low-lying and often swampy district
around the fresh-water lake, the whole region being only elevated
between 20 and 60 feet above the sea. This is an area of decomposing
acid rocks (quartz-porphyries, trachytes).[137] On the other hand in
most of the regions where these materials occur on the surface we have
areas of basic rocks (basalts and basaltic andesites) incrusted in
places with submarine tuffs and foraminiferous clays, the volcanic rocks
undergoing extensive disintegration. Such for instance are the Lekutu,
Sarawanga, Ndreketi, and Lambasa plains. In the Lambasa plains, which
are described in this connection on page 139, we find besides the corals
and flints and nodules of chalcedony, fragments of jasper. In the
Sarawanga and Lekutu lowlands, we find silicified corals and limonite;
but here the crystallised silica of the corals contains a large quantity
of water, whilst in its lesser degree of hardness and in its low
specific gravity it comes near to semi-opal. In these and other
localities, as in the level country around Ndranimako on the right side
of the Yanawai estuary, we find curious concretions of the same kind of
hydrous silica more or less crystalline. These concretions are described
below.

It may be remarked that nearly all the districts in which the silicified
corals and concretions, siliceous minerals, and limonite occur, are
scantily vegetated “talasinga” lands[138] with reddish soil. Except in
the instance of the Kalikoso plains, the swamps and lakes have as a rule
long since disappeared, their sites being alone indicated by the
limonite on the surface. In the Mbua plains, however, there are
occasional small ponds and swamps, and there is no doubt that the
limonite so bountifully represented on the dry districts is still in
process of formation.

Before drawing some general inferences as to the conditions under which
this deposition of silica and iron took place, I will refer to the
characters of the materials thus produced.

The silicified corals include massive corals of the Astraean and
“Porites” kinds and branching specimens of the Madrepore type or habit.
The former are rarely larger than 7 or 8 inches across and are merely
fragments. The latter are always portions of branches, never exceeding 3
or 4 inches in length. In the last case it is sometimes possible to
show, as in the case of a specimen found on the Kalikoso plains, that
before silicification occurred the dead fragment of branching coral had
been extensively eroded by solvent agencies and had been penetrated by
burrowing molluscs. The larger blocks of massive corals have usually
been extensively chipped by the natives in obtaining flints. In past
times they were carried from one place to another, the result being that
occasionally they were brought to me in the mountain-villages, all
showing evidence of their having supplied flints to a past generation.

These corals are as a rule completely silicified. When a massive
specimen is broken across it is not infrequently found that whilst the
coral structure is preserved in its outer part, the inner portion is
composed of a compact seemingly structureless mass of bluish-white or
pale-grey flint, which has the characteristic microscopical appearance
of chalcedony and a specific gravity of 2·59.[139] It is from the more
compact parts of the silicified massive corals that the “worked” flints
found on the surface were obtained, though in some of them, as in the
case of a “scraper” in my collection, the traces of coral structure are
still apparent to the eye. Wichmann observed in the case of the
silicified corals from Fiji that the whole petrifying process appears to
consist in the saturation of the coral with silica, the coral structure
being usually distinct, whilst the septa, often still calcitic, show the
points of the calcite crystals projecting into the chalcedony which
forms the mass. Lime however rarely occurs in the silicified corals of
Vanua Levu. It was only in the case of one or two localities that the
corals displayed any effervescence with an acid. In the microscope slide
the massive specimens appear to be entirely of chalcedonic silica, the
outlines of the cells and of the septa being indicated by ferruginous
material. In a specimen of Porites by my side the crystallization of the
silica has advanced beyond the chalcedonic stage and the coral is
composed entirely of minute quartz-crystals, ·2 to ·4 mm. in size, often
irregular, but sometimes forming doubly-terminated prisms. This has
produced a somewhat crumbling rock, which is easily powdered by the
finger; and in this case, therefore, the complete crystallization of the
silica is resulting in the disintegration of the silicified coral.

The ordinary silicified massive corals of Vanua Levu, where the
replacement by chalcedonic silica is complete, though the structure is
preserved, have a hardness of about 6 and a specific gravity of 2·54,
and yield but little water in the closed tube. Occasionally, however, as
in the Sarawanga plains and in the Lekutu lowlands we find silicified
fragments of branching corals which are easily scratched with a knife
and have a hardness of 3 to 4 and a specific gravity of 2·3. The
fractured surface is milk-white or reddish, and looks like semi-opal.
When powdered and heated in a closed tube, the material loses one fourth
or one fifth of its weight of water, the finest dust (passing away in
the steam) being deposited on the sides of the glass. In the slide there
is displayed a finely granular crypto-crystalline structure with in
places a somewhat coarser quartz-mosaic, whilst chalcedonic quartz fills
minute cracks in the mass. No coral structure is preserved. Numerous
points coloured by iron oxide occur in the section, and minute dust-like
inclusions abound, which are doubtless water-pores. I have described on
a later page certain concretions found associated with these silicified
corals which though formed of the same crypto-crystalline hydrous
silica, are apparently silicified portions of nullipore-rocks.

The fragments of flint that occur commonly on the surface in these
districts are, as above remarked, derived from the hard silicified
coral-masses. Nodules of chalcedony, having all the appearance of having
originated in cavities, are also very frequent. They may take the
mamillary, agate, or onyx form, some of the agates when polished making
beautiful specimens. These nodules are of all sizes up to 3 or 4 inches
across. Some of them are hollow and lined with clear quartz-crystals,
whilst with others the cavity may be completely filled by interlocking
quartz-crystals. The outer surface of one of the agates displays
markings showing in relief casts of the “cups” of a minute-celled coral.

Mingled with the other siliceous materials on the surface of the
Kalikoso and Lambasa plains are found fragments of a whitish
quartz-rock, having a specific gravity of 2·53-2·57, being therefore
markedly lighter than quartzite (2·63-2·67) which it somewhat resembles.
It usually occurs as small hand-specimens; but in the vicinity of
Mbati-ni-kama I found blocks, 12 to 15 inches across, lying in the
river-bed. Under the microscope it displays a fine radio-globular
aggregate of chalcedonic quartz.

Mention has already been made of the siliceous concretions, composed
mainly of hydrous crypto-crystalline silica, which are associated with
the silicified coral fragments formed of the same kind of silica on the
surface of the plains of Mbua, Lekutu, and Sarawanga. They also occur in
the Ndranimako lowlands on the right side of the Yanawai estuary, and in
the more elevated inland districts of the Wainunu and Na Savu
table-lands at elevations of 650 to 770 feet above the sea. They take
the form of irregular nodules, or of flat uneven “cakes,” usually two or
three inches in size. They are as a rule reddish, but sometimes pink and
white. Their hardness is only 3 to 4, and they are easily scratched with
a knife; and when powdered and heated in a closed tube, they lose about
one fourth of their weight of water. Under the microscope they exhibit a
grey crypto-crystalline groundmass showing very finely granular
crystalline silica with the cracks and small cavities filled with more
brightly polarising chalcedonic quartz. But they differ as regards their
other components and also in their mode of occurrence; and it is highly
probable that the history of their origin is not always the same.

Those associated with the silicified corals on the Sarawanga and Lekutu
lowlands show no structure in the slide that gives me a clue as to their
origin; but they may perhaps represent old Nullipore nodules. Those
around Ndranimako are coloured deep red; and whilst some give no
indication as to their source, others are transitional in character, and
display in the sections traces of the vacuolar semi-vitreous basic rock
of which the original fragment was composed. The same red siliceous
concretions form the pebbles and gravel in the stream-beds on the
surface of the Na Savu table-land, 700 feet above the sea. These red
flint-like nodules of Ndranimako and Na Savu somewhat resemble the
jasper of the island; but they are sharply distinguished by their
microscopic characters, by being easily scratched with a knife, and by
the large amount of water which they contain. Rolled stones, which were
found in the shallow stream-courses on the surface of the Wainunu
table-land 750 to 800 feet above the sea, exhibit in the sections, in
spite of the general silicification of the groundmass, the outlines of
the original phenocrysts of felspar, and abundant skeletal magnetite
rods, such as would characterise a semi-vitreous basic rock. It is
evident that in the basaltic districts of the Na Savu and Wainunu
table-lands these concretions have been formed under certain conditions
by the decomposition of the silicates of basic rocks. But these
conditions do not exist now; and I infer that the silicified rocks,
which occur only in fragments on the surface, represent the
silicification that occurred during the emergence of the land ages
since.

Occasionally one comes upon in the mountain districts, as in the
vicinity of Ndrawa, large solitary blocks 2 to 4 feet across of a
whitish chert-like rock which has a hardness of 5 or 6, the harder
variety having a specific gravity of about 2·58 and the softer, which
yields a fair amount of water, a specific gravity of about 2·46. I
noticed such solitary masses also on the Mbua plains. The first-named
locality is dacitic and the last basaltic. They exhibit in the slides a
patchy appearance, showing in some places finely granular
crypto-crystalline silica and in others a coarser mosaic of chalcedonic
quartz. Apart from the absence of any definite coral structure, I can
only surmise that they were originally masses of reef-limestone. Their
elevation even in the mountainous districts was not over 400 or 500 feet
above the sea.

Fragments of jasper, which are associated with nodules of chalcedony and
silicified corals in the Lambasa plains, are also to be found as pebbles
and small blocks in the mountain streams of the Ndrawa, Ndrandramea, and
Lea districts, together with bits of chalcedony and quartz-crystals.
They do not occur, or are of rare occurrence, in the recently emerged
Kalikoso district and probably belong to an earlier stage in the history
of the island’s emergence from the sea. They have a hardness of 6 to 7,
not being scratched by a knife, and a specific gravity of 2·65 to 2·70;
whilst but little water is given off in the closed tube. They are a
variety of chalcedony, rendered opaque by the large quantity of red
oxide of iron that it contains, and are really, therefore, iron-flints.
The microscopical section in one case displays in the clear spaces a
beautiful globular aggregate, each globule having a nucleus of the iron
oxide and giving a black cross in polarised light. In another case the
globular structure is less perfect, and the chalcedonic groundmass is
penetrated by a multitude of fine cracks filled with iron oxide.

The deposits of limonite vary in character in different localities, and
evidently they have not all the same history. The soil of the low-lying
plains around Wai-ni-koro and Kalikoso, and especially in the vicinity
of the fresh-water lake, is often coloured a deep ochreous red. Small
fragments of an earthy yellowish-brown limonite occur on the surface in
quantity and are particularly abundant near the lake. They yield much
water when heated. In some places in this district, as in the country
traversed between Wai-ni-koro and Kalikoso, the surface is strewn with a
number of small round concretions of the size of small marbles (6 to
12 mm.) which are composed of a mixture of carbonate of iron and
limonite, but show no recognisable structures. They are somewhat friable
and give off much water when heated, whilst they effervesce freely in
hot hydrochloric acid. It is probable that some of the earthy limonite
of the Kalikoso district contained originally iron carbonate and has
been produced from concretions such as I have just described.

The variety of limonite found in fragments on the surface of the plains
of Mbua, Lekutu, and Sarawanga, at elevations usually of 100 or 200 feet
above the sea, is a heavy compact kind with a specific gravity of 3 to
3·5, and closely resembling red hematite. Since, however, it is lighter
in weight and still contains a little water, it may be regarded as in
the transition stage. It occurs as portions of cake-like masses varying
usually from a third of an inch to rather over an inch in thickness. As
a rule it is found in localities where no lakes or swamps now exist and
may be associated, as in the Sarawanga and Lekutu plains, with
silicified corals and siliceous concretions; but in some cases, as in
that of the Mbua plains, ponds and swamps are still scantily represented
in the vicinity, and the water of the stagnant streams is deeply
coloured with iron (see page 56).

Ironstone gravel occurs in great quantity strewn over the surface of the
basaltic table-lands, especially in the case of that between the Wainunu
and the Yanawai rivers. The smaller gravel varies usually between one
eighth and one third of an inch in size, the larger fragments being
about an inch. The specific gravity is 3·1 to 3·2. The material forming
the finer gravel dissolves with but little effervescence and scanty
residue in hot hydrochloric acid; it gives off water and is evidently
impure limonite. The larger fragments, 1 to 2 inches in size, represent
the partial conversion into limonite of a basic volcanic rock with much
glass in the groundmass which formed probably the surface of the
basaltic flows of the plateaux. There must be an enormous amount of this
iron-stone in the island. The finer gravel has a concretionary
character, some of the pieces appearing like bits of stick that have
been converted into limonite. It seems to have been formed during the
disintegration of the rock on the moist surface of these densely wooded
basaltic plateaux; the process was not accomplished in ponds or swamps,
but was carried out on ordinary damp ground.

It must be observed in the above connection that the soil in the areas
of basalt and basaltic andesites, which occupy a large portion of the
surface of the island, contains a large amount of fine magnetic
iron-sand. After heavy rains the foot paths glisten with this fine
material which has been washed out on the ground. This is especially the
case in the extensive scantily vegetated “talasinga” regions where the
basaltic rocks are disintegrating for a considerable depth. The
river-sand of these areas, after a little washing, yields about 75 per
cent. of magnetic iron grains which give in some cases a slight titanium
reaction. The amount of magnetic iron-sand in these rivers, as for
instance in the Yanawai and the Wainunu, must be very great. In the beds
of the small sluggish streams on the surface of the Wainunu table-land
the amount is also very large.

Any explanation of the origin of the extensive silicification evidenced
by the occurrence of silicified corals and siliceous concretions on the
surface in various parts of the lower regions of the island will have to
include that of the formation of the limonite fragments so often
accompanying them. The necessary conditions would, I think, be afforded
by an emerging land-surface during the consolidation of the exposed
calcareous muds and the subsequent draining of the new surface. On parts
of the newly formed land, there would follow the successive stages of
sea-water, brackish, and fresh-water swamps, such as are clearly
indicated by the abundance of silicified coral fragments that strew the
surface of the low-lying and often swampy districts around the fresh
water lake of Kalikoso.

In such a locality as that of Kalikoso, there were no doubt at the time
of the emergence large tracts covered with chalky calcareous mud derived
from reef-debris; and it was during the consolidation of this mud in the
recently reclaimed area that the fragments of coral imbedded in it
became silicified. In these cases where the imbedded corals were already
much decayed, it is probable that the empty cavities thus produced were
filled with silica, and that in this manner the nodules of chalcedony
were produced. Here and there a pebble or a larger block of a volcanic
rock would have been inclosed in the mud; and in this case also silica
largely replaced the original material of the stone. I imagine that with
the evaporation of the water in the mud during the drying and
consolidating processes the proportion of silica in solution would
attain a degree of super-saturation and that the silicification would
hence be brought about.

With the consolidation of the mud the deposition of silica ceased; and
in the case of any coral fragments, where the transformation was not
completed, decay would often commence. In the instance of some bits of
coral found imbedded in foraminiferous mud-rock in the Lambasa plains
the process of the change had been suspended, and the fragments were in
a state of decay, and coloured red by iron oxide. If silicification
occurred in a submarine deposit only after it became a portion of an old
land-surface we ought not to find incompletely silicified corals
inclosed in it. For these reasons I do not consider that silicification
would occur in the case of submarine deposits long after they have been
raised above the sea.

On the other hand it would seem that the deposition of silica in the
hard parts of dead organisms does not proceed in the shallow-water
calcareous mud of coral reef coasts previous to emergence. Silicified
corals have never as far as I know been found under such conditions. Nor
could the coral fragments now lying on the Kalikoso plains, often only
elevated some 20 or 30 feet above the sea, have undergone this change
whilst exposed on the land-surface as they now lie. They must have been
inclosed in some material containing abundant free silica; and it is
reasonable to suppose that this material was the chalky mud of the
reef-flats on which they once lived. If this is admitted, then it
follows that since, as above assumed, silicification does not occur in
such a mud either before upheaval or long after it has been raised above
the sea, it must take place in the intermediate period, or in other
words whilst the recently exposed submarine deposits are consolidating
and drying.

Several objections at once occur with reference to this explanation of
the silicification of corals in this island; but much more investigation
is needed to establish any view on the subject. In the Kalikoso plains,
however, we have a critical locality for the pursuit of this inquiry.
Concretions of carbonate of iron and deposits of earthy limonite are
here associated with silicified corals on the surface of a level and
often swampy district around a freshwater lake in a region which is only
elevated 20 to 60 feet above the sea. We are dealing here with an area
of land that has emerged in comparatively recent times as far as the
history of the island is concerned. The element of time is limited, and
the problem is not complicated, as it would be in the case of an old
land-surface, raised some hundreds of feet above the sea, by the
intrusion of many other disturbing agencies. Nature has simplified
matters here for the inquirer.

The evidence of recent emergence with regard to the whole island is
discussed in Chapter II., and need not be again referred to here; whilst
the general description of the Kalikoso district is given in Chapter
XVI. In this connection it may be remarked that before their emergence
the Kalikoso plains were covered by the waters of a large irregular
sea-water lagoon or lake, which though more or less surrounded by hills
had free communication with the sea on the north along the line of the
passages now occupied by the Wai-ni-koro and Langa-langa rivers. Both
massive and branching corals then thrived in the waters of the lagoon.
There is no ground for supposing that during the emergence there was an
intermediate stage characterised by brine-ponds and salt-swamps. The
drainage from the slopes of the mountains to the southward would have
prevented it. Whilst this change of level was in operation, brackish
water collected in the deeper part of the original lagoon, forming a
lake which as evidenced by the present distribution of limonite on the
surface of the plains was then far larger than it is now. As the plains
became exposed large flats covered with chalky mud in which dead corals
were more or less imbedded were bared; and there and then as the drying
and consolidation proceeded silicification took place in the manner
before surmised. This deposit was of no great thickness, and has been
since removed by the denuding agencies, whilst the silicified corals
remain behind.

When in the Solomon Islands I was unable to find the source of the
chalcedonic worked flints of such frequent occurrence in that region. In
my general work on those islands (pp. 77 to 80) reference is made to
this subject. It will probably be shown that there as in Fiji most of
the flints are silicified corals.

In conclusion it may be remarked that those who object to the
explanation of the origin of silicified corals advanced in this chapter
will be able to find support for their alternative hypothesis in many
facts detailed in these pages. Vanua Levu, for instance, abounds in hot
springs; and Mr. Andrews might regard this fact as giving strength to
his view that the silicified corals of Vanua Mbalavu in this group owe
their condition to the agency of superheated water derived from volcanic
rocks, more especially since hot springs are found on the island. Such
an explanation could not, I think, apply to the extensive area of the
Kalikoso plains where the silicified corals are associated with limonite
on the surface of a recently emerged area. If these changes had been
induced by hydrothermal action, one ought to find evidence of this in
those localities in Vanua Levu where the hot springs issue from
foraminiferous clay deposits, as in the vicinity of Vuni-moli; but no
traces of such a transformation came under my notice. Wichmann does not
advance any explanation of the silicification of the corals; but he
considers that the “hornstones,” which he obtained from Fiji, rocks
corresponding to the chert-like rocks described by me on page 355, are
the products of disintegration of the basic andesites. I have already
pointed out that certain siliceous nodules have probably this origin. It
is also likely that some of the jasper of Vanua Levu has been thus
formed.


_Note on a silicified Fern Rhizome._—This is a specimen, about three
inches long, picked up by a native in a stream near Sueni in the centre
of the island. It has the appearance of being a portion of the stem or
rhizome of a tree-fern, and is permeated in its entirety by chalcedonic
quartz, the fibro-cellular structure being still preserved. No other
specimen of the kind came under my notice. The probability of the
occurrence of silicified plant-remains in the pumice-tuffs of the Undu
Promontory is pointed out on page 233.



                              CHAPTER XXVI

                             MAGNETIC ROCKS


THE literature on the subject of the magnetism of rocks is very
extensive,[140] and even if I was capable of doing so, any attempt to
deal generally with this complicated phenomenon would be out of place
here. Zirkel in his characteristically thorough fashion has reviewed the
subject in his general work on petrography, but since the date of the
last publication of that book, 1893-94, the literature has been much
increased and the subject has from time to time been opened up in
scientific periodicals, occasionally in ignorance of the labours of
those that have gone before. Here, the local magnetisation of rocks is
alone considered, the general question of earth magnetism not being
entered into.

According to Zirkel one of the earliest known observations of this
phenomenon was made by Bouguer, the French geographer, whilst he was
engaged in the measurement of a degree of the meridian in the vicinity
of Quito in 1742. Alexander von Humboldt, however, was one of the first
to attract general attention to this subject by the announcement of his
discovery in 1796 of a “great magnetic mountain” in the heart of
Germany. He was then director-general of the mines in two Franconian
principalities; and in order to awaken the interest of German physicists
and mineralogists in this matter, he announced his discovery with an air
of mystery, and did not disclose the locality for many months. He then
placed his specimens in the mining-office at Bayreuth to be sold at so
much by weight for the relief of poor miners. His plan succeeded, and
this young savant who had yet before him his great career, had soon
enlisted the interests of several of the noted scientific men in
Germany, including Werner the mineralogist, Voigt the mathematician,
Blumenbach the naturalist, Charpentier, and others. The amount of
attention that this subject then excited can be inferred from the pages
of the “Intelligenzblatt der Allgemeine Literatur-zeitung” for 1796-1797
and from the contemporary publications. It has been almost forgotten
now, and the matter is indeed often approached “de novo.”

However, although by these means the data became largely increased, no
generally accepted explanation resulted. Opposing views continued at
various times to be advanced; and it has only in recent years come to be
recognised that the magnetic polarity[141] of rocks in exposed
situations, as in the mountain-peak or in the crested spur, often arises
from atmospheric electricity independently of the inductive action of
terrestrial magnetism. This is the conclusion to which the later
evidence given by Zirkel is directed and was that which Oddone and Sella
formed from their study of the magnetic rocks of the Central Alps. It is
not, however, always necessary to suppose that the affected rocks have
been struck by lightning, although Sella and Folgheraiter have shown
that this is the result of such a contact. They may be found, as
indicated by Mr. Harker, in mountainous localities where thunder-storms
are remarkably rare, and where the peaks act, it is suggested, as
natural conductors. It is easy to show, remarks the same author, that no
lapse of time is required for rocks in exposed situations to become
magnetised. The stones of cairns erected a few years before on the
mountain-tops of the Isle of Skye become invariably highly magnetic;
whilst the loose stones lying on the ground display this property to a
much less degree. Nor is it requisite that the rocks affected should be
basic volcanic rocks. It has long been known that granites, trachytes,
&c., can possess magnetic polarity[142]; and the existence of this
quality among acid volcanic rocks is well shown in the case of the
dacites in Vanua Levu, rocks which compose some of the isolated
mountain-peaks.

One finds occasional reference to the highly magnetic character of the
rocks in oceanic islands of volcanic origin, but the nature of the
property is not always described; and it is sometimes not possible to
gather from the data given whether the magnetism affects the whole
mountain mass, when it would be of the regional kind, due probably to
induction, or whether it is the simple magnetic quality that almost all
basic volcanic rocks possess on account of the fine magnetite
disseminated through the rock, or whether there is evidence of a deposit
of magnetite in the vicinity, or whether it is a mere surface phenomenon
confined to the bare rocks of peaks and ridges, when such rocks, whether
gabbro, granite, basalt, trachyte, or dacite, display magnetic polarity.
Dana, with regard to the basaltic mountain of Tahiti, remarks that the
compass was often rendered useless by the local attraction of the rocks,
bearings taken being found to vary two to three points on changing the
position of the instrument.[143] Major Haig says that the compass
becomes perfectly useless anywhere in the neighbourhood of one of the
mountain-masses or extinct craters in Mauritius, and attributes this
effect to the magnetite in the basalt.[144]

On the summit of Mauna Loa in Hawaii, at the edge of the great crater
and in the vicinity of the site where Commodore Wilkes carried out his
pendulum observations in 1840, I found my compass-needle greatly
affected by local attraction, but I neglected to inquire further into
the matter. Judging from my sojourn of twenty-three days on this
mountain-top, thunder-storms are of very rare occurrence there; but the
electric condition of the air is at times very evident, and its
physiological effects are somewhat distressing. My blanket at night
crackled in my hands and emitted sparks, so that I could trace with my
finger the letter A in phosphorescent hues on its surface.

That lightning is directly responsible in some instances for the
magnetic polarity of rocks in mountain-peaks is also well established.
It has been illustrated in an indirect fashion only last year in the
disaster on the Wetterhorn. Rocks partially fused by thunderbolts and
displaying polarity occur on the summit of the Riffelhorn and on one of
the peaks of Monte Rosa, and fulgurites have been also obtained from
Mont Blanc.... It is not always easy to explain, however, isolated cases
of polaric rocks where no signs of fusion occur. Whilst descending into
the Valle del Bove from the Etna Observatory, I picked up four small
volcanic bombs of basic lava, of which one displayed polarity, the poles
being situated at the sides of the bomb. Zirkel quotes the observation
of Naumann on the summit of the volcano of Moryoshi in Japan. Here out
of a number of lava-blocks lying about only one exhibited marked
polarity, whilst the rest showed no signs of it.

Before dealing with the polaric rocks of Vanua Levu, I will refer to two
localities in other parts of the group where magnetic rocks have been
observed. During the Wilkes’ expedition in 1840,[145] Lieutenant
Underwood observed great local attraction at Naikovu, a rock 90 feet
high of volcanic formation lying off the south end of Nairai Island. He
found a “deviation” of 13¼ points (149 degrees) at the top of the rock,
whilst at the foot near the water the needle gave correct bearings. In
the _Sailing Directions for the Pacific Islands_, published in 1900, the
“deflection” at the summit is said to be 87 degrees. It is stated by Mr.
Eakle in his paper (quoted on p. 293) on the rocks collected by the
recent Agassiz expedition that this rock is composed of an
augite-andesite.... I have learned from Mr. Alex. Barrack that there are
some highly magnetic rocks on the west coast of Viti Levu in the
vicinity of Likuri Harbour in the Nandronga district. It is said that
specimens sent down to the colonies were found to contain 95 per cent.
of magnetite.

It is very probable that the results obtained by me for Vanua Levu can
be generally applied to the other large islands of the group. The
observations were made during my various geological journeys and deal
only with certain aspects of this interesting subject.

The first feature in this connection is the frequency with which simple
magnetism is displayed by the acid as well as basic volcanic rocks of
this island. About 95 per cent. of the volcanic rocks collected attract
both ends of the needle.[146] This property of volcanic rocks is well
known, and is to be attributed to the magnetite in the groundmass.[147]
On examining the character of the non-magnetic rocks it appears that
almost all belong to two groups where magnetite might be expected to be
scanty. The first includes the pitchstones or basic glasses, sometimes
fresh, at other times more or less palagonitised. The second comprises
the highly altered basic rocks, where the ferro-magnesian silicates have
been replaced by viridite, calcite, and pyrites. It is not, however, to
be implied that rocks of these two kinds will not sometimes attract the
needle. Many do not, and those in my collection that do so act feebly.

Coming to the magnetic polarity displayed by some of these rocks, when
the ordinary hand-specimen behaves like a magnet in attracting one pole
of the needle and repelling the other, it is to be at first observed
that a rock can become polaric without being previously magnetic. Dr.
Folgheraiter has observed polarity in the case of fragments of ancient
bricks and pottery; and he has described the same effect in the masonry
of a house struck by lightning. In one or two of the Vanua Levu acid
rocks showing polarity this can be also premised since magnetite is
present in very slight degree.

Polarity is very frequent among the volcanic rocks of this island. Out
of 520 specimens in my collection, which was made without any reference
to this matter, 80, or 15 per cent. are polaric. Of these seven-eighths
are basic and the rest are acid rocks; but this proportion is partly
accounted for by the far greater prevalence of basic rocks in the
island. The basic rocks showing polarity include some of the heaviest
olivine-basalts with a specific weight of 3·0, as well as some of the
lighter augite-andesites with specific weight of 2·7. They comprise the
coarse textured dolerite as well as the vitreous pitchstone and include
both scoriaceous and amygdaloidal rocks. The polaric acid rocks are
mostly referable to the dacites, with a specific gravity of 2·5 to 2·6.

Humboldt remarked long ago that there is no direct relation between the
degree of polarity and the specific weight. This is well brought out in
the table subjoined; but it should be at once observed that there is an
indirect relation. Although when we arrange the rocks in a series
according to their specific weight we find no corresponding relation in
the amounts of the polarity, we observe that the extent to which
polarity can be developed is markedly greater in basic than in acid
rocks. From this it may be inferred that the degree of intensity of the
exciting cause required to give polaric powers of a certain value to an
acid rock, like a dacite, would be much greater than that necessary to
endow a basalt with equal powers. We should not expect to find the same
amount of polarity in the bare rocky peaks of two adjacent mountains,
where one was of dacite and the other of basalt; and, other things being
equal, if two mountains had been exposed for ages to the same
conditions, we should regard the polaric powers of the two as nature’s
equivalent values for the work of atmospheric electricity, on the two
rocks in question. We have two such mountains in Vanua Levu in the case
of the adjacent peaks of Ngaingai (2,448 feet) and Navuningumu (1,931
feet) which are about 2¼ miles apart and possess similar bare rocky
pointed summits. I take it that the polaric power of 25° of the dacite
(sp. gr. 2·57) in the first case is equal to the power of 90° of the
basaltic andesite (sp. gr. 2·82) in the other. In the dacitic peak of
Ngaingai and in the basaltic peak of Navuningumu we can measure what
work atmospheric electricity can accomplish in the course of ages in the
magnetisation of rocks. The other conditions being taken as about the
same, the main determining difference is to be found in the
rock-characters.

In the table on the opposite page we have a series of volcanic rocks
placed according to their specific weights, which range from 2·5 to 3·0,
and in the second column are shown their relative polaric powers as
indicated by the number of degrees the north end of the magnetic needle
is repelled by the corresponding pole in the hand-specimen. For this
purpose a magnetic needle 2½ inches in length (strictly speaking 6·5
centimetres) was employed, a card marked in degrees being placed
beneath. The north pole of the stone was placed in contact with the
north end of the needle, and after the needle had become stationary in
its new position a reading was taken.

These polaric rocks came under my notice over most of the island. They
are infrequent in the district between Undu Point and the Wai-ni-koro
River, where, however, acid tuffs are largely exposed; and I did not
find them in the Natewa Peninsula east of Lea, their absence from my
collections made in the Mount Freeland range being remarkable. But it is
probable that this is due to the surface conditions, since dense wood
covers the slopes, and bare rocky peaks are rarely to be seen.

With regard to the influence of locality on the occurrence of polaric
rocks, the results may thus be classified. About one-third are found in
the exposed rocky peaks of hills and mountains. Another third are found
where the rocks are bared in headlands, coast cliffs, inland-bluffs,
ridge-tops, and in the open basaltic plains where trees are scanty. On
the other hand, a third occur in situations, as in wooded districts
where the rock exposure is scanty, when it is not easy to explain the
polarity, unless it was developed in clear districts that have since
become covered with forest.

    _Table showing the Relation between the Specific Gravity and the
                      Polarity of Volcanic Rocks._

 +---------------------+--------------------+------------------------+
 |Character of rock.   |  Specific gravity. |Amount of polarity.[148]|
 +---------------------+--------------------+------------------------+
 |                     |                    |                        |
 |Dark olivine-basalt  |         3·00       |          10°           |
 |Grey    "      "     |         2·94       |          29°           |
 |  "     "      "     |         2·92       |           7°           |
 |Dark    "      "     |         2·90       |          10°           |
 |  "     "      "     |         2·87       |          30°           |
 |Basaltic andesite    |         2·82       |          90°           |
 |   "        "        |         2·77       |           5°           |
 |Pyroxene-andesite    |         2·72       |          38°           |
 |Dacite               |         2·61       |          17°           |
 |  "                  |         2·59       |           5°           |
 |  "                  |         2·57       |          25°           |
 |  "                  |         2·50       |          14°           |
 +---------------------+--------------------+------------------------+

In no place did any evidence of the direct action of lightning come
under my notice. Mr. S. Skinner who kindly looked at a few of these
rocks says that he found no trace of fulgurites in them. It is probable
that here as in the mountains of Skye, as described by Mr. Harker, these
effects are the result of the general influence of atmospheric
electricity independently of the direct agency of lightning. The
frequency of polaric rocks in the highest peaks of the island is very
remarkable. Generally speaking, all the bare summits of the mountains
are polaric. In my experience there is no exception. All the rocks
obtained from the actual summits show polarity. The variety of rocks
thus affected is suggestive; and this chapter may be concluded with a
brief reference to their mode of occurrence on some of the
mountain-peaks.

In Mbatini, 3,437 feet in height, which is the highest mountain of Vanua
Levu, the pyroxene-andesite of which the bare rocky peak is composed is
somewhat weathered and has a polaric or repellent power of 28°.
Specimens of rock obtained below the top show no polarity, the mountain
being well wooded except at the summit. In the adjacent mountain of
Koro-mbasanga,[149] the polaric rocks are limited to those exposed in
the peak which is bared of vegetation. The rocks in question are
tuff-agglomerates, the small blocks of pyroxene-andesite standing out
from the tuff having a polaric force of 14° or 15°. This effect has been
produced in greatest intensity in the isolated peak of Navuningumu
(1,931 feet) in the Ndrandramea region. Here the bare summit is formed
of a semi-vitreous, slightly vesicular, basaltic andesite with a
specific gravity of 2·82 in its present condition.[150] This rock is
powerfully polaric, and rendered the compass useless, the deviation
generally to the westward varying from 20° to 50°. I place its repellent
force at about 90°, hand specimens affecting the magnetic needle at a
distance of 13 or 14 inches. None of the various rocks obtained from the
wooded slopes below displayed polarity.

The neighbouring mountain of Ngaingai is composed entirely of dacite
having a specific weight of 2·57. The highest point of the summit, 2,448
feet above the sea, is bare and rocky, and the stone here is markedly
polaric, the repellent force being about 25°. Specimens from the lower
wooded slopes show no polarity. Near by rises the hill of Ndrandramea,
which is composed in mass of acid andesites or dacitic rocks. The summit
(1,800 feet) is scantily vegetated, and here the somewhat weathered rock
which has a specific weight of 2·44 (probably near 2·5 in the fresh
condition) has a polaric force of 14°. Specimens of a more compact rock
taken from the wooded slopes 300 feet below the summit (sp. gr. 2·58)
and from 700 feet below the top (sp. gr. 2·68) showed no such effect;
but a specimen taken from a mass of agglomerate in the last locality
repels the needle 12°. Its specific gravity is 2·61, and no doubt the
mass had been originally a portion of an exposed cliff-face.[151]

The summit of Mariko (2,890 feet), the Drayton Peak of the chart, is
formed of a rubbly agglomerate of a compact basic andesite. Though it
displays bare rock-faces, the actual peak has a soil-cap at least 18
inches deep and supports small trees and shrubs. Notwithstanding this, I
found when standing on the peak that my compass was very noticeably
affected, the pull being to the eastward, whilst the amount of deviation
increased from 11° to 16° when changing from the sitting to the standing
position. Specimens of blocks from the agglomerate forming a rock-face
10 feet below the summit possessed polaric powers of 12° and 5°. Others
of the same rock exposed in a cliff-face 450 feet below had a weak
repellent power of only 4°.... As in the case of Mariko, the top of
Thambeyu (2,700 feet) is vegetated; and beneath the smaller trees blocks
of polaric rocks lie on the surface. One of these, a pyroxene-andesite
(sp. gr. 2·72), from which I obtained a specimen, has a polaric power of
38°. In another case, that of an amygdaloidal rock of the same
character, the repellent power is 14°.

I might mention several other polaric peaks, but it will be sufficient
to refer to one or two other localities. In the mountainous basaltic
district around Solevu Bay the peaks are usually polaric. Specimens from
the top of Uli-i-matua, 1,100 feet, have a repellent power of 15°. The
three-peaked hill of Koro-tolu-tolu appears to be in the mass of polaric
basalt from the foot to the summit, having a repellent power varying
from 4° to 30°, the most active specimens being obtained from the lower
slopes, which, however, are scantily covered with trees. Samples of the
grey basalt from Koro-i-rea show polaric powers of 3° to 7°.

As examples of the numerous lesser hills with bare rocky polaric summits
I will first take Bare-poll Hill facing Soni-soni Island. This hill is
only about 150 feet above the sea, its top being formed by two large
masses of a basic andesite lava with a glassy groundmass, incrusted with
agglomerate, the whole representing a volcanic “neck.” A specimen of the
rock masses has a repellent force of 22°. Another instance is afforded
by Vatui, a hill 450 feet in height situated south of Mount Sesaleka.
Its summit is capped by a naked mass of tuff-agglomerate pierced by a
dyke 18 inches thick of an olivine-basalt, with a specific gravity of
2·90 and a polaric power of 10°.

A somewhat suggestive example is afforded by the hill of Na Suva-suva,
1,110 feet high, which overlooks Naindi Harbour to the east of Savu-savu
Bay. It is only occupied by trees in its upper part, and a specimen of
the olivine-basalt, of which the hill is composed, that was obtained
from the wooded summit, shows no polarity; whilst another from the
slopes, two-thirds of the way up, which had been cleared of trees, has a
repellent force of 10°. The polarity of the olivine-basalt from the
well-wooded slopes of Ulu-i-ndali, a range 1,100 feet in height on the
east side of the Wainunu estuary, is not so easily explained; the
intensity varies from 8° to 28°. Ngalau-levu, a hill 1,650 feet in
height, rising behind Lea on the south coast of Natewa Bay, is polaric
in its upper portion. Specimens of a hemicrystalline basic andesite,
somewhat scoriaceous, which form the agglomerate of the rocky summit,
have a repellent force of 18°, whilst a similar rock from the
agglomerate of an exposed spur two-thirds of the way up the hill has a
force of as much as 38°. A curious case of polarity is exhibited in a
bare tuff overlooking the Vui-na-Savu River between Rauriko and Vitina.
It is composed of a much weathered whitish trachytic rock, which in
appearance affords no promise of polarity, but has the power of
repelling the magnetic needle 2° to 3°.

                  *       *       *       *       *

    _Note on the Average Polarity (Repellent Power) of the Volcanic
                         Rocks of Vanua Levu._

It would appear from the table given below that the difference in the
average polarity of acid and basic rocks is not very great. The average
for rocks with a specific gravity below 2·7 is about 10°; and that for
heavier rocks is about 14°. The difference mainly lies at the maximum
end of each series, the capacity for extreme polarity being, as before
remarked, markedly greater in the basic rocks.

 +-----------+------------------------+------------+-------------------+
 |           |                        |            |   Polarity.[152]  |
 | Specific  |  Character of rocks.   | Number of  +--------+----------+
 | gravity.  |                        | specimens. | Range. | Average. |
 +-----------+------------------------+------------+--------+----------+
 | 2·50-2·59 | Dacites                |      6     | 2°-25° |  11·6°   |
 |           |                        |            |        |          |
 | 2·60-2·69 | Dacites and            |      5     | 3°-16° |   8·6°   |
 |           | augite-andesites       |            |        |          |
 |           |                        |            |        |          |
 | 2·70-2·79 | Augite-andesites       |      5     | 5°-38° |  13·2°   |
 |           |                        |            |        |          |
 | 2·80-2·89 | Basaltic andesites and |     18     | 3°-90° |  15·0°   |
 |           | olivine-basalts        |            |        |          |
 |           |                        |            |        |          |
 | 2·90-3·00 | Olivine-basalts        |     14     | 3°-29° |  12·7°   |
 +-----------+------------------------+------------+--------+----------+

It is, however, noteworthy, as indicated by the value of the average in
each series that not one of them is a good series. They form curves
which in each case present an extreme maximum variant which is
suggestive of quite another degree of magnetising agency. This is also
illustrated in the combined curve of all the results given above. The
acid as well as the basic series are thus characterised, and the extreme
maximum variants are in each instance afforded by the highest mountain
peaks. It is probable that there is an accelerating ratio of
magnetisation with increased elevation. However that may be, it appears
evident from my observations on the two adjacent peaks of Ngaingai and
Navuningumu that the limits of polarity acquired through atmospheric
electricity without the direct action of lightning would be, as measured
by the scale here employed, four times as great for a dacite (25°) as
for a basaltic andesite (90°).



                             CHAPTER XXVII

                  SOME CONCLUSIONS AND THEIR BEARINGS


VANUA LEVU is a composite island built up during a long period of
emergence, that began probably in the later Tertiary period, by the
union of a number of large and small islands of volcanic formation
representing the products of submarine eruptions. It differs in this
respect from Viti Levu which is much more massive in its profile and
possesses a greater proportion of plutonic rocks. When, however, Viti
Levu comes to be systematically examined, it is likely that traces of
its composite origin will be found. The evidence seems to show that it
is older than Vanua Levu; but they are both situated on the same
submarine platform, the area of which is nearly equal to the combined
areas of the two large islands that rise from it.

This platform, as indicated in the small plan of the group, is limited
by the 100-fathom line in the charts; but since the reefs on their
seaward slopes plunge down precipitously, such a line practically serves
to delineate the margin of the plateau. It has been my object to show on
previous pages[153] that this submarine platform is a basaltic plateau
built up by submarine lava-flows and incrusted with coral reefs and
their deposits. It has been pointed out that this platform passes
gradually, as it proceeds landward, into the low-lying basaltic plains
that constitute the sea-border in the western part of the island, where
the breadth of the submerged plateau is greatest. The basaltic flows of
the plains often display the almost vertical columns of slightly
inclined flows. Their apparent termination at the sea-border, where they
are in places covered over with submarine deposits, cannot, however, be
accepted as their real limits. According to my view they extend several
miles seaward and form the platform, as is shown in the sections on
pages 62 and 107.

[Illustration:

  FIJI ISLANDS.

  FROM ADMIRALTY CHART 780, CORRECTED TO 1901.

  ALL DEPTHS BEYOND 100 FATHOMS ARE COLOURED BLACK.
  HEIGHTS IN FEET, DEPTHS IN FATHOMS.
]

We have in the great basaltic mountain of Seatura, which forms the bulk
of the western end of the island, a probable source of many of these
basaltic flows; and the occurrence inland in the western half of Vanua
Levu of elevated table-lands of basalt like that of Wainunu, which
extend from the centre of the island to near the coast, afford testimony
that the formation of these flows extends over a considerable period of
the island’s history.

It is held by Professor Agassiz that these submarine platforms are the
work of erosion into the flanks of the up-heaved islands.[154] In
Chapter II. it has been pointed out that the eroding agencies are not
actively in operation in our own day, and that there is good reason for
the belief that the process of amalgamation by which Vanua Levu has been
built up during a prolonged period of emergence, is not suspended at the
present time. It is assumed that the uniformity in Nature’s methods has
not been broken. If, however, we have here platforms of erosion, the
coasts of Vanua Levu, as far as my interpretation goes, supply no
evidence of it; and we have to imagine that a period of emergence
extending over a geological age has been followed by a similarly vast
period of erosion without much change in level.

Whatever agencies have been at work, the production of submarine
platforms 10 to 20 miles in width must have been a stupendous operation;
and we shall be obliged to inquire whether plateaux, either submarine or
upheaved, occur in association with volcanic islands in other parts of
the world, and under what conditions they have been formed. At least
four hypotheses have been framed with regard to the submarine platforms
of Fiji. There is first the original theory of subsidence of Darwin; but
Vanua Levu, which presents one long story of emergence, offers nothing
to support this view. There is the growth of a reef seaward on its own
talus, as advanced by Murray. There is the theory of erosion of Agassiz.
There is lastly my own idea of basaltic plateaux incrusted by reefs. We
may therefore inquire as to the evidence afforded by Vanua Levu in
favour of these views. Basaltic flows, in places covered by submarine
deposits, form the low plains at its sea-border, where the platforms are
broadest; and there rises a basaltic mountain of the Mauna Loa type,
occupying most of the western end of the island. No one would be bold
enough to place the limit of these basaltic flows at the water’s edge;
and as is indicated in the sections, they probably extend for miles
under the sea.

If we look for an island which in its extensive palagonite-formations,
in its basaltic table-lands and later basaltic flows, in its huge
mountain-ridges, and in its evidence of submergence, most resembles
Vanua Levu, we seem to find it in Iceland. It is in Iceland, I think,
that we must expect an explanation of many of the puzzling features in
the structure of Vanua Levu.

I pass on now to refer to some of the general points in the geology of
this island, which have been dealt with in detail in the earlier
chapters of this work. With regard first to the distribution of the
volcanic rocks, it may be remarked that my materials do not lend
themselves to making a geological map. The most comprehensive idea of
the principal points in the geological structure will be obtained by
reading the description of the profile given in Chapter I. There is,
however, a method in the distribution of the rocks that may be again
noticed here. The plutonic rocks are very scantily exposed, as is shown
on page 249; and they are not displayed at all in the western half of
the island. The more basic eruptive rocks, the olivine-basalts and
basaltic andesites, are mainly confined to the western half, that is,
west of Nanduri on the north and of the Ndreke-ni-wai River on the
south. Ordinary augite-andesites occur also in the western half; and
together with the hypersthene-augite-andesites they are found over most
of the rest of the island, excluding the north-east portion, east of
Lambasa and Tawaki, where quartz-porphyries, oligoclase-trachytes, and
acid pumice tuffs prevail. The acid andesites, including the
hornblende-andesites and the dacites or felsitic andesites, are best
represented in the Ndrandramea district in the midst of the basic rocks.
They occur in the isolated peaks of Na Raro and Vatu Kaisia and in one
or two other localities, as in the Valanga Range and on the shores of
Natewa Bay in the vicinity of the Salt Lake. These peaks of acid
andesites, as in the instances of Vatu Kaisia and Soloa Levu, are at
times in part overwhelmed or surrounded by the basaltic flows. This
singular feature of bosses of acid rocks in the midst of basaltic fields
offers another point of resemblance between Iceland and Vanua Levu.

The mountain-types vary considerably, the ridge-mountains, however,
being most characteristic of the island. The basaltic mountain of
Seatura, though its lava-flows were evidently in the main submarine,
belongs as before observed to the Mauna Loa type. In its radiating
valleys and gorges and in other characters it recalls the description
given by Dana of the island of Tahiti. The peaks of acid andesites,
represented in the isolated hills and mountains of the Ndrandramea
district, and in the solitary mountains of Vatu Kaisia and Na Raro, are
the necks and stumps of submarine volcanoes dating back to the
pre-basaltic period of the island. It is, however, in the great
mountain-ridges of the central portion of the island, those of Va Lili,
Korotini, Nawavi, Thambeyu, Mbatini, Mariko, &c., that we find, as just
remarked, the most typical features of the internal topography of Vanua
Levu.

Agglomerates overlying palagonite-tuffs and clays, that are usually
foraminiferous and sometimes inclose molluscan shells, clothe the slopes
of these mountain-ridges up to elevations of 2,500 feet and over above
the sea. Most of these great ridges, now more or less covered over by
these submarine deposits, represent lines of submerged vents, of which
only a few raised their summits above the sea in the earliest stages in
the history of the island. At this early period there were no coral
reefs. Some of the ridges present a marked parallel arrangement,
recalling the arrangement of the mountain-ridges and lesser chains of
hills as described by Dr. Johnston-Lavis in the account of his visit to
Iceland.[155] The description of Hekla (as given by Thoroddsen) as “an
oblong ridge which has been fissured in the direction of its length and
bears a row of craters along the fissure,”[156] comes very near to my
conception of the original condition of these great mountain-ridges
before the emergence. Dr. Johnston-Lavis sees in Hekla a type of
volcanic mountain very different from that of Vesuvius and Etna. He
regards it as a ridge marked by a number of parallel ridges and furrows,
and built up along a main fissure with a number of subsidiary parallel
fissures.

The part taken by palagonite in the composition of the finer deposits
over the greater portion of Vanua Levu is another prominent
characteristic of the island. Palagonite, as I have suggested in Chapter
XXIV., is formed probably on the surface of submarine flows of an
ophitic basaltic rock.

The age of the more recent of the deposits of this island, the
fossiliferous tuffs, the pteropod-ooze rocks, and the foraminiferous
muds, cannot be far different from that of the same deposits in other
parts of the group, since it is apparent that the same general movement
of emergence has affected both of the two larger islands. Professor
Martin of Leyden informed Dr. Wichmann that the fossil shells found in
the tuffs of Viti Levu, Ovalau, and other islands were Tertiary but not
older than the Miocene.[157] Dr. Dall, after examining the fossil
mollusks collected by Professor Agassiz from the elevated limestones of
Fiji, confirmed the impression formed by the latter as to their late
Tertiary age. None of the genera were extinct, and the fossils were in
his opinion younger than Eocene and either Miocene or Pliocene.[158] The
Rev. J. E. Tenison-Woods described as extinct Tertiary fossils, some
corals and mollusks from the interior of Ovalau.[159] Mr. H. B. Brady,
basing his conclusions on the character of the foraminifera, assigned a
Post-Tertiary date to the Suva “soapstone” taken at elevations up to 100
feet in that neighbourhood.[160] Professor David referring to some
fossil teeth of Carcharodon and to a fossil Tridacna found at Walu Bay
infers that the deposits are at least as old as Pliocene but not as old
as the earlier Tertiaries.[161] Since, as pointed out by Professor
David, the latest movements of emergence have taken place in recent
geological time, these various observations go to show that whilst the
latest exposure of deposits has occurred in recent time the mass of the
fossiliferous deposits date back to the Pliocene and the Miocene
periods.

According to Wichmann these islands were in a continental condition
during the Palæozoic and Mesozoic periods, and it was only in the later
Tertiary age that the movement of subsidence began that prepared the way
for the formation of the more recent deposits. The submergence during
the Tertiary period and the subsequent emergence are facts that cannot
be gainsaid; but we may ask where is the evidence of the continental
condition during the earlier periods. There is little in the results
obtained from Vanua Levu that directly supports such an hypothesis.
Under such circumstances one ought to have discovered in the deposits of
this island some evidence of this early condition, and there should be
found in the fauna and flora some traces of the original organisms.
According to Hedley there is some indication of a continental condition
in the molluscan fauna, and he quotes Fairmaire as regarding the
Coleoptera as of a continental character; but no one, that I am aware
of, has found any direct evidence of the Pre-Tertiary periods in this
group. It is in harmony with the geological characters to assume that
these islands made their first appearance during the Tertiary epoch.

Coming to the subject of the movements whether of land or sea that led
to the appearance of these islands, we shall not be begging the question
if we speak of their “emergence.” There is no doubt as to there having
been during and since the Tertiary epoch an emergence of some thousands
of feet, allowing for the original depth of the foraminiferous deposits
now found at elevations of over 2,000 feet above the sea. In Chapter II.
it is shown that there is good ground for the belief that these changes
of level have not altogether ceased. Of what nature, we may ask, is this
movement. We have before us the grand conception of Suess that the
emergence of the land in the different phases of geological time has
been produced by the general lowering of the level of the ocean arising
from local subsidences of the earth’s crust. This view in the case of
the recent calcareous formations of the Pacific is applied to the
terraces of the Loyalty Islands;[162] and it follows that it is also
applicable to the elevated calcareous deposits of the islands of the
Western Pacific as a whole, as in the case of the Tongan Islands, the
New Hebrides, the Solomon Group, &c. Such a general change of level
ought to be represented in the large island of Hawaii in the North
Pacific, since it could not be confined to one locality in this ocean.
There is no evidence of emergence, as far as I know, presented by this
island. During my sojourn there, I examined much of its coasts. Now the
antiquity of the flora of this group is sufficiently attested by the
circumstance that it ranks first among the oceanic groups of the Pacific
for the number of endemic plants that it possesses; and the same
conclusions may be drawn from the insects and the birds. There is no
evidence in this group, one of the most ancient of the Pacific
archipelagoes, of that great movement of emergence, which is abundantly
demonstrated over the Western Pacific.

The standpoint is therefore taken that the movement of emergence which
began in the Tertiary period and is probably still in operation is
confined to the southern portion of the tropical Pacific. Speaking of
the time of the Fijian emergence, Professor Agassiz observes that “it is
not unnatural to assume that it was coincident with the elevation of
Northern Queensland, and that the area of elevation included New Guinea,
the islands to the east of it as far as New Caledonia, and as far east
as the most distant of the Paumotus, and extended northward of that line
to include the Gilbert, Ellice, Marshall, and Caroline Islands.”[163]

From the report of Mr. Andrews[164] it is evident that in the Lau
Islands of the Fiji Group volcanic outbreaks have taken place since the
last upheaval. He describes in the case of Mango and other islands the
manner in which cliffs of limestone form inliers in flows of andesitic
lava. In the history of these islands he first distinguishes the period
of calcareous deposits, when the bedded limestones forming the submarine
plateau were laid down. Then followed a period of volcanism during which
masses of volcanic materials were erupted along the axis of elevation.
Alternating epochs of upheaval and stable equilibrium ensued, during the
last of which the reefs grew outwards and formed the terraces now so
characteristic of the profiles of the islands. After the last upheaval
the volcanic forces became again active. There is much of special
interest in the account given by Mr. Andrews of the Lau Group. The
blocks of limestone included in the volcanic agglomerates distinguish
the Lau detrital rocks from those of Vanua Levu. There is no evidence
that coral reefs existed during the early stages of the emergence of
Vanua Levu to be obtained from the submarine formations found on the
higher levels, 1,000 to 2,500 feet above the sea.

The period of emergence for this island may be divided into an earlier
and into a later stage, the last corresponding to the age of emergence
of the Lau Islands. The earlier stage, which may be termed the “Pre-Lau”
stage, is represented by the deposits of the higher slopes of Vanua
Levu, that is above 1,000 or 1,200 feet. This is really the critical
epoch in the history of this group, and assuming that the movement of
emergence has been fairly uniform over the archipelago we cannot but be
astonished at the absence of all traces of ancient reefs in the earlier
stage.

We may infer from the observations of Mr. Lister[165] that the islands
of the Tonga Group represent the Lau stage of the emergence. They are
similar in height and in general geological structure to the islands of
Lau, that of Eua, for instance, which has an elevation of 1,100 feet,
being formed of reef-limestones overlying volcanic tuffs. Dykes
penetrate the tuffs but do not enter the incrusting calcareous strata.
Mr. Harker,[166] after examining the collections of Mr. Lister, remarks
that all the rocks excepting those from Falcon Island appear to be of
submarine formation. The volcanic material, he adds, seems to have been
almost exclusively of fragmental character. It would be rash, it is
remarked, to refer all the rocks to a Recent age, and some of them may
be found to go back far into Tertiary times.

My division of the long period occupied by the emergence of the Fiji
Islands into two stages, the Lau stage corresponding to elevations of
less than 1,000 or 1,200 feet, and the Pre-Lau stage which includes the
earlier evidence of emergence found at heights exceeding these
elevations and ranging up to 2,000 or 3,000 feet, may perhaps be
applicable to other regions of emergence.

As bearing on the question of the isolation and antiquity of the Pacific
Islands the following approximate results for the Hawaiian, Fijian, and
Tongan floras may be here quoted.[167] These data are liable to
correction; but they are near enough to the truth to be very suggestive.
Of peculiar genera of flowering plants and ferns the Hawaiian Islands
possess about 40, the Fiji Group about 16, and the Tongan Islands none.
Of endemic species of flowering plants there are about 80 per cent. in
Hawaii, about 50 per cent. in Fiji, and 3 or 4 per cent. in Tonga.
Granting that there is much to be done yet in the investigation of these
floras, it would be underrating the brilliant results of the labours of
Hillebrand and Seemann to characterise their work as sampling. Let us
suppose, however, that the floras of Hawaii, Fiji, and Tonga have been
only sampled, the data above given would be still reliable. It is quite
possible to obtain a botanical equivalent corresponding to the
geological estimates of the relative ages of these islands; and taking
the proportion of endemic plants as our guide, the Lau stage, as
represented by the Tongan Islands, would have a value of 3 or 4, the
Pre-Lau stage now exhibited in the earliest stage of emergence of Vanua
Levu would have a value of 50, and the Hawaiian stage older than all
would have a value of 80. These results are intended as suggestive and I
hope to work out this subject in the second volume. They make the
problem of the relative antiquity of these islands more mysterious than
it even appeared before.

With regard to the vexed question of the light thrown on the past
condition of these islands by the present state of their floras and
faunas, it may be at once observed that my belief in the general
principle that islands have always been islands has not been shaken by
the results of the examination of the geological structure of Vanua
Levu. In a correspondence in _Nature_ about fifteen years ago it was
suggested by me that this is the position we ought to take with regard
to the stocking with plants of the islands of the Southern Ocean, such
as Kerguelen; and I take the same standpoint for the islands of the
Pacific. If the distribution of a particular group of plants or animals
does not seem to accord with the present arrangement of the land, it is
by far the safest plan, even after exhausting all likely modes of
explanation, not to invoke the intervention of geographical changes. New
possibilities of inter-communication, new ways of looking at old facts,
and new discoveries of an unexpected nature come monthly before us in
the progress of scientific research; and I scarcely think that our
knowledge of any one group of organisms is ever sufficiently precise to
justify a recourse to hypothetical alterations in the present relations
of land and sea.

The hypothesis of a Pacific continent,[168] whether it takes a
trans-oceanic form, as advocated by Von Ihering, Hutton, Baur and
others, or whether it is represented by an island-continent isolated in
mesozoic times, as suggested by Pilsbry, receives no support from the
geological characters of Vanua Levu. Nor can I accept as regards Fiji
Mr. Hedley’s theory of the Melanesian Plateau. There is no evidence that
the various islands of the Fiji Group were ever amalgamated and no
indication of a geological nature that they were ever joined to the
Solomon Group. The Fijis, as we see them, have had an independent
history, and the process at work is not one of disruption but of
amalgamation, lesser islands being united to larger islands during the
prolonged period of emergence. Mr. Hedley, however, has some weighty
data on his side more especially zoological; but even here it would be
wise to suspend one’s judgment. Though the great mass of botanical
evidence is as respects Fiji opposed to such connections, the
distribution of Dammara may, however, be fairly claimed on their behalf.

The dilemma into which such discussions lead us is aptly stated by Dr.
Pilsbry. If we do not accept the hypothesis of a Pacific continent, we
have to explain the cessation of the means of transportal in later
geological times, since this is implied in the isolation necessary for
the development of peculiar characters in a fauna or a flora. This was
the dilemma that presented itself to me in studying the origin of the
Fijian plants. Assuming on geological grounds that the insular condition
had been always maintained I had to explain the differentiation in the
inland plants, or in other words to account for the failure of the means
of transportal that once existed. Since this subject bears directly on
the past condition of the Fiji Islands, I may be pardoned for referring
to it here. It belongs properly to the second volume which it is
proposed to devote to the dispersal and distribution of Pacific plants;
but as I contest the pre-existence of a Pacific continent, it is fitting
though not necessary that this difficulty should be faced at once.

If we in imagination combine in a typical island the characters of the
flora presented by islands of different elevation in the Pacific we get
a result of this kind in an island of the height of Hawaii, nearly
14,000 feet. The littoral plants of such an island are found all over
the coasts of the tropical Pacific, and for the explanation of this fact
we look mainly to the agency of the ocean-currents. The plants of the
mountain summit, belonging to the temperate genera of Geranium, Rubus,
Ranunculus, Vaccinium, &c., are represented at least generically on the
tops of the lofty ranges of the Pacific coasts and in the interior of
the continents; and we find the explanation of the wide diffusion of
such plants in the agency of the migrant birds that at no distant time,
if not actually in our own time, were regular visitors to these mountain
regions. The plants of the marsh, of the stream, and of the pond, belong
often to species that occur in similar stations over a great portion of
the world, such as species of Drosera, Ruppia, Potamogeton, &c.; and
here the agency of wild duck and other waterfowl may be observed in
active operation.

But when in such an island we regard the intermediate region between the
uplands and the coast, usually the forest-zone, we find an area of
change not only for the plants but also for the birds. It is here that
the new genera of plants have been developed that distinguish the floras
of the Pacific groups each from the others; and here also the migrant
bird, having from some cause changed its ways, has given rise to new
varieties and to new species. It is with this loss of the migratory
powers of the birds of the forest-zone that I connect many of the
important differences between the forest-floras of the different groups
of the Pacific. At one time, it would seem, birds were far more active
agents in dispersing seeds and fruits over these archipelagoes than they
are at present; but it is not held that this is concerned with the
extermination and extinction of the birds of these islands in the
present day. The change dates far back and is far-reaching in its
effects. It is assumed in this argument that the alpine plant and the
plant of the pond and of the sea-shore preserve their characters by
reason of the means of free dispersal that they still enjoy; and it is
inferred that the plant of the forest-zone has varied more because
opportunities of transportal of its kind no longer are afforded. Many a
line of ancient migration is now broken.

It is suggested that in the past when birds were more generalised in
type they were much more migratory in habits and that limitation of
range has been associated with specialisation. The plants dispersed by
the birds have undergone a parallel series of changes. At first widely
distributed, as in the more generalised types of birds, they became
localised in proportion as the birds to which they owed their means of
dispersal lost their migratory ways; and both plant and bird began to
vary. There is, I am convinced, a profound connection between birds and
plants, of which we now perceive only the last of a long series of
changes. This subject will be dealt with at length in the volume on
plant-dispersal; and it is only referred to here to illustrate my
contention that we have yet much to learn before it would be safe to
look to hypothetical changes of sea and land to explain difficulties in
distribution.



                                APPENDIX


_Note on the Stone-Axes._—Two of these polished stone-axes from a
collection made in Vanua Levu were selected for sections. One is
light-green and smooth. The other has a very different appearance, being
blackish and rather rough, its smooth surface having been apparently
lost by lying in a stream-course or in wet ground for a long period.
Both, however, are made of the same type of basaltic rock, the specific
gravity in one case being 2·93, in the other 2·97. It is an aphanitic
basalt with scanty olivine containing little or no residual glass and
referred to genus 40 of the olivine-basalts. It is by no means a common
type of basalt in Vanua Levu, and I cannot refer it to any particular
locality on account of the peculiarities it presents when contrasted
with rocks of the same genus. The olivine is very scanty and small, and
in one of the specimens is represented only by pseudomorphs. The
felspar-lathes vary usually from ·05 to ·2 mm. in length, and the augite
granules which are very abundant are ·01 or ·02 mm. in diameter. There
is an occasional small phenocryst of augite. The rock shows little or no
alteration and cannot be characterised as a greenstone. The greenish hue
of one axe is due to weathering; but its extension into the internal
black portion of the tool is not appreciable.

                  *       *       *       *       *

_Note on the ascent of the tide up the Ndreketi River._—On July 20th and
21st, 1899, by observing the surface density it was ascertained that at
high-water the sea-water reached Navundi a mile or two below Mbatiri. At
low-tide it reached about half-way between Kanathangi and Navundi. The
moon was in her quarters.

                  *       *       *       *       *

_Note on the “talasinga” districts._—This subject will be discussed in
the second volume.



                                 INDEX


NOTE.—It has been deemed best to follow the example of the Admiralty
surveys in the spelling of native names. In this book, therefore—

                        Mb = the Fijian B
                        Th = the Fijian C
                        ND = the Fijian D
                        NG = the Fijian G and Q


 Abbreviations, for rock descriptions, 236

 Acicastello, Sicily, basalt and palagonite, 347

 Acid and basic rocks, regions of, 219, 374

 Agassiz, Prof. A., 294, 373, 376, 377

 Agates, 138, 139, 227, 353, 354

 Agglomerates; _see_ Volcanic agglomerates

 Algæ in hot springs, 24, 25, 33, 38

 Alps, magnetic rocks of the, 362, 363

 Altered rocks; _see_ Propylites, etc.

 Andesites, acid, 98-108, 112, 123-127, 193, 194;
   relative frequency of, 235;
   classification of, 293-306;
   distribution of, 374;
   peaks in basaltic flows, 104, 115, 116, 374;
   altered, 105, 297;
   columnar, 102, 104;
   pre-basaltic, 375

 Andrews, Mr. E. C., 7, 22, 294, 350, 378

 Artesian reservoirs, 39

 Augite, crystals of, in tuffs, 45, 182, 193

 Augite-andesites, 51, 162, 199, 204, 209, 232, 235, 294, etc.;
   classification of, 239, 245, 246, 266-284;
   distribution of, 374;
   aphanitic, 117-120, 125, 162, 168, 279;
   relative frequency of, 235;
   _see_ Basaltic andesites

 Avuka, range, 179

 Axes, stone, structure of, 383


 Bare-poll, peak, near Soni-soni, 93

 Barrack, Mr. A. H., 27, 123

 Barrack, Mr. Alex., 364

 Barratt, Mr., 68

 Barrier-reef, great, of Fiji; _see_ under Submarine plateau

 [169]Basaltic andesites, 47, 56, 64, 75, 108, 123, 137, 147, 148, 160,
    164, 190, 204, 206, 208, 288;
   classification of, 239, 266, 278;
   distribution of, 374

 Basaltic rocks, extensive disintegration of, 57, 64, 72, 129

 Basaltic flows surrounding hills of acid andesite, 104, 115, 116, 374

 Basaltic plains and plateaux, 6, 55, 62, 82, 107, 128-135, 373

 Basaltic submarine flows, 338, 342, 344, 346, 347, 372, 375;
   _see_ under Submarine plateau and Basaltic plains

 Basalts, columnar, 3, 63, 78, 83, 84, 85, 123, 129, 133, 147, 170, 173,
    203, 260, 284

 Basalts, ophitic, 50, 74, 114, 133, 137, 138, 155, 159, 162, 191, 204,
    213, 252, 256, 347;
   the ophitic habit, 237, 238;
   ophitic sub-orders, 236, 241-245;
   ophitic genera, 256, 262, 272-276, 283

 Basalts, olivine, relation of black and grey, 89, 90;
   classification of, 239, 241-244, 252-265;
   relative frequency of, 235;
   distribution of, 374

 Basalts, grey olivine, 65-67, 73-75, 77, 89, 253-255, 257, 261, 263

 Basalts, highly basic, 258

 Basalts, of Acicastello, 347;
   of Mbua and Ndama plains, 58, 134, 262, 278;
   of Ndreketi plains, 133, 134;
   of Sarawanga plains, 129, 134, 262;
   of Savu-savu peninsula, 190, 288;
   of Seatura, 63-72, 85;
   of Solevu Bay, 75-77;
   of Wainunu, 85;
   of Ulu-i-ndali, 89

 Basalts associated with palagonite, 347

 Basic glass, 312, 338;
   _see_ Pitchstone, Crush-tuffs, Hyalomelan-tuffs

 Bastite, 182, 297

 Baur, Mr., 380

 Blyth, Mr., 123

 Brady, Mr. H. B., 322, 376

 Breccias; _see_ Volcanic agglomerates

 Bronzite, 182

 Bromlow, Dr., 29

 Büchner, Dr. Max, 22

 Bulling, Mr., 32, 36, 232

 Bunsen, on palagonite, 343, 344


 Carbonate of iron, concretions of, 227, 351, 356

 Carcharodon, 376

 Chalcedony, 13, 138, 162, 183, 199, 226-22, 350-355;
   _see_ Agate, Onyx, Flints

 Chalmers, Mr., 233

 Charts, old and new compared, 18, 19

 Chert, 355

 Classification of volcanic rocks, 235

 Coleoptera of Fiji, 377

 Columnar structure; _see_ Basalt, Acid andesite, Dacite,
    Oligoclase-trachyte

 Combe, Commander, 15

 Cooper, Mr. H. S., 29

 Coral reefs, upheaved, 7-12, 189, 200, 201, 318;
   their absence in the higher levels, 7, 8, 12, 19, 375

 Craters, traces of, 44, 52, 67, 80, 166, 186, 192, 195, 202

 Crush-tuffs, 55, 94, 149, 157, 341;
   general description, 334;
   _see_ below

 Crushing of basic glass;
   in veins, 340, 341;
   on surface of a submarine flow, 93, 338;
   in matrix of pitchstone-agglomerate and in rubbly pitchstone, 94,
      142, 145, 157, 313;
   its connection with palagonite, 93, 339, 340-342, 346

 Cumming, Miss Gordon, 22, 25, 29


 Dacites, 3, 5, 100, 108, 235, 294, 304;
   columnar, 101, 102;
   definition of the term, 295;
   classification and characters, 240, 302;
   _see_ Acid andesites

 Dall, Dr., 376

 Dana, Prof., 3, 10, 11, 16, 21, 26, 72, 84, 129, 135, 218, 309, 363

 Darwin, Mr., on barrier-reefs, 373

 Datum-mark, 195

 David, Prof., 376

 Deep-sea deposits, 337

 Delanasau, 68

 Dillon’s Rock, 45

 Diorite, 182, 193, 235, 249, 251

 Doelter, on hornblende paramorphism, 308

 Doleritic, use and definition of the term, 236, 238, 259, 274

 Dolomite, 7

 Drayton, peak; _see_ Mariko

 Dykes, 51, 54, 63, 68-72, 78, 81, 142, 144, 148, 155, 156, 163, 164,
    170, 171, 184, 199, 202, 209, 216, 220, 233, 234, 267, 268, 270,
    277, 280, 282;
   their two sets of felspar-lathes, 238


 Eakle, Mr., 293, 294

 Earthquakes, 37

 Emergence of Vanua Levu, 1, 7-20, 321, 376-379;
   age of, 376

 Etna, mount, coast springs of, 39;
   dykes in Valle del Bove, 237;
   magnetic bomb, 364


 Fairmaire, M., 377

 Faro, island, Solomon Group, 2

 Fawn, harbour, 9, 200, 318

 Felsitic andesites, 106, 108, 295, 300;
   _see_ Acid andesites and Dacites

 Felsitic groundmass, as employed in classification, 236, 239, 240, 249

 Felsitic orders, 249, 291, 297, 300, 302

 Felspar-lathes in classification, 236, 241

 Fern, tree, silicified caudex of, 360

 Fish-scales, fossil, 154

 Flints, 13, 81, 83, 138, 139, 222, 226, 350-360

 Floating islands, 225

 Floras of Fiji, Hawaii, and Tonga, 379

 Flow-arrangement in classification, 236-238

 Folgheraiter, Dr., 362, 365

 Foraminiferal limestones, 130-132, 202, 318

 Foraminiferous deposits (tuffs, muds, clays), 96, 109, 130, 134, 136,
    139, 149, 154, 161, 170, 198, 205;
   description of, 321-333;
   altered, 324-326;
   thickness of, 156;
   _see_ Globigerina deposits and Palagonite-tuffs

 Fossilised trees, 233

 Freeland, Mount, 2, 6, 203-206, 269, 274


 Gabbros, 180, 182, 184, 211, 235;
   classification of, 239, 240;
   characters, 249, 250;
   _see_ Plutonic rocks

 Geikie, Sir A., 375

 Giant sedge; _see_ Scirpodendron

 Globigerina deposits, 10, 55, 131, 158, 177, 187, 189, 190, 221,
    321-326, 344;
   _see_ Foraminiferous deposits

 Globigerina limestone, 319

 Gold, alluvial, in Vanua Levu, 116

 Granular pyroxene, the sub-orders characterised by, 236, 241, &c.

 Greenstones; _see_ Propylites

 Groundmass, characters of, used in classification, 236-238


 Haig, Major, on magnetic rocks, 363

 Hale Peak, 210, 212

 Hanusz, on floating islands, 226

 Harker, Mr., 362, 367, 379

 Harman’s Point, 191

 Hawaii, 2, 38, 85, 363;
   flora of, 379;
   coast springs of, 38

 Hedley, Mr., 376, 380

 Hekla, Mount, 375

 Holmes, Mr., 68

 Hornblende, magmatic paramorphism of, 293, 299, 301, 303;
   process described, 306

 Hornblende-andesites, 91, 184, 193, 194, 201, 235, 294;
   _see_ Acid andesites and hornblende-hypersthene andesites

 Hornblende-gabbro, 184, 249, 250

 Hornblende-hypersthene-andesites, 240, 298-302

 Horne, Mr., 10, 21, 22, 25, 34, 55, 141, 143, 194, 195, 203, 225;
   _see_ the preface

 Hornstone, 350

 Hot springs, general description of, 21-42;
   list of, 40;
   analysis, 28;
   distribution in Vanua Levu, 36, 138, 233

 Humboldt, A. von, on magnetic rocks, 361, 365

 Hussak, on magmatic paramorphism, 308

 Hutton, 380

 Hyalomelan tuffs, 47, 80, 110, 333, 334

 Hydrothermal metamorphism; _see_ Solfataric

 Hypersthene; _see_ Pyroxene, rhombic

 Hypersthene-andesites, 294, 296

 Hypersthene-augite-andesites, 5, 52 147, 161, 164, 168, 171, 173-175,
    178, 179, 182, 186, 190, 199, 201, 203, 208, 211, 230;
   classification, 240, 247, 248;
   characters of the orders and sub-orders, 285-292;
   relative frequency, 235;
   distribution, 374


 Iceland, Vanua Levu compared to, 374, 375

 Iddings, Mr., 306

 Iron ore; _see_ Limonite

 Iron sand, magnetic, 83, 106, 357

 Ironstone gravel, 356

 Islands, permanence of, 380, 381


 Jackson, Dr. C. T., 28

 Jasper, 13, 121, 139, 199, 350, 351, 355

 Johnston-Lavis, Dr., 347, 375


 Kalakala, Mount, 103, 305

 Kalikoso District, 10, 15, 224-228, 350, 356, 358

 Kandavu, 293, 306;
   hot springs of, 22

 Kavula, 65, 66

 Kia Island, 2

 Kioa Island, 2

 Kiombo Coast, 92

 Kleinschmidt, 22, 25, 293, 350

 Koro, significance as a prefix; _see_ Place-Names

 Koro-i-rea Hill, 75, 77

 Korolevu Hill, 45;
   natural section near, 48

 Korolevu River, 62, 64

 Koroma, Mount, 3, 51, 285

 Koro-mbasanga Mountain, 166-169, 289;
   name wrongly applied in Admiralty charts, 5, 172

 Koro-navuta, 135

 Koro-ni-valu; _see_ under Towns

 Koro-ni-yalewa Mountain, 117

 Koro-tambu Mountain, 167, 171

 Korotasere, 208

 Korotini Bluff, 156

 Korotini Range or Tableland, 5, 153-165, 167, 325

 Koro-utari, 163

 Koro-wiri, 139

 Kumbulau Peninsula, 90-95

 Kuru-kuru District, 228


 Lambasa coast, 218

 Lambasa Plains, 138, 351

 Lambasa River, 15, 138

 Landslips, effects of, 111, 178, 327

 Langa-langa River, 220, 225

 Lango-lango River, 122

 Lau Group, 7, 294, 378

 Lava-flows, indications of sub-aerial, 52, 71, 119, 133, 152, 187, 190,
    213, 232

 Lea District, 199

 Lekumbi Point, 12, 19, 60

 Lekutu District, coast of, 11, 50, 273;
   plains, 128, 351, 353, 356;
   promontory, 16, 18;
   river, 18, 62, 65

 Limestones, recrystallisation of, 131;
   coral, 318;
   _see_ Foraminiferal limestones

 Limonite, deposits of, 56, 132, 138, 226, 228, 351, 352, 356, 359

 Lister, Mr., 378

 Liversidge, Prof., 29

 Liwa-liwa, 119

 Loma-loma Ridge, 141

 Lovo Valley, 169-173

 Lovutu, 156


 Macdonald, Dr., 21

 Magma lakelets, 47, 71, 92, 273, 276, 277;
   description of, 339-342, 346, 347

 Magmatic paramorphism; _see_ Hornblende

 Magnetic iron sand, 83, 106, 357

 Magnetic peaks and rocks, 77, 108, 174, 186, 361-371

 Mako-mako Hill, 103

 Mali Island, 2, 218

 Mali Point, 218

 Malolo Island, 294

 Mangrove belt, bare tracts in, 11, 14;
   relation to reef-flat, 13;
   rate of growth, 15, 19

 Mangrove islands, growth of, 16, 17

 Mariko, range and peak of, 5, 173, 185-189, 289, 368

 Martin, Prof., of Leyden, 376

 Masusu District, 84

 Mauna Kea, 2

 Mauna Loa, 2, 3, 363

 Mauritius, magnetic rocks of, 363

 Mbale-mbale District, 143, 313;
   river, 150

 Mbatini Mountain, 5, 166, 172-174, 367

 Mbati-ni-kama, hot springs of, 33

 Mbatiri, 134, 324

 Mbenutha Cliffs, 109-111, 323

 Mbona-lailai Mountain, 101

 Mbua District, 3, 36, 37, 47;
   bay, 18;
   coast, 12

 Mbua and Ndama Plains, 55-58, 356

 Mbua-Lekutu Divide, 55, 297

 Mbua shell-bed, 12, 58

 Mbuthai-sau Valley, 218, 219

 Mbutu-mbutu River, 79, 279

 Middle Point, 137

 Mountain ridges and their structure; _see_ under Ridge-mountains

 Mountain-towns; _see_ Towns

 Muanaira, 198

 Mumu Peak, 108-110

 Murray, Sir J., 337, 373


 Naikovu Rock, 364

 Nailotha Mountain, 6, 214-216, 310

 Naindi Bay, 8, 189, 191, 195, 318

 Naindi Gap, 189, 192

 Naithekoro, 190

 Naithombothombo, point, 54;
   range, 229

 Naivaka, 2, 3, 11, 18, 43-45, 261

 Na-kalou, 133

 Na Kama, Savu-savu, 25;
   Lambasa, 32

 Nakambuta District, 148-150

 Nakarambo, 208

 Na Kula, valley, 184;
   range, 230

 Nambuna District, 106

 Nambuni Spur, 144

 Nambuonu, hot springs, 32

 Nandi Bay, 78

 Nandi Gorge, 69, 78, 278

 Nandongo, island, 16, 17;
   town, 216;
   hot springs, 33

 Nandronandranu district, 117-121

 Nandroro District, 66

 Nandua District, 86, 320, 344

 Nanduri District, 11, 14, 135, 136

 Nangara-ravi Cave, 141

 Nangara-vutu, 205

 Nangorongoro Peak; _see_ Ngaingai

 Na Raro Gap, 127

 Na Raro Mountain, 2, 5, 123-127, 296, 301, 305

 Narawai District, 66

 Nareilangi, 124

 Narengali District, 140, 147, 149

 Narikosa Point, 220

 Na Salia, 151

 Na Savu, tableland, 79-81;
   falls, 79, 279

 Na Seyanga, 108

 Na Sinu, 146

 Na Suva Range, 64

 Na Suva-suva Hill, 192, 369

 Natasa Bay, 209

 Natewa Bay, north-coast, 9, 208, 209, 291

 Natewa Peninsula, 6, 9, 197-206

 Nativi, 50

 Natoarau, hot springs, 23;
   river, 158

 Na Tokalau, 91

 Natoto Hill, 229

 Natua District, 134, 149, 323

 Natuvo or Natuvu, hot springs, 33, 209

 Naumann on polaric lavas, 364

 Navakaravi Hot Springs, 34

 Navakavura, 96

 Na Vatu Islet, 94

 Navetau, 205

 Naviavia Islet, 8

 Navingiri, 46

 Navuni, 202;
   hot springs, 35

 Navuningumu, 108-112, 303, 368

 Nawavi, range, 135;
   hot springs near, 31

 Nawi or Na Wi, island, 26, 192;
   hamlet, 211

 Ndaku-ndaku, bay, 208;
   hot springs, 34

 Ndama, river and valley, 62, 67, 68, 71;
   plains, 55-58

 Ndavutu, district and river, 87;
   for hot springs; _see_ Wainunu

 Ndawathumi, 64, 80

 Ndevo district, 205;
   hot springs, 35

 Ndoendamu Mountain, 209, 212

 Ndrandramea, district, 2, 3, 83, 296;
   map, 99;
   description, 98-112

 Ndrandramea Mount, 102, 296, 300, 304, 368

 Ndranimako, 96, 322, 351

 Ndrawa, district, 120, 281;
   river, 118, 120

 Ndreke-ni-wai, Natewa Bay, 200, 203;
   hot springs, 34

 Ndreke-ni-wai, Savu-savu Bay, 150, 152

 Ndreketi, river, 15, 128, 132, 383;
   plains, 132, 273, 351

 Ndriti Basin, 67-72, 268, 270, 282

 Ndrukau Mountain, 213

 New Hebrides, 1

 Ngaingai Mountain, 100, 296, 302, 304, 368

 Ngala Mountain; _see_ Freeland

 Ngalau-levu Range, 6, 199, 200, 370

 Ngangaturuturu Cliffs, 119

 Ngau Island, hot springs of, 22

 Ngawa River, 138

 Ngelemumu, 180, 219

 Ngone Hill, 183

 Nukumbolo, district, 151, 161, 162;
   hot springs, 24

 Nukunase or Nuku-ngase, 50, 270

 Nukumbalavu, 190

 Nuku-ndamu, 232

 Numbu, 227

 Numbu-ni-a-vula, 176


 Oddone on magnetic rocks, 362

 Oligoclase-trachytes, 6, 207, 219, 229;
   description of, 308;
   distribution, 374;
   altered, 214-216;
   columnar, 215, 220, 230, 231, 233

 Olivine rocks, classification of, 239

 Olivine-basalts; _see_ Basalts, olivine

 Ono Island, hot springs, 22;
   acid andesite, 293;
   flints, 350

 Onyx, 139, 227, 228, 353

 Opal, 162, 163, 183, 351, 353

 Ophitic basalts, their relation to palagonite, 347

 Ophitic structure, as used in classification, 236, 237, 238

 Ophitic sub-orders and genera, synopsis of, 241-248;
   description of, 256, 272-276, 283

 Orthophyric groundmass, as used in classification, 236, 237, 239, 240

 Orthophyric orders and genera, 248, 290, 296, 297, 299

 Ovalau, 294, 350


 Palagonite, chapter on, 337-349;
   _see_ also Palagonite-tuffs, Crush-tuffs, Crushing of basic glass,
      Pitchstone &c.

 Palagonite of Acicastello, 347

 Palagonite tuffs, classification and characters, 317-336;
   zeolitic, 334;
   marls, 335, 344;
   modes of occurrence, 5, 48, 53, 80, 81, 95, 96, 117, 118-122, 130,
      131, 141, 143, 145, 148, 156-161, 169, 177, 193, 198, 202, 213

 Palagonite, hydration and degradation of, 329, 348

 Phenocrysts, their use in classification, 236

 Pickering, Mr., 35

 Pieper, Dr. O., 28

 Pilsbry, Dr., 380

 Pitchstone agglomerates and rubbly pitchstones, petrological
    characters, 312, 313;
   evidence of crushing and its connection with palagonite, 92-94, 142,
      145, 157, 312, 313, 334, 340-342, 346 (_see_ Palagonite,
      Crush-tuffs, Crushing of basic glass);
   mode of occurrence, 105, 108, 142, 157, 169, 229, 230, 309

 [170]Place-names, meaning of Fijian, 75, 79, 102, 119, 151, 172

 Platania, Prof., 119, 347

 Plutonic rocks, general description, 249-251;
   relative frequency, 235;
   distribution, 249, 374;
   mode of occurrence, 180, 182, 184, 185, 193, 211

 Polarity of magnetic rocks, 366-370

 Porphyrites, 136, 175, 181, 197, 199, 204, 211, 261, 268, 274, 299;
   belong to many orders, 236

 Prismatic pyroxene of groundmass, its use in classification, 236,
    241-248;
   sub-orders and genera, 265, 270-272, 283, 287, 289, 298, 300, 302

 Profiles of Vanua Levu, 3-6, 62, 83, 107, 113, 153, 167, 173

 Propylites, 68-72, 106, 147, 162, 181, 191, 199, 204, 214, 215,
    268-270, 282, 297;
   origin of, 69, 72, 191

 Pteropod-ooze deposits, description of, 320;
   mode of occurrence, 84, 86, 109, 139, 201, 205, 344

 Pumice-tuffs, acid, 6, 207, 218-223, 229-233;
   general description, 336;
   special descriptions, 218, 220, 231

 Pumice-tuffs, basic, 119, 333;
   _see_ Hyalomelan-tuffs

 Pyroxene of groundmass, as a basis of classification, 236;
   _see_ Granular pyroxene, Prismatic pyroxene, Ophitic structure, and
      Synopsis

 Pyroxene, rhombic, characters of, 285, 306;
   intergrowths with monoclinic, 266, 306

 Pyroxene, derivation from hornblende, 306


 Quartz, crystals of, 106, 191, 354;
   veins of, 106, 116

 Quartz-andesites; _see_ Dacites

 Quartz-porphyries, mode of occurrence, 215, 219, 220, 226, 227,
    229-233;
   relative frequency, 235;
   general description, 309-311;
   distribution, 6, 207, 374

 Quartz-rock, 139, 351, 354


 Rainfall, 30, 68, 120

 Rambi, island, 2;
   hot springs, 22

 Ravi-koro mountain, 159

 Ravi-ravi, 94

 Ravuka, 120;
   hot spring, 31

 Renard, Prof., 293, 306, 338, 344

 Rewa District in Vanua Levu, 95, 96

 Rewa River, Viti Levu, changes at mouth, 16

 Rhyolites, 209;
   _see_ Quartz-porphyries

 Rhyolite-glass, 220, 311

 Rhyolitic-tuffs; _see_ Pumice-tuffs, acid

 Ridge-mountains, their general appearance, 2, 6, 146, 153, 185, 210,
    374;
   their structure and mode of origin, 75, 145, 156, 165, 166, 171, 172,
      177-180, 182, 188, 202, 210-212, 216, 234;
   final conclusion, 375

 Rivers; _see_ under Ndreketi, Lambasa Sarawanga, Wainikoro, &c.

 Rivers, eroding power of, 62

 Rocholl, Mr. H., 29

 Rosenbusch, Prof., 306, 344

 Ruku-ruku Bay, 53, 269


 Salt Lake District, 2, 6, 9, 192-196

 Sarawanga Plains, 15, 129-132, 351, 356

 Sarawanga River, 15, 62, 129-131

 Satulaki, 176, 268

 Savarekareka Bay, 190, 326

 Savulu, 118

 Savu-riti Mountain, 210, 212

 Savu-savu Hot Springs, 21, 25-30, 189

 Savu-savu Peninsula, 189-192, 288

 Sawa-ndrondro, 185

 Scirpodendron costatum (Giant Sedge), 79, 83

 Scoriaceous lava; _see_ Lava flows

 Sealevu District, 146, 155, 156

 Sealevu Divide, 136

 Seatovo Range, 73-75

 Seatura, mountain, 2, 3, 56, 253, 261, 374;
   general description of, 61-72;
   old town of, 67

 Sections across Vanua Levu, 62, 107;
   _see_ also Profiles

 Seemann, Dr., 55

 Sella, on magnetic rocks, 362

 Semi-opal, 351, 353

 Sesaleka Mountain, 3, 12, 53

 Siliceous concretions, 81, 83, 96, 132, 135, 351-355

 Siliceous rock, blocks of, 126, 355

 Siliceous sinter, 24, 25, 32, 33, 37, 42

 Silicification, conditions of, 358

 Silicified corals, 10, 13, 81, 132, 135, 138, 139, 207, 221, 226-228;
   theory of their origin, 228, 357;
   general account of, 350-360

 Silicified fern rhizome, 360

 Silicified nullipores, 353, 354

 Singa-singa, 110

 Singatoka River, Viti Levu, 7

 Skinner, Mr. S., 367

 Skye, Isle of, magnetic rocks, 362

 Smallwood, Mr., 196

 Smythe, Colonel, 22, 195

 Soapstones; _see_ Foraminiferous deposits

 Sokena Ridge, 167, 169, 172

 Solevu Bay, 75-78, 253

 Solfataric action on rocks, 52, 69, 72, 191

 Soloa-levu Mountain, 103-105, 115, 296, 305, 312, 374

 Solomon Islands, 1, 2, 294, 359

 Songombiau, 220

 Soni-soni Island, 93, 94

 Soro-levu Mountain, 172-174

 Spence, Mr. F., 193

 Spheroidal weathering in basalts, 57, 129

 Stromboli, 214, 315

 Submarine basaltic flows and eruptions; _see_ Basaltic submarine flows

 Submarine plateau or platform of Fiji, 15, 18, 19, 56, 62, 72, 107,
    372;
   different explanations of, 373

 Submarine tuffs, 326-336

 Sueni District, 163

 Suess, Prof., on thermal springs, 39;
   on changes in the sea-level, 20, 377

 Suva soapstone, 322, 376

 Synopsis of classification of volcanic rocks, 239-249


 Tachylyte or basic glass, 312, 337, 341, 343;
   _see_ Basic glass, Pitchstone, Hyalomelan, &c.

 Tahiti, 3, 72, 84, 363

 Talasinga Districts, 55, 57, 64, 128, 132, 133, 224, 352, 383

 Tambia, district, 137;
   hot springs, 32

 Tambu-lotu District, 104, 105

 Tathelevu, 198

 Tavia, mountain, 121;
   ranges, 121-123

 Tavua, 65

 Tawaki District, 209, 229, 230

 Tembe, 213, 214

 Tembe-ni-ndio District, 130, 131, 319

 Tenison-Woods, Rev., 376

 Thambeyu, mountain, 5, 167, 289, 315, 326;
   description of, 176-179

 Thawaro Peak, 230

 Thermal springs; _see_ Hot springs

 Thiele, Mr., 21

 Thoka-singa Mountain, 103, 249, 302, 305

 Thombo-thombo; _see_ Naithombothombo

 Thomson, Mr. J. P., 31, 135, 210, 225

 Thongea, hot springs, 22;
   basalt, 85

 Thoroddsen, on Hekla, 375

 Thuku, Mount, 6, 231, 308-310

 Thulanga; _see_ Uthulanga

 Thurston Range, 5, 167, 176;
   _see_ Thambeyu

 Tonga Group, 1, 378, 379

 Tongalevu District, 62, 64, 279

 Towns, old sites of mountain, 53, 67, 101, 102, 108, 156, 170

 Trachytes; _see_ Oligoclase-trachytes

 Tuffs, chapter on, 317-336;
   foraminiferous, 326-333;
   altered, 184, 187, 190, 332;
   dacitic, 125, 126;
   _see_ Pumice tuffs, acid and basic; Hyalomelan tuffs; Palagonite
      tuffs; Crush tuffs;
   mode of occurrence, 48, 90, 109, 119, 130, 156, 160, 177, 190, 205,
      209, 215

 Tunuloa District, 205

 Tutu Island, 221


 Ulu-i-matua, 75, 76

 Ulu-i-mbau, 138, 139

 Ulu-i-ndali, 3, 83, 87-90, 253, 370

 Ulu-i-sori, 136

 Underwood, Lieut., 364

 Undu, district and promontory, 6, 10, 36, 228-234, 311, 360

 Upheaval; _see_ Emergence

 Urata, 184, 298

 Uthulanga Ridge, 211, 286 (also named Thulanga)


 Vakalalatha Lake, 15, 225

 Valanga Range, 181-185

 Valavala Bay, 203

 Valeni, 122

 Va Lili, 5, 140-146

 Valleys, origin of, 2, 146, 151, 219

 Vandrani, district, 139, 159;
   hot springs, 32

 Vanua Mbalavu, hot springs, 22

 Variolite, 150, 283, 313

 Vatui, 54, 369

 Vatu Kaisia, 5, 113-116, 296, 301, 305, 374, 375

 Vatu-karoa, 209, 282

 Vatu-karokaro, 54

 Vatu-kawa, river, 151;
   district, 160

 Vatu-kerimasi, 101

 Vatu-lele Bay, 184

 Vatu-levoni, 139

 Vatuloaloa Hot Springs, 31

 Vatu Mata, 103

 Vatu-ndamu, 91

 Vatu-tangiri, 136, 144

 Vatu Vanaya, 101

 Vatu Vono Point, 88, 89

 Vatu-vono District, 121, 122

 Viene District, 198

 Visongo District, 221

 Viti Levu, 7, 18, 350, 364, 372

 Vitina, 223

 Volcanic agglomerates, general description of, 314-316;
   their thickness, 110, 156, 171, 178, 315;
   mode of occurrence, 5, 79, 91, 110, 112, 117, 141, 143, 144, 149,
      155, 156, 161, 169-172, 176-179, 188, 193, 213, 214;
   lying above submarine deposits usually palagonitic and often
      foraminiferous, 110, 141, 143, 149, 168, 169-171, 176-179, 188,
      213, 214;
   _see_ Pitchstone-agglomerates and Landslips

 Volcanic bomb-formation, 46-48

 Volcanic mud deposits; _see_ Foraminiferous deposits

 Volcanic necks, 54, 58, 90, 93, 95, 112, 183, 192, 230, 234, 253, 277,
    283, 286, 375

 Volcanic rocks, classification of, 235;
   distribution of, 374

 Vuinandi Bay, 208

 Vuinandi Gap, 175

 Vuinasanga, district, 145;
   hot springs, 31

 Vui-na-savu, river and district, 222, 223, 225

 Vula Votu Peak, 176

 Vungalei Mountain, 212, 213, 315

 Vuni-ika Bay, 218

 Vunikondi, 232, 233, 282

 Vunimbele, 139;
   hot springs, 33

 Vunimbua, district and river, 182, 183

 Vunimoli, hot springs and district, 33, 138, 139

 Vunisawana, district, 194;
   hot springs, 34

 Vuni-tangaloa, 194

 Vunivuvundi District, 87


 Waikatakata, Natewa Bay, hot spring and district, 34, 203, 275

 Waikawa Mountains, 201, 319

 Wailea, bay and district, 46, 50

 Wailevu River, 15, 138

 Wai Mbasanga, Viti Levu, hot springs, 21

 Waimotu District, 208

 Wai Ndina, Viti Levu, hot springs, 21

 Wainikoro, district, 217, 224-228, 356;
   coast, 219, 308, 310;
   river, 225

 Wai-ni-ngio River, 151, 160

 Wainunu, hot springs, 22;
   rainfall 68;
   river and valley, 62, 82, 83

 Wainunu, plateau or tableland, 3, 82-87, 373;
   _see_ figures on pages 83, 107

 Waisali District, 146, 151

 Waisali Saddle, 146-148

 Waiwai, 143-145

 Waterfalls, 79, 141, 163

 Wawa Levu Mountain, 101, 302, 304

 Weed, Mr., on the origin of siliceous sinter, 38

 Wichmann, Dr. A., on a continental condition of the Fiji Islands, 376;
   on hyalomelan tuff, 334;
   on flints and silicified corals, 350, 352, 360;
   on Kandavu andesites, 293;
   on the absence of quartz-andesites in the South Seas, 294, 309;
   _see_ the preface

 Wilkes, Commodore, 11, 15, 19, 25, 363, etc.

 Wittstock, Mr., 36, 47, 53


 Yanawai coast, 95-97, 122

 Yanawai, river and valley, 113-116, 121

 Yanganga Islands, 2

 Yanutha Point, 123, 284

 Yaroi, 189, 325

 Yasawa Group, 294


 Zeolites, formation of, in palagonite, 338

 Zeolitic palagonite-tuffs, 334

 Zirkel, on intergrowths of rhombic pyroxene, 306;
   on palagonite, 343;
   on magnetic rocks, 361


                                THE END


   R. CLAY AND SONS, LTD., BREAD ST. HILL E.C., AND BUNGAY, SUFFOLK.



                         ---------------------
Footnote 1:

  In the case of the island of Faro in the Solomon Group, I have
  described a similar process of island-building. (_Geology of the
  Solomon Islands_, p. 37.)

Footnote 2:

  In 1897 I spent several months in travelling over this island and
  ascended, sometimes more than once, the three great volcanic
  mountains. Perhaps at some future time I may renew my examination of
  this interesting region.

Footnote 3:

  Strictly speaking Korolevu indicated in the profile would not be
  visible.

Footnote 4:

  Mariko is the native name of the Drayton Peak of the chart. Mbatini is
  the correct name for the Koro Mbasanga of the chart, the true Koro
  Mbasanga lying three miles to the north. Thambeyu is a native name for
  the Mount Thurston Range.

Footnote 5:

  There has been some confusion in the native names of the peaks in this
  part of the island, which I have not been able to remove.

Footnote 6:

  _A Year in Fiji_, 1881, pp. 22, 167.

Footnote 7:

  _Geology of the United States Exploring Expedition_, 1849.

Footnote 8:

  This height has been supplied from memory, as I omitted to refer to
  the exact level of the erosion line in my notes.

Footnote 9:

  They were described to me as dry for a fortnight at a time. I was
  prevented from making more than an occasional visit to them.

Footnote 10:

  _Atlas of the United States Exploring Expedition_, vol. i.,
  Philadelphia, 1850.

Footnote 11:

  This, however, is not the case with the recent changes at the mouth of
  the Rewa River in Viti Levu, where the bare sandy point of Lauthala
  has extended itself seaward between 500 and 600 yards since 1840,
  whilst Port Nukulau has shoaled a fathom in the same period. But I can
  find no evidence of any marked advance in the mangrove margins either
  towards Nukulau or on the Kamba side, the only change recognisable
  being in the bare _sandy_ point of Lauthala, the rapid extension of
  which has been such as to attract the attention of residents, both
  whites and natives. Dana, who was in this locality in 1840, remarks in
  the _Geology of the U.S. Exploring Expedition_, that he had learned
  from a person who had resided there for forty years that during this
  period the deposits had lengthened the river half a mile. When I was
  on the Rewa in 1897 I heard that the natives in old time could see
  Suva Point from Rewa. This is probably a native legend connected with
  the modern extension of Lauthala Point. (The charts compared in making
  the above measurement of the recent advance of this point were the
  plan of the Rewa Roads by Wilkes, in 1840, and the Admiralty charts
  1757 and 905, the former of which was based on Lieut. Dawson’s survey
  in 1875, the last being corrected to 1897.)

Footnote 12:

  Between Mathuata Island and the coast a change is indicated from 9-10
  fathoms to 8-9 fathoms, north of Motua Island 12-13 to 11-12, and
  between Nangano and Thakavi 16 to 14 fathoms.

Footnote 13:

  By referring to the chart it will be seen that extensive mud-flats
  occur at the mouths of the Sarawanga and Ndreketi rivers, where the
  land-margin is slowly advancing.

Footnote 14:

  _United States Exploring Expedition_, vol. x.; _Geology_, by J. D.
  Dana, p. 343.

Footnote 15:

  _A Year in Fiji_, by John Horne, London, 1881, p. 163.

Footnote 16:

  Journal, _Royal Geographical Society_, 1857, vol. 27.

Footnote 17:

  _Scottish Geographical Magazine_, August, 1891.

Footnote 18:

  _Journal des Museum Godeffroy_, heft 14, Hamburg, 1879.

Footnote 19:

  Dr. Max Büchner also refers to this spring in his _Reise durch den
  Stillen Ozean_, 1878.

Footnote 20:

  _Bulletin Museum Comparative Zoology_, Harvard, vol. 38; Geolog.
  Series V., No 1, Nov. 1900.

Footnote 21:

  Amongst the other descriptions of these springs I may refer to that of
  Kleinschmidt in the work quoted on p. 22, to that of Miss Gordon
  Cumming in _At Home in Fiji_, to that of Horne in his _Year in Fiji_,
  &c. They are sketched in the descriptions of Kleinschmidt, Miss
  Cumming, and Commodore Wilkes. The analyses are given on a later page
  together with the references.

Footnote 22:

  _Pacific Islands, Sailing Directions_, vol. ii., Central Groups, 1900,
  p. 185.

Footnote 23:

  From what I remember the usual exposure at low-water in 1898 was less
  than a foot. I have little doubt as to the identity of the locality.
  This rock is one of the “sights” of the place at the present time. It
  would be interesting for a resident to compare carefully its present
  condition with that as described by Wilkes. Dana in the work quoted on
  p. 10, refers to this rock as a knoll of basalt; but he never visited
  the locality and only obtained his account from the officers of
  Wilkes.

Footnote 24:

  _Narrative of the United States Exploring Expedition_, III., 199, by
  Commodore Wilkes. See also Dana’s _Geology_ of the same expedition.

Footnote 25:

  _Journal des Museum Godeffroy_, heft 14, Hamburg, 1879.

Footnote 26:

  _Islands of the Pacific_, by H. Stonehewer Cooper, 1888 edition.

Footnote 27:

  _Journal Royal Society_, New South Wales, 1880, vol. 14. Miss Gordon
  Cumming in _At Home in Fiji_ gives the same analysis but differently
  stated.

Footnote 28:

  To avoid error, I have given the results of each without converting
  them to a common standard. The numbers in brackets are taken from the
  form of Prof. Liversidge’s analysis given in Miss Gordon Cumming’s
  book.

Footnote 29:

  _United States Exploring Expedition_, vol. 10, Geology.

Footnote 30:

  _Proceedings_, Queensland Branch, Geographical Society, Australia,
  vol. 1. 1886.

Footnote 31:

  I took the temperature at monthly intervals between October, 1896, and
  September, 1897. The mean annual temperature of the air in the shade
  would be about 64° at an elevation of between 3,000 and 4,000 feet.

Footnote 32:

  At Ewa there are pumping plants capable of supplying 75 million
  gallons a day, the water being drawn entirely from artesian wells.
  (_Report on Hawaii_, by Dr. Stubbs, bulletin 95, 1901; U.S. Department
  of Agriculture.)

Footnote 33:

  This hill is figured in Wilkes’ narrative under the name of Dillon’s
  Rock (vol. 3, p. 235). This, however, is not the Dillon’s Rock of his
  chart, where the name is given to a rock on the west side of the
  entrance to Wailea Bay.

Footnote 34:

  See remarks on “crush-tuffs” on p. 334.

Footnote 35:

  Species not identified.

Footnote 36:

  In one of my traverses I crossed a level district extending a mile
  N.E. of Ndriti without changing my elevation.

Footnote 37:

  At Delanasau, on the north or dry coast of the island, the average
  rainfall, according to many years’ observations by Mr. Holmes, is
  about 115 inches. At Wainunu, near the wet or south coast, the
  observations of Mr. Barratt and others extending over 16 years give an
  average of 160 inches. In the mountains this would be nearly doubled.

Footnote 38:

  This question, which has so often been raised with respect to the
  propylites, will probably receive a different answer from different
  localities. The matter is further discussed on later pages.

Footnote 39:

  The dyke-rock has a specific gravity of 2·7; but is slightly
  vesicular. It shows a few small plagioclase phenocrysts in a
  groundmass of felspar-lathes, augite grains and prisms, magnetite, and
  a little brown interstitial glass. The felspar-lathes average ·14 mm.
  in length and are for the most part not parallel. Secondary calcite
  occurs in the groundmass, and the powdered rock effervesces a little
  in an acid.

  The rock forming the offshoot of the dyke differs only from the parent
  rock in its more vitreous character. Although the felspars and augites
  of the groundmass are fairly developed, the residual glass is much
  more copious, and in places where it has segregated, forming
  “lakelets,” it has been subjected to an alteration often observed in
  palagonite when there are concentric alternating zones of a
  tan-coloured fairly refractive material and calcite.

  The reddish scoriaceous lava in contact with the dyke shows no
  phenocrysts. The groundmass displays more or less parallel
  felspar-lathes, ·1 mm. long, augite grains, and much magnetite. The
  residual glass is fair in quantity; but is mostly gathered into
  “lakelets” of brown altered glass with sometimes calcite in the
  centre.

  The vitreous border of the dyke is composed of a dark glass quite
  opaque in the outer portion, but clearer and showing incipient
  crystallisation in the inner portion.

Footnote 40:

  _Characteristics of Volcanoes_, 1890.

Footnote 41:

  Referred to genus 16 of the olivine-basalts.

Footnote 42:

  This subject is discussed in Chapter XXVI.

Footnote 43:

  “Na Savu” is the Fijian for waterfall. The complete name of this fall
  is “Na Savu ni nuku.”

Footnote 44:

  The flinty concretions are described on page 354, and the iron sand on
  p. 356.

Footnote 45:

  _Geology of the United States Exploring Expedition._

Footnote 46:

  A similar arrangement was observed in the columnar basalt of Kauai in
  the Hawaiian Islands. It is presumed that these Hawaiian flows are
  sub-aërial.

Footnote 47:

  The unaltered glass, which incloses a few plagioclase phenocrysts, has
  a specific gravity of 2·7, and is readily fusible.

Footnote 48:

  They are described on p. 322.

Footnote 49:

  This absence of a healthy forest-growth, such as occurs on the level
  summit of the neighbouring Soloa Levu and in all like situations, has
  probably some geological significance.

Footnote 50:

  These tuffs are probably submarine. They will be found described with
  tuffs of the same character on p. 333.

Footnote 51:

  The track attains an elevation of about 1,300 feet, but the top of the
  watershed is two or three hundred feet lower.

Footnote 52:

  It belongs to the 3rd order of the hornblende-hypersthene-andesites
  described on p. 299.

Footnote 53:

  Occasional views of its summit only are obtained from the eastward, as
  from the Ndrandramea mountains and their vicinity.

Footnote 54:

  Alluvial gold has long been known to occur in the bed of the Yanawai
  below Vatu Kaisia; but it has never been found in paying quantity.

Footnote 55:

  Under the microscope it is shown to be granular in structure,
  exhibiting a mosaic of irregular quartz grains.

Footnote 56:

  The blocks of the agglomerate in this last locality are from one to
  three feet across.

Footnote 57:

  It displays in the groundmass augite prisms in flow-arrangement, and
  is referred to genus 20 of the augite-andesites.

Footnote 58:

  Referred to genera 16 and 20 of the augite-andesites.

Footnote 59:

  These foraminiferal limestones are described on p. 319.

Footnote 60:

  _Proceedings_, Queensland Branch, Geographical Society of Australasia,
  Brisbane, 1886, vol. i.

Footnote 61:

  _Geology of the United States Exploring Expedition._

Footnote 62:

  _Pacific Islands_, vol. ii. 1900.

Footnote 63:

  It is referred to the 5th sub-order (genus 18) of the
  hypersthene-augite-andesites characterised by prismatic pyroxene and
  more or less parallel felspar lathes in the groundmass, as described
  on p. 289. It displays abundant opaque porphyritic plagioclase giving
  extinctions of oligoclase-andesine. The pyroxene phenocrysts have dark
  alteration-borders. There is a little altered interstitial glass.
  Spec. grav. 2·55.

Footnote 64:

  I did not ascend to the top of Ulu-i-mbau. It is, however, evidently
  composed of basic andesitic rocks, occasionally amygdaloidal. On its
  slopes up to at least 600 feet above the sea occur agglomerate-tuffs
  and finer submarine tuffs, as above described, overlying
  foraminiferous clays, a submergence of quite 500 feet being indicated
  by the investing deposits.

Footnote 65:

  I did not find any foraminiferal shells or other organic remains
  either in this tuff or in the similar tuffs occurring on the adjacent
  slope of Va-lili up to 1,100 feet. My specimens, however, are very
  small.

Footnote 66:

  It rises in the background of the view.

Footnote 67:

  They are described under sample E on p. 332.

Footnote 68:

  Referred to genus 37 of the olivine-basalts.

Footnote 69:

  On the right side of the river close to Vatu-kawa there are some
  cliffs displaying a section of the mountainous spur, referred to on p.
  151, that separates the valleys of the Mbale-mbale and Vatu-kawa
  rivers, an exposure quite apart from the rocks exhibited on the
  adjacent southern slopes of the main range. These cliffs are formed of
  bedded grey tuffs marked by single layers of blocks 6 to 8 inches
  across and dipping about 30° S.S.W. The tuffs in their texture are not
  unlike sub-aerial tuff-deposits. They contain no lime and are composed
  of basic materials with a little palagonite. They seem to indicate
  some subsidiary vent, close to the present village of Vatu-kawa, which
  may have been active shortly before or during the emergence of this
  district.

Footnote 70:

  These altered tuffs on the southern slope of this range are described
  on p. 332.

Footnote 71:

  Referred to genus 16, species A, sub-species 1, of the
  augite-andesites.

Footnote 72:

  Referred to genus 9, sub-genus A, of the augite-andesites.

Footnote 73:

  Referred to genus 9, sub-genus B, of the augite-andesites.

Footnote 74:

  Both these rocks belong to the hypersthene-augite andesites, showing
  phenocrysts of both monoclinic and rhombic pyroxene. The first belongs
  to the orthophyric order described on p. 290; whilst the second
  belongs to the second order (genus 13, p. 287) where the felspars of
  the groundmass are lathe-like and in flow arrangement.

Footnote 75:

  The highest dyke trends N. 48° E. and is inclined from the vertical
  about 15° N.W. The dyke, 5 or 6 yards below it, trends N. 30°E. and is
  vertical. The dyke, 50 yards farther down, trends N. 35° E. and is
  inclined from the vertical about 5° N.W. The inclination was only
  estimated. The bearings are true.

Footnote 76:

  Both the types are referred to genus 1 of the augite-andesites, the
  olivine, when present, being quite insufficient to give a character to
  the rock. They however belong to different species according to the
  length of the felspar-lathes, which in the doleritic rocks averages
  ·2 mm. and in the other type ·08 mm.

Footnote 77:

  It is pointed out on p. 5 that this name is wrongly applied in the
  Admiralty charts to Mount Mbatini, a mountain about three miles south
  of it.

Footnote 78:

  Referred to genus 13 of the augite-andesites.

Footnote 79:

  They are blackish and somewhat compact (sp. gr. 2·67-2·71) and have
  very small felspar-lathes less than ·1 mm. long. They contain both
  rhombic and monoclinic pyroxene, and are referred to genera 1 and 13
  of the hypersthene-augite-andesites.

Footnote 80:

  I discovered this error in a rather practical fashion by ascending the
  wrong mountain. The natives were engaged to take me to Koro-mbasanga
  and they performed their task, my aneroid and compass soon indicating
  that I was not on the highest peak of the island, but on a lesser peak
  three miles north of it.

Footnote 81:

  Vula Votu is the name of a peak lying to the east. Ngoinangai is a
  forked mountain still further east.

Footnote 82:

  A kind of “edible” bird’s-nest is found in this cave.

Footnote 83:

  This rock is described on p. 251.

Footnote 84:

  It is described under Sample C on p. 325.

Footnote 85:

  It is granular, but fuses in the blowpipe flame into a clear glass and
  gelatinises in HCl. Probably a form of natrolite.

Footnote 86:

  It is described under Sample D on p. 326.

Footnote 87:

  _Ten Months in Fiji_, London, 1864.

Footnote 88:

  _A Year in Fiji_, pp. 154, 169; London, 1881.

Footnote 89:

  It displays an abundance of small phenocrysts of plagioclase, augite,
  and olivine partly serpentinised, in a groundmass composed in the main
  of coarse augite grains (·025 mm. in size) and of felspar microliths
  (·07 mm. in length) in smaller proportion, with little if any residual
  glass. Specific gravity 2·98. It is near the Waikawa basalt, referred
  to on p. 202, and is placed in the same genus (13) of the olivine
  class.

Footnote 90:

  They are described on p. 269 under the non-porphyritic sub-genus of
  genus 2 of the augite-andesites.

Footnote 91:

  These rocks are in most cases referred to the orthophyric and felsitic
  orders of the hypersthene-augite andesites. The rocks of the
  last-named order prevail, and form the type of the group, as described
  on p. 291.

Footnote 92:

  It belongs to genus 37 of the olivine class. The felspar-lathes
  average 0·2 mm. in length, and there is a little altered interstitial
  glass.

Footnote 93:

  It is referred to genus 16, species D, of the augite-andesites. The
  felspar-lathes have an average length of ·3 mm.

Footnote 94:

  _Proceedings_, Queensland Branch, Geographical Society of Australia,
  vol. i.; 1886.

Footnote 95:

  Referred to genus 9 of the augite-andesites.

Footnote 96:

  The general characters of these rocks are described on p. 308.

Footnote 97:

  _Geology of the United States Exploring Expedition._

Footnote 98:

  _A Year in Fiji_, p. 22.

Footnote 99:

  Described on p. 310.

Footnote 100:

  These rocks are described on p. 308.

Footnote 101:

  It contains small phenocrysts of plagioclase (medium andesine), and of
  augite and rhombic pyroxene, and is referred to genus 1 of the
  hypersthene-augite andesites.

Footnote 102:

  See paper quoted on p. 31. It is noteworthy that Mr. Horne refers only
  to a single floating island.

Footnote 103:

  _Journal_, Royal Geographical Society, June, 1894.

Footnote 104:

  Described on p. 309.

Footnote 105:

  Described on p. 310.

Footnote 106:

  Referred to genus 13 of the augite-andesites. The felspar-lathes
  average ·1 mm. in length, and there is a little interstitial glass.

Footnote 107:

  Referred to genus 16 of the augite-andesites. There are two sets of
  felspar-lathes in the groundmass; the larger, ·23 mm. long, are more
  or less parallel; the smaller, ·04 mm. long, form a plexus.

Footnote 108:

  Each division would theoretically also possess an orthophyric and a
  felsitic order; but these orders are not represented in my collection
  and need only be mentioned.

Footnote 109:

  Not represented in the collection.

Footnote 110:

  I apply the term “diorite” to granitoid rocks formed entirely of
  plagioclase and hornblende.

Footnote 111:

  According to the size of the plagioclase phenocrysts, whether
  averaging less than or more than 3 mm. in size, these rocks may be
  divided into a non-porphyritic and a porphyritic sub-genus.

Footnote 112:

  _Petrographie des Viti Archipels_; Miner. und Petrogr. Mittheil; band
  v, heft 1, Wien, 1882.

Footnote 113:

  _Physics and Chemistry_, II. Report Scient. Results; H.M.S.
  Challenger; London, 1889.

Footnote 114:

  _Petrographical Notes on the Fiji Islands_; Proceed. Amer. Acad. Arts
  and Sciences; vol. 34; no. 21; May, 1899.

Footnote 115:

  _Notes on the Limestones and General Geology of the Fiji Islands_,
  Bull. Mus. Comp. Zool, vol. 5. Geolog. Ser. vol. 5, no. 1, Cambridge,
  Mass., U.S.A. Nov. 1900.

Footnote 116:

  Wichmann describes rocks from the cliffs of the Singatoka river and
  from Ovalau.

Footnote 117:

  _Geology of the Solomon Islands_, by H. B. Guppy, 1887, pp. 6, 36.

Footnote 118:

  The term “felsitic andesite” is suitable for this microfelsitic type.

Footnote 119:

  I have borrowed this term from Rosenbusch’s _Microscopical
  Physiography of the Rock-making Minerals_, translated by Iddings.

Footnote 120:

  _Challenger Reports_, Physics and Chemistry II.

Footnote 121:

  _Neues Jahrb. fur Mineralogie_, 1884.

Footnote 122:

  For their mode of occurrence, see pp. 215, 220, 230-233.

Footnote 123:

  Highly altered rocks of this class are exposed at the base of Mount
  Nailotha as described on p. 215.

Footnote 124:

  See work quoted on p. 218.

Footnote 125:

  See p. 230.

Footnote 126:

  _Quart. Journ. Geolog. Soc._ xliv. 1888.

Footnote 127:

  Zirkel’s _Petrographie_, iii., 694.

Footnote 128:

  This basalt is not fusible in the ordinary blow-pipe flame.

Footnote 129:

  In this connection see the description of the Soloa-levu pitchstone on
  p. 312.

Footnote 130:

  See the note at the end of this chapter.

Footnote 131:

  Quoted in Zirkel’s _Petrographie_, iii., 689.

Footnote 132:

  I have visited this locality on several occasions with the special
  object of studying the relation of the basalt to the associated
  palagonite-tuffs and clays. A general discussion of this question
  would be out of place here; but I may remark that the conclusion
  arrived at by me is that these deposits are not sedimentary but are
  entirely the result of the disintegration of palagonite _in situ_.
  This is quite opposed to the view of their sedimentary origin held by
  Dr. Johnston-Lavis, Prof. Platania, and other Italian geologists....
  The basalt is scoriaceous, semi-vitreous, and semi-ophitic, and
  closely approaches the type of basalt above defined.

Footnote 133:

  _Reisen auf den Viti-Inseln_, as quoted on p. 22.

Footnote 134:

  _Petrographie des Viti-Archipels_, quoted on p. 293.

Footnote 135:

  See work quoted on p. 378.

Footnote 136:

  _Solomon Islands and their Natives_, by H. B. Guppy, p. 78.

Footnote 137:

  The region is described on pp. 224-228.

Footnote 138:

  For the meaning of “talasinga” see p. 55.

Footnote 139:

  The portion exhibiting the coral structure has a specific gravity of
  2·54.

Footnote 140:

  A good list of references to the early German authorities on the
  subject is given in the _American Journal of Science and Arts_ for
  1831, vol. 20.... Zirkel in his _Lehrbuch der Petrographie_ (1893,
  vol. i. p. 565) gives most of these and many more recent.... Harker in
  his paper below named refers to a review of the earlier literature in
  _Verh. naturh. Vereins_. Bonn, 1851, vol. 8, and to a more complete
  bibliography by Meli in _Boll. Soc. Geol. Ital._ 1881, vol. 9....
  British Association Report in 1889 by Professors Rücker and Thorpe on
  the Magnetic State of the British Isles.... _Nature_ for August and
  September, 1894, &c.... Harker on magnetic disturbances in the Isle of
  Skye, _Proc. Cambr. Philos. Soc._ vol. 10, part 5.... Skinner in
  _Proc. Cambr. Philos. Soc._ May, 1894.... Clark in _Journ. Roy.
  Instit._ Cornwall, 1890-93.... Folgheraiter in _Frammenti concernanti
  la geofisica_, Rome: referred to in _Nature_, July 27, 1899, and Nov.
  8, 1900.

Footnote 141:

  Nearly all volcanic rocks at all basic are magnetic, owing to the
  constant presence of magnetite; but magnetic polarity, when the
  rock-fragment has a negative and a positive pole, is not directly
  concerned in volcanic rocks with the mineral composition.

Footnote 142:

  Some of the earliest observations were made on granites and trachytes.

Footnote 143:

  _Geology of the United States Exploring Expedition_, 1849, p. 294.

Footnote 144:

  _Quarterly journal Geological Society_, vol. li., p. 469.

Footnote 145:

  Wilkes’ _Narrative of the U.S. Exploring Expedition_, iii., 185.

Footnote 146:

  Of the tuffs and clays, almost all submarine and often containing
  tests of foraminifera and sometimes molluscan shells, about 90 per
  cent. exhibit simple magnetism in a slight degree, but out of nearly
  100 specimens tested none show polarity.

Footnote 147:

  On p. 357 will be found some notes on the magnetic iron sand that
  occurs in great abundance in river and stream beds.

Footnote 148:

  These values represent the number of degrees that the magnetic needle
  is repelled. The method is described above. A note on the average
  amount of polarity found in all my polaric rocks is given at the end
  of the chapter. The term “dacite” _is here an equivalent_ of “felsitic
  andesite.”

Footnote 149:

  This name has been wrongly applied in the Admiralty chart to the
  mountain of Mbatini. Koro-mbasanga, 2,500 feet, lies three miles to
  the north.

Footnote 150:

  This rock is described on p. 109. There is no exceptional development
  of magnetite for a basic rock in the groundmass.

Footnote 151:

  Unfortunately, I have no data for the peaks of Na Raro and Vatu
  Kaisia, except that specimens obtained below the summits are
  non-polaric. In the case of Na Raro I did not retain the specimen
  obtained at the top; whilst in my ascent of Vatu Kaisia I did not
  quite reach the summit.

Footnote 152:

  The mode of measurement is described on p. 366.

Footnote 153:

  See pp. 2, 15, 18, 56, 62, 72, &c.

Footnote 154:

  _The Islands and Coral Reefs of Fiji_, Bull. Mus. Comp. Zool. Harv.
  Coll. vol. 33, 1899.

Footnote 155:

  _Scott. Geogr. Mag._ 1895.

Footnote 156:

  _Ancient Volcanoes of Great Britain_, by Sir A. Geikie, 1897, ii. 260.

Footnote 157:

  See Wichmann in _Min. und Petrog. Mitth._ band v. heft 1.

Footnote 158:

  _Amer. Journ. Sci._ VI. 165, 1898. See also Agassiz on the _Islands
  and Coral Reefs of Fiji_, before quoted.

Footnote 159:

  _Proc. Linn. Soc._ N.S.W. 1879-80, p. 358.

Footnote 160:

  _Quart. Journ. Geolog. Soc._ vol. 44, 1888.

Footnote 161:

  See Preface to the report of Mr. Andrews quoted on a later page.

Footnote 162:

  _Das Antlitz der Erde_, French edition by E. de Margerie, ii. 534.

Footnote 163:

  See the paper before quoted on the coral reefs of Fiji.

Footnote 164:

  _Bull. Mus. Comp. Zool._ Harv. Coll. vol. 38. _Geolog. Ser._ vol. 5,
  no. 1, 1900. _On the Limestones and General Geology of the Fiji
  Islands_, by E. C. Andrews.

Footnote 165:

  _Quart. Journ. Geolog. Soc._ vol. 47, p. 590, 1891. See also Mr.
  Harker’s paper below quoted.

Footnote 166:

  _Geolog. Mag._ June, 1891.

Footnote 167:

  Seemann’s _Flora Vitiensis_, Horne’s _Year in Fiji_, Hillebrand’s
  _Flora of the Hawaiian Islands_, Hemsley’s “Flora of the Tonga
  Islands” in _Journal Linnean Society_, Botany, vol. 30.

Footnote 168:

  See Hutton _Proc. Linn. Soc._ N.S.W. 1896, Baur, _Amer. Nat._ 1897,
  Pilsbry, _Proc. Nat. Sci. Philad._ 1900, Hedley, _Proc. Linn. Soc._
  N.S.W. 1892, 1899, &c.

Footnote 169:

  The term “basalt” is here used in a general sense to include
  olivine-basalts, basaltic andesites, and other basic types of the
  augite-andesites and hypersthene-augite andesites.

Footnote 170:

  The usual signification of “koro” as a prefix or part of names of
  hills and mountains is a “prominence” or “projection.” It is a mistake
  on my part to assume that in such cases it is as a rule equivalent to
  a town or village.



                         ---------------------

                           Transcriber's Note

There is some incosistent spelling and hyphenation in this book which
has not been normalized.

Some corrections have been made to the text. In particular, punctuation
was corrected. Additionally, the following changes have been made:

                   p. 10 beween -> between
                   p. 75 ·15 m. long -> ·15 mm. long
                   p. 114 ts -> its
                   p. 208 aud -> and
                   p. 244 Pyroxyene -> Pyroxene
                   p. 270 adoped -> adopted
                   p. 345 tea-estat -> tea-estate





*** End of this Doctrine Publishing Corporation Digital Book "Observations of a Naturalist in the Pacific Between 1896 and 1899, Volume 1 - Vanua Levu, Fiji" ***

Doctrine Publishing Corporation provides digitized public domain materials.
Public domain books belong to the public and we are merely their custodians.
This effort is time consuming and expensive, so in order to keep providing
this resource, we have taken steps to prevent abuse by commercial parties,
including placing technical restrictions on automated querying.

We also ask that you:

+ Make non-commercial use of the files We designed Doctrine Publishing
Corporation's ISYS search for use by individuals, and we request that you
use these files for personal, non-commercial purposes.

+ Refrain from automated querying Do not send automated queries of any sort
to Doctrine Publishing's system: If you are conducting research on machine
translation, optical character recognition or other areas where access to a
large amount of text is helpful, please contact us. We encourage the use of
public domain materials for these purposes and may be able to help.

+ Keep it legal -  Whatever your use, remember that you are responsible for
ensuring that what you are doing is legal. Do not assume that just because
we believe a book is in the public domain for users in the United States,
that the work is also in the public domain for users in other countries.
Whether a book is still in copyright varies from country to country, and we
can't offer guidance on whether any specific use of any specific book is
allowed. Please do not assume that a book's appearance in Doctrine Publishing
ISYS search  means it can be used in any manner anywhere in the world.
Copyright infringement liability can be quite severe.

About ISYS® Search Software
Established in 1988, ISYS Search Software is a global supplier of enterprise
search solutions for business and government.  The company's award-winning
software suite offers a broad range of search, navigation and discovery
solutions for desktop search, intranet search, SharePoint search and embedded
search applications.  ISYS has been deployed by thousands of organizations
operating in a variety of industries, including government, legal, law
enforcement, financial services, healthcare and recruitment.



Home