Full text of "Scientific reports"

^AUSTRALASIAN ANTARCTIC EXPEDITION

1911 - 1914.

UNDER THE LEADERSHIP OF SIR DOUGLAS MAWSON. D.S B.E.

SCIENTIFIC REPORTS.

Series A. -Geography, Physiography, Glaciology, Oceanography,

and Geology.

VOL III. PART 1.-^

THE

METAMORPHIC ROCKS OF ADELIE LAND

SECTION I.

F. L. STILLWELL, D.Sc.

WITH THIRTY-FIVE PL

PRICE. TWO GUINEAS.

ISSUED MARCH 25th. 1918.

PLAN OF PUBLICATION.

SERIES A.
GEOGRAPHY, OCEANOGRAPHY, GEOLOGY, GLACIOLOGY.

SERIES B.
METKOROLOCY. AURORAL OBSERVATIONS, "WIRELESS" OBSERVATIONS, MAGNETICS, TIDES.

SERIES C.
ZOOLOGY, BOTANY, BACTERIOLOGY.

REPORTS ALREADY ISSUED.

SERIES C. PEIOE .

Vol. Part. a . d .

HI. 1. FISHES. By Mr. EDGAR E. WAITB, F.L.S., South Australian Museum, Adelaide 8 6

IV. 1. MOLLUSCA : PELECYPODA AND GASTROPODA.

By Mr. C. HEDLEY, F.L.S., Australian Museum, Sydney 8 6

IV. -J. MOLLUSCA : CEPHALOPODA. By Dr. S. STILLMAN BEERY, Redlands, California 3 6

1. ARAOHNIDA FROM MACQUABIK ISLAND.

By Mr. W. J. RAINBOW, F.E.S., Australian Museum, Sydney I n

BRACHYURA. By Miss MARY J. RATHBUN, U.S. National Museum, Washington 1

I'EPODA. By Dr. (J. STEWARDS BRADY, F.R.S 5 <;

. VIXK'KIU AND HALOCYPRID/E. By Dr. G. STEWARDSON BRADY, F.R.S , 2

AUSTRALASIAN ANTARCTIC EXPEDITION

1911 - 1914.

UNDER THE LEADERSHIP OF SIR DOUGLAS MAWSON, D.Sc.. B.E.

SCIENTIFIC REPORTS.

Series A. -Geography, Physiography, Glaciology, Oceanography,

and Geology.

VOL. III. PART 1.
THE

METAMORPHIC ROCKS OF ADELIE LAND

SECTION I.

F. L. STILLWELL, D.Sc.,
CORRIGENDA.

Page 13. In second last line of second paragraph for " Evidence " read " Evidence."

Page 29. Table I. In column 1, for iron ore 3-0 read iron ore 2-7

In column 10, for mica 10-8 read mica 10-5

In column 12, for hornblende 56-3 read hornblende .... 60-0

epidote *7 epidote 1-0

iron ore -7 iron ore -6

Page 39. In footnote, for J. C. H. Mengaye read J. C. H. Mingaye.

Page 48. The first portion of the second line should read " in varying degrees on those of an adjacent

zone."
Page 65. In the first reference at the foot of the page, for Journ. Geol., vol. 3, p. 1, read Journ. Geol.,

voL 23, p. 1.

Page 63. In the list of projection values, for I. .0-20 read f . .20-0.
Page 87. In the group values of Rainy Lake gneiss, for 8. .74-9 read 8. .74-0.
Page 94. In seventh line from bottom, for (p. 39) read (p. 89).
Page 95. In fourth line of middle paragraph, for (p. 107) read (p. 105).
Page 96. In sixteenth line from bottom, for (p. 45) read (p. 47).
Page 117. In ninth line from top,/or (p. 41) read (p. 141).
Page 195. In sixteenth line from bottom, for (p. 121) read (p. 221).

fat

v.3

CONTENTS.

CHAP. PAGE.

Introduction 7

I. Summarised Account of the Metamorphic Rock Types found in situ

in Adelie Land 9

II. Physiography of Cape Denison 15

Agents of Denudation 15

Origin of Valleys 18

Moraines 19

Note on a Consolidated Beach Sand, by F. Chapman 21

Note on Morainic Material, by F. Chapman 22

III. Metamorphosed Dyke Series of Cape Denison 23

1. Nomenclature 23

2. Field Characters 25

3. Petrographical Characters 27

Summary 37

4. Crystalloblastic Order 40

5. Chemical Characters 41

Grubenmann's Classification of the Crystalline Schists .... 43

The Classification of the Cape Denison Amphibolites 45

6. Metamorphosed Xenoliths '. ...:... 48

Saussuritic Type 48

Gneissic Type 51

Origin of Meta-xenoliths 53

7. The Origin of the Amphibolite Series 55

8. The Origin of certain Clots in the Metamorphosed Dykes and

Metamorphic Differentiation 58

9. Description of the Coarsely Crystalline Basic Patches in the

Granodiorite Gneiss 65

4 CONTENTS.

CHAP. PAGE -

10. Origin of the Coarsely Crystalline Basic Patches and Meta-

morphic Diffusion 70

11. Further Examples of Metamorphic Diffusion 72

Junction Specimens 72

Composite Gneiss 73

12. Further Examples of Metamorphic Differentiation 76

Hornblende, Sphene, Magnetite, Felspar 76

13. Review and Discussion of Field Phenomena 79

IV. The Granodiorite Gneiss of Cape Denison 84

The Aplite Gneisses of Cape Denison 89

Interpretations of Certain Variations in the Granodiorite Gneiss 91

V. Correlation with similar Rocks of other Areas and the application of
the Conceptions of Metamorphic Differentiation and Metamorphic

Diffusion 93

1. General 93

2. North- West Highlands of Scotland 94

3. Haliburton and Bancroft Areas, Canada 98

4. Highlands of New Jersey 103

The Assimilation Theory 105

5. The Lizard, Cornwall 106

Summary 117

6. Chemical Composition as a Criterion in Determining the

Origin of Metamorphic Rocks 118

7. Conclusion 121

VI. The Mackellar Islets 122

VII. Cape Hunter 124

VIII. The Madigan Nunatak .' 128

Plagioclase Pyroxene Gneiss 128

Hypersthene Alkali Felspar Gneiss 133

IX. Aurora Peak 138

Hornblende Plagioclase Pyroxene Gneiss 138

Hypersthene Alkali Felspar Gneiss 139

CONTENTS. 5

CHAP. PAGE.

X. The Cape Gray Promontory and Still well Island 144

Description of Localities 1 44

The Garnet Gneisses 146

Cape Gray 146

Garnet Point 147

Cape Pigeon Rocks 151

Stillwell Island 151

Chemical Characters 152

The Acid Hypersthenic Gneisses 154

Stillwell Island 155

Cape Pigeon Rocks 159

Summary 1 04

XI. The Cape Gray Metamorphosed Dyke Series 168

Cape Gray 1 68

Stillwell Island 171

Cape Pigeon Rocks 177

Garnet Island 181

Chemical Characters 183

Summary 186

Correlation with other Areas 188

XII. Relation between the Rocks at Cape Gray, Madigan Nunatak, and

Aurora Peak 190

The Charnockite Series of India 193

Fermor's Hypothesis of an Infraplutonic Zone 198

The Kodurite Series 198

XIII. The General Problem of Transference of Material during Metamorphism 200

Evidence of Migration in Geological Literature 200

The Process of Migration 203

Solution 203

Solid Diffusion 204

Force of Crystallisation 208

X I V. Description of Plates 210

General Index 221

Number Index of Rock Specimens 229

THE METAMORPHIC ROCKS OF ADELIE LAND.

SECTION I.

By F. L. STILLWELL, D.Sc. (Geologist to the Australasian Antarctic Expedition),

UmvusaiTY OF MELBOURNE.

INTRODUCTION.

Adelie Land is a portion of the Antarctic Continent which lies in the region
surrounding Long. 143 and Lat. 67. It consists for the most part of a huge ice-covered
plateau which rises rather steeply from the coast and reaches a height of over 6,000ft.
at 300 miles inland. It appears, at first sight from the ship, as the side of a vast dome-
shaped shield of ice, which descends to the ice cliffs or ice barrier on the seaward aide
and rises to about 1,500ft. on the southern horizon.

The ice cliffs present a vertical face varying between 80ft. and 120ft. in height,
and delimits the boundary of the land ice sheet. In many cases the cliffs form the edge
of floating glacier tongues or marginal shelf ice, but as the Aurora traced the coast line
from Cape de la Motte to Commonwealth Bay, rock could be frequently seen at the
base of the cliffs. Yet these exposures remained quite inaccessible to us, and only
those outcrops which rise 100ft. or more above sea level and which break the monotonous
line of ice cliffs could be reached by sea or land. These latter outcrops are rare, and
while they can be readily seen from the ship they are only found and approached with
difficulty from the landward side. Rarer still are the nunataks, or islands of rock in
the snow fields, and our knowledge of the rocks of the hinterland remains very largely
dependent on the study of the glacial debris on the moraines.

Three small rocky promontories exist along the 60-mile stretch of coast line of
Commonwealth Bay a broad open bay about 40 miles across the headlands. Of
these the middle one is Cape Denison, on which the Main Base of the Australasian
Antarctic Expedition was situated. The western rocks, always visible from Cape
Denison, form Cape Hunter. The bay is studded in the centre by the Mackellar Islands,
a group of low-lying islands due north of Cape Denison. East of the bay is the Cape
Gray Promontory, thickly fringed with the numerous small rocky islets which
constitute the Way Archipelago. These islets mostly lie in a 2-mile zone around the
edge of the ice cap, and one of the largest is Stillwell Island, on which a landing was
made. On the eastern side of the Cape Gray Promontory, facing Watt Bay, are the
rock outcrops which have been called Garnet Point and Cape Pigeon Rocks.

Madigan Nunatak and Aurora Peak, the two remaining rock exposures dealt with
in this thesis, are widely separated nunataks. Madigan Nunatak is 2,400ft. above sea
level and lies on the ridge which gradually slopes down to Cape Gray, 18 J miles to the

8 AUSTKALASIAN ANTAECTIC EXPEDITION.

north. Aurora Peak is the more easterly and lies on the west side of the Mertz Glacier,
1,750ft. above sea level and about 60 miles distant from Cape Denison.

The rock specimens and the field data of these nine rocky areas were obtained
by the following :

Cape Denison Large rock collection made during the winters of 1912

and 1913 by Sir Douglas Mawson and Stillwell.

Cape Hunter Visited in December, 1913, by Sir Douglas Mawson.

Gt. Mackellar Island Visited in December, 1913, by Sir Douglas Mawson.

Stillwell Island .... Visited in December, 1913, by Sir Douglas Mawson.

Cape Gray Visited in summer, 1912-13, by Stillwell's sledging party.

Garnet Point Visited in summer, 1912-13, by Stillwell's sledging party.

Cape Pigeon Kocks . . Visited in summer, 191 2-1 3, by Stillwell's sledging party.

Madigan Nunatak . . Visited in summer, 1912-13, by Stillwell's sledging party.

Aurora Peak Visited in summer, 1912-13, by Madigan's sledging party.

The total rock collection of the Australasian Antarctic Expedition is very large,
but the bulk of it comes from the glacial moraines at Cape Denison, and is not treated
in the following. The specimens secured by the sledging parties had necessarily to be
limited in size and number. Madigan collected the specimens at Aurora Peak, and
Laseron was a valuable assistant in looking after the rocks obtained by my sledging
party. No geological specimens were obtained by the other three sledging parties,
except a stony meteorite, a very extraordinary find made by Bickerton's sledging party
on the ice plateau.

We have only been able, up to the present, to deal fully with the rocks which were
found in situ. This part is an essential preliminary to the study of rock types from the
moraines. Even this portion could not have been concluded without the active
co-operation of friends and supporters of the A.A.E. We are greatly indebted to
Mr. Herman, Director of the Victorian Geological Survey, who sanctioned the assistance
of the Victorian Geological Survey Laboratory. The 15 rock analyses that are now
presented are the work of the analysts in this laboratory, working under the supervision
of P. G. W. Bayly.

Though a good portion of the work has been done while associated with the Adelaide
University, it was commenced and finished at the Melbourne University. Its progress
throughout has depended wholly on the assistance afforded by the Geological Department
of the Melbourne University and by Professor Skeats. The " sinews of war " have here
been provided in a very large number of excellent rock sections and much useful
criticism has been levelled during discussion at some of the conclusions.

The illustrations at Cape Denison were obtained by Hurley, the official photo-
grapher. Laseron acted as photographer on my sledging trip, -and obtained some very
fine results. The photographs of the rock specimens and the microphotographs of
the sections have been prepared by myself with the apparatus and facilities in Melbourne.

Finally, we must record the active sympathy of the leader of the Expedition, Sir
Douglas Mawson, who entrusted me with the work.

CHAPTER I.

SUMMARISED ACCOUNT OF THE METAMORPHIC ROCKS FOUND IN SITU

IN ADELIE LAND.

In all the outcrops in Adelie Land accessible to us, varying types of crystalline
schists have been found. The chief types may be enumerated as follows :

Locality. Crystalline Schist. Pre-existing Rock Type.

Cape Hunter Phyllite Clay sediment

Cape Denison Granodiorite gneiss Granodiorite

Aplite gneisses Aplites

Amphibolite series Dolerites

Gt. Mackellar Is. . . Granite gneiss Granite

Amphibolite Dolerite

Madigan Nunatak . Plagioclase pyroxene gneiss Dolerite

Hypersthene alkali felspar gneiss Granite

Aurora Peak Hornblende plagioclase pyroxene Dolerite

gneiss

Garnet hypersthene alkali felspar Granodiorite

gneiss

Cape Gray Plagioclase pyroxene gneiss Dolerite

Garnet cordierite gneiss Clay sediment

Garnet Point Amphibolite Dolerite

Cyanite biotite gneiss Clay sediment

Garnet felspar gneiss Sediment (probably)

Stillwell Island . . . Plagioclase pyroxene gneisses . . . Dolerite

Amphibolites Dolerite

Acid hypersthene gneisses Granitic veins and

diorite

Garnet felspar gneiss Sediment (probably)

Cape Pigeon Rocks Amphibolites, etc Dolerite

Garnet gneisses Sediment (probably)

Garnet hypersthene biotite gneiss Diorite (?)

In all cases except Cape Hunter there are two main rock types an acid type and
a basic type. The basic type represents the metamorphic equivalent of a basic igneous
rock that has intruded the granitic rock or the sedimentary rock before the metamor-
phism. In all cases the primary dyke origin of the basic type has been established with
certainty.

At the most westerly area, Cape Hunter, there is an old sedimentary series, now
represented by a phyllite. Nine miles east of Cape Hunter lies the granodiorite gneiss of

10 AUSTRALASIAN ANTARCTIC EXPEDITION.

Cape Denison and its associated basic dykes. Twenty miles further east again, the meta-
morphosed sediments appear again at Cape Gray in the form of garnet cordierite gneisses
and garnet felspar gneisses. Inland from Cape Gray granitic areas have existed at
Madigan Nunatak and Aurora Peak. These facts are illustrated in the accompanying
section (fig. 1).

In the section the granitic mass at Aurora Peak is assumed to be intrusive into
the original sediments, because altered granitic dykes appear on Stillwell Island and
Cape Pigeon Rocks cutting the garnetiferous gneisses. The altered granitic dykes
are not necessarily to be associated with the primary granitic masses of Aurora Peak
and Madigan Nunatak. The presence of tourmaline in the Cape Hunter phyllites is the
only evidence available for representing the intrusive nature of the Cape Denison grano-
diorite gneiss.

The outcrops dealt with are isolated areas extending over 60 miles of country.
Over this great distance it is only to be expected that varying conditions of meta-
morphism would be found. On the west, at Cape Hunter, we have dominant epi-zone
metamorphism. At Cape Denison the metamorphic conditions are intermediate between
those of the epi zone and meso zone of metamorphism, with variation in both directions.
The amphibolites which are completely recrystallised rocks sometimes approach the
character of meso zone rocks and sometimes are more like epi zone rocks. On the Cape
Gray Promontory, where the garnetiferous gneisses abound, and at Madigan Nunatak
and Aurora Peak evidence of kata zone metamorphism is found in all cases. At Madigan
Nunatak very remarkable epi zone metamorphism is superimposed upon the kata zone
metamorphism, while between Madigan Nunatak and Cape Gray and at Aurora Peak
meso zone metamorphism is superimposed upon the kata zone metamorphism. There
is, therefore, quite a distinct regional distribution of the metamorphic products of
Grubenmann's three metamorphic zones and an argument in support of the general
conception produced.

The detailed studies of the dyke series at Cape Denison and at Cape Gray have
produced some extraordinary results. In these, considerable use has been made of
the Rosiwal method of volumetric rock analysis for the purpose of obtaining relative
mineral composition of different specimens. The absolute mineral composition could
only be obtained by three determinations in three planes at right-angles a process
too tedious to be of any service. Yet the mineral composition of schists can be com-
pared with advantage in rock sections which have been cut from a constant direction
relative to the plane of schistosity.

The Cape Denison dyke series ranges from epidote biotite schists at Azimuth Hill,
on the western part of Cape Denison, through biotite amphibolites to amphibolites
on the eastern side. Lawsonite has been detected in several cases, and lawsonite
amphibolites have been described. In the mineral composition of the series it is found
that the percentages of epidote and biotite vary sympathetically and inversely with

.x x

X x X
X x x

X X

H
o

M
03

12 AUSTRALASIAN ANTARCTIC EXPEDITION.

the percentage of hornblende. There are, however, exceptions both at Cape Denison
and Mackellar Island, but these can be explained by varying metamorphic conditions.
The chemical composition of the epidote biotite schists differs distinctly from that of
the amphibolite, though both possess the general characteristics of a basic igneous
rock. The chemical composition of the biotite amphibolites is calculated to be near
the mean of the two end members of the series.

Extraordinary examples of metamorphosed xenoliths have been found in an
amphibolite dyke at Cape Denison, and include both the cognate and accidental types.
These bodies were caught up by the invading dyke and set in position before the develop-
ment of the amphibolite characters by metamorphism. In some instances they still
preserve a very remarkable angular outline, but they are now an integral part of the
metamorphic rock, and the foliation passes through them irrespective of their outline.
The cognate xenoliths now possess the same recrystallisation products as the dyke.
The accidental xenoliths are undoubtedly foreign to the primary dyke, and bear relation
to the enclosing granodiorite gneiss.

Quite distinct from these xenolithic bodies are certain varying rock types which
are enclosed in the amphibolite dykes. These include epidosites, chlorite rock, biotite
hornblende schists, and biotite felspar schists. These rocks bear definite relation to
the development of the metamorphic characters, and can only be adequately explained
by an hypothesis depending on the metamorphism. Their mineral composition shows
that their chemical composition is quite distinct from that of the amphibolite from which
they were derived. The proved difference in chemical composition involves the migra-
tion and segregation of material under metamorphic conditions, and is called metamorphic
differentiation.

In addition to the sharply-walled dykes at Cape Denison there are isolated patches
of amphibolite completely enclosed by the granodiorite gneiss. Argument is deduced
to show that these amphibolites, differing only in grain size, have once been part of
the dyke series. This means that these " inclusions " are younger than the enclosing
gneiss, and the dyke channels have been rendered discontinuous during the meta-
morphism. In some cases the boundary between amphibolite and granodiorite gneiss
has been destroyed and replaced by a gradual transition. The destruction of a pre-
existing boundary by the migration of material across it during metamorphism is called
metamorphic diffusion. In all cases of transition there is no evidence of refusion.

The processes of metamorphic differentiation and diffusion are very limited in
range and produce the exceptional types. Nevertheless, they have been found to be
applicable to localized portions of many widespread areas of crystalline schists. In
several cases they produce results directly opposed to previous conclusions. The
assumed evidence, for example, on which is based the conclusion that amphibolites
can be produced by the extreme metamorphism of limestones in Canada, is rendered
invalid. The evidence demands the reconsideration of all conclusions in areas of
crystalline schists which are based on the evidence of transition.

THE METAMORPHIC ROCKS OF ADEL1E LAND.-8T1LLWELL. 13

In the Cape Gray dyke series the metamorphism of the primary dolerite has occurred
under very different conditions from those which operated at Cape Denison. The
mineral changes have occurred under conditions which have not destroyed the dyke
form. The dykes branch and send forth little tongues into the garnet gneisses, in the
same manner as in unmetamorphosed regions. Minerals and structures of the primary
dolerite may be recognised, but in all cases the metamorphic features dominate the
igneous character. The mineral changes can be directly traced and prove to be very
interesting. The relic pyroxene of the dolerite is usually found to be crowded with
minute, dusty inclusions of ilmenite (schiller inclusions). On recrystallisation the
pyroxene may pass into a granular aggregate of clear secondary pyroxene, including
both hypersthene and augite, while the minute ilmenite inclusions coalesce into crystal
units. If the pyroxene then passes over to hornblende, amphibolites are produced.
The pyroxene may also react with the anorthite molecule of the labradorite and produce
garnet and quartz. The formation of the latter may be preceded by the development
of a diablastic intergrowth of felspar with vermicular pyroxene which may extend as
a reaction rim around the pyroxene crystals. If part of the pyroxene is amphibolised
at the same time, a garnet amphibolite is produced.

The grain size of the recrystallised rock depends entirely upon the metamorphic
conditions. An early stage of the recrystallisation of a dolerite may be the production
of a fine-grained aggregate consisting of felspar and augite. A later stage involves the
growth of large crystals at the expense of smaller ones, and the production of a
moderately coarse grained rock. Rosiwal analyses have demonstrated the similarity in
percentage in mineral composition between a fine-grained type at Cape Gray, a coarser
type at Aurora Peak, and still coarser type at Madigan Nunatak. As the dyke origin
of the Cape Gray type is certain, sound evidence is thus brought forward concerning the
origin of the plagioclase pyroxene gneisses at Aurora Peak and Madigan Nunatak. In
this case, and in many others in this series, there are similarities between these Antarctic
rocks and the rocks called pyroxene granulites in Saxony, and norite in India. Evidence
is therefore provided for the metamorphic nature of these rocks.

The acid hypersthenic gneisses of Madigan Nunatak, Aurora Peak, Stillwell Island,
and Cape Pigeon Rocks possess affinities due to similar conditions of complete recrystal-
lisation. The primary igneous nature of the first two is determined by their chemical
composition, which closely resembles that of granites, while the igneous nature of one
example from the third occurrence is indicated by its dyke form. These rocks are found
to be very closely related to the hypersthene rocks in India, which have been described as
charnockite, and which have been looked upon as igneous rocks that have consolidated
under phenomenal conditions. The metamorphic nature of the Antarctic rocks is
maintained, and argument is found to show that the charnockites are also metamorphic
rocks which have been completely recrystallised under deep-seated metamorphic
conditions.

14 AUSTKALASIAN ANTAKCTIC EXPEDITION.

Apart from the metamorphosed dyke series the Cape Gray Promontory is noted
for its highly garnetiferous gneisses. The garnets are most remarkable at Garnet Point
and at Still well Island, where they are found l^in. to 2in. in diameter, and give the
outcrops a curious mottled appearance. These large garnets are always partly altered
to biotite and quartz, which appear in cracks and around the edges of the crystals.
Cordierite, sillimanite, and corundum are found in different specimens and indicate the
high alumina content. Very beautiful pleochroic haloes are found in some of the biotites
of these gneisses. Some show an inner nucleus and an outer corona, while some change
the colour of the biotite in their sphere of action from a pale green to a brown.

At Garnet Point and the Cape Pigeon Rocks dykes containing prominent garnet
and felspar cut the garnet gneisses. A single specimen from one of these has indicated
its relation to the hypersthenic felspar gneisses on Stillwell Island. Very extraordinary
variation has been discovered in the mineral content of these hypersthene garnet rocks,
which correspond to the intermediate members of the charnockite series. At one end
of a specimen Sin. long, garnet is present, and has been produced by the reaction between
biotite and plagioclase and quartz, while, in a section cut from the opposite end of the
specimen, garnet and all evidence of reaction are absent. In other cases biotite and
quartz result from reaction between hypersthene and orthoclase, while in one case
the garnet seems to be formed directly from the hypersthene. Garnet-forming
conditions may therefore be highly localised, and all the evidence that has been collected
is antagonistic to Fermor's conception of an infra-plutonic zone a deep zone in the
earth's crust which is supposed to be characterised by garnets. The evidence reminds
us that the metamorphic zones are not denned by certain depths in the earth's crust,
but by a set of physico-chemical conditions.

The study of the examples in Adelie Land of destroyed igneous boundaries, of the
metamorphic differentiation products, of the mineral changes, and of the development
of large crystals during recrystallisation leads us to believe that solid diffusion in rocks
is an important factor that needs to be considered in the detailed study of the develop-
ment of the crystalline schists.

CHAPTER II.
THE PHYSIOGRAPHY OF CAPE DENISON.

The rocky promontory of Cape Denison covers approximately half a square mile.
It forms a roughly triangular area with a base of three-quarters of a mile on which the
rocks rise to a height of 140ft. above sea level, when they disappear beneath the glacier
ice. On the east and the west the uniform ice cliffs of Commonwealth Bay give way
to rocky cliffs, which are first of similar height but which descend to sea level as they
continue north (Plate XIV., fig. 1). The shore line is indented, and one small bay
is 400yds. deep and 100yds. broad at the mouth, and, as it broadens towards the head,
it forms an excellent boat harbour. This boat harbour is actually an extension of the
valley depression in which the hut is situated a miniature drowned valley.

The rocks remain uncovered throughout the whole year. A little more than the
average area is exposed by the summer thaw, and the winter snow drifts do not bury
much on account of the incessant wind.

AGENTS OP DENUDATION.

The promontory may be described as a miniature mountain area. It is rugged,
and possesses steep rock faces and sharp ledges (Plate XVI., fig. 2). It is carved by
four parallel valleys, and the intervening ridges are crowned by numerous small peaks
(Plate XV., fig. 1). The sculpturing is the combined effect of different factors which
we tabulate and discuss in order, as follows :

1. Glacier Action.

2. Frost Action.

3. Water Action.

4. Wind Action aud Atmospheric Weathering.

5. Shore Ice Action.

6. Nature and Structure of the Rock Mass.

1. Glacier Action. The area reveals abundant evidence of glaciation. Glacial
erratics are promiscuously scattered everywhere. Apart from the well-defined moraines
the distribution is as abundant on the higher ridges as in the lower valleys, and the
erratics are frequently seen perched in curious positions on the highest crags. Some
of the erratics have polished faces and rounded edges, and some show glacial striae, while
others are quite subangular. Some weigh several tons, and many consist of foreign
rock types. Polished, striated, and grooved surfaces of the " in situ " rock are to be
seen. Very highly polished surfaces (Plate XVIII., fig. 1, Plate XXII., fig. 2) are quite
characteristic of the peripheral area below the 40ft. contour level. Above this belt

16 AUSTRALASIAN ANTARCTIC EXPEDITION.

the broad outlines of rounded bosses of rock are always indicative of glacial planing,
but in minute detail the surface is usually found to be minutely pitted and roughened
by the unequal erosion of the constituent minerals. Nevertheless polished and striated
surfaces can be discovered, and one block about 9ft. square with long, well-marked,
parallel striae was photographed (Plate XVIII., fig. 2). The striae trend N. 32 E.

The existence of a lower zone of relatively polished rock and a higher zone of
relatively unpolished rock is a noteworthy feature. It is believed that the roughnesses
on the surface of the rocks is due to abrasion of millions of snow grains. In this case
the upper zone must have been exposed for a longer time than the lower (below the
40ft. contour), which has been relatively protected. But the shore line is capped for
almost the whole year by an ice foot, which is stationary and protective, and wave
erosion has only opportunity for limited action. The ice foot is usually about 15ft.
high, and may provide a simple explanation of the preservation of the well-polished
surfaces on the border zone of land and water. Then if we assume a slight relative
and recent uplift we may easily explain a relatively protected zone up to 40ft.

The highest point of the Mackellar Islands is about 40ft. above sea level, and the
notes made by Sir Douglas Mawson show that the character of their surface is the same
as the peripheral area of Cape Denison. Though no highly polished surfaces or striae
were found, the general surface of the islands is flat and in part very smooth, with the
prominences well rounded. Roughnesses can be explained by very recent disintegration.

Glacial plucking is evident. When the glacier has passed over a sloping resistant
rock face the lee side is often found to be steep, because the rock has been plucked out
by the onward travel of the ice.

Lakes. Five small glacial lakes are present in this area, and theii position is shown
in the locality plan, and their manner of occurrence is illustrated in Plate XXXI., fig. 2.
Four of these are almost round in shape and 30yds. to 40yds. broad. Lake II. is not
quite so broad, and its length is more than three times its breadth (Plate XVII., fig. 1).
We mention this point because the direction of the glacial movement has not
corresponded with the long axis of the lake. Lakes III. and V. (Plate XX., fig. 1)
are situated on valley floors, while Lake IV. (Plate XX., fig. 2) is on the highest level
about 120ft. above sea level. Lake IV. is bounded on the northern side by a rock
wall thickly banked with morainic material (Plate XXL, fig. 2). The lakes average
about 20ft. in depth, and they represent depressions gouged out of the rock floor by the
forward movement of the ice.

Valleys. There are four parallel, broad-bottomed, shallow valleys. In parts
their sides are moderately steep, and their trend is N. 15 W., which is very nearly
coincident with the strike of the rocks (N. 4| W.) (Plate XV., fig. 2, Plate XVII., fig. 2).
This direction makes an angle of 45 with the direction of the trend of the glacial striae,
and the origin of these valleys is an interesting problem which will be introduced after
other sculpturing factors have been discussed.

THE METAMORPHIC ROCKS OF ADELIE LAND 8TILLWELL. 17

2. Frost Action. Frost has played a subordinate part in the sculpturing of this
area. In the summer we may get daily thawing and freezing. The thaw water finds its
way down rock joints, and, refreezing, tends to force open the joints and perhaps cause
fresh fracture. Boulders with planes of ready percolation, e.g., bedding planes, may
be seen completely shattered into flakes parallel to the bedding plane. Areas were
photographed (Plate XX., fig. 4) which consist of angular boulders twisted and jumbled
into a very confused state as the result of displacement by frost. These areas are most
marked where the drainage of thaw water is slow and impeded. In this way frost
action has helped to demolish the little mountain peaks and ridges.

3. Water Action. The prevailing winds blow from the south, and, being very cold
and constant, do not permit much thaw. Given, however, calm weather, a clear sky,
and a bright sun, the thaw is rapid. Such days are not frequent. On one occasion in
December, 1912, the thaw water, draining off the moraines, was dammed back for a
while by ice. When the barrier gave way about mid-day a small cataract rushed down
the side of the valley in which the hut was situated, and in about three hours the stream
carved a channel through 6ft. of glacier ice down to bed rock. Stones and pebbles were
rolled along the course with great energy.

Running water is uncommon, and is usually too insignificant in quantity to have
much bearing on the rock sculpturing.

Marine erosion also must be slight on the mainland, for the weather is constantly
" off-shore," and an ocean swell is rarely seen, even though there is open sea throughout
the whole year. Except for about one month during the year the shore is protected
by an ice foot.

4. Wind Action. The violence of the wind in Adelie Land is impressive by reason
of its possibilities. The constant blast of snow grains against the rock makes all surfaces
minutely rough and pitted. The effect of the abrasion on timber became very marked
on the roof of the hut and on all pieces of exposed box timber. The softer constituent
minerals in the gneiss, such as mica, are worn away much more rapidly than the harder
quartz and felspar. The etching is very prominent in specimens of the limestone
schists found on the moraines, and the harder portions are raised in strong relief.
Nevertheless, it is quite certain that the rocks are relatively unweathered and fairly
recently uncovered in comparison with the Madigan Nunatak. The wind affects the
topography by preventing soil deposition ; the surfaces of the rocks and moraines
are swept clean of all fine rock detritus that is not held down by the weight of boulders
or embedded in ice, and this detritus finds permanent lodging only in the cracks and
joint planes of the rocks. It is interesting to recall that occasionally in the winter
pebbles were heard to strike the roof of the hut just as if some one outside were throwing
stones. Grit may always be discovered when one displaces some of the larger boulders
on the moraines.

SeriM A, VoL m., Part 1 B

18 AUSTRALASIAN ANTARCTIC EXPEDITION.

5. Shore Ice. Sea ice did not remain in Commonwealth Bay for more than a couple
of days during 1912 and 1913 ; in the boat harbour, however, the ice was able to remain
and thicken. If sea ice did form at winter quarters, it would be held in by the Mackellar
Islands, and results similar to those for lake ice described by Chamberlin and Salisbury*
might be possible. When the ice has once formed, a further lowering of temperature
will cause contraction ; and the water will then rise along the contracted margins and
immediately freeze. If the water is shallow, ice may become attached to rocks and
boulders on the bottom. A subsequent rise in temperature causes the ice to expand,
and the marginal layers, with their burden of enclosed boulders, may be pushed some
distance up the shore ; final melting will then leave a bank of boulders. It is possible
that similar action has contributed to the formation of the " lower moraines " on
Cape Denison.

6. Nature of the Rock. Here, as usual, the hard resistant nature of the rock is
important. Foliation planes and cross jointing, and the junction planes of the gneiss
with the amphibolites have facilitated frost action. The amphibolites are usually
softer than the grey granodiorite gneiss, and in consequence there are frequent
depressions due to this fact. In one instance (Plate XVI., fig. 1) the reverse is true,
and the dyke rock appears as a low thin wall. The detailed structures and compositions
are discussed in a subsequent chapter.

OBIGIN OF THE VALLEYS.
The valleys above mentioned may have resulted from

(a) The travel of the main ice sheet.

(6) Valley glaciers existing after the partial recession of the ice sheet.

(d) Wind action.

(e) Pre-glacial strike valleys.

(a) The ice sheet has certainly been more extensive. Assuming that the ice sheet
does erode its under surface in its forward movement, we naturally expect valleys to
be carved approximately parallel to the direction of travel ; but still differential move-
ments are known to exist in a large ice sheet, and a small valley depression diverging
45 might result from such. In the present case, the four parallel valleys would require
four different parallel sets of differential movements. This does not seem possible
when we realise that we are dealing with a very small area only three-quarters of a
mile broad.

(6) The glacier rises very steeply to the south of the rocks, approximately 1,000ft.
in three miles and 1,500ft. in five and a half miles. The valleys are in the direction
of greatest slope, and hence the formation of these valleys by subsidiary valley glaciers
is quite possible. It is to be noted that three of the four valleys now contain ice

* " Earth Processes," p. 389.

THE METAMORPHIC ROCKS OF ADELIE LAND 8TILLWELL. 19

permanently. In one case the ice was under observation for nine months and no move-
ment capable of measurement took place. The fourth and ice-free valley has the steepest
grade, and the sides are straight and in places steep, and, in a general way, indicate
glacial movement in the direction on the valleys. The valley walls which slope with
the dip are rounded and smooth, while those which slope against the dip are torn up
and ploughed very rough. Morainic material is in one case distributed laterally along
the walls, while the glacier ice itself is partly coloured by the presence of the subglacial
material. If valley glaciers of this type exist, a set of minor ice movements are intro-
duced subsequent to, and in a different direction from, the movements which produce
the striae observed N. 32 W. We were unable to pick up any trace of a second set of
striae, while in the lower part of one valley, where it opens out to a broad area, striae
can be seen trending in the constant direction N. 32 W.

The upper limit of the subglacial material of the ice sheet (marked by a dotted
line on the locality plan) descends to lower levels at the head of valleys, suggesting
that morainic material has been pushed further on in the direction of the valleys. Hence
in spite of the absence of the second set of striae, secondary minor movements of the
valley glacier type are suggested. Such movement is negligible at the present time.
A combination of (a) and (b) may probably produce the present result.

(c) Thaw Water Action. This is a possibility that suggests itself immediately to
one that has not visited the area. As above stated, thaw water action is, on the whole,
feeble. The lakes are thawed out only for two or three weeks during the year. The
character of the valleys is markedly not that produced by water streams.

(d) Wind Action. The constant direction of the wind corresponds with the direction
of the valleys, and this fact suggests wind-scoured depressions. The unsymmetrical
and varying slopes on the valley walls are hardly consistent with the hypothesis, as the
rock is of uniform hardness. Further, the undercutting, which is so characteristic of
wind erosion, is, in the main, absent. An odd boulder shows wind erosion, but only in
one small spot in situ did we observe undercutting. The rocks are remarkably fresh
and unweathered, and the lapse of time since their uncovering cannot be great, and
cannot be sufficient for wind excavation of the valleys.

(e) Pre -glacial Strike Valleys. This theory will be accepted if all other possibilities
are rejected. At present the short shallow valleys possess nothing that suggests water
action. All such traces might disappear in the subsequent modification of the valleys
during glaciation.

MORAINES.

Not only is our small area sprinkled with erratics, but also well marked moraines
have been left by the retreating ice front in more or less parallel banks. These moraines
are entirely the product of the basal load of the glacier. No surficial or interglacial
material is present in the terminal section.

20 AUSTRALASIAN ANTARCTIC EXPEDITION.

The ice front of the glacier rises very steeply to the south, and, for a vertical thickness
of 40ft. above the rock floor, the ice is coloured brown by the presence of subglacial
rock detritus (Plate XXXI. , fig. 1). Similarly to the east and the west of Cape Denison,
whenever the ice cliffs rest on a rock basement, brown-coloured ice could be seen for
estimated thickness of 15ft. or 20ft. (Plate XIV., fig. 2). The ice sheet, therefore, has
at this point eroded the underlying surface. In this connection it is interesting to recall
the observation of a sledge party, 20 miles to the east, to the effect that rock outcrops
at this distance away are found to be nearly devoid of morainic material, though the ice
sheet is continuous and unbroken. A paucity of erratics has also been noted by Sir
Douglas Mawson at Cape Hunter, nine miles west of Cape Denison. The area of abundant
glacial debris is, therefore, limited in Commonwealth Bay.

The subglacial material (Plate XXX., fig. 1) consists of fine rock meal, grit,
pebbles, and boulders, some of the latter weighing many tons. For the greater part
of the year it is all part of a hard-frozen zone, but the summer thaw loosens a good deal
of surface material. The yearly ablation also liberates a certain amount. The finer
material is, as above stated, swept away by the winds, and no glacial soil or gravel,
except in isolated cases, can remain exposed. The exposed surface thus presents an
aggregate of boulders of all sizes and shapes.

The moraines are, for the most part, thickly banked up in lines parallel and near
to the glacial front. They contain a great variety of rocks which are not found in situ,
and are, therefore, known to exist hidden beneath the ice cap. Among these rock types
are crystalline limestones, lime silicate schists, sandstones and quartzites, granites,
dolerites, vein and pegmatitic material, and schists and gneisses in great variety (Plate
XIX., fig. 2). In general there is a great similarity between many of our specimens
and those reported from the Ross Sea area by the Scott and Shackleton expeditions.

In the field it was found convenient to recognise an " upper moraine " and a " lower
moraine." The " upper moraines " are true moraines, formed in the manner indicated
above. It is the usual type of deposit with many of the stones much worn, though
not often well rounded like river pebbles. Some look like rock chips with the salient
angles and edges worn off. Many are subangular with plane and bevelled faces, and
a small percentage (possibly % per cent.) are striated. The " lower moraines " are not
strictly moraines at all. They do not occur above the 40ft. contour level and contain
a very large percentage of local rock.

The abundant variety of rock types on the " upper moraines " makes them very
different from the "lower moraines." In the latter the boulders are more rounded
(Plate XIX., fig. 1), though still not altogether like river or marine boulders ; many
have plane, bevelled, gouged, and polished faces, but very few are striated. Further,
some degree of sorting into sizes seems to have been accomplished. Some banks are
dominated by boulders averaging one foot in diameter ; while others consist, in the

THE METAMORPHIC ROCKS OF ADELIE LAND.-STILLWELL. 21

main, of pebbles about Sin. in diameter. The pebbles are, on the whole, coarse ; but
fine gravel and sand are always found when one turns the boulders over. The " lower
moraines " are found in banks and bars which are parallel to the coast, and which follow,
more or less, all the indentations of the contour. They are well seen in the lower parts
of the valleys, and in one case three moraine bars appear in succession between the
40ft. level and sea level, producing a terraced appearance. In the panoramic view of
Cape Denison looking east (Plate XXX., fig. 2), the " lower moraine " has the general
appearance of a beach deposit.

The presence of these " lower moraines " is, no doubt, associated with the presence
of the zone of relatively polished rock in the peripheral area below the 40ft. contour
level. It has been mentioned that this zone is represented on the Mackellar Islands,
where Sir Douglas Mawson has also noticed the presence of patches of roughly-rounded
boulders of local varieties of gneiss which are similar to those on the " lower moraines "
on the mainland.

There is no doubt that some of the boulders have been glaciated as well as water
worn ; but the abundance of local rock indicates a local origin. Their relation to the
contours and to the outline of the coast is evidence of a marine origin. The grading
into sizes must have been accomplished by sea water, though shore ice may have assisted
to bank the material up into terraces. They probably represent glacial debris which
has been subjected to wave erosion and which has become largely diluted with the local
detritus produced in the ordinary course of marine erosion. They are the equivalent
of the raised beaches in normal climates, and, therefore, indicate, like the zone of polished
rock, a recent, slight relative uplift.

The beach origin of the " lower moraines " is confirmed by the finding in them of
a small piece of grit containing shells. This single specimen (No. 702) was found close
by the hut and has been examined by Mr. F. Chapman, A.L.S., Palaeontologist to the
National Museum, Melbourne. Mr. Chapman has also examined samples of the sands
obtained from both the typical moraines and the " lower moraines." His reports are
now added.

NOTE ON A' CONSOLIDATED BEACH SAND FROM CAPE DENISON.

By F. CHAPMAN, A.L.S.

Macroscopic Appearance. The rock is of a grey colour, with a roughened weathered
surface (Plate XIII., figs. 1 and 2). It measures about 5*5cm. by 4-3cm. The texture
is coarse and gritty, and closely resembles a consolidated beach sand such as may be met
with in all latitudes under favourable conditions. To my own knowledge the nearest
rock in appearance to this is a specimen I collected many years ago from the coast of
Ilfraconibe, in Devonshire. In that instance the rock was formed by the concreting
action of percolating water charged with dissolved CO S reacting on the shelly particles
of the beach. The present specimen shows strong effervescence with HC1.

22 AUSTRALASIAN ANTARCTIC EXPEDITION.

Microscopic Characters. In thin slices under a moderate power of the microscope
(2in. obj.) the rock is seen to consist mainly of angular quartz grains intermixed with a
few particles of igneous rock, more or less basic, and with occasional twinned and zoned
felspars. Brown palagonitic glass showing perlitic structure occurs in the rock, and
also numerous crystals of augite. The average diameter of the larger grains is 0-4mm.
(Plate XIII., fig. 3).

Organic Particles. Remains of foraminifera tests are seen ; and one fine cross
section of a milioline, of the Miliolina subrotunda type, is present. A fragment of an
indeterminate coral appears in the hand specimen, and its examination by means of
the micro slide leaves no room for doubt as to its relationship to that group. The
coenenchyma is cavernous and traversed by strong pillars, and there is abundant evidence
of the " dark-line " structure of a recent coral. It appears to be an arborescent form
(Plate XIII., fig. 4).

An echinoderm plate is seen in section, distinguished by its typical perforate
character and calcitic structure.

Matrix. This consists of a fine detrital dust and minutely crystalline calcareous
cement. The larger fragments are generally evenly spaced out in the mass.

Conclusions. The present specimen is a consolidated beach sand, consisting largely
of cleanly broken particles of terrigenous material of the nature of basic and sub-acid
igneous rocks, evidently derived from the immediate locality ; these, together with
littoral and coral zone organisms, are cemented in a matrix composed of a mixture of
fine igneous rock detritus and calcareous deposited material.

NOTE ON MORAINIC MATERIAL, CAPE DENISON.
By P. CHAPMAN, A.L.S.

Morainic Mud. From Lower Moraines found underneath boulders. About 35ft.
above sea level.

This consists of angular fragments and sand grains derived from hornblende gneiss
and schists, with some material, probably from a granitic source. The fine washings
contain numerous crystals of micaceous, pyroxeuic, hornblendic, and felspathic minerals.
No organic remains were noticed.

Morainic Detrilw. Sample from Upper Moraines, taken from the ice, about 120ft.
above sea level, Commonwealth Bay.

Coarse and fine detritus of hornblende gneiss and granitic rocks containing pink
and white felspars. Fine washings include numerous small crystals of ferro-magnesian
silicates and felspars. No organic remains were seen.

CHAPTER III.
THE METAMORPHOSED DYKE SERIES OF CAPE DENISON.

1 . NOMENCLATURE.

The nomenclature generally adopted in the following pages is that taught in the
" Die Kristallinen Schiefer."* Grubenmann's treatise deals with schists and gneisses
which have developed by the thermo-dynamical alteration of pre-existing rocks. He
has told us what rocks belong to the crystalline schists, and in these the effects produced
by the alterations completely determine the character of the rock.

Various meanings have been attached to the term " metamorphic rock," and we
are in general agreement with the criticism of this term given by Crookf. Crook points
out that Van Hise's use of the term is too broad, as it is made to include all rocks,
and that there is no satisfactory definition of the term ; either it is made too incomplete
or too comprehensive in its meaning, and he desires to exclude a group of metamorphic
rocks from the fundamental rock classification into igneous, sedimentary, and
metamorphic. The old group of metamorphic rocks is replaced by several groups,
including a group of thermo-dynamically altered rocks which are unfused and unmodified
by exudations. This group seems to correspond with the Kristallinen Schiefer. If
we adopt this subdivision, are we to exclude the term " metamorphic " from our nomen-
clature ? We are inclined to think that the term is too deep rooted and too convenient
to permit this. It is therefore necessary to state that in the following account of the
metamorphic rocks of Adelie Land we use the term in the limited sense indicated in
Grubenmann's work and adopted by Johnston and NiggliJ. We are dealing with
rocks which come under the heading of the crystalline schists, and when we refer to a
rock as metamorphic we imply that it falls into the group of the crystalline schists.

Its use in this way comes partly from the want of a general rock term for those
members of the crystalline schists which possess a massive texture in contrast to the
schistose or gneissic texture. The terms schist and gneiss are not fundamentally
different, and the same processes produce the foliation of the gneissic granites and the
foliation of the sedimentary gneisses. A gneiss can be looked upon as a schist with an
imperfect schistose structure ; and should this structure become too insignificant to
be noticeable, what shall we call the rock ? Sederholm has considered that it is
impossible to give the term " gneiss " a limited meaning, and he believes that it must
remain a comprehensive name with a very wide significance. We find that this wide

Die KmUUinen Schiefer." U. Orubenmann, Berlin, vol. I., 1904, voL IL, 1907.

f " The Genetic Classification of Rocks and Ore Deposits," T. Crook, Hin. Mag., vol. XVII., p. 69.

J " Principles Underlying Metamorphic Processes," Johnston ft Niggli, Journ. Oeol., vol. XXI., p. 481.

| " Om granit och gnei," J. J. Sederholm, Eng. Summary, BulL Com. Geol. Fin., No. 23, 1907, p. 109.

24 AUSTRALASIAN ANTARCTIC EXPEDITION.

significance may have real use, and we have used the term " gneiss " as a general rock
name for members of the crystalline schists which have no special name and which
may or may not possess a schistose structure ; for example, a certain rock has been
styled pyroxene granulite by the Germans, pyroxene gneiss by the French, and a norite
by the Indian Geological Survey. We do not use the term " pyroxene granulite "
because the granulitic structure is a very common structure in almost all classes of the
crystalline schists. We do not use the term " norite " because it is applied to a special
variety of igneous rocks. We apply the term gneiss in each case, and distinguish as a
plagioclase-pyroxene-gneiss, or a pyroxene-alkali-felspar-gneiss, in spite of the fact that
no schistosity may be evident.

As we make large use of the term " amphibolite " we quote the following from
L. Henzer * : " By older writers a pure amphibole rock is occasionally called
amphibolite, while the garnet-amphibolites, felspar-amphibolites, or zoisite-amphibolites
were called amphibole schists, greenstone, hornfels, etc. According to Zirkel, the
typical amphibolite consists only of hornblende. Chiefly through Rosenbusch the
term has become firmly established in the literature in recent years. Rock types with
amphibole as their main constituents (hornblende-schists or hornblende-fels and
actinolite schists) are distinguished from the actual amphibolite whose mineral content
is essentially hornblende and plagioclase, though the latter can be replaced partly or
wholly by zoisite, epidote, garnet, or scapolite." This is the usage which has been
adopted by Grubenmann and which gives the term amphibolite a precise chemical and
mineralogical meaning.

The need of the general term, which we supply for ourselves in the term " gneiss,"
is evidenced in a paper published by Loewinson-Lessingf in 1905, which discusses the
classification and nomenclature of the amphibole rocks belonging to the crystalline
schists. In the study of the crystalline schists of the River Tagil, in the Middle Urals,
Loewinson-Lessing was struck with the close connection between the schistose and
massive members of the crystalline schists. He found massive schlieren in the midst
of schistose rocks and argued that the term " schist " was scarcely fitting for the non-
schistose rocks. But he meets his difficulty by introducing terms which emphasise the
likeness of the massive types to igneous rocks e.g., paradiorite, amphibole, para-
gabbro, etc. These terms obscure the observed close connection of these rocks with
the crystalline schists. They unduly accentuate differences which have no genetic
bearing and place them equivalent with characters which do have genetic meaning.
But yet they are intended to limit and add precision to the term " amphibolite." An
amphibolite, according to Loewinson-Lessing, is a rock whose essential constituent is
amphibole alone. This violates the conclusion expressed in the preceding quotation
from a paper to which Loewinson-Lessing makes no reference. A rock composed
of amphibole and plagioclase is called " paradiorite " or " amphibole para-gabbro " if

* " Bin Beitrag zur Kenntnis der Eklogite und Amphibolite," Laura Hezner, Wien, 1903, p. 5.

t " Ueber Klassifikation und Nomenklatur der zur Formation der Kristallinischen Schiefer gehorigen Amphibolgesteine,"
von F. Loewinson-Lessing, Centralblatt Mineralogie Geologie, 1905, p. 407.

THE MBTAMORPHIC ROCKS OF ADELIB LAND. 8TILLWELL. 25

it has a massive structure, but if schistose, it is called " diorite gneiss " or " amphibole
gabbro schist." If we accepted this nomenclature we should have to give the names
" paradiorite " and " diorite gneiss " to two rocks of similar origin but with slight
variation in structure. If quartz should be introduced into this rock by some meta-
morphic process the rock would become a " para-granodiorite " or a " granodiorite
gneiss " ; but we will show that such a rock has nothing whatever to do with
granodiorite or its gneissic modification. It is obvious, therefore, that this nomenclature
can have little value in any genetic study.

In the nomenclature that we have adopted we use the term " amphibolite schist "
when the amphibolite has a marked schistose structure, the term " biotite amphibolite "
when the hornblende is partly replaced by biotite, and the term " quartz amphibolite "
when the amphibolite has been modified by the addition of quartz.

In our study rocks have been found to contain structures similar in appearance
to the micrographic and micropegmatitic intergrowths that are formed by the
crystallisation of a eutectic mixture in igneous rocks. In numerous instances this
" micrographic " intergrowth has a direct metamorphic origin. Consequently, we
have not used the term " micrographic " or " micropegmatitic," which may be
conveniently retained for igneous structures. We have described the structure a
diablastic structure.

2. FIELD CHARACTERS.

The Cape Denison granodiorite gneiss is laminated by a rock type which appears
as a series of parallel black bands of amphibolites and epidote biotite schists. It is a
foliated rock type and its foliation is parallel to the trend of the bands as well as to the
foliation of the gneiss. The cleavage planes are usually vertical. The bands are usually
between 18in. and 2ft. wide, and being jet black in colour they present strong contrast
to the grey gneiss on the bare rock floor. When the gneiss assumes a dark colour relations
are less obvious. The junctions between the bands and the grey gneiss are typically
sharp, well defined, and straight, and in general appearance they suggest the field
relations of a system of parallel dykes which have intruded the gneiss.

One band may be continuous across the area, but more often it is broken. In the
latter case the band wedges out and disappears for a while and then reappears some
distance further along the strike. Sometimes a band opens out into a bulge and then
occupies a width of 30ft. or more. A case was noted where a band which is continuous
for some distance suddenly breaks off and does not reappear along the line of strike,
while a band commences in a similarly sudden manner 40yds. due west. A white quartz
vein starts from one broken end and runs toward the other, but it peters out before
reaching it. One immediately suggests that this band has been faulted, and the fault
fracture filled by quartz. But if all the bands were contemporaneously formed, such
faulting is highly localised, as the next adjacent band on the east is continuous.

AUSTEALASIAN ANTARCTIC EXPEDITION.

The bands are not uniformly spaced across the area. They may be 50yds. or 100yds.
apart, or they may be separated by only a foot. Three bands close together can be seen
on the valley floor in Plate XXII., fig. 1. In one case two bands appear to unite, but
the surface rock is disturbed and the junction so fractured by frost that the observation
is a little indefinite. A single band, however, may run out into a series of parallel threads
which enclose layers of gneiss between them, and these threads may unite. Detached
fragments of the black rock are sometimes observed to be completely enclosed in the
gneiss. Such fragments are either in close proximity to a band or else replace the band
in a discontinuous outcrop (fig. 2). These schlieren always possess a lenticular outline.

'qratnoalioril-e

arnphlb-oltf'e

Fig. 2.

DIAGRAMMATIC SKETCH SHOWING A LENTICULAR
"INCLUSION" OF AMPHIBOLITE LYING CLOSE TO
AND PARALLEL TO THE MAIN DYKE CHANNEL.

In one example that was noted in both plan and section (fig. 3) the foliation of the gneiss
is bent around it, i.e., the foliation does not continue through both black and grey rocks
indiscriminately, though the black rock is schistose itself. This example was about
6ft. long and about l^ft. in diameter. These schlieren may be penetrated by quartz
felspar veins.

The black rock is generally softer than the enclosing gneiss and yields more readily
to the mechanical weathering. Consequently the bands frequently travel along slight
depressions. In Plate XXII., fig. 1, where we have several bands close together, we find
them on a valley floor. We often find that a band forms a gully-way. The reverse,
however, is true in the case of a band which passes the memorial cross on Azimuth Hill.
Here the band has been more resistant than the granodiorite gneiss and stands as a thin
wall 3ft. above the level of the gneiss (Plate XVI., fig. 1). The latter is a fine-grained
type and contains more mica than usual.

THE METAMORPHIC ROCKS OF ADEL1E LAND.-STILLWELL.

Green epidote and less commonly purple fluorite are often developed along the
junction planes between the gneiss and the amphibolites. These minerals are also
found along joint planes in both granodiorite gneiss and amphibolite. Small quartz
segregation veins may be found in the bands, and these may carry excellent crystals
of epidote 2jin. and Sin. in length.

Fig. 3.

DIAGRAMMATIC SKETCH SHOWING A LENTICULAK

" INCLUSION " OF AMPHIBOLTTE WHICH is BEEN

PARTLY IN PLAN AND PARTLY IN SECTION.

In one instance (No. 629) where a band has opened into a bulge a number of
inclusions of white, grey, and pink colour are found enclosed in the amphibolite. These
inclusions appear across the whole outcrop, but are present in greatest numbers along
the western side of the band. They are considered to be xenoliths and will receive
full description later.

3. PETROGRAPHICAL CHARACTERS.

The rock type presents great variation in texture, structure, and mineral content.
The colour is usually jet black, more rarely grey black. The texture, in some examples,
is highly schistose, and the dominating flaky constituent is either biotite or hornblende.
In fine-grained types the texture becomes slaty schistose. In some examples the texture
becomes approximately massive. Indeed, the texture may vary in a short distance
from massive to schistose in one and the same band. The predominating structure

28 AUSTKALASIAN ANTAKCTIC EXPEDITION.

is granoblastic, which, in the massive types, might be called gabbro structure. Some-
times there is a tendency to a nematoblastic (fibrous) structure and poikiloblastic and
relic structures are also present.

Mineral Composition.

Hornblende, biotite, and felspar are the commonest mineral constituents that can be
recognised in the hand specimen. Sphene, ilmenite, pyrite, epidote, and lawsonite
are occasionally visible to the naked eye. Quartz, chlorite, calcite, apatite, fluorite,
allanite, and rutile have been determined microscopically. In general, the mineral
composition appears uniform along any one band, but this has not been completely
tested microscopically. The variation, however, in the mineral content of the different
bands is so marked and interesting that it has been considered advisable to give
quantitative expression to it. One cannot apply the Rosiwal method and determine
the relative mineral volumes in a schistose rock with the same ease that is possible in
igneous rocks. We can fairly assume in igneous rocks a similarity along three directions
at right-angles, but such assumption is far from correct in schists. Lamellar constituents,
like biotites, have relatively large surface area in planes parallel to the schistosity. Hence
we have to make the Rosiwal measurements in three sections cut in three planes at right-
angles. This involves the preparation of three thin rock sections of each specimen
and three times the labor of measurement that is necessary for an igneous rock. In
order to obtain some definite idea of the variation in different planes, two sections of
specimen No. 153 were prepared, one at right-angles to the plane of schistosity and the
other parallel to it. Rosiwal counts were then made on each slide and the figures are
given in Table 1. Biotite and hornblende increase their volume by one-third, the epidote
increases in smaller proportion, while felspar decreases its volume by one-third. Not-
withstanding these differences, it was considered that a relative quantitative expression
of the mineral volumes would be obtained from a single section provided that each
section of the rocks to be compared is cut in the same direction. The sections studied
in each case are cut at right-angles to the plane of schistosity and typical examples
selected for treatment. The figures obtained are sufficiently accurate to be serviceable
in expressing the relative variation of the constituents in this suite of rocks of common
origin and of more or less common history.

Later study of the crystalline schists from Stillwell Island and Cape Pigeon Rocks
shows that the estimation of the mineral content from a single slide may be absolutely
misleading in regard to the nature of the rock. Small specimens of gneiss apparently
uniform in the hand specimen may contain considerable garnet in one part and none
in another part. When, however, the rocks are of comparatively uniform mineral
composition, as in the case of this dyke series, the results are useful.

The figures are given in Table 1, where each example represents a different band,
and these are arranged in the order of outcrops met in traversing the area from west
to east. No. 1 is the most westerly. The heading " felspar " includes all the colourless
constituents the clear albite, the cloudy felspar, quartz, and apatite, lawsonite, and

THE METAMORPHIC ROCKS OF ADELIE LAND. STILLWELL.

calcite when the last three are not separately determined. Quartz is never abundant and
nearly always very subordinate ; as it is very difficult to distinguish it in all cases from
the clear secondary felspar, very little is lost in value and a great saving in time is effected
by reckoning it in with the felspar. Apatite, lawsonite, and calcite are separately
measured when sufficiently abundant, otherwise their presence is indicated by p. The
heading " iron ore " includes magnetite (or ilmenite) and pyrite as pyrite is very sporadic
and occasional. " Mica " includes biotite and chlorite, which are frequently intergrown.

TABLE 1.

10.

11.

12.

13.

Field No.

153

412

720

630

629

631

637

634

634A

635

Felspar . . .

46-1

31-2

32-3

25-8

12-4

54-5

27-3

28-5

35-3

31-8

13-9

27-4

23-1

Mica

35-9

48-8

7-8

33-0

16-8

15-9

1-2

7-5

4-2

10-8

5-6

8-0

10-7

Hornblende

2-5

4-0

47-6

36-1

60-3

19-8

69-7

61-8

49-8

56-3

77-3

56-3

61-8

Epidote . .

8-2

9-5

6-8

3-0

1-9

5-0

1-8

Sphene . . .
Iron Ore . .

3-2
3-0

2-2
3-4

3-5

P-

1-5
3

3-2
6

9
4

1-4
4

4-1
6-0

1-5

\2-6

Apatite . .

1-4

P-

1-0

P-

{p.

P-

Lawsonite.

P-

1-5

6-7

P-

1-5

3-0

Calcite ...

P-

1. A fine-grained micaceous type which has proved more resistant to weathering than the surrounding
granodiorite gneiss.

2. The same rock as 1, but measured in a section parallel to the schistosity.

3. An example that appears as a system of irregular schlieren rather than as a distinct band.

4. A normal band which contained irregular quartz segregation veins carrying large epidote crystals.
The quartz in-filling is subsequent to the band and the foliation.

5. A well-defined band from which No. 720 is the single specimen in the collection.

6. An abnormal band containing one of the remarkable biotitic patches. The sample was collected
as the " normal " part of the band, a few feet away from biotite clot.

7. A typical massive type taken from the band which contains the xenolith.

8. Normal band.

9. A sample of a schliere of dark rock in the granodiorite gneiss. The schliere is impregnated with
felspar-quartz veins which have participated in the folding and are contorted.

10. Normal band.

11. A band situated between No. 10 and No. 12. These three bands lie very close together and are
not separated by more than a yard from each other.

12. Normal band.

13. Massive amphibolite which is here inserted for comparison. It grades out in parts into felspathic
gneiss and in part into pure hornblendic rocks.

The table indicates that we have three main varieties in this suite of rocks

1. Rocks containing dominant biotite (e.g., No. 153, which will be described

as an epidote biotite schist).

2. Rocks containing biotite and hornblende in approximately equal proportions

(e.g., Nos. 412, 630, the biotite amphibolites).

3. Rocks containing dominant hornblende (e.g.. Nos. 629, 631, etc., the

amphibolites).

30 AUSTRALASIAN ANTARCTIC EXPEDITION.

We can notice in general that the biotite-rich varieties belong to the western side
of Cape Denison and that the hornblende-rich varieties belong to the eastern side. These
two groups are separated geographically, with the exception of No. 5, by bands in which
biotite and hornblende are both important. The exception, No. 5, loses its importance
when we reflect that it is not an example of a typical band. This uniform mineral
variation will be considered later to be a reflection of the variation in chemical
composition. The geographical nature of these variations is important, and can have
but two possible explanations. The explanation may be a metamorphic one, and
we may consider the reason to be a uniformly varying set of metamorphic conditions
across Cape Denison combined with a uniform degree of migration of material
corresponding with differences in chemical composition. The detailed examination
indicates that we really get considerable variation in the conditions in localised patches,
and thus scarcely supports this explanation. The alternative consists in regarding the
whole as a differentiated series of primary basic dykes. In either case it does signify
that we are not dealing with bands that have been repeated by folding.

With a high percentage of mica there is always found a high percentage of epidote.
The only important exception to this rule is again No. 5. No. 720 is a partial exception,
but a relatively low percentage of epidote is compensated by a high percentage of hydrous
lime silicate, lawsonite. Nos. 720 and 634A possess a percentage of ferromagnesian
minerals very considerably above the average while the felspar percentage is
correspondingly low. It is possible that they are chance specimens which are not
strictly normal of the bands they represent. Unfortunately, there is no systematic
set of specimens collected for the purpose of studying the longitudinal variation in
any particular band. Only in exceptional cases is the variation noticeable in the hand
specimen. In cases where more than one specimen has been examined from one band
some degree of variation is always noticeable.

The table renders it evident that the majority of the examples are rich in sphene,
but the accessories show considerable variation. Apatite is relatively high in some
cases and practically absent in others. The iron ores show much greater variation, and
vary from 6 per cent, to instances where it is almost negligible. The complete discussion
of this table of mineral constituents must await the presentation of the microscopical
and chemical characters.

Microscopical Characters.

No. 153. This rock is rather fine grained and schistose, but it is not so particularly
fissile as some of the hornblende varieties. The glint of the mica is obvious on the
cleavage faces. In thin section the constituents conform to a linear arrangement
and produce the crystallisation schistosity (Plate I., fig. 3). The biotite is strongly
pleochroic in sections showing cleavage from a light straw to a deep brown colour ;
it is reddish brown in basal sections. Pleochroic haloes are sometimes found around
inclusions which appear to be sphene. Hornblende is sometimes included in the biotite

THE METAMORPHIC ROCKS OF ADELIE LAND. STILLWELL. 31

and sometimes the biotite in the hornblende. The felspar appears in rounded and
indented grains. A few scattered grains only are clouded. It is generally perfectly
clear and transparent, and there seem to be two types of plagioclase, and possibly also
orthoclase. The bulk of the felspar is untwinned. A few grains with lamellar twinning
give values between 20 and 27 for the extinction angle. This is best interpreted
as andesine. From ground-up powder of the rock refractive index was found to be less
than 1-542 in some cases, but mostly between 1-542 and 1-551. In the section the
refractive index was rarely found below that of Canada balsam. Hence the bulk is
probably andesine with a few grains of either orthoclase or albite. Quartz is present
as a very minor constituent and has an appearance very similar to the clear felspar.
It can only be distinguished with certainty from the felspar by the use of convergent
light, and considerable search is required to find a uniaxial figure. The hornblende is
definite and strongly pleochroic. Its colour scheme is X, greenish yellow ; Y, bright
green ; Z, bluish green. It has an extinction up to 15 in prismatic sections. Sphene,
exceeding 3 per cent., is conspicuous in wedge-shaped crystals and in grains. It conforms
to the schistosity, and frequently encloses an idioblastic crystal of magnetite. It is
also occasionally associated with rutile. Epidote (8-9 per cent.) is more abundant
in this rock than in any other member of the series except when found in segregations.
It appears in pleochroic lemon-yellow crystals, commonly idioblastic, and shows the
brilliant polarisation colours of the third order. It does not possess the same dark
border of colour as sphene, and the polarisation colour of sphene usually reaches the
high order whites. In rare cases it is found intergrown with zoisite (or clinozoisite).
Colourless apatite is a notable accessory, and its crystals are at times comparable in
size to the sphene grains. Magnetite is present in the same proportion as sphene and
is sometimes idioblastic, apart from the sphene enclosures. Cubes of pyrite are very
sporadic. Fluorite is not infrequently found in irregular isotropic blue grains, but
it may be intergrown with the biotite after the manner of chlorite. No. 153 may be
called an epidote biotite schist.

No. 5. This specimen has been referred to as abnormal. It possesses a slightly
coarser grain size than the average. The crystallisation schistosity is less marked
and there is a tendency to a massive texture. The structure at the same time becomes
more granoblastic. Apatite, epidote, and sphene especially participate in the increased
grain size and now appear in grains comparable in size with the felspar and hornblende
crystals. The percentage of epidote is noticeably high in comparison with the mica
content, and we can recall a field note stating that the precise locality of No. 6
is especially rich in epidote. Epidote and, to a less degree, fluorite are especially
abundant in seams and joint planes at this point. A further characteristic of the
sample is the replacement of biotite by the green pleochroic chlorite which frequently
gives the ultra blue polarisations colour. The chlorite plates are usually associated
with and penetrated poikiloblastically by well-formed epidote, clear felspar (which
looks like quartz), and stray grains of magnetite. The green hornblende is abundant
and idioblastic in sections, showing two cleavages. The terminal faces of prismatic
sections are not developed. At times the hornblende is streaked and fringed by much

32 AUSTRALASIAN ANTARCTIC EXPEDITION.

paler hornblende. Such hornblende becomes very distinct between crossed nicols
in virtue of its bright interference colours. Chlorite is associated in some parts with
the hornblende in a way suggesting that the chlorite gradually passes into hornblende.
In other parts there is a transition between epidote and hornblende. The greater part
of the felspar is cloudy and saussuritised. Arising out of the saussuritised part the
secondary plagioclase can be seen. Epidote may be found in the saussuritised mass
as well as magnetite, scales of hematite, and micaceous products. In the same connection
radial aggregates of a fibrous cloudy mineral are found with very low refraction, double
refraction about the same as quartz and oblique extinction with a small extinction
angle. This is probably stilbite, one of the zeolite group, which is much better developed
in the next slide. Sphene possesses rather a deeper clove-brown colour than usual,
and is almost free from the magnetite core. No. 5 may be called an amphibolite.

When the joint planes, which are lined with epidote, are exposed by weathering,
they form a rock wall brilliantly green in colour. Some of the hand specimens are faced
with the green epidote about 1mm. thick. In these specimens the rock can be seen
to become very distinctly richer in epidote as the face of the joint plane is approached.
There is a segregation towards a plane in this case, whereas in a subsequent example,
the epidosite in the band No. 629, the segregation is apparently towards a centre.

Some of the joint planes are characterised not so much by the green epidote or by
fluorite as by a radiate, fibrous mineral of pinkish-white colour. A thin section was
cut as close as possible to the face of one of these joint planes in order to include the
fibrous mineral, and has been found to be very interesting. Not only are the relation
of epidote, chlorite, and hornblende clearer than in the preceding slide, but stilbite,
fluorite, and lawsonite are present. The radiating fibrous aggregates are prominent in
the section and the mineral is mostly clear and colourless, though cloudy in part. Its
refractive index is less than Canada balsam, and its polarisation colours are a little
higher than those of felspar. It has an extinction angle up to 10, and, as the prism
axis is the direction of the fastest ray, it can be determined as stilbite. Grains of
nuorite are not infrequent. Sometimes they show a trace of blue colour, and they
may be crowded with minute needles of a green mineral with oblique extinction,
probably pale hornblende. The development of chlorite and epidote from hornblende
is more noticeable. The green hornblende first passes into colourless hornblende, which
may be replaced by an aggregate of epidote and chlorite. Grains of fluorite may be
found in these aggregates. The hornblende also undergoes a change through colourless
hornblende, with oblique extinction, into the brightly polarising lawsonite, with its
straight extinction. Aggregates of lawsonite may fringe the radial groups of stilbite.
Some of the lawsonite aggregates resemble scapolite in appearance, but the refractive
index is too high, and wherever an interference figure is obtained it is biaxial in character.
There is here, perhaps, more clear felspar than in the normal slide. The refractive
index of this clear felspar is sometimes above and sometimes below Canada balsam, but
it is always very close to it. The clear felspar, therefore, belongs to the albite end of
the lime soda series.

THE METAMORPHIC ROCK8 OF ADBLIE LAND. 8TILLWELL. 33

No. 412. In this case the hornblende and biotite exist in practically equal
proportions. There is a well-developed schistosity, and the mica flakes give a bright
sheen to the cleavage surface. The grain size is about normal and a little larger than
that of No. 153. In thin section (Plate I., fig. 2) the hornblende and biotite possess the
usual characteristics. The greater proportion of the felspar is quite clear and transparent,
the lesser portion is saussuritised as in No. 5. Calcite is present in coarse granular
crystals, and has probably developed in the alteration of the felspar. Epidote is very
distinct in pleochroic crystals and may be intergrown with the biotite or with the
hornblende. In both cases it may exert its crystal form against the hornblende and
the biotite. A small percentage of lawsonite is recorded in the rock, and it is usually
interlaminated between the cleavage veins of the biotite. It might be mistaken for
colourless epidote, but a difference is obvious when the two minerals are brought into
the same field of view. Lawsonite attains its best development in Nos. 720 and 635,
and its characters will be fully defined from those sections. Occasionally purple
fluorite is threaded with the biotite in the same manner as in No. 153. Sphene is
practically absent, and there are only very few particles of iron ore and crystals of
apatite. No. 412 may be called a biotite amphibolite schist.

No. 720. The example is a lustrous schistose rock in which hornblende, mica, and
some colourless felspar or lawsonite are visible to the naked eye. The mica has a golden-
brown colour.

In thin section (Plate I., fig. 6) the hornblende has its normal appearance, but the
mica content consists of intergrown biotite and chlorite in which the latter is dominant.
The biotite tends towards a biscuit-brown colour that is best developed in No. 635,
and it is pleochroic from this brown colour almost to colourless. The chlorite is also
pleochroic through shades of pale green to colourless. Hence, in some positions of
the polariser, the intergrowth of biotite and chlorite may look homogeneous. The
composite character is readily observed by rotating the stage or by crossing the nicols.
The chlorite possesses bright ultra blue polarisation colours, and the biotite shows
brilliant second and third order colours. The " felspar " percentage consists entirely
of cloudy brightly polarising aggregates. Clear felspar is absent. The association of
saussuritised felspar and lawsonite seems to be very characteristic. The lawsonite is
abundant and may be intergrown with the mica, less frequently with hornblende, or
it may appear in laths with parallel arrangement contributing to the schistosity of the
rock. When best developed the laths are clear and colourless with prominent parallel
cleavage. Sometimes it is a little cloudy and some lawsonite areas enclose pieces of
saussurite, thereby indicating the connection. The lobate outline is conspicuous when
intergrown with biotite, but the laths mostly possess straight sides when intergrown
with hornblende. A slight lobateness is sometimes detected against the hornblende,
but is never prominent. A section has been noted where the end portions of a lawsonite
crystal are bounded by hornblende crystals, while the middle third abuts against chlorite.
The junction against the hornblende is linear at both ends, but the middle portion is

Seriei A, VoL m.. Part 1 C

34 AUSTRALASIAN ANTARCTIC BXPEDITION.

curved against the chlorite. Granular epidote is present and can be distinguished from
the lawsonite by its colour and pleochroism. Sphene is a moderately abundant accessory.
The rock is the best example obtained in situ of a lawsonite amphibolite.

No. 630. This sample comes from one of the thicker outcrops. The outcrop is
remarkable in containing a patch of rock which appears to be mostly biotite in the hand
specimen. The difference noticed underfoot in passing from the ordinary rock to the
biotite is similar to the difference between a pavement and a carpet. The hand
specimen is, perhaps, not so dark coloured as the other examples. The measured
section is cut at right angles to the plane of schistosity but not at right angles to
the direction of stretching. This, however, has not affected the general character of the
individual.

The mineral composition shows an extraordinary percentage of the colourless
components, felspar and quartz, but examination shows that there is much more quartz
than usual. The quartz is occasionally in large grains, clear, and often without wavy
extinction, and so grouped as to suggest segregation or absorption of secondary silica
into the rock. This quartz gives the rock an abnormal silica percentage. The clear
felspar is about equal in amount to the saussuritised felspar, but their distribution is
quite irregular. Clear felspar is dominant in some parts of the slide and saussuritised
felspar in other parts. Some comparatively clear felspar with broad lamellae can be
found. Extinction angles up to 37 can be measured, and labradorite is therefore
present. Sometimes the calcic plagioclase is partly saussuritised, and the saussuritisa-
tion is more intense along one alternate set of lamellae. Such examples suggest that
it is a relic felspar. Apart from this calcic plagioclase is another plagioclase, always
in clear rounded interlocking grain, which is interpreted as an oligoclase or andesine.
The biotite is curious. Some biotites are pleochroic in cross sections from a light-straw
colour to a dark brown, while other biotite crystals are pleochroic from a pale-greenish
yellow to a deep-emerald green. That there is no great difference between the brown
and green varieties is shown by the way the brown and green may be laminated in one
and the same crystal. The polarisation colours of the green part are too high for green
chlorite. In some parts, when in association with the cloudy felspar, the green mica
is a little paler and more like chlorite. Here lenticles showing the ultra blue polarisation
colour may be found. Epidote is associated with the green biotite in the same way
as it is associated with chlorite in No. 5. It seems that we have here the change from
chlorite to biotite during the process of recrystallisation. The chlorite passes over
into biotite with the absorption of alkali first by deepening its green colour and later
by changing colour to brown. The hornblende is normal. Epidote, sphene, and apatite
are relatively abundant. Occasionally the grains of epidote are found with a nucleus
of calcite and magnetite. The calcite has been formed from the decomposition of a
primary mineral, and at this centre of recrystallisation there has been more calcium
and iron available than actually necessary for the production of epidote. Epidote
sometimes forms a pale border to a red-brown mineral, pleochroic with high refractive

THE METAMORPHIC ROCKS OF ADELIE LAND. 8TILLWELL. 35

index and high double refraction. By comparison with other occurrences from this
area we record this mineral as allanite. The rock may be called a quartz biotite
amphibolite.

No. 629. This specimen is a normal example of the hornblende rich varieties,
and it is especially interesting as it comes from the band which contains the xenoliths.
It contains the second highest percentage of hornblende, almost 70 per cent., in the series.
It is one of the massive types, and white felspar and dark hornblende are the only con-
stituents which are macroscopically distinguishable. As is usual with the massive
varieties the structure becomes granoblastic (Plate I., fig. 1), and the average absolute
grain size is approximately -22mm. The hornblende is green with the bluish-green tint
parallel to Z. As in No. 5, the hornblende is sometimes bordered and streaked with
paler coloured hornblende. Prism faces are well developed, but prismatic sections
always have a broken and ragged appearance. Sometimes the hornblende forms a
skeletal framework for felspar and quartz, and then a sieve structure (siebstructur)
appears. Like the hornblende, the felspar is in granular individuals. The majority
of it is clear and transparent, and there are two types present. Some clear calcic felspar
is found with an extinction of 35 measured from the trace of the broad lamellae and is
labradorite. The lamellae of the felspar are often confused and intermittent and dis-
crimination is difficult. A more sodic felspar is present which has a refractive index
less than 1-551, and sometimes greater than and sometimes less than 1-538. This felspar
is, therefore, becoming albitic. A quantity of the saussuritised felspar is present, and the
secondary clear felspar can be seen arising from the turbid mass. Green chlorite appears
in occasional skeletal plates showing the ultra-blue colour. It shows a transition to
biotite. Sphene, with its magnetite core, is a common accessory. Epidote and pyrite
are sparsely scattered through the rock. Lawsonite appears in microscopic veins in
some sections. This section is made from a specimen taken from the centre of the
outcrop. Sections made from specimens containing the xenoliths at the side of the
exposure show more lawsonite. These lawsonite veins are sometimes composite (Plate
I., fig. 5). There may be an epidote lining on the walls with lawsonite in the centre,
or there may be lawsonite on the walls with calcite in the middle. The abnormally high
percentage of hornblende and the low mica percentage are possibly to be associated with
the patches of epidosite and chlorite which were noted here. The rock is an amphibolite.

No. 631. There is another example of the rocks with dominant hornblende.
Though the mica percentage is small the crystallisation schistosity is well marked.
The granoblastic structure is evident. The felspar is nearly all clear and transparent
and the proportion of saussurite is small. Occasional grains of quartz may be found.
The hornblende conforms to the crystallisation schistosity and some crystals enclose
felspar and blebs of quartz in a typically sieve-like manner. Brown biotite and sphene
are normal. Magnetite is scarce but in large grains bordered with sphene. Some
include prismatic sections of red-brown rutile. Minute crystals of apatite are enclosed
in all other minerals and lawsonite is rarely intergrown with biotite. The rock is an
amphibolite.

36 AUSTBALASIAN ANTAKCT1C EXPEDITION.

No. 637. This example possesses weak schistosity and approaches a massive
texture. Strings of white material can be seen in the hand specimen conforming to
the schistosity. The measured section is not quite normal to the schistosity, and this
is a factor in producing a slightly higher felspar percentage than is usual. Further,
the example is obtained from a schliere impregnated with felspar-quartz veins, and it
is likely that some of the visible white material is the same as in the veins. This would
also raise the felspar percentage above the normal, and this appears confirmed by the
presence of large grains of quartz in the section. The rock is granoblastic and very
little saussuritised felspar is present it is nearly all secondary clear felspar. The
hornblende percentage decreases with the increased felspar percentage, but still the
similarity of this rock to other examples is very obvious. The hornblende colour is
normal, except in an isolated grain which gives bright blue pleochroism, indicating a
tendency to the formation of glaucophane. This is interesting, as lawsonite has been
looked upon as a frequent constituent of the glaucophane schists. Brown biotite is
present in skeletal plates penetrated poikiloblastically by large crystals of sphene,
magnetite, epidote, and felspar. The biotite appears in patches throughout the rock.
Epidote is in small quantities, but the grains included in the biotite are well developed.
Sphene is more abundant in this case than in other members of the series, though its
average grain size is not so large. The core of magnetite is prominent and may be
idioblastic. The magnetite as well as the sphene attains its maximum percentage
here. The magnetite crystals are frequently idioblastic, and may be so large that
we only get a thin veneer of sphene on them, and the sphene rim may not be continuous
around their circumference. Minute apatites are present. The rock is an amphibolite.

No. 634. In this example some degree of schistosity is produced by a parallel
arrangement of hornblende prisms. Such arrangement is almost perfect in an adjacent
band, No. 633, which is exceedingly fissile. No. 634 is granoblastic, and the abundant
rounded and embayed grains of clear plagioclase enhance this character. Though the
lamellar twinning is not well developed, some crystals with broad lamellae can be found
to give an extinction of 36, indicating labradorite. The second sodic plagioclase is
again present, and very occasional quartz grains can be found. The hornblende is
normal but with more indented outline than in No. 635. The biotite is brown, and
rarely intergrown with lawsonite, as the latter is very scarce. The iron ore consists of
scattered grains of magnetite and pyrite. Sphene is absent. The rock is an amphibolite
schist.

No. 634A. This is a highly schistose specimen with abundant hornblende.
Glistening mica with a light golden-brown colour is noticeable on the schist surface.
The very large percentage of hornblende is the striking feature of this rock, and in section
possesses the usual characters. The felspar is mostly cloudy and saussuritised, and only
occasionally does a clear grain arise from it. The small amount of clear felspar in this
rock is in contrast to the large amount of clear felspar in the neighbouring dyke, No. 634,
a yard or so away from it. The mica consists of intergrown biotite and chlorite, of
which the former is more abundant. The biotite possesses the same curious brown

THE METAMORPHIC ROCKS OF ADELIE LAND. STILLWELL. 37

colour as in Nos. 720 and 635. Lawsonite appears in the same manner as in No. 720.
The iron ore consists of magnetite and pyrite with some reddish hematite. The rock
may be called a lawsonite amphibolite schist.

No. 635. This example is a typical one with the parallel arrangement of hornblende
crystals. The hornblende is the most abundant mineral and is developed in long prisms.
Like No. 634A, the felspar is in strong contrast to No. 634, where it is mostly clear,
while here it is nearly all saussuritised. Occasional clear fragments arise among the
saussurite products. Grains of quartz, epidote, and mica can be distinguished in the
decomposed mass. The biotite is not abundant and contributes little to the schistosity
of the rock. Its colour is not quite normal. Basal sections possess a biscuit brown
colour, and cross sections are pleochroic from reddish brown to colourless. Consequently
the biotite possesses a curious bleached appearance. Lawsonite is well developed and
reaches 3 per cent., and was identified in this section before the preparation of the thin
sections of Nos. 720 and 634A. This colourless mineral is frequently found in parallel
growth with the biotite (Plate I., fig. 4). Sometimes it is so developed after this manner
that the biotite appears to be merely threaded in along its cleavage planes. Its form
is usually lobated, and the biotite plates, in consequence, bend around its contour.
Its cleavage is well developed and parallel to the elongation of the crystal and the cleavage
of the biotite. Apart from its association with biotite, it may be found in grains among
the saussuritised masses, and it also appears in small microscopic segregation veins.
The outline of the sections is frequently granular, but in the veins it may be rectangular,
and these show second order polarisation colours. Some sections, with low polarisation
colours, have a distinct tendency to a rhombic outline. Wherever the cleavage or
crystalline form is observable the mineral is found to have straight extinction. Further,
its biaxial character can be verified, and hence it is orthorhombic. Its optical character
is positive. Refractive index is high, higher than biotite and hornblende with which
it appears in contact. The birefringence is high, and second and third order colours
are seen. Some sections, however, possess quite low first order colours, indicating
that two of the three principal refractive indices are not much different. These
characters have fixed the identification as lawsonite. The lawsonite is not unlike a
colourless epidote in appearance, but the latter would not have uniform straight
extinction. Besides, epidote is present in isolated pleochroic grains of pale yellow
colour and may be compared with lawsonite in the same field of view. The iron ore
consists of scattered grains of magnetite and pyrite. Sphene is practically absent,
and apatite is found in occasional needles. No. 635 is a lawsonite amphibolite schist.

Summary of Microscopical Characters.

The dominating constituent throughout the series is the dark-coloured ferromag-
nesian mineral, either biotite or hornblende, or both. The biotite is most important in
the mica schists and the hornblende is most important in the amphibolites. The biotite
is fresh and clear, and the Z colour is normally brown, varying to a reddish brown.
In exceptional cases it is green, and both types may be found in the one section. Indeed,

38 AUSTRALASIAN ANTARCTIC EXPEDITION.

both types may appear in one and the same crystal, and the green biotite appears to
be a transition stage between green chlorite and brown biotite. The micaceous
constituent, in two of the described examples, is green chlorite. The chlorite plates
always, and the biotite sometimes, present a poikiloblastic structure. The included
minerals are epidote, felspar, sphene, and magnetite. Pleochroic haloes may surround
inclusions in biotite.

The green hornblende is also fresh and clear. It is characteristically bluish green
in the Z direction, indicating an admixture of the glaucophane molecule, and in one
case a bright-blue grain of glaucophane appeared. More commonly there are streaks
and fringes of pale hornblende through the more deeply coloured hornblende. Inclusions
are not abundant in the hornblende. Sometimes rounded blebs of quartz and felspar
produce the sieve structure. The prism and clino pinacoid faces are the best developed,
and cross sections with two cleavages may be idioblastic. At times chlorite and epidote
seem to develop from the hornblende.

In contrast to the fresh biotite and hornblende is the felspar, which may be quite
turbid and decomposed. The decomposition has been referred to throughout as
saussuritisation as epidote, lawsonite, zeolites, calcite, a secondary plagioclase, and
micaceous products have been recognised. , It is the same type of alteration as appears
on a large scale among the xenoliths which are to be described later. In some cases a
clear transparent felspar dominates, and in others the decomposed felspar, but both
often appear together. The discrimination of the plagioclase is not simple, because
lamellar twinning is not well developed. The lamellae are often confused and inter-
mittent, and many sections are not twinned at all. Such untwinned sections can be
proved by their refractive index not to be orthoclase. Labradorite, with broad lamellae
and large extinction angle, has been detected. This labradorite is at times partly
saussuritised. A second sodic plagioclase has been found with a smaller extinction
angle, fine lamellae, and lower refractive index. This second plagioclase is sometimes
andesine and sometimes nearer oligoclase. It is a product of recrystallisation in
which saussuritisation is but a stage, and the composition varies with the conditions
of recrystallisation.

Quartz is a minor constituent and detected in several cases. It is abundant in
examples where an ingress can be suspected.

Epidote appears throughout the series, but is more important in the mica schists
and mica amphibolites. Generally, its percentage varies with the percentage of mica.
It is found in characteristic honey-yellow or yellow-green grains frequently with crystal
outlines. It may also be found in aggregates of little grains which may appear like
small heaps. It may be intergrown with biotite or hornblende, and it may appear as
poikiloblastic grains in chlorites, biotite, or hornblende. It may rarely be intergrown
in parallel with zoisite (or clinozoisite). Even when the percentage of epidote is very
small we may still find large grains enclosed in biotite or chlorite. The epidote crystals
are occasionally found with nuclei of calcite and allanite.

THE METAMORPHIC ROCKS OF ADELIE LAND STILL WELL. 39

Sphene is an abundant accessory mineral in most members of the series. It is
frequently idioblastic and normally contains a crystal of magnetite as a nucleus.
Specimen No. 5, where sphene does not possess this nucleus, is exceptional. The
magnetite development varies from small crystals to large individuals, comparable in
size with the hornblende, which are only coated with a thin rim of sphene. In the
latter case the iron ore may total 6 per cant. Magnetite may also appear without the
sphene. The occurrence is similar to the " titanomorphitokranz " of the German
petrographers*, where sphene has recrystallised from decomposed ilmenite in
amphibolites. The magnetite has been described as such because of the highly magnetic
and polarised character of some separated grains, and because no violet colour was
obtained when an HC1 solution was reduced with tinf. In some sections a little red-
brown rutile is associated with the magnetite. Pyrite is occasional and sporadic.

Colourless apatite becomes an important accessory in some cases. Its host is
usually, though not necessarily, the felspar. The crystals may be minute, but in its
best development we may get crystals comparable with the grain size of the rock.

Lawsonite is an interesting constituent. Being a saussuritisation product it obtains
its best development in rocks with abundant saussuritised felspar, though it may also
form from hornblende. It is partly intergrown with biotite, and the lobated character
of the laminae is characteristic. It is partly in grains and partly in thin microscopical
veins. Occasionally the vein walls are lined with epidote while lawsonite forms the
main vein filling. In such cases the epidote has crystallised before the lawsonite. Both
the refractive index and the birefringence of the epidote are noticeably greater than
those of lawsonite in such instances. Calcite may form vein filling with the lawsonite,
and then lawsonite is found along the wall while calcite forms the centre. Calcite also
occasionally appears in coarse granular crystals. Fluorite appears in grains, or is
intergrown with biotite in the same manner as chlorite or lawsonite, while stilbite is
well developed in one example. Stilbite, fluorite, lawsonite, and epidote may be found
along joint planes and microscopic veins.

The rock types may be summarised thus

No. 153. Epidote biotite schist.

No. 5. Amphibolite.

No. 412. Biotite amphibolite schist.

No. 720. Lawsonite mica amphibolite schist.

No. 630. Quartz biotite amphibolite.

No. 629. Amphibolite.

No. 631. Amphibolite.

No. 637. Amphibolite.

No. 634. Amphibolite schist.

No. 634A. Lawsonite amphibolite schist.

No. 635. Lawsonite amphibolite schist.

Op. cit., voL I, p. 74.

t Since writing the above, some samples of iron ore from Cape Denison have been analysed by J. C. H. Mengaye, of the
New South Wale* Mine* Department. It has been found that the magnetic properties do not vary proportionately with
the titanium content. The more magnetic samples may have the higher TiOj value.

40 AUSTRALASIAN ANTARCTIC EXPEDITION.

4. CRYSTALLOBLASTIC ORDER.

The crystalloblastic order, as defined by Grubenmann*, is determined by the form
development of the crystal grains. A mineral placed in the crystalloblastic order will
assert its crystal form against all minerals that follow it in the sequence. The order is
based on a fundamentally different conception to that on which the order of
crystallisation of minerals in igneous rocks is based. In igneous rocks a definite order
is obtained based on the principles of solubility and mass action which prevail in a
rock magma. When, however, a set of conditions prevail which will stamp the special
metamorphic characters on a rock mass, the whole alteration takes place while the
rock remains solid. The new minerals resulting from the new set of conditions will
arise at practically the same time. They will grow together, and any one mineral may
be included in, or surrounded by, any other mineral. In doing so some minerals will
exert their crystalline form against the adjacent grains of other minerals. For example,
we may find felspar included in epidote, and we may also find epidote included in felspar.
The felspar included in epidote will be rounded in outline while the epidote included in
felspar will very likely possess crystalline boundaries. Epidote, therefore, stands
above felspar in the crystalloblastic orderf.

The crystalloblastic order is of value because the minerals possessing crystalline
form in schists are those which possess the greatest crystallisation force. The speed of
crystallisation may also be a factor in assisting or hindering crystalline form.

Grubenmann's teaching in this manner is not accepted by Leith and MeadJ. These
authors are inclined to imagine that crystal habit or crystal dimensions influence the
development of the new minerals, and believe that the mineral constituents are not of
equal rank and do show a definite order of crystallisation. It is to be pointed out that
Leith and Mead make no attempt to discriminate between sets of physico-chemical
conditions grouped together in Van Hise's zone of anamorphism, and, therefore, do
not recognise that a metamorphic rock may carry the impress of two or more sets of
conditions. They have not appreciated the fact that Grubenmann has attempted
to form a crystalloblastic order for each defined set of conditions, i.e., for each zone.
The crystal habits of the new minerals are but a reflection of these superimposed
conditions, and the size of crystals in metamorphic rocks is a variable factor without
important genetic connection. They have not shown that their order of crystallisation
means anything more than an order of application of successive sets of metamorphic
conditions.

The crystalloblastic sequence, as far as it can be observed in this suite of rocks, is
magnetite, sphene, epidote, hornblende and lawsonite, biotite and chlorite, felspar and
quartz.

* Op. oit., voL I., p. 73.

t The expression of this interpretation of Grubenmann's crystalloblastio order is necessary, because the term has been
in other meaning by Lahee in his paper " Cryetalloblastic Order and Minerals," in the Journal of Geology, vol. 22 pp.
>15. Lahee (p. 514) has confused order of origin with the crystalloblaatic order.
J " Metamorphic Geology," C. K. Leith & W. J. Mead, New York, 1915, p. 187.

THE METAMORPHIC ROCKS OF ADELIB LAND. 8T1LLWELL. 41

The magnetite is placed above the sphene because the common magnetite nucleus
of the sphene sometimes appears idioblastic. Large crystals of magnetite, when not
enclosed in sphene. may be xenoblastic and penetrated by idioblastic epidote. Biotite
flakes, with their usual ragged ends, may penetrate the ragged ends of hornblende prisms,
but the idioblastic cross sections of hornblende exert their form against biotite.
Lawsonite is placed above biotite because it exerts its lobate outline against biotite.
Sometimes lawsonite shows crystalline boundary against hornblende and sometimes
the reverse is seen. Biotite and chlorite are inseparable ; so also are quartz and felspar,
as quartz is always a very minor quantity.

5. CHEMICAL CHARACTERS OF THE CAPE DENISON AMPHIBOLITES.

In order to determine the chemical characters of the series, examples of two extreme
members were selected for analysis. Nos. 153 and 629 have been, therefore, analysed
in the Victorian Geological Survey Laboratory by A. G. Hall, under the supervision of
P. G. W. Bayly. No. 153 is an epidote bidtite schist with biotite developed almost to the
exclusion of hornblende. No. 629 is an amphibolite in which hornblende dominates
very largely over the mica, and it was chosen for analysis because its outcrop contains
the remarkable xenoliths. Actually, its hornblende content is a little higher and its
mica content a little lower than in the most typical examples. Such variation finds
its explanation in the metamorphic differentiation that has occurred in this band.

No. 153. No. 629.

Si0 2 52-73 .. 48-74

A1 2 S 13-99 . . 13-64

Fe 2 s 4-31 .. 3-31

FeO 9-19 .. 9-98

MgO 3-27 .. 7-12

CaO 5-98 .. 10-34

Na 2 1-73 .. 1-96

K 2 2-98 .. 0-83

H 2 0+ 1-72 .. 1-95

H 2 0- 0-10 .. 0-11

C0 2 strong trace . . trace

Ti0 2 2-14 .. 1-26

ZrOj nil .. nil

P 2 5 0-90 .. 0-14

S0 8 nil .. nil

Cl 0-08 .. o-04

S 0-04 .. 0-06

CrA 0-03 . . 0-05

MnO 0-39 . . 0-35

NiO, CoO 0-03 . . 0-01

CoO trace trace

42 AUSTRALASIAN ANTARCTIC EXPEDITION.

No. 153. No. 629.

BaO 0-05 . . nil

Li 2 trace . . trace

= C1 0-02 .. 0-01

= 8 . 0-01 0-02

Total . 99-63 .. 99-86

Specific gravity 2-953 . . 3-030

These two analyses bear important resemblances. The silica percentages are
relatively low and bear approximately the same ratio to the alumina. The total iron
is almost identical in the two cases. Both percentages of magnesia are lower than the
percentages of lime, and, further, the ratio of the magnesia to the lime is the same in
each case. The soda percentages are not far different, and the water content is similar.
Both are rich in titanium, and, in general, the similarity is- sufficiently strong to emphasise
the field observation that the two samples are of common origin.

At the same time the differences are interesting and important when compared
with the relative mineral compositions expressed in Table 1. The high mica percentage
in No. 153 involves a higher silica percentage and a noticeably higher percentage of
total alkalies with potash in greater amount. The high hornblende percentage in No.
629 involves the correspondingly lower silica, the much higher percentages of magnesia
and lime, and the much lower alkali total. The amounts of felspar are approximately
the same in each case, and hence the alkali percentage of No. 629 gives approximately
the amount of alkali in the felspar, and the extra amount in No. 153 can be attributed
to the mica. There is considerably greater quantity of iron ore in No. 153, and its
amount of Fe 2 3 is correspondingly greater. There is no corresponding variation
with FeO, as varying quantities of FeO are required for the ferromagnesian constituent.
The larger amount of sphene in No. 153 is also partly responsible for its higher titanium
percentage, but the differing percentages of P 2 5 are precisely reflected by the differing
percentages of apatite. The chlorine is probably associated with the apatite, and
thus appears in greater amount in No. 153. The sulphur is derived from the very
occasional grains of pyrite. Cr 2 3 , MnO, NiO, and CoO are no doubt contained in
the ferromagnesian. Finally, No. 153 is notable for its definite percentage of barium.

Since No. 153 expresses the composition of the bands with high mica content on
the western side of Cape Denison and No. 629 gives the composition of the amphibolites
on the east, the intervening bands, containing varying proportions of biotite and
hornblende, can confidently be expected to possess a chemical composition within the
limits of these two extremes. The actual variation could, indeed, be approximately
estimated from the mineral content expressed in Table 1 e.g., the greater the mica
percentage the nearer will the silica percentage approach that of No. 153. The percentage
of the minor constituents, like Ti0 2 and P 2 5 , will vary in much the same manner
as the corresponding accessory minerals.

THE MBTAMORPHIC ROCK8 OF ADBLIE LAND. 8TILLWELL. 43

Both analyses bear marked resemblances to analyses of basic igneous rocks. While
this could be illustrated by comparison with numerous examples, it is well illustrated
by assuming the rocks to be igneous and then treating them in accordance with the
principles of the American Classification of Rocks. No. 153 is then a member of the
division Class II., Dosalane ; Order 4, Austrare ; Rang 3, Tonalase ; Sub-Rang 3,
Harzose. The examples of this division quoted by Washington* are chiefly afforded
by granodiorites, diorites, andesites, and porphyrites. No. 629 falls into Class III.,
Salfemane ; Order 5, Gallare ; Rang 4, Auvergnase ; Sub-Rang 3, Auvergnose. The
examples quoted of this division include mainly diabases, gabbros, basalts, some
porphyrites, and camptonites. Both rocks, judged, therefore, from their chemical
composition, are likely to be metamorphosed basic igneous dykes. No. 153, being
more siliceous, probably approached rather towards a porphyrite, while No. 629 would
have probably tended to typical diabase or dolerite. The association with a meta-
morphosed granitic mass suggests their original character as basic lamprophyres ; but
the small percentage of the alkalies renders it unlikely. Metamorphosed lamprophyres
or lamproschiste are recorded from Garbh Allt, a mile S.E. of Glencaloie Lodgef, and
in the analysis the alkali percentage is as high as 6-95. The corresponding percentages
of these Cape Denison rocks are 4-71 and 2-79.

It is now necessary to examine these analyses with the view of classifying the rocks
in Grubenmann's classification of the crystalline schists. Before doing so, however,
we give a resume of the method of classification as little or no use of it has hitherto
been made in the English language. In the present state of our knowledge of the
crystalline schists this classification has considerable value. It has been put forward
to organise our knowledge, but a more complete understanding of the metamorphic
processes and their products will cause, at least, modification.

Grubenmann's Classification of the Crystalline Schists.

Grubenmann has classified the crystalline schists primarily on a chemical basis.
The chemical data yield him 12 groups, each of which is divided into three sub-groups
which are based upon the typical features associated with the physico-chemical
conditions of his three zones of metamorphism. He has pointed out that classification
on any other basis, e.g., mineral composition, mode of origin, original character, etc.,
will not succeed in bringing similar crystalline schists together, and, at the same time,
maintain their marked individuality which distinguishes them from the igneous and
sedimentary rocks. The variation, for example, of mineral content in this suite of rocks
under consideration, which bear strong chemical analogies, are similar in origin, and
have been subjected approximately to similar metamorphic conditions, is evident from
Table I. Mineral content is, therefore, useless as a classificatory basis if the classification

" Chemical Analyses of Igneous Rocks," H. 8. Washington. Professional Paper, No. 14, U.S. Geol. Sunr., 1903.

t " The Geology of Ben \Vyvi, Cam Chuinneag. Inchbae, and the surrounding Country." Memoir Geol. Surv. Scot.,
No. 93, 1912, p. 125.

44 AUSTEALAS1AN ANTARCTIC EXPEDITION.

is to succeed in grouping together similar species. The following are the groups appearing
in his classification :

1. Alkali felspar gneisses. 7. Chloromelanite rocks.

2. Aluminium silicate gneisses. 8. Quartzite rocks.

3. Lime soda felspar gneisses. 9. Lime silicate rocks.

4. Eclogite and amphibolites. 10. Marmorites.

5. Magnesium silicate schists. 11. Iron oxide rocks.

6. Jadeite rocks. 12. Aluminium oxide rocks.

Each group has its kata, meso, or epi division based on the characteristics of the
lowest zone, the middle zone, and the highest zone of metamorphism. Each division
again consists of families whose number depends on the number of known types of
schists contained in the division.

The classification is made quantitative by the use of an adaptation of Ozann's
treatment of a chemical analysis. The analysis is first modified so that the Ti0 2 and
P 2 5 are reduced to equivalent amounts of Si0 2 percentage ; the Fe 2 3 , MnO, Cr 2 O 3 ,
NiO, CoO, are reduced to, and then added to the FeO percentage, the BaO, SrO, to the
CaO, and the water neglected. The values of the seven constituents are then reduced
to their molecular proportions, which, in turn, are reduced to molecular percentages.*
From the molecular percentages seven group values, designated S, A, C, F, M, T, K, are
obtained in the following manner :

S denotes the Si0 2 in molecular proportion.

A is the similar sum of K 2 and Na 2 which is combined with A1 2 3 in the

1 : 1 proportion.

C is the CaO combined with A1 2 3 in the 1 : 1 proportion.
F is the sum of FeO and MgO and that part of CaO which is not absorbed

in the 1 : 1 proportion with A1 2 3 .
M is the residual CaO used in F.
T is the residual A1 2 3 not absorbed in the 1 : 1 proportion with K 2 0, Na 2 0,

and CaO.

K is the value of the quotient . =F,

6A + 2C + F

The values S, K, A, C, F are used exactly with Ozann's meaning ; K, however,
is only important here in determining the degree of acidity of the crystalline schist.
M gives the absolute amount of CaO in F and is useful in dealing with lime silicate rocks
signifying their sedimentary origin. T is necessary to express the high A1 2 3 content
in some gneisses, especially those derived from clay sediments. No term is introduced
to express the relation of the alkalies to one another as it is undesirable in the present
state of our knowledge of the crystalline schists. A classification at present can only
deal with the broader features.

* The extra step, explained by Grubenmann, op. cit., vol. II., p. 12, of reducing the seven values to percentage values
before determining the molecular proportions, is superfluous.

THE METAMORPHIC ROCKS OF ADELIE LAND. 8TILLWELL.

The group values together, not individually, represent the chemical characteristics
of each main group of the classification. For each main group there is a set of mean
group values with a definite range of variation.

The group values for any example are represented graphically by points in Ozann's
triangular projection. If the factor 20 is used, each side of the equilateral triangle is
divided into 20 and lines parallel to the sides are drawn through each division. Perpen-
diculars are drawn from the angular points on to the sides and the projection values
a, c, f, are measured from the base along the perpendiculars. The projection values
are calculated thus

20A 20C 20F

a =

C + F'

c =

A + C + F

The result of this is that differing groups of schists occupy more or less distinct areas
in the triangle, and the position of a schist on the projection may give a means of indicating
the origin, igneous or sedimentary.

The dassificatory Position of the Cape Denison Amphibolites.

If the analyses of rocks Nos. 153 and 629 be treated in this manner, we obtain the
following results :

No. 153.

No. 629.

Reduced
Analysis.

Molecular
Proportion.

Molecular
Percentage.

Reduced
Analysis.

Molecular
Proportion.

Molecular
Percentage.

SiO,

65-09

918

61-6

49-80

830

53-1

AUG.

13-99

137

9-2

13-64

133

8-6

FeO

13-49

188

12-6

13-36

186

12-0

CaO

6-00

107

7-2

10-34

185

11-9

MgO

3-27

6-5

7-12

178

11-5

K.O

2-98

2-1

0-83

0-6

Na g O

1-73

1-8

1-96

2-0

96-56

1,491

100-0

97-06

1,555

100-0

Group Values.

No. 163

61-6

3-9

5-3

20-0

1-9

No. 629

53-4

2-6

6-0

29-4

5-9

AUSTRALASIAN ANTARCTIC EXPEDITION.
Projection Values after Ozann.

a =

c =

20A

A + C + F

200

A + C + F

20F

No. 153.
2-7

3-6
13-7

No. 629.
1-4

3-1
15-5

A + C + F

Examination of these group values enables one to place both rocks among the
eclogites and amphibolites of Group IV. No. 629 is a typical amphibolite not far
removed from the mean group value. No. 153 approaches the plagioclase gneisses of

^VNAAAAAA

VWYWVV

AAAA/WV

AAAAAAAA/V\IA/WVWV\AA
a c

Fig. 4.

3. Mean Value of Group III., the Plagioclase Gneisses.

4. Mean Value of Group IV., the Eclogites and Amphibolites.
153. Epidote Biotite Schist, Cape Denison.

629. Amphibolite, Cape Denison.

Group III., and the group values S and M actually fall within the variation limits of
this group ; but, nevertheless, its position on the projection is much closer to the mean
position of Group IV. than to the mean position of Group III. Since both examples
fall into Group IV., we are able to assert that the whole suite of rocks considered falls
into the same group.

In order to determine the subdivision of Group IV. it is necessary to recall the
microscopical characters. When we do so we find that the rocks do not wholly present
the characteristics of either the epi division or the meso division. In cases where the

THE METAMORPHIC ROCKS OF ADELIE LAND . 8TILLWELL.

clear felspar is albitic and other plagioclase is much saussuritised, where biotite is replaced
by chlorite, where epidote partly replaces the calcic plagioclase thereby absorbing a
good deal of the lime content, where calcite also absorbs some of the lime content as in
No. 412, and where lawsonite is present, we have features of the epi division. Where,
however, we find considerable quantity of recrystallised clear andesine, biotite without
chlorite, and abundant clear hornblende with only rare transition to epidote, to chlorite,
or to glaucophane, we have dominant meso division features. Yet it is to be noted
that abundant saussurite and lawsonite is found with clear hornblende, saussurite with
clear felspar which is not albite, chlorite and epidote with biotite in the same section.
It, therefore, appears that the series has to be considered as more representative of the
transition types between the meso and epi divisions.

o
c;

<fl

mete-ten otil-h a

: <

amp hi Mi Ae

^.
^

Fig. 5.

DIAGRAMMATIC SKETCH or THE AMPHIBOLITE DYKE No. 629 WITH

THE SCATTERED META-XENOLITHS AND THE CLOTS OF CHLORITE

ROCK AND EPIDOSITE.

No. 153 is an epidote biotite schist and has suffered higher recrystallisation than the
epidote chlorite schist in family A of the epi division. With the temperature and the
uniform pressure approaching that of the middle zone, the chlorite has passed over to
biotite. The actual transition is found in No. 630, where the chlorite passes first into
green biotite and the latter into brown biotite. That the epidote remains after the
chlorite has changed to biotite is due to the fact that epidote can retain its water at a
much higher temperature than chlorite. Grubenmann points out that such individual

48 AUSTRALASIAN ANTARCTIC EXPEDITION.

characteristics of mineral and rocks necessarily cause the features of one zone to encroach
in varying degrees on an adjacent.* The abundance of biotite in any of these specimens
probably means, therefore, a previous abundance of chlorite. This abundance of
chlorite, in turn, means abundant chloritisation of the pyroxenes of the primary diabase,
which may have occurred either in normal weathering or in the upper parts of the epi-
zone. As, however, hornblende as well as biotite could develop from chloritised pyroxene,
the amount of biotite cannot be considered an index of the amount of primary
chloritisation.

6. METAMORPHOSED XENOLITHS (META-XENOLITHS).

One band of amphibolite (No. 629), outcropping near the centre of the Cape Denison
area, is phenomenal in containing a large number of xenoliths.f These xenoliths possess
the same metamorphic character as their host, and may be distinguished as
" metamorphosed xenoliths." We propose, for convenience, to abbreviate " meta-
morphosed xenolith " to " meta-xenolith."

The particular band appears as a broad bulge, about 4yds. wide, issuing from
underneath the ice sheet, and after continuing for about 15yds. or 20yds. it narrows
down to a band of average width. The meta-xenoliths are scattered through the whole
outcrop, but are most abundant along the western edge of the bulge (fig. 5). They
consist of white, grey, pale-green, or pale-pink masses which are never more than a few
inches long, and which produce strong contrast in colour to the black amphibolite host.

There are two distinct types of material among these meta-xenoliths, and they
may be distinguished as

(1) Saussuritic type.

(2) Gneissic type.

These two types will be subsequently found to correspond to the cognate and
accidental xenoliths of normal igneous rocks.

1. Saussuritic Type.

The saussuritic type includes the pale-green and pale-pink masses, which may
be again subdivided into

(a) Those composed wholly of saussurite the individual type.

(b) Those composed of an aggregate of saussurite and hornblende the

composite type.

(a) The Individual Type of Meta-xenolith. The meta-xenoliths composed wholly
of saussurite} may retain the original outline of a primary felspar crystal. The largest

* Op. oit., vol. I., pp. 70, 71.

t We use the term " Xenolith " in the same sense that it is applied to igneous rocks. Grubenmann does not provide a
special equivalent in his system of nomenclature for the crystalline schists.

t The term " Saussurite " is used in the same sense as given by Weinshenck (Petrographic Methods, trans Clarke p 336)
and by Flett ("Geology of the Lizard and Meneage," Mem. Brit. Geol. Surv., 1912).

THE METAMORPHIC ROCKS OF ADEL1E LAND. STILLWELL. 49

example in the collection of such a crystal is lin. broad, and shows the re-entrant angle
of a simple twin (Plate X., fig. 6). In other cases the saussurite masses are both rounded
and angular. The largest rounded mass among the specimens in the collection is 2in.
in diameter. A remarkable example of an angular mass of saussurite is shown on Plate
X., fig. 5. Here the section is a perfect triangle, with the sides measuring 2in., If in.,
and If in. The boundary is macroscopically sharp, except for a minor length which
is a little ragged at the left hand corner and which is scarcely noticeable in the
photograph. A small amount of hornblende and epidote is macroscopically visible
in this example. In all cases the junction between the saussurite and the amphibolite
is normally sharp, irrespective of the crystalline, angular, or rounded nature of the
contour. Some examples, which are illustrated on Plate IX., fig. 4, consist of
amphibolite uniformly and thickly studded with small patches of saussurite averaging
Jin. in diameter. A boulder found on the lower moraine a little north of the outcrop
is used as the diagram, but similar examples collected in situ are in the collection. The
appearance is that of a porphyroblastic amphibolite, though there is considerable
variation in size. Such would be a likely explanation were they not only found in
association with the better defined meta-xenoliths.

Macroscopically the saussurite is a compact, stony mass, in which one can sometimes
distinguish black specks of hornblende, green crystals of epidote, and, more rarely,
white patches of calcite. Thin sections of this type reveal the crystalline aggregate
known as saussurite. The larger xenoliths have produced relatively coarse crystalline
aggregates wherein identification of the constituents has become possible (Plate II.,
fig. 4).

A good portion of the aggregate is always a cloudy, brightly polarising mass similar
to the saussuritised felspar, to which reference has been made in dealing with the previous
rock types. At times a system of parallel lines, defined by thin lines of hematite or
limonite, are observed in parallel light, and these represent traces of the broad lamellae
of the primary felspar. In rare instances relics of the primary felspar itself are found.
In such cases the bulk of the crystal has been saussuritised, and only a few clear lamellae
are left. A section was found normal to these primary lamellae and gave an extinction
angle of 44, measured from the lamellae bands. The primary felspar is highly calcic
and near the anorthite end of the series.

In the confused aggregate epidote is prominent, and the large grains can be
recognised at once by the brilliant polarisation colours. It is a very pale epidote with
feeble pleochroism in the thin section and with the (001) and the (100) cleavages well
developed. The optic axial plane is normal to the cleavages as usual, and from the
(001) cleavage the extinction angle is approximately 25, and from the (100) it is approxi-
mately straight. The outline of the large pieces is usually granular, but sections showing
two cleavages with crystal boundaries may be found. Sometimes the larger grains
are bordered with finely granular masses of a rather darker epidote. The latter is
much more abundant in some sections than in others, and it seems to mark a stage in

i A, VoL m.. Part 1 D

50 AUSTKALASIAN ANTARCTIC EXPEDITION.

the decrystallisation of the felspar. Associated with the epidote is clinozoisite and
zoisite, which are often intergrown in the one crystal. Clinozoisite is the more abundant,
and is readily detected by the ultra blue polarisation colour. Like the epidote, two
cleavages are present, and it is found to have approximately straight extinction with
reference to one cleavage and a large extinction with reference to the other. In several
instances the optic axial plane was determined to be perpendicular to the cleavage,
and the small curvature of the bar in the interference figure in sections normal to an
optic axis indicates a large optic axial angle. Clinozoisite may also be bordered by
finely granular material. The zoisite, with its bluish-grey polarisation colour, is distinct
from the clinozoisite and the epidote. Wherever determined the optic axial plane is
normal to the cleavage indicating the variety zoisite ft.

Lawsonite has been found in some sections in large individuals and in small veins.
In the development of some of the crystals a brown micaceous mineral, like poorly-
developed biotite, has been thrown out along the cleavage planes. This fact may be
interpreted as evidence of the contemporaneous development of the intergrown lawsonite
and biotite reported in the lawsonite amphibolites.

Green pleochroic chlorite is present, and may show the usual anomalous polarisation
colour or a pale-greenish-white colour between crossed nicols. The amount of chlorite
is very small in some cases and in others the aggregates may be radial. A white
mica is present, which is probably muscovite. When best developed it is clear and
colourless, with cleavage and the usual absorption. Sometimes it presents a confused
and ragged appearance with the laths set in a criss-cross manner.

In similar association to the white mica is scapolite, with its low refractive index
and brilliant polarisation colours. It has been identified by its uniaxial and negative
character. How much of the brightly polarising mass is scapolite and how much is
white mica must remain an unsettled question. Calcite is sometimes found in plates
of irregular outline, and a small segregation of calcite is present in one instance. Grains
of pyrite and magnetite are nearly always present, and in several cases the pyrite is
visible macroscopically.

In this mineral aggregate there is sometimes a clear felspar which is either untwinned
or finely lamellar twinned. The refractive index is moderately low, but always above
Canada balsam. In No. 628 (5) the extinction angle goes up to 18 when measured in
sections with cleavage but without twinning. It is therefore interpreted as either
oligoclase or andesine. In this instance the felspar of the adjoining amphibolite is quite
clear and recrystallised, and seems to be identical with the clear felspar in the saussurite.
The latter is sometimes fringed with the clear felspar which then comes in contact with
the clear felspar of the amphibolite. An extinction angle of 17 can be measured among
the grains in the amphibolite. Hence, if both are identical the determination must be
andesine.

THE MBTAMORPHIC ROCKS OF ADELIE LAND. 8TILLWELL. 51

This mineral aggregate, even apart from the primary felspar, is conclusive that we
are dealing with the decomposition products of a highly calcic felspar. The crystalline
outline, as far as it is observable, agrees with a felspar. The secondary felspar
in saussurite is usually recorded as albite, but the composition of the felspar depends
on the conditions under which saussuritisation takes place. The conditions under which
muscovite, scapolite, zoisite, and clinozoisite form are not those under which secondary
albite can form. Among the minerals identified above, lawsonite and chlorite are
unimportant or absent in the best or coarsest crystalline aggregates.

(b) The Composite Type of Meta-xenolith. The composite type of meta-xenolith
is formed of a number of saussurite " crystals " set in a hornblende matrix. These
aggregates may have an irregular, rounded, or angular outline which frequently appears
macroscopically sharp, but it is not necessarily so. A diagrammatic example of one
of these aggregates is given on Plate IX., fig. 3. In this case the boundary is sharp
and some of the saussurite masses have preserved the shape of the primary felspar
crystals. The coarse grained character of the primary xenolith is here quite evident.
A remarkable angular example is illustrated on Plate IX., fig. 2. These two " diagrams "
were collected from the lower moraines, which consist almost entirely of local rock,
at a point a few yards north of the occurrence in situ. A rectangular example obtained
in situ is shown on Plate IX., fig. 1. In all these cases the apparently sharp boundary
is actually a line of interlocking saussurite and amphibolite", and the saussurite aggregates
throughout the block are set in a mass of interlocking, granular hornblende. In type
the saussurite is similar to that described in the individual type of meta-xenolith, but
here in the smaller crystals recrystallisation has not been so intense. Further, in the
specimens that have been examined, lawsonite and chlorite and calcite are more
abundant than in previous cases. In addition, sphene is sometimes found in these
aggregates. The hornblende is precisely similar to the hornblende in the amphibolites,
and the clusters of granular hornblende may readily represent the decrystallisation
products of a large primary augite.

There is every reason to believe that these composite meta-xenoliths represent
the relics of clots of coarse-grained rock of the same composition as the primary dolerite.
No minerals, except those which have become recognisable in the saussuritic aggregates,
are found in the clots that are not found in the amphibolite proper. The clots consisted
chiefly of coarse felspar and coarse augite. The coarse felspar is now a saussurite complex
and the coarse augite is now an aggregate of granular hornblende.

2. Gneisaic Type of Meta-xenolith.

The gneissic type of meta-xenolith shows considerable variation in colour, shape,
and size in the hand specimen. They are indiscriminately mixed with the saussuritic
meta-xenoliths. Some examples possess a grey colour and so bear strong resemblance
in the hand specimen to the grey granodiorite that surrounds the amphibolite. Other

52 AUSTRALASIAN ANTARCTIC EXPEDITION.

examples have a pure white colour, and others, again, have a vitreous grey colour
which is suggestive of a colour change during recrystallisation.

The shape, in many instances, is clearly angular and fragmental, and the corners
may be well preserved. Frequently the gneissic fragment is drawn out into a lenticular
shape in the direction of the schistosity (Plate X., fig. 1). In this example the lenticular
bodies are not symmetrical to the schistose plane. A side view of the same specimen
is shown (Plate X., fig. 2). Here a meta-xenohth at the upper right hand corner is almost
triangular in outline, yet the schistosity of the rock can be distinctly seen to follow
through the inclusion from the amphibolite irrespective of its shape. Hence the
amphibolite and the fragment must have formed a single unit before the reception
of the metamorphic impress. In fewer cases the cross section is elliptical and, therefore,
symmetrical to the schistosity ; in such examples recrystallisation and rearrangement
are evident. The back and front views of another specimen are illustrated on Plate
X., figs. 3 and 4, where the gneissic meta-xenoliths are not lenticular but possess an
angular and variable shape.

The outline of the gneissic inclusion is often clear and sharp, though we may find
it slightly embayed. There are instances, however, where the entire boundary is lost
and replaced by a transition between the amphibolite and the white gneiss. An
inclusion is also observed where part of the boundary is sharp and part indistinct. This
indefinite boundary might be accounted for by postulating chemical action between
the xenolith and the host before the metamorphism. Such, however, is not a necessary
hypothesis, because evidence will be produced later which leads us to discount the normal
face value of transitions in metamorphic rocks, and to believe that such transitions
can arise during the progress of the metamorphism. That there has been an adjustment
of molecular equilibrium along the junction during metamorphism seems evidenced
by the lines of amphibolite which may be sometimes seen threading their way from the
host in the direction of the schistosity of the inclusion. Though the junctions may be
sharp there is perfect crystalline continuity and an interlocking of crystals across them.

In thin section (No. 628-3) the fragments are found to be clear granoblastic
aggregates of quartz and felspar (Plate II., figs. 1 and 2). The grains have a tendency
to be rounded or elliptical, and are of moderately even size, averaging about -16mm.
in diameter. Actually each grain has irregular outline, is much embayed, and always
interlocks with its neighbour. Undulose extinction is marked in the quartz and some-
times in the felspar. The felspar often possesses lamellar twinning, and as its refractive
index is near that of quartz, and sometimes above, it is identified as andesine. Small
crystals of biotite and chlorite are distributed through the mass and show a tendency
to parallel arrangement. Hornblende is present in rather larger crystals, and epidote,
clinozosite, and other saussuritic products are scattered in groups with a tendency to
linear distribution. Magnetite and pyrite are accessories.

In other cases (No. 628-6) porphyroblasts of quartz and felspar are found. They
possess the lenticular cross section, and are relics of the primary minerals. The quartz

THE METAMORPHIC ROCKS OF ADELIE LAND. ST1LLWELL. 53

porphyroblasts have just commenced to develop a broken granular appearance and
show the intermediate stages in the destruction of a large primary crystal. One quartz
porphyroblast is a fractured granulitic aggregate, though it still retains its entity in
both ordinary and polarised light. In another case fracturing has not occurred, and
the central portion of a porphyroblast still shows unbroken strings of linear inclusions,
though incipient granulitisation appears between crossed nicols. These strings run
diagonally across the plane of schistosity. The section is elliptical and the ends of the
longer diameter consist of a granular interlocking quartz aggregate in which each grain
possesses different optical orientation. It is an excellent example of the result of
solution at the points of maximum pressure with simultaneous deposition at the points
of minimum pressure in the plane at right angles to the direction of pressure. In the
fractured quartz porphyroblast secondary minerals like chlorite, epidote, and calcite
now appear along the fractures. The felspar porphyroblasts are also elliptical. Their
calcic nature is evident by the saussuritic products in which chlorite and epidote are
definitely recognisable. The centre of one porphyroblast is a granular aggregate,
produced by the breaking down of the primary felspar, which consists chiefly of clear
secondary felspar with lower refractive index with some epidote, chlorite, and calcite.
The remainder of the felspar porphyroblast, apart from the granular nucleus, has also
suffered decrystallisation and now presents a " peg " structure. Small rounded blebs
of secondary felspar appear in contrast to the primary felspar in polarised light.

These porphyroblasts of quartz and felspar are set in a much finer granoblastic
aggregate of quartz, clear felspar, and saussuritised felspar with sporadic grains of
magnetite and pyrite, epidote, chlorite, hornblende, and sphene. The typical grain
is here elongated in the direction of the schistosity, giving evidence of a certain amount
of crystallisation schistosity. Idioblastic crystals of apatite are included in the quartz.
Besides the granular individuals of saussurite in this section there are lenticles of
saussurite from a neighbouring saussuritic meta-xenolith. The cloudy appearance
has occasionally disappeared and there is left a mass of epidote and chlorite. Some
of the layers, rich in saussurite, can be traced directly into the enclosing amphibolite,
and some contain sphene and hornblende.

The gneissic meta-xenoliths, therefore, possess characters which are essentially
foreign to the amphibolite host. They possess affinities to the surrounding gneiss
though they seem to show a slightly greater degree of recrystallisation.

Origin of the Meta-xenoliths.

1. Saussuritic Type. The individual variety of saussurite inclusions have been
derived from the decomposition of a felspar. The primary felspars, particularly those
with crystal outline, may have been phenocrysts of intra-telluric origin brought up with
the injection of the dyke magma. But the presence of the angular and rounded masses

54 AUSTRALASIAN ANTARCTIC EXPEDITION.

of saussurite show that we are not dealing with a porphyritic dyke rock, while the
irregular and local distribution is strong evidence of cognate xenoliths. No one can
suppose that the meta-xenolith in Plate X., fig. 5, could be anything but a fragment
of a pre-existing felspar crystal. Large calcic felspars do develop in amphibolites
under the metamorphic conditions of the kata zone or the lower meso zone. An
example of this nature (No. 212) was found among the boulders on the moraine and
the porphyroblasts (Plate IX., fig. 5) do not bear a trace of decomposition in thin
section. Such crystals would become saussuritised if subjected for a sufficient length
of time to the conditions of the epi zone. We have found no evidence to suggest that
any of the amphibolites found in situ on Cape Denison have been subjected to the
kata zone conditions, and no such hypothesis would explain the extraordinary local,
irregular, and unsymmetrical distribution.

The composite variety of saussuritic meta-xenolith is also best explained as a
metamorphosed cognate xenolith. It is not likely that they are unabsorbed residuals
of primary rock which has survived the metamorphism. Their boundary is too definite
and they actually bear the same metamorphic impress as the amphibolite itself. They
are fragments of a rock of the same composition as the amphibolite, but of much
coarser grain than the primary dolerite. They have been cognate xenoliths brought up
from the magma reservoir, and probably represent differentiation products produced
by crystallisation in that reservoir.

2. The Gneissic Type. It has been shown that this type of inclusion is foreign to
the enclosing amphibolite, but that it is related to the granodiorite gneiss which
surrounds the amphibolite. Their unsymmetrical character and arrangement can be
accepted as definite evidence that they attained their present situation before the
reception of metamorphic characters. There is no alternative but to consider them
as "accidental xenoliths" or fragments which have no genetic relation to the enclosing
amphibolite, and which have been caught up during the injection of the primary dolerite
dyke. Xenoliths have been frequently reported in the basaltic and doleritic dykes, and
such was the original nature of this amphibolite host.

Significance of the Meta-xenoliths.

The consideration of these different kinds of meta-xenoliths collected from the
same small area shows conclusively that their host is an igneous rock intrusive into
the surrounding granodiorite. There can be no question of bedded tuff.

The marked angularity of some of the fragments means that the xenoliths did not
travel far along the dyke channel. The saussuritic type must have come from the
magma reservoir, and therefore the present surface must be close to the original magma
reservoir. The gneissic xenoliths may have been fractured from the walls of the dyke
channel and not necessarily from the roof of the magma reservoir.

THE MBTAMORPHIC ROCKS OF ADELIE LAND. 8TILLWELL. 55

Analogous Occurrences.

Xenoliths have been frequently reported in normal dyke rocks, and a summary
of a number of such occurrences in Europe and America has been recently made by
Powers*. Xenoliths in dykes which are now represented by metamorphic rocks, are
much less frequent. Flettf has recorded examples from the Lizard district where
dykes of gabbro schist and gabbro pegmatite contain inclusions of serpentine, and where
epidiorite dykes contain inclusions of red granite.

I am not aware of the recognition of xenoliths within amphibolite dykes. In
Beinn Lair and Meall Mheinnidh, of the Loch Maree and Gairloch District, C. T. CloughJ
has reported certain zones of hornblende schist which contain lenticles of a dirty white
opaque susbstance which Teall identified as saussurite. A considerable number of
these lenticles are more than 1ft. long, while some exceed 3ft. Their long axes lie
parallel with each other in some patches, while in others they do not. The long axes
are independent of the foliation of the schist. The lenticles have an irregular
distribution, and an isolated instance is recorded at a distance of 60yds. from any
others. Clough, finding difficulty of explanation, decided that they more probably
represent concretions in an igneous rock before its conversion into schist, and that
they may have been originally nearly spherical and analogous to spherulites.

It is quite likely that this occurrence in Scotland is analogous to the occurrence
of saussuritic meta-xenoliths at Cape Denison. There is no marked angularity of the
fragments in the Scottish instance, neither is there more than one type of fragment
recorded, nor is the dyke-like nature of the host obvious. Nevertheless it is quite
possible that the Scottish saussurites were cognate xenoliths brought to their present
position by an invading magma before the development of metamorphic action. From
quite independent sources Clough considers that there is little doubt that the
hornblende schists were intrusive rocks.

7. THE ORIGIN OF THE AMPHIBOLITE SERIES.

We have now presented the field, microscopical, and chemical characters of the
amphibolite series, and we may now summarise the evidence bearing upon its origin.

Field Evidence. In the first instance field observations strongly suggested that
this suite of rocks constituted a parallel system of intrusive dykes. The uniform width,
the frequent sharp line junction, the linear trend, and their persistency are valuable
criteria. Fresh from the study of a parallel system of dykes|| the likeness to such was
found to be highly suggestive. Bulges or swellings in the dyke channels had been

" The Origin of Inclusion* in Dyke*," 8. Power*, Journ. Oeol., vol. 3, p. 1.

t " The Geology of the Lizard and Meneage," J. 8. Flett ft J. B. Hill, Mem. Geol. Burr. Gt. Britain, Sheet 359, 1912,
pp. 94-128.

J " The Geological Structure of the North-Weit Highlands of Scotland," Mem. Geol. Surv. Gt. Britain, 1907, p. 243.

{ Op. cit., p. 240.

|| " Preliminary Note* on the Monchiquite Dykes of the BendiRo Gold Field," Proc. Roy. 8oc. Viet,, 1911, p. 1.

56 AUSTRALASIAN ANTARCTIC EXPEDITION.

seen in the Bendigo mines, and a broken surface outcrop in a metamorphic series is not
unfavorable when such can appear in the unaltered series at Bendigo. The detached
fragments of the dykes which are encircled by gneiss, and which could be mistaken
for inclusions caught up by the invading magma, are undoubtedly related to and belong
to the dyke magma.

Xenolith Evidence. The discovery of metamorphosed xenoliths in one outcropping
band of amphibolite is very important evidence of igneous and intrusive origin. The
fragments of gneiss with sharp boundaries and with marked likeness to the surrounding
gneiss possess a composition fundamentally different from that of the amphibolite host.
The saussuritic type of meta-xenolith is one that might be expected to come from the
magma reservoir from which the dykes issued. Knowing the granitic -nature of the
surrounding gneiss, it is impossible to conceive these xenoliths as undigested fragments
of an igneous or of any other pre-existing rock.

Structural Evidence. No relic of any kind of sedimentary structure is to be found.
The typical granoblastic structure is in this case more suggestive of igneous origin.

Mineralogical Evidence. It would be difficult to account for the suite of minerals,
particularly the abundant saussuritised felspar, the relic felspar, and some well-formed
apatite crystals, on any other hypothesis than that of igneous origin. The uniform
variation in mineralogical composition of the different members of the series, which is
illustrated in Table I., and which reflects uniform variation in chemical composition,
indicates an igneous differentiated rock series.

Chemical Evidence. The chemical analyses bring forward strong evidence of
derivation from doleritic rocks. The analyses are similar in all essential points with
analyses of diabases or dolerites. The definite grouping, on quantitative data, among
the amphibolite group of the crystalline schists, is further evidence when we recall that
many members of this group have arisen from diabasic dykes*.

The total evidence is thus conclusive that this suite of rocks from Cape Denison,
conformable to the general foliation of the country, is the metamorphosed equivalent
of a system of parallel igneous dykes. The dykes have intruded the granite prior to
the development of the foliation. The granodiorite and dykes have then suffered the
same metamorphic conditions with varying amounts of recrystallisation. The
surrounding granodiorite excludes any possibility of the amphibolites representing
altered bedded tuffs.

The nature of the primary dyke corresponds with a diabase or a dolerite whose
mineral composition has been calcic felspar (labradorite), pyroxene, biotite, ilmenite,
and apatite. No trace of serpentine is found, and, as serpentine can be preserved
under epi zone conditions, it is concluded that no olivine was present in the primary.

* " Die Kristallinen Schiefer," vol. II., p. 94. " Data of Geochemistry," F. W. Clarke, Bull. 330, U.S.A. Geol. Surv.,

p. 508.

THE METAMORPHIC ROCKS OF ADELIE LAND. 8TILLWELL. 67

The calcic felspar has been saussuritised, rarely persisting as a relic. The
saussuritisation here is not a simple weathering process. Normal surface weathering
is absent in Adelie Land, and the mineral products in saussurite are perfectly fresh,
even after exposure at the surface. Members of the epidote family, chlorite, lawsonite,
occasional zeolites, a secondary white mica, scapolite, and a secondary sodic felspar,
have been recognised in the saussurite. Part of the saussuritised felspar has
recrystallised, and in some cases we get clear andesine formed. The pyroxene has been
completely changed. It is replaced by clear hornblende in the amphibolites, by biotite
and epidote in the biotite epidote schists, and by both hornblende and biotite with
associated epidote in the biotite amphibolites. In discussing the zonal changes, it
has been considered that the biotite has developed through a chloritic stage, and that
the chlorite was derived directly from the pyroxene. It is probable, however, that
there was a little primary biotite in the diabase. The amount of mica in the
amphibolites is approximately constant, and therefore cannot be dependent on a
varying amount of chloritisation of the pyroxene. Further, the chlorite in these
members appears regularly in large broken plates, which are sparsely distributed, and
which are always penetrated poikiloblastically by epidote, together with clear quartz
or felspar and iron ore. Such chlorite can be considered as produced in the
decrystallisation of primary biotite*. The primary ilmenite has been altered to
leucoxene, which has recrystallised as sphene, or it has decomposed into sphene and
magnetite according to the equation given by Van Hisef. It is doubtful whether all
the titanium is dissociated from the iron though the occasional presence of rutile rather
suggests so. Certainly the larger grains are little magnets. The apatite has remained
unchanged.

Either during or subsequent to the metamorphism fracturing occurred and the
fractures have been filled with quartz, felspar, epidote, lawsonite, and calcite. Such
epidote and lawsonite, etc., may be subsequent to the epidote and lawsonite in the
schists, but cannot be used as an argument to show that all the epidote and lawsonite
is formed subsequent to the schistosity. The epidote that takes definite part in the
foliation must be considered as a primary metamorphic mineral of the same standing
as biotite or hornblende. The epidote percentage has been shown to vary
sympathetically with the biotite percentage which, in turn, varies inversely with the
hornblende percentage. Further, the biotite or chlorite may be moulded on to perfect
crystals of epidote in a manner which is impossible on a theory of subsequent
epidotisation. The mineral-filled fractures do show that the rocks have been in a zone
containing water. As fracturing may occur under the conditions of excessive stress
in the epi zone of metamorphism, and as water may be present in this zone, there is
no need to dissociate these minute fractures from the metamorphic characters.

In this manner, then, a diabase or dolerite containing calcic felspar, pyroxene,
biotite, ilmenite, and apatite has been converted into an epidote biotite schist or an

* Van Hue, " TroatiM on MeUmorphum," Hon. 47, U.S.A. Geol. Surv., p. 341.
fOp. oit., p. 227.

58 AUSTEALASIAN ANTAECTIC EXPEDITION.

amphibolite containing saussuritised felspar, sodic felspar, hornblende, biotite, chlorite,
epidote, magnetite, sphene, pyrite, apatite, and rarely rutile and fluorite. These changes
have, at times, been accompanied by the addition or transfer of material, and, in some
cases, it is very important and leads to a theory of metamorphic differentiation.

8. ORIGIN OF CERTAIN CLOTS IN THE DYKES. METAMORPHIC DIFFERENTIATION.

Though we have determined the origin of the amphibolite dykes, there remains
for explanation the curious schlieren of biotite which were found in two dykes and were
mentioned in the field characters. These appeared like segregations in the dykes and
could be completely surrounded by the apparently normal dyke rock, though there
were no sharp boundaries. Schlieren of chlorite and epidosite were found in similar
circumstances (fig. 5). In each of these three cases the schlieren occur within portions
of sharply walled dykes. At first sight these seem to find explanation by postulating
primary magmatic xenoliths, composed possibly of augite or olivine, whose individuality
has been preserved throughout the metamorphism.

We consider first the biotite schlieren. One of the biotite schlieren, No. 4 (Plate
II., fig. 6), has been found to possess the following mineral composition :

Biotite 64-9

Hornblende 32-2

Quartz 1-7

Muscovite 0-6

Lawsonite 0-4

Epidote 0-2

Apatite, Sphene present

The rock is highly schistose and shows a number of angular folds. The angle
made by the sides of the folds is 30. The biotite forms practically two-thirds of the
rock, while hornblende nearly completes the remaining third. The biotite is brown,
well crystallised with numerous pleochroic spots. Epidote is only rarely associated
with the biotite. The hornblende is intergrown in parallel position with the biotite,
and cross sections are idioblastic against the biotite. It has a more pronounced
prismatic habit than in the normal amphibolites. Its colour is different and appears
to follow the scheme X very pale yellowish green, Y green, Z bluish green. The
colour is not so intense as usual, indicating less iron in its composition. The hornblende
only rarely contains inclusions of quartz, apatite, or sphene. A small amount of biotite
is replaced by colourless muscovite^ and occasionally the biotite is intergrown with
lawsonite. Quartz is irregularly distributed, but rather seems to concentrate in the
axes of the miniature folds.

THE METAMORPHIC ROCKS OF ADELIE LAND. 8TILLWELL. 59

The chemical composition of this rock is found by J. C. Watson to be

Si0 2 43-12

A1 2 0, 12-74

Fe 2 8 1-35

FeO 10-14

MgO 17-13

CaO 4-70

Na 2 0-26

K 2 6-08

H 2 + 3-07

H 2 - 0-02

C0 2 nil

Ti0 2 1-35

P 2 5 trace

SO S nil

Cl 0-06

MnO 0-13

NiO, CoO trace

CoO p.

Li 2 nil

Total . 100-15

Sp. Gr. at 4 C 3-012

If this composition is compared with that of the normal amphibolite, No. 629,
strong points of difference are noticed, and these correspond with the mineralogical
differences. There is 5 per cent, less silica, but the most striking differences are found
in the percentages of CaO, MgO, and K 2 0, and NajO. The amount of K 2 is more
than seven times greater, and the Na 2 about seven times smaller. There is two and
a half times as much MgO, while the CaO has decreased by a half. The large percentages
of MgO and K 2 correspond with the high percentage of biotite. There is no important
difference in the alumina, total iron, or titanium.

The Ozann group values are S = 4-60, A = 4-3, C = 3-5, F = 37-4, M = 1-8,
T = 0, K = -6.

The projection values are a = 1-9, c = 1-5, / = 16-6.

These values, considered collectively, place the rock among the magnesium silicate
schists, Group V., though the high value of A is exceptional for this group. The position
in the triangular diagram is shown in fig. 6.

The composition of these biotite hornblende schlieren seems impossible for any
primary magma tic xenolith that can be postulated. Neither augite nor olivine can

60 AUSTRALASIAN ANTARCTIC EXPEDITION.

yield the high potash percentage of biotite, and, if we postulate sufficient felspar to
supply the alkali, there would be insufficient magnesium or iron for the hornblende
and biotite. If we ignore this difficulty and still assume an augite xenolith, we raise
further difficulty in recalling that hornblende, or biotite with epidote, is the normal
metamorphic equivalent of augite in this series. The amount of epidote in this schliere

is scarcely appreciable.

AAAAAAA/VVXAAAAAAA

AAAAAAAAAAAAAA

A A AA/ \A A AAA AAAA

\AAAAA / VWXA AAAAA

Fig. 6.

629. Amphibolite, Cape Denison.
640. Chlorite schist, Cape Denison.
415. Epidosite, Cape Denison.
4. Biotite hornblende schist, Cape Denison.

The original dolerite may have contained some biotite and, therefore, it might be
conceived as possible that the schlieren are the metamorphosed equivalent of primary
segregations composed of two-thirds biotite and one-third augite. Such would be a
very extraordinary xenolith, and I am not aware that we have any information of such
a type of cognate xenolith. Here it is quite apt to remark that it is fundamentally
wrong to insist on explaining curious metamorphic features by reference to abnormalities
in the primary rock, igneous or sedimentary. The biotite and the hornblende throughout
the series have been developed during the metamorphism, and it is quite reasonable
to view these schlieren as true metamorphic products. The beautiful parallel arrange-
ment of the biotite and the hornblende strongly suggests that this rock owes its origin
to the metamorphism and nothing else.

A biotite schliere is recorded in the band from which specimen No. 630 was collected
as the normal rock of the band. The schliere occurred in a broad bulge 12ft. or 15ft. wide,

THE METAMORPHIC ROCKS OF ADELIE LAND 8TILLWELL. 61

and No. 630 was picked up not more than 2yds. or 3yds. away from it. Unfortunately
there is no example of this schliere in the rock collection, but it is quite certain from ite
soft character that it contained a large percentage of biotite. Another specimen (No.
630A), however, was obtained from this spot which is also of curious composition. It
is coated with black biotite, but it is hard and contains a good deal of felspar. The rock
is not of such even composition as the biotite hornblende schist. The felspar occasionally
appears as a porphyroblast, or tends to aggregate and form lighter coloured patches.
The mineral composition of the ground mass of this rock, determined in a section cut
at right angles to the schistosity, is

Biotite 44-4

Felspar 51-8

Epidote 2-8

Apatite 1-0

Sphene and magnetite present but less than !.

The rock is, therefore, essentially an aggregate of biotite and felspar. The specimen
shows a certain amount of mechanical deformation, but this again is subsequent to
the development of the biotite and felspar. Some of the biotite is twisted and shows
attrition, while some of the felspar is granulated. The felspar is andesine, and a portion
is saussuritised. This saussuritisation may have developed in the subsequent crushing.
The epidote crystals commonly contain a core of allanite. Quartz is absent. The
mineral composition bears some resemblance to that of the band No. 153 (Table I., No. 1).
In this case there is no hornblende or sphene and less epidote but more felspar and
biotite. We can, therefore, picture its chemical composition with more silica and
alkalies and less FeO, MgO, and CaO than No. 153 (p. 21). It is thus certain that the
composition of No. 630A, as well as the composition of the biotite hornblende rock No. 4,
differs considerably from that of the normal amphibolite. The conclusion is unavoidable
that there has been a rearrangement of chemical composition during metamorphism.

The composition of No. 630, the supposed normal rock of this band, is also
abnormal. Table I., No. 6, shows that the colourless constituents in No. 630 are double
those in No. 412, an example to which it is otherwise strikingly similar. This large
excess in No. 630 is due to the numerous grains of clear, uncrushed quartz, a mineral
which is nearly absent in all the normal bands. The microscopical structure of this
quartz is essentially different from the quartz in the adjacent granodiorite gneiss.
Whereas the latter shows abundant cataclasis, strain polarisation, and participation
in the mortar structure, the quartz in the amphibolite is perfectly clear, uncrushed,
and with strain polarisation weak or absent. In other respects the minerals in No.
630 are similar to the minerals in the biotite amphibolite, No. 412. The abnormal
percentage of quartz seems to me to be connected with the abnormal formations of
the biotite hornblende schliere and the biotite felspar rock. From evidence which will
be given later, we might look upon the biotite felspar rock as a metamorphic hybrid
produced by the intermingling of gneiss and amphibolite in the solid state, because a

62 AUSTRALASIAN ANTARCTIC EXPEDITION.

fragment of gneiss may readily have been caught up in the injection of the dyke. But
in the case of the biotite hornblende rock we must picture during metamorphism a
transference of material which results in the formation of segregations within the dykes.
The formation of a segregation is equivalent to a differentiation in situ, which we
propose to refer to as " metamorphic differentiation."

On such a hypothesis we find a ready explanation for the schlieren of chlorite and
epidosite. These two schlieren occurred in the same broad outcrop from which No. 629
and the meta-xenoliths were collected. In size they are less than 2ft. in their longest
direction. It has been observed from Table I., No. 7, that No. 629 is abnormally low
in mica, and this fact can be correlated on this hypothesis with the observed segregation
of chlorite. Microscopical examination shows that the chlorite rock is composed
entirely of chlorite except for a few very minute grains of magnetite and quartz. The
chlorite is green in colour, with very low polarisation colour, but it does not show the
blue interference colour common with penninite. In contrast to the biotite hornblende
schlieren, the chlorite rock has an approximately massive structure like its host.
Further, it is to be noted that where the dominant mica is biotite in the No. 630 band,
the mica schliere is composed of biotite. In No. 629, where the dominant mica is
chlorite, the mica schliere is composed of chlorite, yet the outcrops of Nos. 630 and
629 are less than 40yds. apart.

The schliere of epidosite occurs 2yds. away from the schliere of chlorite. Its shape
tended to be rounded and, like the previous schlieren, no boundaries against the
amphibolite were observed. The hand specimens of the epidosite are massive, and the
mineral composition of a thin section is

Felspar 28-4

Epidote 65-1

Hornblende 2-5

Sphene 3-8

Iron ore 0-2

Biotite, chlorite, and apatite are present.

The thin section is illustrated on Plate II., fig. 5. The proportion of felspar is very
close to the felspar percentage (27-3) of the amphibolite host No. 629, and its character
is the same. The hornblende of the amphibolite is almost completely replaced by
epidote in this rock. The epidote is well crystallised, has well-developed cleavage,
and its characteristic pleochroism. It may contain inclusions of ragged hornblende,
and it can also be observed replacing relic hornblende crystals. The transition is almost
complete, but examples can be found where irregular remnants of hornblende with
optical continuity are scattered through an epidote crystal. If cleavage be observed
in one relic fragment, it is also observed in the associated group. The hornblende
possesses a stronger bluish-green colour than in the normal amphibolite, and cross
sections still exert their form against the epidote. Sphene is very prominent, and large

THE METAMORPHIC ROCKS OF ADELIE LAND. STILLWKU.

crystals may be included in the epidote. Rarely fragments of biotite with alteration
to chlorite are found in the epidote. The iron ore consists of magnetite with alteration
to hematite.

The results of the analyses of the chlorite rock and the epidosite made by J. C..
Watson in the Victorian Geological Survey Laboratory are as follows :

Si0 2

I.
24-96

II.

45-49

III.
25-40

ALO. .

20-76

19-50

22-80

Fe,0 3

3-24

9-13

2-86

FeO

21-86

0-64

17-77

McO

18-18

0-45

19-09

CaO

nil

16-88

nil

Na 2

nil

2-66

nil

K 2

nil

0-30

nil

H 2 + ....
H 2 - ....
CO,

11-45
0-19
nil

0-08
1-25
nil

12-21

Ti0 2

0-20

2-29

P 2 5

nil

0-87

so s

tr.

0-02

MnO . . .

0-05

0-25

NiO.CoO...
CoO

tr.
nil

0-02
nil

Ld,0..

nil

st. tr.

. . tr.

Total ....

100-89

99-63

. . 100-38

Sp. Gr. ..

2-938

3-118

2-835

Group Values.

29-4

66-3

0-5

0-20

n. .

53-9

3-2

9-9

19-9

10-8

1-9

6-0

12-1

Projection Values.

I. Chlorite rock. Cape Denison.
II. Epidosite. Cape Denison.
Ill Chlorite. Washington, D.C. " Rock Minerals" Iddings, p 472.

64 AUSTRALASIAN ANTARCTIC EXPEDITION.

Both these analyses are again very different from that of the amphibolite host.
The analysis of the chlorite is very close to that of a pure chlorite, as is seen
by comparison with the analysis of a prochlorite quoted from Idding's " Rock Minerals."
Its group values place it among the chlorite schists, Group V., though the projection
values do not separate the rock from the magnetite schists of Group XL

The very high values of FeO and MgO in the chlorite rock are notable in
comparison with the very low values in the epidosite, while the reverse is true with
regard to CaO. The total lime and magnesia is practically the same in the epidosite
and in the amphibolite, No. 629, and not much different to the magnesia percentage
in the chlorite rock. The total alkalies in the epidosite are also approximately the same
as in the amphibolite, with a large excess of soda in both cases. The latter point
corresponds with the observed fact that the amount of felspar is the same in both rocks,
and that the formation of the epidosite occurs with the replacement of hornblende
by epidote. There is also a notable increase of titanium in the epidosite, corresponding
to the increased percentage of sphene in the epidosite.

All the group values of the epidosite, except M, agree with those of Group IX.,
the lime silicate rocks. But though the value of M is below the stated limits for this
group, there can be no doubt that this epidosite should be included in the group of
epidosites which appear in the epi division 'of Group IX.

The projection values of these two rocks are plotted in fig. 6, and it is to be noticed
that they fall symmetrically on either side of the position of No. 629.

Hence from the microscopical and chemical study of these rocks we consider that
the epidosite has been derived from the amphibolite during the recrystallisation, and not
from a pre-existing magma clot. The same is no doubt true of the chlorite rock, and
the conclusion is again forced upon us that there has been chemical migration and
rearrangement during metamorphism. It is the type of exchange that we intend to
refer to as metamorphic differentiation.

Geological literature provides many examples where epidosites have been observed
in association with amphibolites or hornblende schists. In one instance in the Lizard
area Flett has supposed them * to be due to chemical segregation during metamorphism,
and our conclusion is a similar one.

If one still urges that the biotite hornblende schlieren may be the result of meta-
morphism of a primary igneous xenolith in a dolerite dyke he is now confronted with the
difficulty of explaining why the schlieren have the composition of biotite hornblende
in one place, of biotite felspar in a second, of chlorite in a third, and of epidosite in a
fourth. Finally he must explain why these four types of schlieren appear as primary
metamorphic products, and yet are all essentially different from the relics of the
primary cognate and accidental xenoliths that have already been described from the
same outcrop of No. 629.

* " Geology of the Lizard and Meneage," Flett & Hill, p. 50.

THE MBTAMORPHIC ROCKS OF ADELIE LAND. 8TILLWELL. 65

We cannot at present indicate the conditions which permit metamorphic
differentiation in localised portions of certain bands, while the majority of bands remain
undifferentiated. The conditions may possibly arise from some combination of solid
diffusion with that force of crystallisation of the specific mineral which determines
its position in the crystalloblastic order. If it be objected that solid diffusion is a process
of much too limited range, and of too infinitesimal a rate, then it must be remembered
that the whole record of geological time is available. The presence of water may be
an assisting factor it is necessary material in the formation of epidote and chlorite
but we cannot assume mere migration by solution while we insist that the formations
occurred under the influence of strong stress. We can only surmise that the points of
metamorphic differentiation have in some way been the focus of special conditions of
stress or uniform pressure which, combined with other special mineral forming
conditions, have caused an abnormal development of that special mineral. A condition
of relatively low hydrostatic pressure and stress might, on the one hand, permit the more
ready transfer of molecules ; but, on the other hand, a condition of low hydrostatic
pressure and strong stress might favor the process of solid diffusion.

Metamorphic diffusion and differentiation are essentially processes of limited range.
They occur in a rock which is, to all intents and purposes, solid, and molecular movement
is hindered. Their products can never attain the dimensions of the products of
magmatic differentiation.

9. DESCRIPTION OP THE COARSELY CRYSTALLINE BASIC PATCHES IN THE

GRANODIORITE GNEISS.

Apart from the well-defined series of metamorphosed dykes that have just been
described, there exist a number of outcrops of hornblendic rock whose origin has only
become evident on investigation. In the field the dykes are distinct in that they have
maintained their sharp junctions and their linear trend, even though their surfact: out-
crop may be broken. The hornblendic rocks now under consideration present a contrast
and have scarcely any definite shape, and appear as irregular dark-coloured clots in the
grey gneiss. They possess a rough lenticular outline and tail out in the direction of
foliation, but the boundaries may be indefinite when the dark rock passes gradually
out into the grey rock. These dark rocks are often characterised by a uniformly coarser
grain and the average diameter of the mineral grains may reach one and a half times
that in the normal amphibolites. The rock type is not constant, and one may find
patches of almost pure hornblende rock, or massive amphibolite or hornblende and biotite
gneisses, which may pass through varying stages into the normal granodiorite gneiss.
Sometimes one can macroscopically distinguish brown sphene crystals up to |in. long
as well as pyrite or magnetite.

The indefinite boundary, the coarse granularity, and a relatively massive texture
suggested in the field that they would yield evidence of primary consolidation of the
same nature as the granitic rock. Later study, alone, has shown that the coarse

SeriM A, VoL m.. Prt 1 B

AUSTRALASIAN ANTARCTIC EXPEDITION.

granularity has been produced by secondary or metamorphic crystallisation, and that
they are indeed a part of the dyke series. Hence a revised study of the field relations
would have been profitable had circumstances permitted it. Small examples of basic
schlieren can be seen in the illustrations of polished rock (Plates XV1IL, fig. 1 ; XXII.,
fig. 2).

Petrographical Characters.

We deal in detail with three examples of this type which bear the field numbers
of 9, 13, and 10. No. 9 belongs to the massive type, and Nos. 13 and 10 to the schistose
types. These specimens were collected from hornblendic patches which passed by
transition into normal gneiss. Rosiwal measurements have been made of thin sections
of these rocks in the same manner as before, with the following results : Columns
13A and 10A are the recalculated compositions of Nos. 13 and 10 when the quartz has
been disregarded.

No. 9.

No. 13.

No. ISA.

No. 10.

No. 10A.

Quartz

29-2

23-4

Felspar

23-1

22-2

31-4

34-3

44-8

Mica

10-7

15-0

21-1

32-2

42-0

Hornblende

61-8

31-9

45-1

Epidote . .

1-8

1-7

2-4

7-6

10-0

Sphene. . ...

2-6

0-8

1-0

Iron Ore

0-7

0-9

Apatite

1-0

1-3

No. 9. The specimen was collected near the magnetograph house. It is dark,
massive, coarse grained, showing abundant platy hornblende and dull felspar. Grains
of pyrite and sphene are occasionally seen.

In thin section the rock is coarse and granoblastic. The average absolute grain
size of the hornblende is approximately l-5mm. ; but in other specimens from the
same locality the hornblende crystals are as much as 4mm. and 5mm. broad. Horn-
blende, which forms nearly two-thirds of the rock, is found in granular crystals without
terminal faces. The prism faces and cleavage are well developed as usual. Its colour
scheme is X greenish yellow, Y bright green, Z bluish green. It contains abundant
inclusions of biotite, sphene, ilmenite, and epidote. Parallel strings of small sphene
inclusions are common in sections parallel to the cleavage.

The 23-1 per cent, of felspathic material forms the colourless constituents of the
rock, and consists partly of turbid saussuritised felspar and partly of clear felspar. The
saussurite yields a brightly polarising aggregate which, under close examination, opens
up into mica, epidote, chlorite, and clear felspar. There are no traces of cataclasis.
The clear felspar is less in amount than half the total felspar. Part is untwinned and

THE iMETAMORPHIC ROCKS OF ADELIE LAND. 8TILLWELL. 67

part possesses both albite and pericline types of lamellar twinning. A simple twin
with lamallse in both halves was found to give extinction angles of 14 in one set and 15
in the other set. Hence we designate the clear felspar albite.

The biotite is well formed and often appears as inclusions in the hornblende. It
is mainly the normal brown biotite, but it is often intergrown with a green biotite.
Some of the green appears to be chlorite, and normal chlorite with its low, anomalous
blue polarisation colour is present in the section. The brown biotite seems to be
developing from the green biotite, which in turn comes from the biotite. The biotite
and chlorite appear under one head in the quantitative statement, as it is not always
possible to assert the line of demarcation. Sphene is relatively abundant, and some
grains are comparable in size with the hornblende, while others are* minute inclusions
in the hornblende. Most grains are anhedral, and only a few possess the characteristic
wedge-shaped outline. Pleochroism is strong in thick sections, and many of the
crystals possess a magnetite nucleus. As almost all the magnetite occurs in this manner,
the sphene-magnetite individuals were treated together in the Rosiwal measurement.
In some cases the rim of sphene is made up of a number of sphene grains with different
optical orientation. Some of the iron ore is ilmenite, as it is associated with its whitish
alteration product, leucoxene, but when the leucoxene recrystallises as sphene, the
ilmenite may change to magnetite. Occasionally there is a reddish-brown mineral
which is taken to be rutile. Epidote appears in colourless or honey-yellow pleochroic
grains. It is sometimes included in the hornblende, sometimes interlaminated with
biotite, and sometimes found as individual grains surrounded by the felspar
decomposition products. Pyrite in small scattered cubes and apatite are present.

The rock is thus seen to correspond very closely with the description of a typical
amphibolite of the dyke series. The same type of hornblende, the same saussuritised
felspar, and the same clear felspar, and also the same peculiarities of the mica are found
in both cases. The quantitative expression of the mineral composition is now valuable
for comparison with the mineral compositions of the amphibolite dykes in Table I.
The composition of No. 9 is quoted in Table I. for this purpose. The strong similarity
towards types like Nos. 631 and 635 becomes obvious, and as the texture of No. 9 is
approximately massive, any error due to the schistosity is very small. The felspar
content is the smallest of the series, but only by a very small amount ; but the
hornblende percentage is the same as a normal amphibolite, and so also is the mica.
The mineral composition is therefore quantitatively as well as qualitatively essentially
the same as a normal amphibolite with dominant hornblende produced in the
metamorphism of a dolerite dyke. The chemical composition must also be the same.
Specimen No. 9 is therefore identical in kind with the examples of the undoubted dykes
series, and the conclusion is unavoidable that it is part of the same series. The apparent
difference is due to the fact that secondary crystallisation has proceeded under more
favorable circumstances and larger crystals have been formed. This larger granularity
signifies nothing in primary origin.

68 AUSTEALASIAN ANTARCTIC EXPEDITION.

No. 13. Specimen No. 13 is a transition type, and was collected from a similar
outcrop to No. 9, but a linear trend was more noticeable. Its boundary with the gneiss
is indefinite. It is a rock with much the same granularity as the typical granodiorite
gneiss, and this is larger that that of the average amphibolite. Its colour is intermediate
between the black amphibolite and the grey gneiss. A coarse crystallisation
schistosity is rendered prominent by dark bands of hornblende and white bands of
felspar and quartz. This schistosity, of course, reduces the accuracy of the statement
of the mineral composition in Table II.

In thin section we find the schistose bands consist of quartz, of hornblende with
biotite and epidote, and of saussuritic aggregates. The hornblende is identical in type
to that of No. 9. Chloritisation of the hornblende is not uncommon. Biotite is
associated with the hornblende bands, and is found with both a green and a brown color,
interlaminated together as before. The brown is more abundant than the green,
which is again an intermediate stage between green chlorite and brown biotite. Epidote
is very frequently associated with biotite, and is illustrative of a previous conclusion that
biotite and epidote are equivalent zonal products of hornblende, and that the biotite
appears when there is a supply of potash. The epidote expressed in the quantitative
statement in Table II. is that amount of epidote which occurs in this association. A
larger amount of epidote appears among the cloudy saussurite, and has been included
therein in the measurement.

The felspar percentage expresses the amount of saussuritic aggregates which
include all the cloudy material under the low power objective. Some of the cloudy
parts remain dense and unresolvable. Part, however, can be resolved into epidote
and a colourless well-formed mica, which is possibly paragonite. Most of this epidote
is in fine granular aggregates. Clinozoisite or zoisite is also present. The colourless
mica shows strong absorption, and is similar in appearance to muscovite. Rough
measurement has indicated that it forms at least one-ninth (|) of the saussuritic aggre-
gates, and since the primary felspar is here as in previous cases a calcic plagioclase,
we cannot refer it to a potash mica without providing a source for the potash and a
means of escape for the soda. The aggregates consist chiefly of epidote and colourless
mica, with some chlorite and biotite. No secondary clear albite has been determined
with certainty, and quartz grains appear among the cloudy masses, and hence the
colourless mica may have absorbed the soda from the felspar. The percentage of
biotite is considerable, and this means an absorption of considerable potash. It is
reasonable, therefore, to strongly suspect the presence of paragonite mica.

Quartz is abundant, and provides the chief distinguishing feature from the typical
amphibolites. Entering as it does into the crystallisation schistosity it cannot be
looked upon as a quartz-veining subsequent to those processes which impressed the rock
with the individuality of the schist. It is as essentially part of the schist as the
hornblende layers or the saussurite layers. It is clear, and the larger grains invariably

THE MBTAMORPHIC ROCKS OF ADBLIE LAND. STILLWELL. 69

show strain polarisation. The grains are interlocking, and not infrequently possess
a lenticular shape due to solution at the points of greatest pressure and simultaneous
deposition at points of minimum pressure. It is quite different in character to the
quartz in the granodiorite gneiss. Apatite is an accessory mineral. Reddish hematite
occurs among the saussurite, but grains of magnetite are scarce. The rock may be
named an hornblende gneiss.

The quantitative expression of this mineral composition in Table II. emphasises
the difference between No. 13 and No. 9. Marked as this difference is, the microscopical
description brings forward points of resemblance. The hornblende and mica are similar
in both cases, and the saussuritised felspar is quite abundant considering the high
silica percentage of the rock. If we neglect the quartz and recalculate the mineral
composition we obtain 31-4 per cent, felspar, 21-1 per cent mica, 45-1 per cent, horn-
blende, and 2-4 per cent, epidote. Then we find that the proportion of felspar (saussurite)
to the ferromagnesian (hornblende and mica) is very similar to that of the typical
amphibolites. Yet, in appearance and in the abundant quartz it possesses some likeness
to the granodiorite gneiss. The examination of this rock, therefore, provides micro-
scopical evidence to support the field observation that there is a gradual passage from
this basic patch into the enveloping granodiorite gneiss.

No. 10. Specimen No. 10 is a different type collected from the same small area
as No. 13. Glistening biotite is abundant on the cleavage surfaces, but nevertheless
the rock has a tendency to a massive texture as a result of its very fine-grained character.

In thin section the rock consists of biotite, saussuritic felspar, clear felspar, quartz,
epidote, apatite, magnetite, and zircon. The biotite is the pale brown variety and has
a noticeable parallel arrangement. Scattered patches of chlorite may be found which
are often accompanied by iron ore. The felspar consists of twinned and untwinned
felspar and cloudy saussurite. The quartz shows considerable cataclasis and strain
polarisation effects. It appears as parallel layers in the section conformable with the
layers of saussurite and biotite. The grains are clear except for occasional inclusions
of apatite, and there is a noticeable absence of the linear inclusions that appear in the
quartz of the granodiorite gneiss. Epidote is relatively abundant and sometimes forms
large well-shaped individuals and sometimes it is finely granulated. It is frequently
included in the biotite, but the finely granular epidote may form a rim to a biotite
crystal. Sphene and apatite are accessory minerals and ilmenite is present in occasional
large crystals. Hornblende is absent. The rock may be described as a biotite gneiss.

The quantitative mineral composition is expressed in Table II. In the large
percentage of quartz it resembles No. 13 ; but this percentage is lower than that of
No. 13. It is not expected that these percentages would show any other similarity
than correspondence between two extremes. The ferromagnesian total, however, is
not much different. We notice again the sympathetic variation of the percentages of
biotite and epidote.

70 AU8TEALASIAN ANTARCTIC EXPEDITION.

If we assume, as in the preceding case, that silica is the chief mineral addition
to the original dyke rock, and the mineral composition be recalculated to 100 per cent,
after neglecting the quartz, we obtain the figures in column 10A. These figures bear
some resemblance to the composition of No. 153, in Table I. The proportion of felspar
to ferromagnesian is much the same in both cases ; but the felspar of No. 10 is nearly
all saussurite, whereas the felspar of No. 153 is perfectly clear. No. 10 thus appears
to be related to the epidote biotite schists in the same way that the hornblende gneiss
No. 13 is related to the normal amphibolites.

10. ORIGIN OF THE COARSELY CRYSTALLINE BASIC PATCHES.

The origin of these dark hornblendic and biotitic rocks which are enveloped in
the granodiorite gneiss, and which have been designated the coarse-grained types,
is a very interesting question. It has been shown that the massive amphibolite from
these patches is identical, except for larger grain size, with the amphibolites which have
been established as altered dykes. It is also plain that there is true transition from
this amphibolite through hornblende gneiss or biotite gneiss to the granodiorite gneiss.
Accepting the face value of these gradual transitions, we might say that these " basic "
patches have been derived out of the granodiorite itself. We might conceive of a
magmatic differentiation which was initiated in the granodiorite magma which became
frozen before the differentiation process was complete. A sudden cessation of the
differentiation forces has left a gradual apparent transition between the amphibolite
and the granodiorite. Such an argument completely ignores the observed similarity
of the textural, structural, mineralogical, and chemical relations of related rock types
at Cape Denison, and at the same time we miss the recognition of a true metamorphic
process.

Since the coarse-grained type is so precisely similar to the amphibolite dyke series
which has been proved to be the metamorphosed equivalent of diabasic dykes, it is
extremely likely that the No. 9 type of amphibolite has been derived from a primary
rock of similar nature. They occur approximately along the extension of well defined
dykes, and hence it is extremely likely, and as definite as it is possible to be, that the
primary rock of the No. 9 type was part of the intruded series of dykes. We have
described fragments of the dyke series proper which have been torn away from the
dyke channel and now appear completely enclosed in the gneiss. Discontinuity of the
hornblendic clots, irregularity or isolation are, therefore, matters of little weight. These
detached fragments of the established dykes have escaped the more intense meta-
morphism which produced the larger grain size of No. 9, and they have been able to
preserve their sharp outline against the gneiss.

There is no special reason, however, why a pre-existing junction between two rock
types must be preserved during metamorphism. We have maintained in our hypothesis
of metamorphic differentiation that a limited migration may occur in the solid rock

THE METAMORPH1C ROCKS OF ADELIE LAND. ST1LLWELL. 71

under special conditions of metamorphism. If such migration occurs across a pre-
existing junction there must, a priori, be a strong tendency to efface that junction.
If we imagine a diffusion of some of the amphibolite material into the granodiorite
gneiss, or some of the gneissic material into the amphibolite, we would get the former
junction replaced by the gradual transition observed. The transition types would
be mixtures of amphibolites and granodiorite gneiss, and would correspond to the types
No. 13 and No. 10. Such a theory is in agreement with the observations, and we will
speak of the process, for convenience, as metamorphic diffusion. Diffusion products,
like the hornblende and biotite gneisses, are, therefore, looked upon as metamorphic
hybrid rocks.

Solid diffusion has been suggested before to account for the perfectly gradual
passage of granitoid rocks into surrounding schists. Greenly* endeavoured to compare
such phenomena with the laboratory experiments of Roberts-Austen on the diffusion
of gold into lead. Greenly, however, postulates a mixing of a granite magma and the
neighbouring sedimentary rocks, a conception which involves not true solid diffusion, but
merely a mechanical percolation of the surrounding schists by highly fluid magma. Deschf
therefore pointed out that the term " diffusion " had been loosely employed. Another
claim for solid diffusion is mentioned by ElsdenJ in the observations of Trener on the
contact phenomena of Cima d'Asta, but the same objection again holds. While it
has been usual in these cases to suppose that the mixing takes place at the time of
intrusion, I do not know of evidence to show that a degree of mixing has not occurred
after complete consolidation ; and, if this is so, these cases may be examples of solid
diffusion.

The difficulty lies in the proof of the solid nature of the rocks before the mixing.
At Cape Denison the granodiorite must have been solid before it could be fractured
and penetrated by the primary dolerite dykes ; and the presence of the meta-xenoliths
in the amphibolites indicates the consolidation of the dykes before their metamorphism.
Further, we cannot suppose that a thin sheet of dyke magma would remain fluid for
a sufficient length of time to permit the mechanical percolation that is possible in the
case of a large, deep-seated, slowly-cooling plutonic mass. Hence at Cape Denison we
consider that solid diffusion, in the strict sense of the term, has operated. The term
" metamorphic diffusion " implies that diffusion has occurred in the solid state.

Metamorphic diffusion is not restricted to the amphibolite gneiss junction at Cape
Denison. It also appears along the junction of the aplitic gneisses with the granodiorite
gneiss. The examples quoted tend to show that quartz is a mineral that is readily
diffused, but other mineral molecules like hornblende and biotite can be so transferred.

Basic segregations are common in many granitic masses and may be relatively
rich in either biotite or hornblende. If these were recrystalh'sed under conditions

" Diffusion of Granite into Schuta," Greenly, GeoL Mg., vol. 10, dec. 4, N.S., p. 207.
t "Report on Diffusion in Solid*," C. H. Detoh, Brit. AM. Report (Dundee, 1912), p. 348.
t " Principle* of Chemical Geology," J. V. EUden, p. 2.

72 AUSTRALASIAN ANTARCTIC EXPEDITION.

permitting metamorphic diffusion it may be imagined that hybrid rocks, similar to the
hornblende and biotite gneisses, might arise. Consequently such types may possibly
be discovered in isolated masses in regions where there are no traces of the existence
of dykes. At Cape Denison, however, the evidence seems clear that they are connected
with dykes.

It will be subsequently shown that the most intricate dyke structures can be
preserved during the metamorphism at Cape Gray, where the recrystallisation has
occurred under kata zone conditions, in which the pressure factor is chiefly hydrostatic.
At Cape Denison the pressure factor in the metamorphism is chiefly stress, and hence
the destruction of the dyke structures and the migration of material is to be connected
with the dominating stress.

11. FURTHER EXAMPLES OP METAMORPHIC DIFFUSION.
Junction Specimens.

The above interpretation of the biotite gneisses as metamorphic hybrid rocks,
produced by an intermingling of two diverse rock types by solid diffusion, is upheld
by the examination of specimen No. 372, found on the moraines at Cape Denison. This
specimen was collected as a diagrammatic example of the normal " sharp " junction
between the amphibolite and the gneiss. One-half of the specimen (Plate XII., fig. 5)
is black amphibolite, and the other half is grey granitic gneiss. The junction, however,
is not sharp, and how far this applies to all the dyke junctions at Cape Denison is not
known. As the specimen was not found in situ it is not possible to assert that the
amphibolite represents a portion of a dyke originally intrusive into the granitic gneiss.
But as both the amphibolite and the gneiss are analogous to specimens found in situ
it is very probable that such is the case.

In a section of the granitic portion of the specimen it is found that the cataclasis,
so marked in most of the typical granodiorite gneisses at Cape Denison, is absent. The
epi zone metamorphism, however, is signified by the amount of saussuritised or
sericitised felspar and by the chloritisation of the biotite. The amount of
ferromagnesian minerals is less than in the typical example No. 11, but it is not
noticeably less than in other examples from Cape Denison. In addition to the cloudy
felspar there is a considerable quantity of clear, recrystallised felspar which, with the
quartz, possesses the crystalloblastic structure. Some of the clear felspar is untwinned,
but some of the twinned crystals have been determined to be oligoclase-andesine. The
biotite is brown, but is largely altered to green chlorite. There is in addition scattered
epidote, allanite, lawsonite, pyrite, and apatite. The general characters and the
composition of the felspar indicate the relation to the granodiorite gneisses.

A section across the junction reveals the presence of a zone of biotite gneiss,
approximately 1 c.m. wide, between the amphibolite and the granodiorite gneiss. The

THE METAMORPHIC ROCKS OF ADELIE LAND. 8TILLWELL. 73

transition from the latter into the biotite gneiss is fairly rapid. There is no variation
in grain size or in structure or configuration of the crystal grains across the apparent
junction. The zone of biotite gneiss consists of abundant brown biotite set among
grains of clear felspar, cloudy felspar, and quartz. The character of the plagioclase
is the same as in the gneiss. There is less quartz in the biotite zone than in the grey
gneiss, and plagioclase occupies a greater percentage of the colourless material. Epidote
is practically absent.

The transition from the zone of biotite gneiss to amphibolite is more gradual than
its passage into the granodiorite gneiss. In the amphibolite the hornblende largely
replaces the biotite, but the relative amount of biotite is probably sufficient to call
the rock biotite amphibolite. The hornblende has, on the average, a larger grain size
than the biotite, but its character is quite similar to the hornblende in the amphibolite
dykes. Quartz still appears in small pieces in the hornblende area ; but by far the
greater portion of the colourless constituents consists of saussuritised felspar. A little
lawsonite is intergrown with the biotite.

In this case it is perfectly clear that a zone of biotite gneiss has developed along
the contact between granitic gneiss and amphibolite. But it has not been produced
by simple contact metamorphism, nor by assimilation, and we believe that it is another
example of a metamorphic hybrid produced by solid diffusion.

No. 160 is another specimen from the moraines (Plate XII., fig. 6) which is
diagrammatic of the manner in which the black amphibolites cut the granitic gneiss.
In this case the gneiss is more basic than the Cape Denison granodiorite gneiss. Though
there is still abundant quartz in it, labradorite has been identified among the plagioclase
and hornblende is much more abundant than biotite. It may be distinguished as a
hornblende gneiss. The amphibolite consists of hornblende and plagioclase with a little
biotite. Though the granoblastic structure is noticeable on both parts of the rock,
the junction is a line of interlocking crystals and is comparatively sharp.

Composite Gneiss.

We have now to consider areas at Cape Denison in which the amphibolite bands
seem to open out into a series of thin parallel threads interwoven with the granodiorite
gneiss. The boundaries of the threads are often indefinite, so that some doubt existed
in the field as to whether they were related to the amphibolites. These are areas of
composite gneiss. A specimen (No. 144) from one of these thin interwoven bands is a
dark-coloured massive rock, a little coarser than the normal amphibolite. The hand
specimen shows abundant glistening biotite and small felspar porphyroblasts are
distributed through it.

In section it is found to be a crystalline aggregate similar in type to the biotite felspar
gneisses, No. 630A and No. 10. It consists chiefly of biotite and felspar in much the

74 AUSTKALASIAN ANTAECT1C EXPEDITION.

same proportion as in No. 630A. Epidote is often associated with the biotite, and
there is a subordinate amount of quartz; sphene, apatite, magnetite, and pyrite are
present. The felspar is nearly all perfectly clear, and in some cases two sets of lamellar
twinning are beautifully developed, especially in the porphyroblasts. The maximum
extinction angle that has been measured is 16, and the refraction is very close to, but
always less than, nitrobenzol (1-551). It is always above nelkenol (1-542). We,
therefore, consider it to be an oligoclase-andesine. The brown biotite shows very little
change to chlorite.

In the amount of biotite and in the complete absence of hornblende this rock is
similar to the previous transition type, No. 10, and to the biotite zone described in the
junction specimen No. 372. There is much less quartz in this case than in No. 10,
while clear felspar replaces the cloudy felspar. There is more epidote than in the No.
372 example, but it is very likely that this rock has a similar origin and is a product
of metamorphic hybridisation.

There can be no doubt that the clear character of the felspar is due to the meta-
morphism, and the size of the crystals has, therefore, no great significance. There is,
in fact, considerable variation in the size of the felspars from the same locality as No. 144.
In specimen No. 146-1 the average size of the felspar is about equivalent to the
porphyroblasts in No. 144 (Plate XII., figs. 1, 2, and 3), but the mineral composition
of the rock is similar to that of No. 144. In specimen No. 146-2 there is a still greater
development of the felspars ; but in this case there has not been a uniform development
and some crystals are much larger than others. In the hand specimen there is a
suggestion of a brecciated appearance, but in section there is no evidence at all of
crushing or cataclasis. No variation in the constituent minerals in the different
specimens has been noticed, while the specimens were collected within a yard or two
of one another.

Primary relic felspar can usually be recognised, both in the granodiorite gneiss
and in the amphibolites, either by the mechanical alteration or by the saussuritisation.
On this ground alone we would have difficulty in maintaining an igneous origin for
the porphyroblasts. Further, as neither the amphibolite dykes nor the granodiorite
is porphyritic near the particular outcrop, it is impossible to apply Cole's explanation*
of the origin of porphyritic felspars in the biotitic schists, related to amphibolite, which
are associated with a porphyritic granite gneiss at Barna, County Galway. In Cole's
theory the felspar crystals were present in the original granite magma which threaded
and penetrated the surrounding schists, increasing their Si0 2 percentage. The felspar
phenocrysts, however, could not flow away and became stranded, one by one, in parallel
series in the schists. In the Barna granite the large felspars are orthoclase, while in
Cape Denison examples they are plagioclase ; but in both cases the matrix is chiefly
biotite, and the rock is related to amphibolite. In our case the origin of the large
felspars is connected with the origin of the biotite felspar rock.

* " A Composite Gneiss near Barna (County Galway)," G. A. J. Cole, Q. J.G.S. LXXL, 1916, p. 183.

THE METAMORPHIC ROCKS OF ADELIE LAND 8TILIAVELL. 75

Specimens (No. 145) from the same locality are in the rock collection which show
the junction of the dark biotite threads with the grey granitic tongues. A dark line
of demarcation exists between them in some specimens, but no sharp junction exists.
In other cases (Plate XII., fig. 4), where the biotite felspar rock is seen on either side of
a tongue of grey granitic gneiss, the boundaries are quite indefinite.

In sections of No. 145 there is little to indicate a junction. At one end of a section
we may find biotite and felspar, with but little quartz. At the other end there may be
less biotite and much more quartz, but the distribution of these constituents is not
regular. No change in the character of th'e plagioclase is observed throughout the
slide. The presence of large quartz crystals and the larger grain size of the quartzose
areas are the most noticeable features of the grey rock. We are therefore dealing with
a partially obliterated junction, and we can again consider, as in No. 372 and No. 10,
that we are dealing with a metamorphic hybrid product produced by metamorphic
diffusion. The biotite felspar gneisses are related to the amphibolites which have
been produced from an intrusive rock. A supply of potash and silica from the granitic
rock enables the ferromagnesian content to be expressed in biotite instead of in
hornblende, as in the normal amphibolite. As a dyke disappears into the thin sheets
the number of junction planes is considerably increased, and there is more opportunity
for the subsequent interdiffusion of material. With this opportunity there is a greater
development of the biotite felspar schist, and the intrusive features in the field become
correspondingly more indefinite. Hence the thin dark threads which appear in the
field to be connected with the amphibolite bands are so related, but the basic rock has
been modified by metamorphic diffusion. It is not to be assumed that only the
amphibolite undergoes change during the metamorphism. The granodiorite gneiss
may also be modified. It only so happens that the change from hornblende to biotite
is one that can be readily recognised. A change in the granodiorite, which involves
a decrease in silica and in alkalies, is one that cannot be so readily detected.

No. 424. No. 424 is another example of a biotite felspar gneiss, which was obtained
from a schliere of dark rock in the granodiorite gneiss. The schliere is a few yards away
from a definite band. In addition to the brown biotite and clear felspar, the rock,
like No. 10, contains a good deal of quartz. Epidote is also moderately abundant,
while pyrite, allanite, apatite, and sphene are present. In these accessory minerals
there is a likeness to the dark amphibolites. The junction with the enclosing gneiss
is not sharp, and the quartz may again have entered by diffusion. In any such isolated
instance there is always a possibility that such a rock is the metamorphosed equivalent
of a primary basic segregation in the granodiorite ; but against this supposition there is
the symmetrical relation of the schliere to the planes of foliation of the gneiss. Knowing
the relations of other biotite gneisses, one would favour the inclusion of this lenticle
with the amphibolite series.

No. 411. No. 411 is another example of porphyroblastic felspars which are in
this case set in a biotite amphibolite. It was found not more than 100yds. away from

76 AUSTRALASIAN ANTARCTIC EXPEDITION.

the area of composite gneiss from which No. 144 was taken. It occurred close to the
junction plane of the dyke which was cut by a quartz segregation vein carrying large
crystals of epidote. This quartz vein may have carried felspar. The specimen has a
more noticeably brecciated appearance than No. 146, and whereas in the latter the
felspar is white or transparent, it is here pinkish or greenish white, a colour which
indicates saussurite. The outline of the crystals is not definite, and they approach
to lenticles in character. In section there is no special evidence of crushing, and the
large irregular felspar crystals are set in the amphibolite ground mass. The same
minerals are present as in the normal rock of the band, No. 412, though the large felspar
causes a preponderence of the colourless constituents. Some of the felspar is clear,
and approaches andesine in character, but its refractive index is close to, but less than,
1-551 (nitrobenzol). The bulk of the felspar in the section is saussuritised. A few
blebs of quartz are recognised, and there is perhaps a little more chlorite than in No.
412. As in No. 412, lawsonite is present in small amount.

The altered nature of the large felspar in this case prevents the assertion that they
are metamorphic products. They may have been associated with the accompanying
epidote-bearing vein, or they may be allied to the xenoliths of saussurite described
from the band No. 629.

12. FURTHER EXAMPLES OF METAMORPHIC DIFFERENTIATION AT CAPE DENISON.

If the amphibolite patches, which may have either sharp or indistinct boundaries,
are to be included in the dyke series, we immediately find further samples of
metamorphic differentiation. In the description of these coarse-grained types, reference
has been made to the bands and lenses of pure hornblende associated in the field with
them. That these are also part of the original diabasic magma seems evident, because
we only find them in such association. We think, therefore, that we can consider these
patches of hornblende in the same way as we have considered the biotitic, chloritic, and
epidotic clots which are enclosed in the sharply- walled dykes. As the biotite, etc.,
patches are metamorphic differentiation products, so also are the hornblende patches.
In the one case there has been long continued conditions for the formation of biotite,
and in the other case an analogous set of conditions for the formation of hornblende.
It has been noted throughout the series that the hornblende and the biotite appear as
equivalent zonal products, and if we get the differentiation of one we should reasonably
get the differentiation of the other. Indeed, we have already discovered this in the
biotite hornblende clot. It is true in the case of hornblende that its composition may
be similar to a xenolith of pyroxene crystals in the primary magma, or to an ultrabasic
magma, and a hornblende patch may conceivably develop by the metamorphism of
such a primary xenolith. If this were so, we should reasonably expect to find some
such altered xenoliths among the sharply-defined dykes. The distribution, however,
in layers conformable to the schistosity is sure evidence of at least some transference,
and the frequency and variation of shape, combined with symmetry to the plane of

THE METAMORPHIC ROCKS OF ADEL1E LAND -ST 1 1. 1. \\KLL. 77

schistosity, favour an origin by metamorphic differentiation. The neighbouring areas
to the hornblende differentiates are frequently enriched in felspar when the rock assumes
a lighter colour. That this hornblende differentiation only appears in the less distinctive
amphibolite patches simply means that the conditions for the hornblende differentiation
have been accompanied by conditions permitting metamorphic migration on a greater
scale than in the biotite differentiation.

Differentiation seems to have occurred in two other basic clots. In one, No. 928,
there are exceptionally large percentages of sphene and magnetite. In the section
the measured percentage of sphene is 13'1 per cent. The iron is segregated in coarse
crystals and, as only one or two crystals appear in a section, it is impossible to get an
adequate idea of the proportion of magnetite in the rock from a single section. The
magnetite crystals are as much as Jem. broad and are abundant in hand specimen.
In thin section they always possess the normal sphene rim, and in the large crystals
the sphene rim is very thin. The abundant sphene crystals are large and are mostly
without a magnetite nucleus. Some are twinned and some enclose biotite, but are more
often surrounded by biotite. Biotite, felspar, quartz are also present in the rock. The
biotite is the most abundant mineral and absorbs the ferromagnesian content. No
hornblende is present, but there is a small amount of colourless muscovite. The felspar
is fairly evenly distributed through the slide, but clusters of felspar crystals are
noticeable in the hand specimen. The felspar is perfectly clear and colourless, but
some quartz can be recognised. Small apatite crystals are abundant and there are
odd grains of pyrite and epidote.

No. 928 was collected from the eastern side of Cape Denison, but coarse sphene
rocks were also noticed close by the magnetograph house, the locality of No. 9. The
extraordinary sphene content cannot be due to mere chance. The clot must be
considered as part of the dyke series, and it would be very difficult to account for the
high titanium percentage without an appeal to a metamorphic agency. The sphene
and the magnetite are metamorphic minerals, and we can look upon this rock as an
example of metamorphic differentiation wherein both the sphene and the magnetite
contents have been enriched. The abnormal amounts of sphene and magnetite are
reflected in the high specific gravity (3-10).

Specimen No. 143 is another example of a basic clot in which metamorphic
differentiation has occurred. The specimen is rich in magnetite, and the magnetite
crystals stand out prominently on the weathered surface. They are not quite so large
as in the preceding case and can be seen to be distinctly oval in section. The longest
diameter may be 4mm. and the shortest as much as 2mm. Some seem to have crystal
boundaries and others are more rounded. Surrounding each magnetic bleb is a zone
of white felspar, which can be plainly seen in the hand specimen and is noticeable in
the photograph (Plate IX., fig. 6). Separations of magnetite were made in both this
case and the preceding specimen (No. 928). In both cases tests were made to detect

AUSTKALASIAN ANTAECTIC LXPEDITION.

Ti0 2 by the reduction of HC1 solution with tinfoil. A faint trace of the violet colour
was obtained in the sample from No. 928, but no trace at all from No. 143. In both
cases the mineral is highly magnetic, and the magnetite blebs from No. 143 were found,
when suspended by a silk fibre, to possess polarised magnetism.

Fig. 7.

Sketch of a nodule in the amphibolite No. 143. A crystal of magnetite is
surrounded first by a thin rim of sphene and then by a felspar zone.
Crystals of apatite (A) and sphene (S) are distributed through the felspar
zone which passes into normal amphibolite by the gradual appearance of
hornblende (H) and biotite (B).

In thin section the rock is found to consist of hornblende and biotite in about equal
proportions. The same clear felspar is present in the same proportion as in the normal
members of the dyke series. Sphene is again abundant. The crystals are, perhaps,
more numerous than in No. 928 ; but the average size is probably less than a quarter
of that in No. 928. The magnetite blebs are surrounded by a very thin rim of sphene
(fig. 7). Sometimes the blebs tail out a little in the direction of the schistosity. The
felspar zone around the magnetite consists of a granulitic aggregate of clear felspar
whose grain size is the same as the grain size in the normal part of the rock. The kind

THE METAMORPHIC ROCKS OF ADELIE LAND ST1LLWELL. 79

of felspar in the two portions of the rock is precisely the same. The felspar zone is
marked more by the absence of the biotite and hornblende rather than by the felspar
itself. Small crystals of sphene and apatite are present in the felspar zone. Apatite
crystals and small pieces of felspar are also included in the magnetite. With the gradual
increase in biotite and hornblende the felspar zone passes out into the normal biotite
amphibolite.

In this example the rock is obviously part of the metamorphosed dyke series. It
does not seem possible to account for the zonal structure on any primary igneous
hypothesis. The magnetite crystals with a sphene rim are definite metamorphic
products, and the clear felspar is also a product of the recrystallisation. There is,
therefore, no reason to suppose that an association of these two products is anything
else than a metamorphic structure. The formation of this structure in these circum-
stances involves a migration of certain material. It is, in fact, a small differentiation-
magnetite centres have been enriched in magnetite and the biotite and the hornblende
have been repelled from the felspar zone. The process of metamorphic differentiation
in this case has involved the force of crystallisation.

The magnetite nucleus of sphene crystals is a normal feature in most examples of
the amphibolite series at Cape Denison. The Ti0 2 content of the primary ilmenite
has combined with the felspar, producing sphene and hornblende, and it is, therefore,
readily understood why the sphene surrounds the nucleus of magnetite or relic ilmenite.
The relatively large crystals of magnetite with only a thin and often incomplete rim
of sphene are abnormal in amphibolites Nos. 143, 637, and abnormal conditions must
be pictured during their formation. It is certain that the Ti0 2 content of these examples
is not less than in the normal amphibolites, because they possess a high sphene content.
The abnormal conditions have permitted certain magnetite crystals to enlarge them-
selves by attracting smaller magnetite crystals, and diffusion of magnetite, which is
prevented in the normal case by a sphene shell, has occurred. We, therefore, suppose
that the rate of diffusion of the magnetite molecules in these abnormal cases has been
more rapid than the rate of reaction which produces the sphene. When the supply
of magnetite molecules around any one centre has been nearly exhausted the sphene
rim has become attached to the large crystal.

13. REVIEW AND DISCUSSION OF FIELD CHARACTERS.

It is desirable to review the field characters in the light of the dyke origin of the
bands. This metamorphosed series of dykes differs from a normal parallel system
in the frequency and magnitude of the breaks in the surface outcrops. A normal dyke
channel may here and there swell out into local bulges, but the general appearance of
the bulges at Cape Denison, and the sharp, irregular way in which the bulge may
terminate, seem to indicate that the bulges are not normal dyke swellings. In following
the trend of the dyke we find no dyke in many places where we expect dyke, and, in

80 AUSTKALASIAN ANTARCTIC EXPEDITION.

other places, we find more dyke than we might reasonably expect. The field appearance
suggests that the dyke walls have been squeezed together by a pressure of varying
intensity at different points of the dyke plane. Where the pressure has been greatest
the dyke wall might have closed together, and where the pressure has been least the
dyke rock has formed a bulge. The normal width of a dyke is about 2ft., and the width
of the enlarged outcrops is 9ft. or 12ft. At the same time we find the dykes running
out of the thin parallel threads, and there are detached fragments of amphibolite adjacent
to the dyke channel or along its continuation, which are wholly surrounded by the
granodiorite gneiss. As far as observed, the foliation of the granodiorite gneiss bends
around the contour of these isolated fragments. Sometimes they are precisely similar
to the " canoe-shaped infolds " described in other areas.

There seems to be no reasonable alternative but to consider these " inclusions "
as part of the dyke series. They have been shown to be so similar in character to the
normal dyke, and so dissimilar from the granodiorite gneiss. In metamorphic areas,
therefore, caution is necessary before we can assert the younger or the older age of the
enclosing rocks. With our interpretation the " inclusion " is the younger rock the
reverse of the normal igneous or sedimentary deduction.

These abnormal dyke features demand an attempted explanation, especially as
we will subsequently infer that analogous cases may exist in other areas of metamorphic
rocks. One can, perhaps, imagine that branching offshoots of dyke into the adjacent
gneiss might become detached from the main dyke channel during a period of excessive
stress, and so form isolated fragments that lie adjacent and parallel to the main dykes.
Such, however, provides no mental picture of the manner in which the main dyke has
itself been rendered discontinuous.

Possibly there is an analogy with some curious features in the Ordovician rocks
at Daylesford, Victoria, which have been recorded by T. S. Hart*. These Ordovician
sediments are a steeply folded series, and unequal thickening and thinning of slate
beds between sandstone beds is a common feature. The continuity of the slate
beds is often broken. In a railway cutting near Daylesford slate now appears in
numerous pockets of various shapes and sizes in a hard sandstone. At one place the
pockets possess a prominent linear trend which would correspond in position and
direction to a bed of slate. During the process of folding the slate has behaved towards
the sandstone as a relatively plastic rock. The slate bed has had a thickness comparable
in size with the minor irregularities and small displacements of the adjacent rigid
sandstone, and been squeezed out irregularly so that it is now represented by a number
of isolated fragments. The squeezing out of the slate goes so far sometimes as to show
only occasional slate patches along a definite line of junction of two beds.

This, therefore, is the case of a primary band of solid rock that lost its identity
by the play of stresses which have resulted in nothing beyond folded sediments. Could

* " On some Features of the Ordovician Rooks at Daylesford," T. S. Hart, Proc. Roy. Soc. Vic., vol. XIV.,

N.S. pt. II., p. 167.

THE METAMORPHIC ROCKS OF ADELIE LAND -STILL WELL. 81

the disruption of the amphibolite dyke channels at Cape Denison occur in a like
manner ? Can the primary dolerite dyke, under the more intense conditions which
have resulted in the decrystallisation, be considered a relatively plastic rock alongside
the granodiorite ?

In this respect the only experimental data available are not encouraging. Adams
and Coker* have carried out an investigation into the elastic constants of rocks during
which they determined the cubic compressibility (D = ratio of the stress per unit area
to the cubical strain) of five marbles and limestones, six granites and four basic plutonic
rocks. The average of their results is

D (in inch, pound units).

Marbles and limestones 6,345,000

Granites 4,399,000

Basic intrusives 8,308,000

These results show that the granites are much more compressible than the marbles
or the basic intrusives. The experimenters varied one set of readings over a
temperature range of about 30 C. and found no perceptible difference. The actual
case, however, under temperatures which are very high in comparison to living room
temperatures, may be possibly very different.

These results are the reverse of what our proposed analogy would lead us to expect.
Yet we have the fact before us that the impressed conditions were sufficient to cause
the complete recrystallisation of the dolerite, but only a very imperfect recrystallisation
of the granodiorite. In this sense the basic rock has been more susceptible to the
superimposed conditions.

With these experimental data we must picture the basic dyke as a sheet of hard
rock enclosed in a mass of relatively soft rock, viz., the granodiorite, and we must
endeavour to understand what would happen to the system under the influence of great
stress. If the hardness can be associated with brittleness, then, perhaps, we may picture
the fracturing of the brittle sheet and the production of isolated fragments. That such
fracturing actually occurs is shown by the observations of Adams and Barlow in the
Haliburton and Bancroft areas. These authors figure and describe the initial stages
in the disruption of an amphibolite dyke embedded in crystalline limestonef. The
basic rock, on the experimental evidence, is less compressible than the limestone, and
hence the experiments cannot furnish argumentative data against the disruption of a
basic dyke channel in granodiorite.

We find further in the Kylesku to Loch Broom district, in the North-West Scottish
Highlands,^ that basic dykes have been observed to be wrenched into a series of isolated

* " An Investigation into the Elastic Constants of Rocks," F. D. Adams & E. G. Coker. Pub. 46, Carnegie Inst. Wash.,
June, 1906.

f " Geology of the Haliburton and Bancroft Anas," F. D. Adams ft A. E. Barlow, Mem. 6, Can. Geol. Surv., 1910, fig.
G, p. 160. Plates XXIX., XXX.

t " The Geological Structure of the North-West Highlands," Mem. Geol. Surv. Gt. Britain, 1907, p. 169.

Series A, Vol. m., Part 1 F

AUSTRALASIAN ANTARCTIC EXPEDITION.

lenticles or phacoidal masses embedded in a zone of granulitic gneiss. We have,
therefore, some reason to believe that the thin dyke channels of relatively hard rock
have been rendered discontinuous and irregular in localised areas, in some manner
not unlike that pictured in the case of a relatively thin band of soft shale embedded
in sandstone at Daylesford. The present lenticular outline of most of the fragments
can be ascribed wholly to recrystallisation under stress.

I i ' i

M\i

i I \ *

' ; ' /' - \ \ '' \

' / / / * * \ \ \

/ ////^:oA,

\ > \ ^ \ / / ,

i( \ ^ / ' / '

I \ i ' 7

I ' I

1 ' qne/ss

i ^ i i i i

,; ! : ;

i i i 1 1 1

Fig. 8.

DIAGRAMMATIC REPRESENTATION OF THE MANNER IN WHICH THE
FOLIATION OP THE GRANODIORITE GNEISS BENDS AROUND AN
AMPHIBOLITE INCLUSION, AND THE MANNER IN WHICH THE
FOLIATION OF THE AMPHIBOLITE PASSES DIRECTLY THROUGH A
QUARTZ FELSPAR GNEISS INCLUSION.

The manner in which the foliation of the granodiorite gneiss bends around the
contour of the enclosed fragments of amphibolite is also a question inviting comment
(fig. 8). The same kind of observation has been recorded by Cole, Adams, and others
when it has been considered to demonstrate the stream lines of the gneissic flow around
the inclusion which has been carried along like a log in a stream.* At Cape Denison
the diverted foliation must be considered parallel with the foliation in the gneissic
xenoliths embedded in amphibolite. In the latter the foliation of the amphibolite
continues straight through the xenolith, sometimes quite irrespective of its angular
outline. In the first case a block of amphibolite is embedded in a relatively large mass
of granodiorite, and in the second a piece of granitic gneiss is embedded in a relatively

* Op. oit., p. 74.

THE MBTAMORPHIC ROCKS OF ADELIE LAND 8TILLWELL. 83

large mass of amphibolite. The same general metamorphic conditions have been applied
to each, and now the general foliation is diverted by the amphibolite block and not by
the gneissic block. There is some disparity in size between the two typical cases, but
we cannot see that any such disparity can provide adequate explanation.

It has been stated that the gneissic xenoliths seem to show a greater degree of
recrystallisation than the granodiorite or aplite gneisses. This is probably to be
explained by the degree of recrystallisation of the host. We obtain the following data
from Van Hise * : The change of augite to hornblende is exothermic, and, for an
assumed average composition, the increase in volume is 4-30 per cent., provided all
the resulting compounds are solid ; the change of augite to biotite is exothermic and,
for an assumed average composition, the calculated increase in volume is 17-26 per cent. ;
the change of felspar into each component of saussurite is exothermic and involves
expansion of volume. Hence we can be quite certain that the recrystallisation of the
dolerite which involves these changes has been accompanied by an expansion of volume
and a liberation of heat. The recrystallisation of the granodiorite is not so complete
as the recrystallisation of the dolerite. We may, then, imagine that the small gneissic
xenolith enclosed in the relatively large mass of amphibolite has been exposed to greater
pressure and higher temperature than the main mass of the granodiorite gneiss. As a
result the small gneissic xenolith shows a different degree of recrystallisation than the
granodiorite gneiss.

If a small mass of rock be enclosed within a larger mass of another type and the
whole subjected to metamorphic conditions, then I think it would be generally expected
that the foliation would travel independently through the two types as has happened
in the case of the gneissic xenolith. We would, therefore, be inclined to view the diverted
foliation as the abnormal case, even though it has been more commonly observed. As
the general metamorphic conditions are the same in both cases, the only important
difference lies in the greater expansion in volume of the amphibolite which is directly
due to the chemical rearrangement. In this expansive effect we are forced to conclude
must lie the cause of the diverted foliation.

* " Treatise on Metamorphism," C. R. Van Hise, pp. 277, 278.

CHAPTER IV.
1. THE GRANODIORITE GNEISS AT CAPE DENISON.

The chief rock type at Cape Denison is a coarse-grained, grey-coloured gneiss with
a granitic appearance. It is foliated, and the strike of the foliation is N. 24 W. The
dip of the foliation is at a high angle, sometimes to the east and sometimes to the west.
By observation of these dips the axes of folds seem to be determined, but no evidence
of folding is forthcoming from the study of the black amphibolite bands that traverse
the area parallel to the strike of the foliation. In the description of this gneissic type
we exclude reference here to the patches of dark-coloured gneiss that may appear
enclosed in the granitic gneiss.

In the hand specimen the grey gneiss has a varying amount of schistosity.
Foliation is well marked in some specimens, while only faint in others. Beautiful
examples of contorted gneiss are found in some places where the crystallisation
schistosity is marked by bands of quartz and felspar (Plate XX., fig. 3). Yet the
character of the gneiss is fairly constant across Cape Denison. Quartz, felspar, and
biotite are always visible to the naked eye ; magnetite is sometimes well developed,
and there are patches where the same is true of pyrite. In some parts black vitreous
allanite is obvious and developed in flat prismatic individuals. The largest allanite
crystal obtained is an impertect one, |in. long, Jin. broad, and -]Vn. thick. Apatite
may also be abundant in the same areas as the allanite. Rarely large orthoclase
crystals are found as white or reddish- white porphyroblasts up to 2in. in breadth.

In the hand specimens the normal texture of the rock is dominantly massive, but
a tendency to the schistose types can always be detected. The biotite flakes may bend
round large crystals of quartz and felspar, and then there is a tendency to augen gneiss
and a rough lenticular texture. These lenticles may become flattened and more
granular and then a distinct banded appearance is evident (Plate XL, fig. 5). Also
the parallel bands of quartz and felspar against mica may develop the columnar
appearance of wood gneiss (No. 143 A).

The structure is granoblastic, due to the approximately isometric character of
the quartz and felspar grains. Blasto-granitic structure is present, because the original
big crystals of felspar and quartz in the granite can often be reconstructed in the
cataclastic areas. Cataclastic structures are common when quartz and felspar crystals
have been crushed. Mortar structure is common, but is usually best developed along the
j unction of two felspar crystals. Diablastic structure is frequently seen in the crush areas.

Specimen No. 1 1 has been selected and analysed as the normal type. It was collected
from the site of the main hut at Cape Denison. No. 11 will, therefore, be described
first, and then the other types can be dealt with in a relative manner.

The chief minerals present are quartz, microcline, orthoclase, andesine, perthite,
and biotite. In smaller amounts are epidote, muscovite, sphene, chlorite, and calcite.
As accessories are apatite, zircon, magnetite, pyrite, and hematite.

THE METAMORPHIC ROCKS OF ADEL1E LAND. 8TILLWELL. 85

The quartz is present in irregular, indented grains, and frequently shows marked
cataclasis. Some of the original crystals are replaced by interlocking granular aggregates
with undulose extinction. Well marked strings of linear inclusions frequently pass
through adjacent grains in such aggregates. Microcline, with its characteristic cross
hatching, is abundant, and has developed from the original orthoclase of the granite.
In some cases the transition from orthoclase is incomplete, and clear orthoclase forms
the bulk of the crystal, which possesses a fringe with the cross twinning of microcline.
The microcline exhibits some cataclasis. Perthite or networks of soda plagioclase
and orthoclase are common in large individuals. The orthoclase is in large plates,
and an albitic plagioclase appears in short, broken, more or less parallel strings which
have the higher refractive index. This perthitic intergrowth can be found along lines
of incipient fracture. Diablastic intergrowths of quartz and felspar, or of orthoclase
and plagioclase, similar to the micropegmatitic intergrowths, are common. It is
evident that the intergrowth has a metamorphic origin, because, not only are they
most frequent in the crush areas, but they may be seen, with a rounded outline,
developing parasitically within a plagioclase crystal. Sericite has developed from the
orthoclase and is chiefly to be found in the crush areas. Like secondary biotite, it tends
to wrap itself around primary quartz and felspar. In one case it appears as a zone
between two microcline crystals. It may also appear as a rim on biotite crystals
bending with the biotite. The original plagioclase has a refractive index above basal
quartz and below other grains of quartz, and is, therefore, referred to as andesine. The
application of Becke's bright line method is limited because the edges of the crystals
are frequently crushed. Saussuritic aggregates have, in some cases, developed from the
andesine, but only granular epidote and rounded blebs of secondary felspar can be
distinguished in them.

Biotite is common in the crush areas but it is not confined to them, and some of it
may have been preserved from the original granite. In some cases it tends to wrap
itself around the relic quartz and felspar. Its colour is normally brown, but there
is a subordinate quantity of green. Green chlorite in small amount is interlaminated
with biotite. The biotite is often associated with epidote, sphene, and magnetite.
Sometimes there is a thin rim of granular epidote around the biotite. The rim may
also be sericite, which may develop into muscovite, because the latter is sometimes
associated with the biotite. Muscovite is sparingly present in individuals comparable
in size with the biotite. Pleochroic halos in biotite appear around inclusions of zircon
and sphene. Epidote is present either in pleochroic crystals and grains or in the finely
granular form. It is frequently associated with the biotite.

Sphene is usually granular, but some wedge-shaped crystals are seen. Sometimes
it encloses a magnetite core, but not so frequently as in the amphibolites. Both sphene
and epidote may be completely enclosed in biotite. Granular calcite has been found
and has probably developed with the saussurite. Apatite, zircon, pyrite, and magnetite
are scattered throughout. The apatite may be in large crystals, and the zircon is
noticeable in small well-defined crystals with pointed ends which have clearly never
left their primary host.

AUSTRALASIAN ANTARCTIC EXPEDITION.

Other sections show variation in the degree of metamorphism of the rock. The
crush areas may be less abundant and the andesine felspar better preserved. The
twin lamellae of the andesine may be curved and bent by the pressure. At the same
time there may be less sericite, less perthite, and less of the diablastic structure. In
other cases muscovite may be better developed, or chlorite may replace a portion
of the biotite, while green hornblende may appear.

The microscopical examination, therefore, renders it apparent that this gneiss
is the metamorphosed equivalent of a granite or a granodiorite.

CHEMICAL CHARACTERS.

The following analysis .of the type specimen No. 11 was made in Victorian
Geological Survey Laboratory :

II.

III.

IV.

SiO a

67-10

68 '92

68'62

fifi-7fi

A1 4 S

14-87

15'26

15'70

U-SR

Fe.O,

T14

0-80

1 -fifi

9-04.

FeO

3'76

3-30

1-77

3-75

MgO

rso

1'64

1'28

2-71

CaO

3-47

S-04

9-Kfi

4..co

Na a O

2-56

2-71

K-OR

1.4.4.

K.jO

3'50

2 '93

141

Q.5Q

H 2 0+

0-68

1-04

0-fi6

Ton 0'4.Q

H 4 - .

O'll

0-22

0-10

CO,

Nil

Ti0 2

0'68

0'70

0-2fi

P 2 S

0-20

0'19

0-10

s6. ...

Nil

Clf

0-05

Nil

MnO

Oft?

0-14.

NiO, CoO

Cr 2 8

Nil

Cob. 8

Nil

Li 2

BaO

O-O9

0-03

Total

99-92

100-75

100-19

QQ'fiA

Specific Giavity .

2-725

2-fi88

2.79

Class

I.
I

Order

Rang .

Subrang . .

Magmatic Name .......... Amiatose Amiatose

I. Granodiorite gneiss, Specimen No. 11, hut site, Cape Denison, Adelie Land. Analyst, J. C. Watson-
II. Granodiorite, near Old Sawmill, Hesket, Macedon District No. 35. Analyst, A. Hall*

III. Typical banded gneiss, north side of Hopkin's Bay, Rainy Lake, Canada, t

IV. Biotite gneiss, near Sangobeag, Durcess, Scotland J

" Annual Report of the Secretary of Mines, Victoria, for 1907," p. 61.

The Archaan Geology of Rainy Lake, Restudied," A. C. Lawon/Geol. Surv. Canada, Mem. 40, p. 93.
The Geological Structure of the North-West Highlands of Scotland," Mem. Geol. Surv. Gt. Britain, 1907.

THE METAMOKPHIC ROCKS OF ADELIE LAND. STILLWELL.

The analysis of the Cape Denison gneiss is strikingly similar to that of a grano-
diorite, and an analysis of a Macedon granodiorite is, therefore, inserted for illustration.
The similarity is strong in all essential features, and both rocks occupy the same division
in the American classification. Analyses of a banded biotite granite gneiss from the
Rainy Lake region in Canada, and of a grey biotite gneiss from the Scottish highlands,
are also quoted, and these show general similarities to the Cape Denison gneiss. Such
comparisons, which could be readily multiplied, are interesting in emphasising the
lithological uniformity in the Archaean terraines in all parts of the world. Similar
rocks are known to exist in Australia and in South America and in South Africa.

The ratio of the potash to the soda is abnormal in the Canadian rock, while the
alkali percentage of the Scottish rock is lower than that of the Antarctic rock. The
differences in total alkali percentage are made important in Grubenmann's classification
of schists. The group values and projection values of these three gneisses are :

Rock.

Group Values.

Projection Values
after Osann.

No. 11, Cape Denison

74-0
74-9
72-3

5-1
6-9
3-7

4-0
3-1
5-3

7-8
6-1
9-5

1-0

0-5
0-2

1-6
1-4
1-7

6-0
8-5
4-0

4-8
3-9
5-8

9-2
7-6
10-2

Rainy Lake Gneiss

Banded Gneiss, Scotland

THE CLASSIFICATORY POSITION.

The Cape Denison gneiss occupies a position on Osann 's triangular projection
(fig. 9), which is midway between the mean group values of Groups I. and III. The
Canadian rock enters Group I. and the Scottish rock Group III. The Cape Denison
gneiss should be considered as an intermediate type, and it occupies a position on the
triangular projection halfway between the positions of these Canadian and Scottish
rocks. Since, however, it is the metamorphic equivalent of a granodiorite, and since
granodiorites are well known and definite rock types, it must be acknowledged that
the metamorphic equivalents of granodiorites should be recognised. The Cape Denison
gneiss is, therefore, best named as a granodiorite gneiss.

The mechanical effect of the metamorphism upon the original granodiorite is evident
in the jundulose extinction of the quartz, the cataclasis of the quartz and felspar, the
prominent mortar structure, and the tendency of the colourless minerals to .be arranged
in layers of aggregated fragments. Evidence for the following transformations have
also been noted :

1. Partial decomposition of primary biotite into epidote, sphene, and ilmenite.

2. Partial decomposition of primary biotite into chlorite.

AUSTRALASIAN ANTARCTIC EXPEDITION.

3. Partial saussuritisation of plagioclase.

4. Partial change of orthoclase into sericite.

5. Development of microcline from orthoclase.

6. Development of perthite from plagioclase.

7. Development of secondary biotite from chlorite.

A7 V W W VV\ A /

AAAAA A/ VVWVXAAA A

\AA7V

Fig. 9.

I. Mean position of Group I., the Alkali Felspar Gneisses.
III. Mean position of Group III., the Plagioclase Gneisses.

A. Rainy Lake Gneiss.

B. Scottish banded Gneiss.

C. Granodiorite Gneiss, Cape Denison.

The general survey of these changes indicates that the conditions of Grubenmann's
epi zone of rock metamorphism have been dominant. The last three changes indicate
that there is an approximation to the meso zone conditions, and so also does the
occasional development of a rough crystallisation schistosity. Hence, while we consider
that the rock may be referred to as an epi granodiorite gneiss, the meso zone tendency
should be recognised.

THE METAMORPHIC ROCKS OP ADELIE LAND STILLAVELL. 89

2. THE APLITE GNEISSES.

Associated with the granodiorite gneiss are quartz felspar gneisses which are the
metamorphosed equivalents of pegmatite and aplite veins, which were most probably
connected with the intrusion of the original granodiorite magma. These gneisses are
red or white or grey, and appear as small bosses or veins in the granodiorite gneiss.
The bosses may be a dozen yards or more in width and the foliation cuts through them
independently of the outline of the boss or of the trend of the vein. As the surface
outcrop of the rocks is perfectly fresh and uncovered, it could be observed in the field
that the boundary between the aplite gneiss and the granodiorite gneiss was often
indistinct, and there was frequently a gradual transition between the two. A set of
specimens was accordingly collected across such a boundary and show a gradual change
from the pure white aplite gneiss through pale grey shades to the darker grey granodiorite
gneiss. In this we have another example of metamorphic diffusion. No field evidence
is available concerning the relation of the aplite gneisses to the amphibolites. The small
quartz veins that cut the amphibolites may be correlated with quartz veins which fill
fractures in the granodiorite gneiss definitely subsequent to the development of the
foliation.

No. 10A. Specimen No. 10A is an example of the aplite gneiss and was collected
from a vein about 18in. wide, close by the southern magnetic hut. The trend of the vein
is approximately parallel to the direction of the foliation but was observed in section
to cross it horizontally. In the hand specimen the rock has a pale-grey colour and a
fine granulitic appearance. Quartz and felspar are the chief minerals, but small biotites
are evenly distributed through the rock and produce perceptible schistosity. Occasional
large crystals of allanite appear in the vein and have formed a centre of crystallisation
around which felspar crystals radiate.

In thin section the rock is even grained with granoblastic structure and with
abundant evidence of mashing and granulation. It is composed chiefly of interlocking
crystals of quartz and felspar with smaller amounts of sericite, muscovite, and biotite,
while magnetite, apatite, allanite, monazite appear as accessories.

The quartz appears in rounded, indented, and interlocking grains, and shows
considerable cataclasis. Some of the granular aggregates of quartz have developed
from the primary individuals of the pegmatite. At times there is a partial drawing
out into lenticles and layers. Clear orthoclase is present, but the bulk of the potash
felspar is microcline. Some of the microcline is quite clear and transparent and has
developed from orthoclase as in the granodiorite gneiss. The microcline may appear
as rounded blebs within the quartz crystal. Part may be relics of the original pegmatite
because microcline is a common constituent of such, and some microcline crystals show
strain polarisation and incipient granulitisation. Perthite is present. A small portion
of the felspar has been sericitised and some of the sericite has passed over into muscovite.

90 AUSTRALASIAN ANTARCTIC EXPEDITION.

Diablastic structure is common. That this vermicular interlocking of quartz
and felspar is part of the metamorphic character is evident, because it is most common
in the areas with marked cataclasis, and it appears wholly enclosed within felspar crystals-
In other cases it has developed as a partial fringe to the plagioclase whose original outline
is quite evident, or it may transgress as a bight into the side of a crystal. These features
distinguish this intergrowth from the pegmatitic intergrowth of igneous rocks which is
the crystallisation product of a eutectic mixture, and which is the last to crystallise
in the consolidation of a rock magma. The diablastic structure does not have the
character of a final product, but it has arisen contemporaneously with the other meta-
morphic minerals and structures.

Small crystals of ragged brown biotite are evenly distributed throughout the
slide, and epidote may be associated with it. Odd grains of allanite are present, though
no crystal comparable in size with the large examples exists in this slide. The
development of allanite is, however, quite a feature of this locality. Macroscopically
it has a black, pitchy lustre, and, in some cases, is surrounded by a reddish-brown zone.
In thin section the allanite is found in reddish-brown pleochroic crystals. When
associated with biotite it is surrounded by pleochroic haloes. They are biaxial with
oblique extinction. The double refraction in many cases is high, and there may be
a small amount of zoning. In such cases clinozoisite seems to be developed along its
sides. In other cases it alters to a brownish-yellow amorphous gum-like mass. The
allanite proved to have a refractive index greater than monobromnapthalin (1-648)
and less than iodmethylene (1-740). When equal proportions of these two oils are
mixed, part of the crystal had a refractive index greater than the mixture and part less.
The mean refractive index is, therefore, in the neighbourhood of 1-68, a value which
is on record for allanite. These characters are sufficient to render the identification
fairly certain.

Since the cerium metals are present, and apatite is present as an accessory, it is to
be expected that monazite should be found. Grains are found with a heavy dark border
and with high polarisation colours and with marked similarity to zircons. Oblique
extinction has been noticed, and these small crystals are, therefore, considered to be
monazite. Pleochroic haloes around monazite in biotite are strong. Accessory grains
of magnetite and reddish hematite are fairly common, while clear apatite is less so.

The rock corresponds closely to the family of Glimmerarme meso alkali felspar
gneisses in Group I. of Grubenmann's classification. The development of microcline and
perthite and the diablastic structure rather signifies the meso zone characteristics. It
may, therefore, be described as an alkali felspar gneiss, poor in mica, or an aplite
gneiss developed from an aplite vein under conditions approximating to those of the

meso zone.

No. 150. An example (No. 150) from the pegmatite bosses is very similar in most
respects to the example (No. 10A) collected from a vein. It shows variation in its larger
grain size, and its more massive texture, and in general, it shows stronger epi zone

THE METAMORPHIC ROCKS OF ADELIE LAND. 8TILLWELL. 91

features. Microcline and perthite individuals are present, but there is a greater amount
of sericite. Relic plagioclase lamellae can be recognised in large sericitic masses. The
development of sericite can be found along shear planes in microcline crystals. There
is also some kaolinisation of the orthoclase. Fracturing and granulation of the quartz
and felspar is more prominent than in the preceding example, and so also is mortar
structure. The diablastic structure is not common and mostly in incipient stages.
The mica content is small and includes green chlorite, green biotite, and white muscovite.
The chlorite and biotite are often associated in one individual. Grains of epidote are
associated with the mica. A little calcite is present, and magnetite, apatite, monazite,
and allanite are again accessories. The epi zone characters are here considered dominant,
and the example is described as an epi alkali felspar gneiss, poor in mica, or as an epi-
aplite gneiss.

In further examples the amount of sericite may increase sufficiently to yield sericite
gneiss. Occasionally a relic garnet is found with considerable development of green
chloritic products along the cracks. The percentage of ferromagnesian minerals increases
towards the margin of the bosses as we pass outward through metamorphic diffusion
types into the granodiorite gneiss.

3. INTERPRETATION OF CERTAIN VARIATIONS IN THE
GRANODIORITE GNEISS.

Since the boundaries of the aplitic masses with the granodiorite are in places
destroyed and replaced by metamorphic diffusion types, there is no a priori reason why
such diffusion types should not, under favourable circumstances, extend across the whole
width of the vein. In such contingencies the vein will completely lose its identity
and become part of the main gneissic mass. A study of metamorphic diffusion
specimens indicates that this has actually taken place. Rocks which were collected
in the field as varieties of the granodiorite gneiss are now considered to be diffusion
types. This is particularly the case with gneisses collected from the locality by the
magnetograph house, which is the precise locality from which the conception of meta-
morphic diffusion is developed in the case of the amphibolites.

In this case the gneiss has a lighter colour, due to the absorption of some quartzo-
felspathic material. The composition and the granularity are variable and the texture
is usually more massive. A feature of the locality is the abundance of monazite and
allanite (Plate XI., fig. 6). The allanite is found in exactly the same manner as noted
in the aplitic gneiss, i.e., frequently with a radial arrangement of felspar around it. No
definite gneissic vein is recorded from the precise point where the examples were collected,
and they are very similar to the metamorphic diffusion products on other parts of Cape
Denison.

Specimen No. 60 is an example of the pale grey gneiss of this type. The chief
constituents are quartz, orthoclase, perthite, and microcline. The ferromagnesian

92 AUSTRALASIAN ANTARCTIC EXPEDITION.

constituents form less than 4 per cent, of the slide, whereas the normal ferromagnesian
percentage of the granodiorite gneiss is about 18 per cent. Its silicity would, therefore,
be probably more comparable with the aplite gneiss than with the granodiorite gneiss.
Diablastic interlacings are abundant. Chlorite and epidote are associated with biotite
and muscovite. No hornblende is present. Apatite, monazite, and allanite are acces-
sories, and of these allanite is the best developed.

Specimen No. 154 is a similar example, but possesses coarser grain size. There is
also a larger ferromagnesian percentage than in the previous case, and the biotite appears
in clusters. The irregular distribution of biotite is noticeable in the hand specimen.
Muscovite is again present. Large crystals of quartz, orthoclase, and plagioclase are
the dominating minerals. Some microcline is present, and the orthoclase does not show
much sericitisation. The plagioclase seems to be albite oligoclase or an albite, and
is, therefore, the plagioclase of an aplitic vein rather than the plagioclase of the grano-
diorite gneiss. Apatite is the most abnormal constituent and forms large crystals
which, though not uniformly distributed, contribute 3j per cent, of the slide. Monazite
is also an abundant accessory, so that the P 2 5 content of this sample must be unusually
high. Allanite is very well developed and zircon also seems to be present.

The resemblances to the aplitic gneisses are apparent. This likeness can only be
reconciled with the field evidence by the recognition of metamorphic diffusion and
its obliteration of the individuality of the vein.

CHAPTER V.

CORRELATION AND CRITICISM OF ANALOGOUS AREAS.

1. GENERAL.

Crystalline schists have now been reported from widely separated parts of the
Antarctic continent. Large areas of metamorphic rocks can, therefore, be assumed
to exist under the ice sheet. They have already been regarded as forming the ancient
platform on which the central part of South Victoria Land was built.* The present
knowledge of the distribution of these rocks indicates that they form the platform of the
Great Antarctic Plateau.

In the Ross Sea region the known extent of the gneisses has been extended by
Shackleton's expedition and by Scott's last expedition. They have been proved to exist
in King Edward VII. Land on the east. They are now known to range on the west
to Adelie Land and to Queen Mary Land. Dredgings from the " Challenger " expedition
indicate that they probably extend still further west, and they have been recorded
from West Antarctica.

Streaks of hornblende schist are found associated with the gneisses in the Kukri
Hills, in South Victoria Land, by Ferrar. Amphibolites and pyroxene granulites have
been recorded from the moraines by the Shackleton expedition.! Amphibolites and
hornblende schists have also been mentioned in the description of the rocks obtained
by the " Belgica."J

In Antarctica, as elsewhere, amphibolites are found in manifold forms wherever
the crystalline schists appear over a considerable area. Large areas of crystalline
schists appear in every continent, and any attempt to correlate occurrences immediately
becomes a tabulation of areas of Archaean rocks, and this is unnecessary here. As a
consequence of the lithological similarity of most Archaean terraines it follows that any
theory correctly deduced from one area should immediately find wide application. The
interpretation of one area should materially assist the interpretation in all other areas.
That this has not been the case has been in some measure responsible for the more or
less disorganised condition of the study of metamorphic areas, and for the complexity
that is commonly associated with their study. More particularly, conflicting opinion
has been responsible for Cole's description of amphibolites as puzzling rocks.

* Natural History, voL I., Geology, H. T. Ferrar, Nat. Ant. Ezped., Brit. Miu., p. 25.

tGeol. vol. II., Brit. Ant. Ezp., Mawmon, Walkom.

t ' RerolUU da Voyage du S. Y. Belgica.' " Expedition AnUrctique, Beige, Cteologie II., Teil, Dragomir Surtek, 1912.

f " Rocks and their Origin," Q. A. J. Cole, p. 148.

94 AUSTRALASIAN ANTARCTIC EXPEDITION.

If any theory can replace opposition by harmony such a theory will grow in strength
as each new concordance is produced from the field of geological literature. The
theories of metamorphic differentiation and metamorphic diffusion which appear to
account for certain features at Cape Denison seem to be applicable in other areas. In
some cases they produce interpretations quite different from the published explanations,
and the value of these interpretations depends partly on the assurance that can be
given to the Cape Denison phenomena and partly on the value that can be attached
to Grubenmann's great work, " Die Krystallinen Schiefer." In many cases it has seemed
to us that the description of the products of metamorphic diffuson and metamorphic
differentiation are better than those which have been presented from Cape Denison.
The Cape Denison descriptions have necessarily suffered from our inability to revisit
the area. The field work could not be revised with the progress of the work in the
laboratory, and the conclusions cannot be forced home with the wealth of evidence
that might otherwise have been available.

We propose now to examine the data from some of the other areas from the stand-
point that the Cape Denison study has created. Attention is only given to a few
recent and important publications, and the criticism is offered to stimulate interest
and to direct attention to explanations which have not been hitherto considered.

2. NORTH- WEST HIGHLANDS OF SCOTLAND.

Features at Cape Denison and Cape Gray are reflected in the areas of Lewisian
gneiss in the North-West Highlands of Scotland. The Lewisian gneiss (p. 41)* has been
subdivided into

1. Fundamental Complex.

2. Ultra basic dykes.

3. Basic dykes of dolerite, epidiorite (amphibolite), hornblende schist.

4. A few dykes of exceptional composition.

5. Granites and pegmatites.

Groups 2 and 3 have been found to be associated with the Lewisian gneiss, and yet
intrusive into the Fundamental Complex, and are so referred to the pre-Torridonian.
Teall remarks (p. 39) that in many places the dyke-like character is obvious, as more
or less vertical walls of black rock clearly cut across the gneissic banding. But, in
other places, owing to movements after or during the injection of the dykes, the dyke-
like character is lost and the rocks of the dykes become more or less incorporated with
the earlier complex.

Home remarks (p. 36), in connection with the basic intrusions, that it is of special
importance to note that in the southern tracts, where the dykes are represented by

1 The pages quoted in connection with this area have reference to the following publication : " The Geological Structure
of the North-West Highlands of Scotland," B. N. Peach, J. Home, W. Gunn, C. T. dough, L. W. Hinxman, J. J. H. TealL
Memoirs Geol. Surv. Gt. Britain, 1907.

THE MITAMOBFmC ROCKS OF ADELIE LAND STILLWELL. 95

hornblende schists, which seem to become part of the Fundamental Complex, and where
intrusive junctions are only occasionally met with, biotite gneisses and hornblende
gneisses are characteristically developed. Hence the obvious nature of the dyke masses
at Cape Gray, their less obvious appearance at Cape Denison, and their partial
destruction at Cape Denison, are matched by similar instances in the Scottish area.
Remarkable variation in mineral and structural composition is noted in both areas,
and the dominant types are the same in both cases.

The above remark of Home illustrates the incomplete separation of the second
and third groups from the Fundamental Complex, and it also appears to be evidence
of the development of biotite gneisses and hornblende gneisses in the same way as at
Cape Denison, viz., by the destruction of the walls of the basic dykes by metamorphic
diffusion.

In some cases, as in the Gruinard district (p. 176), it is clearly shown that the
basic dykes form an absolutely different series to that which supplied the early basic
material in the Fundamental Complex. On the other hand, in the description of
the Cape Wrath to Laxford area (p. 107), it is recorded that the grey biotite gneiss, the
hornblende biotite gneiss, and the dark gneiss alternate in bands and areas of
varying breadth, having no sharply defined boundaries, but graduating from the more
acid to the more basic types. These types cannot be distinguished on the field map,
and are therefore considered only as portions of the primary mass. In the description
of the Loch Maree to Gairloch area (p. 195) it is recorded that field distinction is
impossible between the hornblende gneisses with quartz and hornblende gneisses
without quartz on account of their variation. These last two cases afford further
analogy to the Cape Denison rocks, where hornblende gneisses are considered to
be metamorphic products of a mixture of grey gneiss and amphibolite, resulting from
metamorphic diffusion.

Thus it is a perfectly natural result that it should be recorded in the Loch Maree
and Gairloch district (p. 195) that the " early basic rocks " (those which have not
been separated from the Fundamental Complex) are more variable in composition than
the basic dykes. The basic dykes, which happen to be parallel to the direction of
foliation, are only recognised as such when they have escaped metamorphic diffusion,
and a varying amount of diffusion will produce varying results. In the Loch Carron to
Point Sleat (Skye) district the rocks are stated (p. 262) to consist of biotite and horn-
blende gneisses with bands of hornblende schist, which are considered to represent the
basic dykes in the unmodified areas of Lewisian gneiss in Ross and Sutherland. Had
these bands been dislocated, and had they then suffered metamorphic diffusion, they
would certainly present the same features as the " early basic rocks." I do not argue
that there is but one series of pre-Torridonian basic dykes in the North- West Highlands
to which both the so-called " early basic " and the " basic " rocks may be referred.
There are at least two, and possibly more, but I do mean that the " early basic " rocks

96 AUSTRALASIAN ANTARCTIC EXPEDITION.

may possibly represent the torn-up and diffused remnants of some dyke series, and
that they are not to be looked upon as necessarily earlier, without further considera-
tion, than the enclosing gneiss.

In the Loch Laxford to Kylesku area it is stated (p. 134) that gneisses enclose
frequent lenticles and lumps composed entirely of hornblende and pyroxene. Again,
a leading feature of the Fundamental Complex in the Gruinard district (p. 177) is the
extraordinary abundance of knots of basic material in the gneiss. A beautiful,
unfoliated diorite is here recorded, but the most abundant consist of the dark hornblendic
rock. The same type of thing is observed in other places, and they are looked upon
as products of segregation, in common with the acid gneiss, from an intermediate magma,
or as included fragments of an older rock system. Now in the Kylesku to Loch Broom
area (p. 169) it is found that near areas of dominant stress a dyke may be wrenched
into a series of isolated lenticles or " phacoidal masses." The evidence of the Scottish
area, apart from the Cape Denison observations, thus shows that a dyke can be torn
into fragments which may now appear as isolated inclusions. Hence a complete account
of the Scottish Highlands must consider the possibility that many of the hornblende
and pyroxenic clots may be the metamorphosed remains of torn-up fragments of basic
dykes. They may be stated to be differentiation products caught up in the manner
illustrated at Depot Island, South Victoria Land*, by the intrusion of granite, which
have been subsequently modified by metamorphic processes ; but there is no positive
proof, at present, that requires them to be looked upon as " earlier " than the enclosing
gneiss.

Equally evident as the metamorphic diffusion in the North- West Scottish Highlands
is metamorphic differentiation. Rocks are recorded by Teall (p. 45) which consist of
pure hornblende, and which are found as knots, lenticles, and bands. Similar separation
of the biotite is mentioned in the description of the Loch Maree and Gairloch district
(p. 193). Some of these are no doubt similar to the hornblende and biotite patches
which have been considered as metamorphic differentiation products at Cape Denison.

In discussing hornblendites and pyroxenites (p. 45) Teall finds that, by the gradual
increase in the amount of hornblende, the pyroxenites pass into hornblendites. They
form banded masses, and the possibility must arise as to whether both these types are
not metamorphic differentiation types. With suitable metamorphic conditions
pyroxene could very well differentiate itself with equivalent result to the hornblende.
The recorded section (p. 47), where four hornblendite bands and four pyroxenite bands
occur in 4ft. 5in., could well be an example of metamorphic differentiation developed
with alternating conditions. It may, indeed, be viewed as a magnified crystallisation
schistosity. Where Teall describes the rocks containing hornblende and pyroxene
(Group III.B2) he finds that such are related to the pyroxene gneisses. He states
(p. 63) that it is impossible to avoid the conclusion that they have been formed from the

* "Geology." lol. I.. Brit. Ant. Exp., Plate LXXXL, p. 246.

THE METAMORPHIC ROCKS OF ADELIB LAND. 8T1LLWELL. 97

pyroxene gneisses by secondary metamorphic processes. It is interesting to note that
this type of change of pyroxene to hornblende is just that which Grubenmann would
describe in transition from the kata zone conditions to the meso zone conditions.

The phenomena illustrated on Plates VI., XII., XIV., XVIII. of this British Memoir
are exactly paralleled in the Cape Denison area. Plate VI. would there be described
as a stage in the metamorphic differentiation of the constituents of a primary basic
rock. In Plate XII. the hornblende differentiation is more advanced, and an imperfect
banded arrangement appears. Plate XIV. may have originated where a basic dyke
has run out into parallel threads, or it may be again metamorphic differentiation. A
dyke fragment with smaller pieces detached from it, together with a certain amount
of migration, might give rise to an appearance similar to that on Plate XVIII.

After the publication of the preceding memoir in 1907 the Geological Survey of
Great Britain has produced a series of memoirs dealing with the North- West Highlands.
These memoirs provide the explanation of the published quarter sheets, but, like the
1907 memoir, are largely a mass of field data that await the co-ordinating process of
some worker.

In the " Geology of the Seaboard of Mid Argyll " (Memoir No. 36, 1909) I have
noted an interesting (p. 6} description of phacoids of epidiorite in pebbly limestone
matrix. These phacoids are recognised as fragments torn off epidiorite dykes during
crushing, and the limestone is considered to have played the part of a plastic matrix.
This is another illustration of what may happen to a dyke sheet submitted to strong
stress.

In the geology of Glenelg, Lochalsh, and south-east part of Skye (Memoir No. 71
1910), there is an excellent illustration (Plate V.) of a large knot of foliated basic rock
in the Lewisian gneiss at Rudha Gaol. The general appearance of this knot could be
matched among the disrupted dykes at Cape Denison. At Rudha Gaol the basic
inclusions have varying composition, indicating metamorphic diffusion or metamorphic
differentiation, or both. Another illustration (Plate VIII.) of the same memoir provides
a good example of metamorphic differentiation where a basic lenticle is illustrated in
. thinly-banded hornblendic gneiss.

In the geology of the Fannich Mountains and the country around Loch Maree and
Strath Bromm (Memoir No. 92, 1913), we have again reference to the exposures which
have been illustrated in the large 1907 Memoir, and which we have considered to provide
examples of metamorphic diffusion and metamorphic differentiation. The explanations
given in 1907 are adhered to, and Plate VI. is still the picture of a magma of intermediate
composition from which the more distinctly basic and acid rocks are in the process of
formation. Another view is stated which regards the bulk of the rock as a mixture
rock formed by the mingling, probably while in a semi-fluid state, of basic and acid
portions. This latter view is a step nearer to the recognition of metamorphic diffusion
phenomena. There is also recognition (p. 23) of transitions by insensible gradations

SariM A, VoL Hi., Part 1 G

98 AUSTKALASIAN ANTARCTIC EXPEDITION.

from ultra basic rocks to basic rocks. On Plate III. of this memoir there is an
illustration of garnetiferous muscovite biotite gneiss with lenticles of " pegmatite." It
can be pointed out that these lenticles may be fragments of broken pegmatite veins,
but they are quite possibly metamorphic differentiation products, as the same quartzo-
felspathic material is distributed right through the base of the gneiss.

If, then, the phenomena of metamorphic diffusion and metamorphic differentiation
be upheld, and the rock types be traced back with their aid to the primary types, we
must surely arrive much nearer the true history of the Fundamental Complex in the
Highlands. If the Complex be studied from the view point of these theories it seems
possible that some of the apparent complexity will disappear. The rocks must no longer
be approached through the eyes of a mineralogical classification, as attempted by Teall,
which obscures relationships and separates similar rocks. True metamorphic types
must be recognised, and the kata zone, meso zone, and epi zone varieties of the same
type must be correlated together according to Grubenmann's method or to some
analogous system. Metamorphic diffusion types and metamorphic differentiation types
should also occupy divisions in the mental field of view before it would be possible to
present an orderly exposition.

3. HALIBUBTON AND BANCROFT AREAS, CANADA.

Amphibolites are recognised as forming an important part of the Canadian Archaean
rocks, and considerable study has been given to them by Adams and Barlow in the
Haliburton and Bancroft areas in the Province of Ontario. Their work is embodied in
a memoir published by the Geological Survey of Canada in 1910*. Papers containing
their results appeared earlier in the Quarterly Journal of the Geological Society of
Londonf and in the Journal of Geology^. In presenting criticism on that portion of
their work which pertains to amphibolites, attention is only given to the memoir, the
latest and most complete publication.

From a study of this memoir we find that it appears to be claimed that amphibolites
are formed in diverse ways throughout the area. According to these different modes
of origin the amphibolites may be classified in the following manner :

1. Those derived by alteration of basic dykes or similar igneous intrusions

(a) Those which can now be recognised as true dykes.

(6) Those which appear as bands in crystalline dolomite, &c.

2. Those derived by alteration of limestones by action of intrusive granitic magma

(a) Those which appear in a linear manner along the contact of the limestone

masses and the gneiss.
(6) Those which appear as inclusions in the grey gneiss.

" Geology of the Haliburton and Bancroft Areas," F. D. Adams & A. E. Barlow, Geol. Surv. Can. Mem., No. 6.
t " The Laurentian System in Eastern Canada," F. D. Adams. Q.J.G.S. 1908, p. 127.
t " On the Origin of the Amphibolites of the Laurentian Area of Canada," Journ. Geol., 1909, vol. 17, p. 1.

THE METAMORPH1C ROCKS OF ADEL1E LAND. STILLWELL. 99

3. Those derived by the metamorphism of impure bands in the limestone series

(a) Those which are described as " pyroxene hornblende gneiss " or
" pyroxene hornblende granulite."

(6) Those which are described as " feather amphibolite."
(c) Those which contain orthorhombic amphibole.

The amphibolites of igneous dyke origin are recognised when they are found in the
field to cut across the bedded white crystalline limestone. Adams and Barlow find
that the field evidence is essential to recognise the igneous origin with certainty, but
other cases which are macroscopically identical with the established dykes, and which
appear interbanded with crystalline dolomite or crystalline limestone, are also considered
to be probably igneous. A chemical analysis is quoted, and it is stated (p. 161) that
it is highly probable that they were diabases. It is interesting to note that this altered
Canadian diabase and the typical amphibolite (No. 629) from Cape Denison are so
strikingly similar that they occupy the same division of the American classification.

Adams and Barlow have assured themselves that amphibolites are formed by the
second method by a study of the contact phenomena in the border zones of the granite
gneiss. Where the granite has intruded limestone the changes produced are divided
into two classes (p. 87)

1. Alteration of the limestone into masses of scapolite-bearing pyroxene rock.

2. Alteration into pyroxene gneiss or amphibolite.

The No. 1 change is proved by finding all possible transitions between the pure limestone
and the pyroxenite, which is stated to occur (p. 88) at or near the contact with the
granite. With simple contact metamorphism we expect to find in a traverse across the
boundary transition from limestone to pyroxenite and then a sudden change from
pyroxenite to granite. This seems to be indicated, and the varying nature of the product
in a measure supports the theory ; but it needs to be demonstrated that the same
metamorphism which affected the granite after its consolidation would not produce
the observed results in the limestone. The contact alteration is considered to be of the
pneumatolytic type, but it is difficult to understand how such would produce the biotite
rock (p. 93) or the felspar scapolite rock (p. 94). It is stated that the field relations
of these two rocks to the limestone in Harcourt and Dudley (p. 96) render it almost
certain that they are produced by the alteration of the limestone.

The second change is found (p. 97) " where granitic magma shatters the invaded
rocks and floats away the fragments in its moving mass." It may be pointed out that
not only is such an idea opposed to the present day conceptions of the manner in which
holocrystalline rocks of coarse granularity arise, but it only permits heat as the
metamorphic agent. These inclusions are stated to be similar to those which are
described in the granite gneiss at great distance from the junction with the limestone.
It is stated (p. 98) that the field evidence is scarcely susceptible of any interpretation

100 AUSTRALASIAN ANTARCTIC EXPEDITION.

other than that, under the influence of granitic intrusion, the limestone has, in the
zone of most intense action, been altered into an amphibolite. The limestone is found
to gradually pass into the amphibolite by the development in it of certain silicates.

A series of thin sections from a suite of specimens of the amphibolites are examined
(p. 103), which are claimed to illustrate the transitional stages of alteration. A
significant fact was noticed which was found difficult of explanation on the accepted
hypothesis. At one end of the series is a rock containing augite, calcite, and felspar,
and at the other end is a rock containing dominant hornblende with plagioclase and
subordinate augite. It was found that no passage existed between the characteristic
pyroxene of the recrystallised limestone and the characteristic hornblende. The
hornblende and pyroxene found together in the one section are fresh and show no
alteration of one to the other. The absence of transition is important, and shows that
the rocks dealt with are metamorphosed products in which both the augite and the
hornblende are primary metamorphic minerals, not secondary one to the other. In
metamorphic rocks of basic origin hornblende is frequently derived from the alteration
of augite, and in such cases the evidences of direct transition are abundant in thin
section. The occurrence is, in fact, highly suggestive of the phenomena of metamorphic
diffusion. The so-called gradual alteration may very well be a series of metamorphic
diffusion products between the limestone and the amphibolite. Were this same type
of argument accepted it could be shown at Cape Denison that amphibolites are the
product of alteration of granite. Metamorphic diffusion products naturally show
the chemical transition (p. 104), and will also yield an explanation of the microphoto-
graphs on Plates XIV., XV., XVI. of the memoir. The chemical analyses are useful
to show again the constant features of amphibolites. No. 16 (104) again falls into the
same division of the American classification as the typical amphibolite (No. 629) of
the Cape Denison rocks.

With the outlook of metamorphic diffusion one finds no evidence to disprove the
theory that along the junction of granite and limestone there has been a later intrusion
of basic rock, either in the form of a dyke or a boss. During the subsequent
metamorphism of the area the limestone and the basic rock have recrystallised and
the granite changed to gneiss. The enclosed fragments of basic rock in the gneiss might
well be considered as the torn-up fragments of a possible dyke. The transition of basic
rock to limestone is possible under the conditions which give rise to metamorphic
diffusion, and the isolation of blocks of an intrusive rock in the intruded rock is believed
to be an established possibility.

If a dyke mass has appeared along the limestone-granite boundary, it is not
surprising that (p. 97) among the many inclusions of amphibolite a careful search should
only lead to the discovery of one single fragment of coarsely crystalline limestone.
Thick bands of hornblende schist, which are looked upon as originally intrusive rocks,
similarly appear at the sides of, or within, outcrops of altered sedimentary rock* in

* British Geological Survey Memoir, 1907, op. oit., p. 238.

THE METAMORPHIC ROCKS OF ADELIE LAND. 8TILLWELL. 101

the Loch Maree and Gairloch district in the Scottish Highlands. The theory of an
amphibolite intrusion is quite consistent with the general statement (p. 115) that it is
almost a universal rule that the limestone near the contact is filled with various silicates
which have been developed in it, while the inclusions actually present in the granite
near the contact are composed of amphibolite or some allied rock.

The unrecognised presence of metamorphic diffusion is upheld by Adam's and
Barlow's discussion (p. 115) of the question " Has the granite gneiss anywhere actually
dissolved the invaded rock ? " Several instances are quoted to show that the granite
magma has been rendered basic by absorption of amphibolite. The peculiarities of the
gneiss are those which have been received long after consolidation, and the effects of
solution by the primary magma can only be interpreted with the greatest care. In
this case the gneiss is conceived as a molten magma and the conceptions of flowing
gneiss and the confusion of bedding and foliation are dangerously wrong and court
erroneous interpretation. It is stated (p. 117) that where the granite runs into a corner
between two tongues of amphibolite a basic development of the granite is seen also, due,
in all probability, to a partial solution of the invaded amphibolite ; that the products
of solution bear (p. 122) a marked resemblance to the grey gneiss. This " solution "
may readily be another case of that process which has been called metamorphic diffusion.

It is interesting to note (p. 114) that in one area of gneissic granite, that in the
township of Methueu, southern portion of the Kasshabog Lake, amphibolite inclusions
are abundant and appear in a linear belt parallel to the foliation. Again (p. 118) the
limestone, as shown on the Bancroft sheet, has a number of belts of amphibolite parallel
to the strike. This recognition of linear development is suggestive of dyke origin.

With regard to the amphibolites (Group 26) which appear as inclusions in the grey
gneiss, Adams and Barlow state (p. 121) that they are portions of rock forming the walls
or roof of the batholith which had fallen into the granite magma and had partaken of
its subsequent movements. He also adds that there is positive proof that this is the
correct and only explanation in several parts of the area. The positive proof, however,
is not convincing. Even if there were no reason to believe, as affirmed by Adams
and Barlow (p. 122), from the form or composition that they are ever due to magmatic
segregation, there will remain the hypothesis of a broken and disrupted dyke. I fail
to see even how the form, much less its composition, can preclude the hypothesis of
metamorphosed primary magmatic segregation products. We find on page 160, fig G,
an illustrative sketch in the memoir actually showing an amphibolite in the first stages
of disruption, and on page 76, fig. A, we see the characteristic lens shape similar to that
of fragments which have been proved to be part of dykes. Again, fig. B, page 76,
we see again the disruption of the amphibolite inclusions. We therefore see that the
Canadian evidence is sufficient, apart from the evidence from Cape Denison or the
Scottish Highlands, to show that a detached fragment of amphibolite, enclosed by
gneiss, is not necessarily to be regarded as earlier than the invading granite in the
metamorphic areas. This fact lends considerable support to the theory that many of

102 AUSTBALASIAN ANTAKCTIC EXPEDITION.

the Canadian amphibolites are portions of intrusive igneous rocks, frequently in the
form of dykes. It is very interesting, therefore, to read the footnote (p. 121) which
states that B. Frosterous, in his work in Finland, finds that the amphibolites which are
characteristic associates of the granite gneiss of Southern Finland are probably for the
most part altered dyke rocks. How far the dyke theory is applicable in Canada
cannot be at present determined.

There remains the third class of amphibolite which are considered to be derived
from the metamorphism of impure bands in the limestone series. The cause of the
metamorphism is assumed to be (p. 164) " undoubtedly the granite lying below and
exerting its action upward." The evidence relied upon is the interbanded character
(p. 165) of thin amphibolite bands in crystalline limestone on the Hastings Road on
lots 31 and 57, near the village of Ormsby. To the north of this the limestone bands
disappear and the amphibolite covers a great area. This evidence could be adequately
interpreted by the supposition of a primary intrusive mass which sends out tongues
into the surrounding limestone. Hence it is difficult to see how the evidence can carry
the burden of proof placed upon it by Adams and Barlow.

The " feather amphibolite," for which a sedimentary origin is claimed, seems to be
a different type of rock from the above amphibolites which the authors have called
the granular amphibolites. It is questionable whether the " feather amphibolite "
is a typical amphibolite. No analysis is given of the rock, so that it is impossible to
strictly correlate it with the normal amphibolites and to see whether it, like them, falls
into Grubeumann's group of amphibolites. The hand specimen (Plate XXXVII.)
shows striking differences, and to group the rock types illustrated in the memoir by
microphotographs (Plates IV., XXXVIII.) under the one generic term " amphibolite "
without modification is scarcely justified. The mode of occurrence, too, is different,
because the " feather amphibolite " is not found (p. 62) as inclusions in the granite
gneiss. The scientific term "amphibolite" will have much decreased value if made to
include dissimilar things.

The amphibolite containing orthorhombic amphibole, which is described as a product
of the extreme alteration of limestone by a granitic magma, is also an abnormal type
and grouped among the amphibolites without a consideration of its chemical composition.
It contains abundant gedrite and garnet associated with cordierite, quartz, biotite,
iron ore, rutile, and sillimanite. No felspar is present, and the rock could be better
described as a cordierite-bearing garnet gedrite schist. Though it is not stated whether
the rock is schistose or massive, the microphotograph (Plate XXXIX.) shows a schistose
character. Anthophyllite has been found in dyke rocks in the Lewisian gneiss*, but
this mineral composition suggests a chemical composition quite unlike an igneous rock.
No actual evidence of the nature of the origin of this rock is stated, but its composition
could be expected to be that which would result by the recrystallisation of an impure
magnesian limestone.

'^British Geological Survey Memoir, 1907, op. oit., p. 49.

THE METAMORPHIC ROCKS OF ADEL1E LAND. STILL WELL. 103

In another part of the memoir we have noticed a description (p. 127) of the nodular
granite of Pine Lake, township of Cardiff. The granite is metamorphic, and shows
in part curious nodules which average 2in. to Sin. in diameter. The nodules are
composed chiefly of quartz, muscovite, and sillimanite. Muscovite and sillimanite
are especially metamorphic minerals, and it is therefore likely that the nodules are
formed under metamorphic conditions. The chemical composition of the nodules is
decidedly not that of an igneous rock, and yet they occur in granite ! It seems,
therefore, feasible to appeal to metamorphic differentiation. The nodules have been
described as sometimes aggregating and forming foliated veins, and in this case the whole
vein must be looked upon as a metamorphic differentiate. The instance is in some
respects analogous to the courts of crystallisation that have been described in No. 143
from Cape Denison, but in the Canadian instance the product is coarser and more readily
recognised as metamorphic.

Summary.

Summing up, it seems that the positive statements by Adams and Barlow in the
memoir, and by Adams in his published summaries, concerning the origin of amphibolites.
are not sufficient. It cannot be considered as proved that normal amphibolites can be
formed by the alteration and recrystallisation of impure calcareous sediments. The
evidence that has been presented in detail can be shown to be explicable on the
supposition that the amphibolites are recrystallised basic intrusive rocks. It is suggested
that unrecognised examples of metamorphic diffusion have been interpreted as proof
of the change of limestone to amphibolite. The amphibolite inclusions in the gneiss
have not been shown not to be the alteration of primary basic differentiation magma
products or the isolated fragments of fractured and disrupted basic igneous dykes.
The authors have stated (p. 62) that it is a remarkable fact that the amphibolites
originating in the two very diverse manners often resemble one another so closely that
it is impossible to tell them apart. Such resemblance is to be expected on our alternative
hypothesis of igneous intrusion.

4. HIGHLANDS OF NEW JERSEY.

The results of the Cape Denison observations are also applicable to the phenomena
described by Fenner in the Highlands of New Jersey*. Fenner describes hornblendic
bands which may show (p. 598) remarkable continuity and parallelism and which may
pinch out for a distance, recontinue after an interval, and appear as knots or inclusions.
There is no mashing or granulation. There is frequently a sharp contact between the
" inclusions " and granite, although an interlocking of crystals may occur. Fenner
states (p. 601) that it appears in some cases that the basic minerals at the immediate
contact have become involved in the granitic magma without losing their parallelism,
so that a perfect transition is produced from the hornblende or biotite gneiss with

" Mode of Formation of Certain Gneisses in the Highlands of New Jersey," C. N. Fenner, Journ. Geol., voL XXII.,
pp. 594-612, 694-702. Under this head numbers in brackets refer to this publication.

104 AUSTEALASIAN ANTARCTIC EXPEDITION.

prominent foliation, through types in which, with increasing proportion of granite, no
parallelism of structure can be preserved. The direction of the bands is parallel to the
schistosity of the granite gneiss, which may be schistose as well as massive. In places
(p. 602) the dark minerals appear to have been taken up or digested by the magma
and to have crystallised out again in large blades.

Here, again, we find the interpretation of phenomena among gneisses coloured by
the conception that everything happened when the granite was molten. The gneissic
characteristics have been impressed after the consolidation of the granite by the influence
of stress. The processes which result in crystalline schists are entirely distinct from
those which result in normal igneous rocks. Solution, as we know it in igneous magmas
and liquids, is inapplicable to bodies which are to all intents and purposes solid. The
so-called " basic " minerals which are found exhibiting parallelism are metamorphic
minerals which have arisen in response to the impressed external conditions which
have caused the gneissic characters. The gradual transition, observed by Fenner, is
simply significant of metamorphic diffusion, and it is obvious in the field because
the two rocks in contact have strong difference in colour. In perfect accord with
the theory of metamorphic diffusion, he states (p. 604) that where there is the largest
amount of dark basic rock the adjacent granite contains the greatest quantity of
dark silicates, and where the inclusions are rare the granite is very light coloured
and nearly free from ferromagnesian minerals. From this observation he rightly
concludes that the dark minerals in the massive granite are derived from the basic
rock. The term " granite " is persistently used throughout the paper, but it is
questionable how far it is correct to do so, for the normal rock of the series seems to
be a granitoid gneiss. But Fenner is impressed with the conception that the granite
is intrusive into the rocks of basic composition with laminated structure after the
manner of lit-par-lit injection. He has tried to examine the process by which a
thinly fluid granitic magma could be injected between the layers of an original
sedimentary rock. In doing so he finds difficulty in understanding how these thin
walls of original rock could remain intact during injection and, at the same time, allow
transfusion of material.

Like many other workers Fenner has placed considerable value on the evidence of
the so-called inclusions and the transfusion. As has been pointed out in other cases,
" inclusions " in metamorphic areas do not necessarily signify an earlier age than the
enclosing primary granite. It does signify an age earlier than the development of the
metamorphic characters, but not earlier than the granite magma. There is nothing
in the evidence presented to show that these dark hornblende bands and " inclusions "
are not the metamorphosed product of thin basic dykes which, in the first place,
intruded the granite, and which, in the subsequent metamorphism, have had their
boundaries partially destroyed by metamorphic diffusion, and which have been fractured
and broken so that fragments can now appear detached and isolated as if they were
inclusions caught up by an invading magma. New mineral formations have

THE METAMORPHIC ROCKS OF ADELIE LAND. STILLWELL. 105

resulted from the stress, and the mashing and granulitisation of the primary constituents
are absent. In this manner, then, the rocks of the Highlands of the Hudson in
North- Western New Jersey can be correlated with occurrences at Cape Denison and
in other parts of the world.

The Assimilation Theory.

The assimilation theory for the production of hornblende schist, biotite and horn-
blende gneisses, which has been illustrated in the preceding paper by Fenner, has had
wide application. It has been recently emphasised by Cole, who has quoted a large
number of references, and who considers that the undermining and weakening of the
earth's crust by molten magma is the only interpretation of the widespread phenomena.*
Cole states: " Again and again strongly banded gneisses occur in which granitic material
alternates with sheets of hornblendic or biotitic schist. The biotitic varieties can often
be traced back into amphibolites. In places lumps of these amphibolites are seen,
streaked out at their margins, and providing a clear explanation of the dark bands
throughout the gneiss. This swallowing up of a mantle of basic material by a very
different and highly siliceous magma rising from below is seen to be a world-wide
feature, wherever we find the lower crust-layers brought up within reach of observation."
Further on he continues : ' We see the highly metamorphosed material further
attacked by the great cauldrons under it and becoming seamed with intersected veins.
Block after block has been caught, as it were, in the act of foundering into the depths.
In the gradual absorption of these blocks, and their penetration by insidious streaks
of granite, we see pictured on a few square yards of surface the destruction of a conti-
nental floor." Such is the catastrophic manner in which the metamorphic phenomena
are accounted for by assimilationists. It needs to be pointed out that as soon as any
of the features are demonstrated to be the result of metamorphic action as opposed to
igneous action, the theory is rudely shaken. If another explanation, e.g., metamorphic
diffusion, be found for the supposed gradual assimilation, a modified statement becomes
necessary. If, further, some of the supposed invaded crust be actually shown to be
younger in age than the granitic magma, the theory must completely crumble unless
recast. The widespread nature of the evidence is no more than the widespread
occurrence of lithologically uniform areas of the crystalline schists. We must include
Cape Denison among such areas of crystalline schists, and there we find no reason to
appeal to molten magmas for explanation of the observed phenomena.

Col*, Pre. Add, Sect. C.. B.A.A.S. Manchester, 1916.

106 AUSTRALASIAN ANTARCTIC EXPEDITION.

5. GEOLOGY OP THE LIZARD AND MENEAGE. Flett and Hill.*

This memoir contains an account of the metamorphic region of the Southern Lizard
surveyed and described by Dr. Flett. In this area an extraordinary number of rock
types are developed which have been the object of study of a large number of workers.
The earlier work has been summarised by Flett and incorporated in his own work, so
that a very full and clear description of the rocks and the rock problems is presented.
One, however, turns to the memoir to discover if any recognition has been made of true
metamorphic types, and the general impression obtained from the memoir is that the
rock problems have been treated rather from the standpoint of igneous rocks and of
rock magmas. It appears that the metamorphic character of the rock is considered
to be of subsidiary importance in comparison with its primary nature, and the value
of some of his conclusions is thereby lessened.

The serpentine, one finds, is looked upon as a modified igneous rock. It is continually
referred to as the intrusive body, and there is only occasional reference to the peridotite
from which it is derived. The serpentine in all cases is considered (p. 80) as a
weathered or decomposition product of olivine. This is the recognised origin of
serpentine in many cases, and for this reason Grubenmann says']" that the position of
serpentine among the crystalline schists, from which normal weather products are
excluded, is doubtful and uncertain. Nevertheless, Grubenmann considers that
certain occurrences must be included therein, and, therefore, creates the serpentine
family in his classification of the crystalline schists. It seems to us that Flett has
produced strong evidence for the similar inclusion of the Lizard serpentine.

We cannot agree with Flett that there is any reason (p. 97) to think that the main
serpentinisation of the Lizard peridotites took place at a comparatively late period of
their history, though there may be some development of serpentine in subsequent
weathering, as there is even in the example of primary serpentine recorded by
WeinschenckJ, and referred to by Flett (p. 97). Subsequent veins of chrysotile which
are unaffected by the schistosity seem to me to have very little bearing on the matter.
At Cape Denison there are numerous quartz segregation veins, independent of the
foliation, which may contain large and beautiful crystals of epidote, while epidote may
also appear along any joint plane of the gneisses and schists. These formations of
epidote are clearly subsequent to the development of foliation, yet no one would assume
from this that the formation of epidote in the amphibolite and schists was subsequent
to the development of foliation. In fact the epidote in some schists takes definite part
in the foliation, and has been proved to be a definite part of the mineral composition
of the amphibolite series.

* "Memoirs of the Geological Survey. England and Wales." Sheet 359, 1912. The numbers placed in brackets in
the following have reference to pages in this memoir unless a special reference is given,
t Op. cit., vol. II., p. 113.
t " Spezielle Gesteinskunde," 2nd edit., 1907. p. 184.

THE METAMORPHIC ROCKS OF ADELIE LAND. 8TILLWELL. 107

It is recorded (p. 20) that there are certain zones or belts in the serpentine which
have a rudely concentric arrangement. This zoned character is stated (p. 21) to be
" clear evidence that the serpentine is an intrusive stock that welled up and forced
outwards the surrounding schists." In this it seems that the very metamorphic imprint
of the mass is turned into evidence of intrusion. The metamorphic character of the
Lizard serpentine is more or less affirmed by Flett when he says (p. 70) that the
microscopic examination shows that very few specimens can be described as normal
igneous rocks ; that the normal poikilitic association of olivine and pyroxene is absent
except for traces in the least modified bastite serpentine or " weathered Iherzolite " ;
that the large pyroxene crystals are commonly broken or have their cleavage planes
twisted ; that tremolite is found to increase in quantity hand in hand with the
development of foliation and augen structure. As tremolite is a well-known
metamorphic mineral, the zone of tremolite serpentine may very well be but one phase
of the metamorphic product due to varying metamorphic conditions. The augen
structure, the foliation, and the schistose character that is described (p. 69) is strong
evidence of the metamorphic, not weathered, character of the rock. The absence of
foliation and schistosity in some parts of the serpentine body is no evidence to the
contrary, as massive textures are common among the crystalline schists. Further,
the serpentine has been demonstrated by Flett (p. 120) to be earlier in age* than the
group of rocks styled " Kennack Gneisses," which bear the very marked individuality
of typical crystalline schists. These schist characters have been impressed by certain
external metamorphosing conditions, and it is not reasonable to suppose that
the peridotite, surrounded now by metamorphic rocks, has escaped the whole of these
forces. If, then, the serpentine be acknowledged to be a primary metamorphic product
in any occurrence at all, it is reasonable to concede that the Lizard serpentine is likewise
a metamorphic product, and should therefore be treated primarily as such. If the
water content of serpentine is considered a barrier to the hypothesis, it must be
remembered that Grubenmann postulates considerable water in his epi zone of meta-
morphism in which hydrous minerals like chlorite are common.

Throughout the memoir there are frequent references to fluxion banding in the
serpentine in the so-called gabbros, and in the Kennack gneisses. Now, fluxion
structures are true igneous structures developed by movement in the magma during
consolidation. If they appear in metamorphic rocks they can only do so as
relic structures, and if the decrystallisation during metamorphism is intense, fluxion
structure will be very difficult to recognise. It therefore becomes incumbent to examine
the evidence put forward in order to discover if a different aspect will create a different
interpretation. Before doing so, however, we will more fully explain the standpoint
from which the question is approached.

* In a later publication, " The Crystalline Rocks of the Lizard," issued as a pamphlet by Bowes and Bowes, Cambridge,
Prof. Bonney doos not accept this opinion of Flett's and insists on his own former interpretation, viz., that the Kennack
gneisses are older than the serpentine. Whether Bonney be correct or not, the argument remains unaffected, as far as the
Survey memoir is concerned.

108 AUSTRALASIAN ANTARCTIC EXPEDITION.

Though Becke has explained in 1892 that dynamo-metamorphism may involve
either complete recrystallisation, resembling contact metamorphism, or granulation,
there arose a tendency in the following years to associate only mechanical structures
with pressure effects. This tendency produced Weinschenck's attack on the use of the
term " dynamo-metamorphism " which is declared to be vague, and to connote the
personal interpretation of the user. Weinschenck's attack is directed mainly against
the idea of purely mechanical transformation of rocks. Grubenmann* has approved
of this criticism, but he opposes the introduction by Weinschenck of the terms " piezo-
crystallisation " and "piezo-contact-metamorphism." " Piezo-crystallisation " is the
crystallisation which proceeds in a magma under the influence of a tangential thrust
and produces primary pressure banding. ' Piezo-contact-metamorphism " is the
influence exerted by the intrusive magma and its attendant gases or vapours on
the enveloping rocks during " piezo-crystallisation." The ideas are introduced by
Weinschenck to explain the large area of metamorphic schists surrounding the central
massif of the European Alps. Grubenmann maintains that these terms are unnecessary,
as the kind of metamorphism is explained as soon as the fundamental physico-chemical
factors are denned. The special terms are superfluous provided we consider collectively
the temperature, the uniform pressure, the stress (non-uniform pressure), and the factor
of the individual substance at the time of the metamorphism. In this Grubenmann
is dealing with rocks which undergo no essential change in composition during
metamorphism. Metasomatic changes and all rocks whose composition is changed by
igneous exhalations and heated waters are not considered. Johnston and Nigglif
have reached the same conclusion in their exposition of the general principles underlying
metamorphic processes.

Grubenmann, Van Hise, and others have attempted to classify these conditions
of temperature, etc., by reference to zones of metamorphism when each zone is
characterised by special conditions which grade into the special conditions of the
neighbouring zone. Crook has adversely criticised the value of the conception of
metamorphic zones as given by Van Hise, partly because the crustal zones fail to provide
a basis of genetic classification of rocks either in a general or metamorphic sense, and
partly because he believes in the " paramount importance of igneous intrusions as agents
of metamorphism "J. Provided, however, that the zones are made sufficiently
definite, and are not made dependent on depth within the earth's crust, they are very
useful in defining the sets of conditions under which changes occur.

Rock flowage is a term that frequently appears in portions of the literature on
metamorphic geology, but no use is made of it in our discussion. It is, no doubt, a
useful term in structural geology, when, for structural considerations, the earth's crust
is divided into a zone of fracture and a zone of rock flowage. Deformation of the rocks
occurs in the former by fracturing, and in the latter by rock flowage involving a

* " Die Kristallinen Schiefer," vol. I., p. 46.

t " The General Principles underlying Metamorphic Processes," Johnston and Niggli, Journ. Geol., voL XXI., p. 63.

t " The Genetic Classification of Rocks and Ore Deposits," T. Crook. Min. Mag., vol. XVII., July, 1914, p. 55.

THE METAMORPHIC ROCKS OF ADELIE LAND BTILLWEI.L 109

permanent change of form without conspicuous fracture. This permanent change is
supposed to be accomplished by interior readjustments of rock substances by chemical,
mineral, and mechanical changes, and produces the slaty cleavage and schistose
structures. The latter are included by Leith under the one term " flow cleavage "*.

We find, however, that there are objections to the use of these terms in a treatment
of metamorphic rocks. It is desirable to analyse and distinguish more carefully the
interior readjustments which are combined in rock flowage. We need to distinguish
between slaty cleavage and crystalline schistosity, which are separated by the degree
of the all-important recrystallisation, while all the physico-chemical conditions of
recrystallisation are not included in the zone of rock flowage. A zone of rock flowage
implies a zone where there is a dominating stress combined with a hydrostatic pressure,
and it will not include a zone of very high hydrostatic pressure with less important
stress ; and under such conditions we picture certain recrystallisations. As a general
term " rock flowage " includes too little and as a restricted term too much.

Primary gneissic banding or primary pressure banding has been frequently recorded
on the margins of igneous masses. They are terms which are frequently supposed to
involve movement in the rock magma. We consider, however, that the primary pressure
banding, apart from injection banding, can be considered as a metamorphic texture
without any appeal to fluxion or moving magma, and that it is identical with the
schistose structure produced by recrystallisation under stress. The coarse granularity
and holo-crystalline character of even-grained plutonic rocks indicate that there has
been no movement in the magma during its consolidation. Large crystals do not grow
uniformly in moving solutions, and it is difficult to see how the symmetrical arrangement
of the mineral constituents in a schistose margin can be produced by movement in a
viscous, semi-solid rock. For the injection of rock magma against the weight of the
overburden of enveloping rocks we must postulate big orogenic forces. If these forces
continue after the magma has been brought to rest they are distributed through the
magma only as a hydrostatic or uniform pressure. Cooling and consolidation proceed
under the uniform pressure as in any normal case. Though the crystallisation may
be uniform throughout the whole mass, we imagine that the outer margin will become
solid before the centre. If, after the development of this solid crust, the pressure causing
intrusion be still maintained in the molten portion, the hydrostatic pressure in the molten
portion will be exerted normally on all parts of this crust, which is then affected as if
subjected to a stress (non-uniform pressure) or a squeeze (fig. 10). This stress, combined
with the other factors of temperature, etc., produce, not movement, but the stable
molecular rearrangement and the gneissic banding in the manner most recently
expounded by Grubenmannf, Johnston, and NiggliJ, and proved experimentally by
Wright. This stress may also possibly be produced by the expansion caused by the

* " Structural Geology," C. K. Leith, Constable & Co., 1914, p. 76.

f Op. cit., vol. I., p. 42.

t Op. cit., p. 610.

{ " Schistosity by Crystallisation," F. E. Wright, Amer. Journ. Sci., vol. 22, p. 224.

AUSTRALASIAN ANTARCTIC EXPEDITION.

crystallisation of the central portion. When the pressure is relieved the crystallisation
proceeds normally and the centre becomes massive. Non-uniform pressure is essential
to produce the gneissic banding, and it cannot be applied to a liquid. Therefore, not
until the magma has become frozen and solid can the gneissic character be impressed
upon it. We find no satisfaction in Van Hise's statement* that " the parallel orientation
of minerals in the original gneisses formed from magmas is due to differential stress
during the primary crystallisation of the rocks."

-U+.l+.J.+'-U ~~-^ +"-,_ "*" ~*"_U-|--L_~ t ~-l_" ( ~

+ -J- 4- ~t- -t- -t- -h -p -f- -f- -H
4--H4- 4-4--I-4-4-4--4-

4- 4- + -h .JxdUJ- 4- 4- 4- >
4- 4- 4-

Fig. 10.

DIAGRAM SHOWING THE FORMATION OF A SCHISTOSE MARGIN WITHOUT

ANY APPEAL TO MOVEMENT IN THE MAGMA. HYDROSTATIC PRESSURE

IN THE MAGMA is EXERTED AGAINST THE MARGINAL RIM AS NORMAL

STRESSES, WHICH INDUCE RECRYSTALLISATION AND SCHISTOSITY.

In Wright's experiments he aimed at producing crystallisation, and with it
schistosity, from solution under strain. He does it by heating a solid, viz., glass, under
strain. He looks upon glass as an undercooled liquid, but it can equally well be named
a solid solution, as glasses fall within Van HofE's definition of a solid solution. He
really shows that crystallisation within a solid, not crystallisation in a liquid under
strain, produces the schistosity. The difference between the liquid and the solid
becomes, of course, a fine point when viscous substances are being considered, though
the difficulty could be arbitrarily settled by looking upon a viscous mass as solid as
soon as it can take a stress. It cannot be inferred from his experiment that schistosity
can be produced in the first crystallisation of a cooling magma.

From these considerations it is necessary to strongly oppose the use of the terms
" flowing gneiss " and " fluxion gneiss " which constantly appear in geological literature
dealing with metamorphic rocks. To be consistent it is also necessary to oppose the
introduction of the term " injection foliation," proposed by Flett, because it likewise
carries false meaning. A term " injecting banding " may be appropriate for the banding
which Geikiet considers is produced by the injection of aplitic magma into dark schists

* " Treatise on Metamorphism," C. R. Van Hise, p. 782.
t " Text Book of Geology," A. Geikie, p. 256.

THE METAMORPHIC ROCKS OF ADELIE LAND.-STILL\VELL. Ill

with biotite and hornblende. Banding of a similar nature is found at Cape Denison,
where thin threads of basic magma intruded the granodiorite and have partially main-
tained their integrity during the metamorphism. Such banding is an igneous structure
which the metamorphism has not effaced, and is analagous to a lit-par-lit injection on a
small scale. Both fluxion banding and foliation are insisted upon by Flett in the Lizard
serpentine, in the gabbros, in the Kennack gneisses, and in the granite gneiss. But
there does not seem to be any great difference between the two except a mineralogical
one. The fluxion banding (p. 22) is marked by olivine and pyroxene, and the foliation
by serpentine, and it is stated that they are parallel to one another and are so intimately
connected that they seem to be parts of one phenomenon and must be closely allied in
origin. We also notice that the figure referred to both on page 22 and page 68 as an
illustration of fluxion banding is titled " The foliation banding in the serpentine." When
we consider the Lizard serpentine to be a metamorphic rock belonging to the epi zone
of metamorphism, it is probable that the so-called fluxion banding and foliation are
part of one phenomenon.

" Injection foliation " is the conception which Flett puts forward to account for
the remarkable foliation of the system of gabbros dykes that penetrate the serpentine.
He states (p. 94) that the foliation of the dykes is in nearly all cases parallel to the
margin of the dykes, whatever be its course ; that it is equally well marked in the
horizontal, vertical, and inclined dykes, and when one foliated dyke cuts another each
has its own direction of foliation ; that consequently the schistosity at once suggests
fluxion movement ; that where a dyke bends the schistosity bends with it, and if a dyke
forks, each branch is foliated parallel to its length. It is therefore supposed (p. 23) that
" the dyke rock was forced upwards in a plastic state under severe pressures, and the
foliation was produced as the injection went on, being really an injection foliation."
If this is so, the mineral constituents are arranged parallel to the direction of pressure
instead of at right angles to it. Further, if non-uniform pressure is an essential factor
in the production of gneissic structures, then it is not right to connect the foliation of
these Lizard dykes with their infilling with liquid magma. Not till the dyke matter
has become solid can the non-uniform pressure affect it, and we can only avoid the
difficulty by permitting the plastic magma to have the properties of both a liquid and
a solid. This seems to us unreasonable, and we therefore think that the super-induced
foliation must find some explanation independent of any movement along the dyke
channels. The heat of the primary magma may be a contributing factor, but its
supposition is unnecessary, as other sources of heat can be found. Rejecting, therefore,
the theory of injection foliation, we tender the following for consideration.

We have explained that there is good reason to think that the Lizard serpentine
has been developed from a peridotite by metamorphic processes. The development of
serpentine by meteoric waters, if present, is only subsidiary. The serpentine has been
forced to develop during geological time under the physico-chemical conditions in the
depths of the earth crust that exist in Grubenmann's epi zone of metamorphism or

112

AUSTRALASIAN ANTARCTIC EXPEDITION.

in Van Hise's zone of katamorphism. External energy has been impressed upon the
system, and olivine and pyroxene have passed over into serpentine, liberating a large
quantity of heat and molecular energy. The accompanying figures, quoted from Van
Hise's tables,* show that this type of alteration in peridotites is accompanied by the
formation of minerals of lower specific gravity and larger molecular volume.

Specific Gravity.

Molecular Volume.

Olivine ...

3-4

50-40

Enstatite ....

3-2

31-22

Ausite . .

3-4

63-38

Bastite

2-6

118-11

Tram nlit.p. ...'........... ................. , , ,

3-0

138-44

Serpentine . .

2-57

107-2

According to Van Hise's interpretations of the reactions, the following figures express
the increase in volume of the individual systems :

Mineral Change.

Heat
Change.

Volume Increase Per Cent.

Serpentine from enstatite . .

+ K

14-25 or 38-26, according as all compounds do or do
as solids

not separate

Serpentine from diopside . .

+ K

56-32 or 0-44, according as all compounds do or do
as solids

not separate

Bastite from enstatite

+ K

22-77 or 46-87, according as all compounds do or do
as solids

not separate

Tremolite from diopside . .

+ K

5-68 or 10-15, according as all compounds do or do
as solids

not separate

Serpentine from olivine . . .

+ K

29-26 or 15-19, according as all compounds do or do
as solids

not separate

Tremolite from olivine

-K

37-13 or 12-43, according to composition of olivine
12-29 volume decrease

In all cases except the last, which Flett has only stated as a possibility in the Lizard
serpentine, there is increase in volume and liberation of heat. Hence, in a large mass
of rock like the Lizard serpentine, there will be enormous expanding forces exerted on
the enveloping rocks and on any dyke sheets that happen to traverse the serpentine.
These forces will act as compressive stresses approximately normal to the dyke plane
whatever may happen to be its direction. The serpentinisation thus causes the squeezing
of the dyke rock. Large quantities of heat are liberated at the same time, and the
compression is met by the molecular rearrangement of the constituents with the
development of those forms and shapes which are most stable against the imposed

* " Treatise on iletamorphism," p. 196.

THE METAMOKl'HIC ROCKS OF ADELIE LAND. STILL WELL. 113

conditions. As a necessary consequence gneissic banding appears parallel to the trend
of the dyke. The metamorphic character of the so-called gabbro dykes is therefore
caused in the first place by the same external metamorphic forces which are producing
serpentinisation, and in the second place by the simultaneous internal metamorphic
forces that develop during serpentinisation. Irregularity is, therefore, to be expected,
and it becomes easy to understand why it should be commonly observed (p. 98) that
the gabbro dykes are wrenched off by irregular planes of movement. The general
theory would cause the impression recorded by Flett (p. 98) that these veins have been
caught up in powerful but irregular movement ; that they have been softer and more
plastic than the peridotite which surrounded them and have yielded to stresses ; that
there has been an internal shearing which has crushed the minerals and set up a rough
foliation. It also yields explanation why there should sometimes be a development of
foliation (p. 96) in the serpentine walls parallel to a dyke junction. This foliation
in the serpentine is always parallel to the foliation in the dyke, is very similar to it in
character, and in some cases becomes gradually lost as one passes outward from the
edge of the dyke.

Apart from external pressure the expansion of the peridotite will be more or
less symmetrical in all directions from the centre of the mass, and so there will arise
approximately hydrostatic pressure in the centre and non-uniform pressure towards
the margins. Such will be more or less the case, but the external pressure will tend to
destroy the symmetry and move the region of hydrostatic pressure away from the
centre. There will therefore arise concentric zones of similar pressure throughout
the mass. Where the hydrostatic pressure prevails there can arise the coarsely
crystalline massive product, and where non-uniform pressure dominates there can
arise the schistose character. Consequently there may arise the approximately zonal
arrangement of tremolite serpentine around the bastite serpentine as noted by Flett
(p. 20). The dunite serpentine may be the result of a marginal facies in the primary
peridotite. It has a lower alumina and lime content than either the tremolite or bastite
serpentine. The microscope shows that it consists entirely of olivine and its alteration
products. As an absorption of heat is required for the formation of tremolite from
olivine, we have only to suppose that the heat factor was not sufficiently strong on
the margin to yield a tremolite schist, and serpentine without tremolite would be
found. If, then, we are to give the Lizard serpentine a metamorphic history of this
kind, we cannot accept Flett's statement (p. 68) that fluxion banding in the Lizard
serpentine is a very common phenomenon until the metamorphic processes have been
fully considered and the laminae which are marked by olivine, or by olivine and pyroxene,
or by olivine and tremolite, have been shown not to be comparable with Grubenmann's
crystallisation schistosity texture. It would be very remarkable if fluxion banding is
always parallel to the subsequently induced foliation which, in turn, is parallel both
to banding and foliation of the adjacent hornblende schists and to the actual line of
junction of the serpentine and the hornblende schists. Apart from fluxion banding
this parallelism is a necessary consequence of the suggested explanation.

Seriee A, Vol. ra., Part 1 H

114 AUSTRALASIAN ANTARCTIC EXPEDITION.

It is very interesting to note that Flett finds (p. 74) that there is a transition between,
or an intermixture of, the serpentine and the adjacent hornblende schist at the junction
at Pol Cornick ; that there is little evidence of crushing to be found in the slides cut
from the junction ; that there is a little development of tremolite, and the minerals
are exceptionally fresh. It seems reasonable to infer that a primary metamorphic
product has arisen at this point as the result of metamorphic diffusion, with the destruc-
tion of the hornblende schist-serpentine boundary. This means that the hornblende
schist and serpentine have suffered together similar metamorphic processes, and hence
we have further justification in treating the serpentine mass as a metamorphic product
rather than as a weathered peridotite.

We may now, perhaps, go further and make reference to the inter-banding of schist
and serpentine described (p. 75) as due to folding. Canoe-shaped infolds of fine, rotten
hornblende schist are mentioned. These remind us of the lenticular inclusions of the
amphibolite dykes in the granodiorite gneiss that have been torn off the main dyke
channels at Cape Denison during the metamorphism of the area. Such " infolds "
have been referred to as " inclusions " by Adams and Barlow in the Haliburton and
Bancroft area in Canada, and considered to be evidence of the lesser age of the enclosing
rocks.* As we have found that this is not necessarily the case, we must apply great
caution in the interpretation of the folding in metamorphic areas until the field
phenomena are better understood. More especially as it is observed by Flett (p. 99)
that no part of the gabbro has been folded, and all the movement seems to have taken
the form of internal shearing in a large unwieldy mass which would not fold.

Flett distinguishes two series of hornblende schists in the Lizard, and both are
determined as metamorphosed igneous rocks. One series is spoken of as the Landewed-
nack schists (p. 46) and the other the Traboe schists (p. 50). The Laridewednack schists
are considered to be older than the Man of War gneisses, and the Traboe schists are
younger, though Flett acknowledges their inter-relation is difficult to make out. Apart
from differences in weathering, the Traboe schists are distinguished from the Landewed-
nack schists (p. 51) by the following characters :

1. Paucity of epidote.

2. Its relation by folding and transition to the serpentine, for no case is known

where the Traboe schists occur at any considerable distance from the
margin of the serpentine.

Now the presence or absence of epidote in an hornblende schist depends upon the
temperature factor in the metamorphism. Epidote will only form if the temperature
is not too high to drive the water out of the molecule. The inference is that the Traboe
schists have been metamorphosed at a higher temperature than the Landewednack
schists a fact which has no bearing on relative age. If we neglect the doubtful evidence
of folding there only remains the fact of transition which, it is maintained, can be

" Geology of the Haliburton and Bancroft Areas, Ont.," Adams and Barlow. Mem. No. 6 Can. Geol. Surv., p. 62.

THE METAMORPHIC ROCKS OF ADEL1E LAND. 8TILLWELL. 115

interpreted as an example of metamorphic diffusion. With this explanation it is to be
expected that the Traboe type of schist should only be found near the margin of the
serpentine whose metamorphism has been shown to develop additional heat. Hence,
on the present available evidence the Traboe and the Landewednack schists can only be
looked upon as slightly different facies of the same primary rock. There is at present no
valid reason for discrimination in age. Further, the possibility that the Traboe schists
represent an igneous rock intrusive into the serpentine has not been disproved.

Accompanying the Laudewednack schists (p. 50) are streaks and nodules of epidosite,
and it is very interesting to note that Flett looks upon these as due to some kind of
segregation during metamorphism. They are similar to the Cape Denison epidosite,
which we have called a metamorphic differentiation product. Epidosites also occur
abundantly (p. 36) among the green schists and granulites of the old Lizard Head series.
Here Flett considers them as facies of the other rocks rather than types entitled to
recognition as a distinct group, and states that some of them are segregations, nodules,
and vein-like masses in the hornblende schists produced either by weathering before
shearing, or by chemical segregation during movement ; that others are probably due
to weathering of a fine type of hornblende schist and hornblende granulite ; that others
are quartzose granulites in which epidote may represent volcanic detritus or ashes,
or may be a secondary infiltration during metamorphism. It seems to me that such
an aspect is only possible when metamorphic rocks are denied their own special
individuality, because epidosites do not fit into any sedimentary or igneous rock group.
In this case again the epidosites can be explained as metamorphic differentiation
products.

The gabbro dykes (p. 81, et seq.), whose foliation we have contended is in no way
connected with their injection, occur only in the serpentine. In addition to the dykes
there are intrusive bosses, the largest of which is the Crousa Downs Gabbro. These
gabbros were grouped by Teall into

1. Gabbro Schists.

2. Flaser Gabbro.

3. Normal Types.

Flett also considers them in this manner. The treatment from the standpoint of igneous
rocks is evident here in the nomenclature. Though the dykes are known to be
metamorphic, the igneous rock term " gabbro " is applied. The gabbro schists are
coarse-grained saussuritic hornblende schists which are a well-known type developed
from dolerite or diabase. The present coarse-grained character is not necessarily evidence
of the original coarse granularity of the gabbro. The flaser gabbro, which is the
prevailing type, is also a metamorphic rock, because it includes all those gabbros which
exhibit distinct evidence of crushing and recrystalhsation. The normal gabbros are
restricted to the neighbourhood of Coverack, and as one recognises the general
metamorphism of the Lizard, and also as a weak " fluxion banding " is mentioned,

116 AUSTKALASIAN ANTARCTIC EXPEDITION.

one must question the term " normal gabbro " for the least altered members of the
series. In the description of the normal gabbro it is stated that the felspar contains
cracks due to incipient fracture and crushing, as well as cloudy spots of saussurite ; that
the augite appears as diallage and brown hornblende is often associated with diallage ;
that the olivine " weathers " to serpentine and magnetite, and sometimes to dense
aggregates of talc ; that hypersthene is rare and only seen as thin borders to clusters
of olivine ; that there are reaction rims of diallage or hypersthene around the olivine
and of a fibrous radiate mineral between felspar and olivine ; that there is a typical
gabbroid structure. It is well to remember that Grubenmann has stated * that the
gabbroid structure may be a variety of the granoblastic structure. The presence of these
characters inclines us to view the so-called " normal gabbro " not as an absolutely
unmodified igneous rock. Now Grubenmann recognises | the difficulty in the separation
of certain crystalline schists with massive texture from igneous rocks. But, however,
if the gabbros are pre-serpentinisation in age, and if the schistositv be produced in
the manner suggested, we would scarcely expect normal gabbro to remain as such,
and we might expect part of the gabbro to be involved in pressures of the hydrostatic
type. It may, therefore, be better to accept the small amount of evidence and
recognise the affinity of the " normal gabbros," as well as the flaser gabbro and the
gabbro schists, to the metamorphic types. If, on the other hand, the igneous character
be maintained as dominant, the small area at Coverack must not be looked upon as
normal, but as a relic of the original gabbro. The metamorphic, not the igneous, state
is here the truly normal character. The troctolite at Coverack possesses saussuritised
felspar, serpentinised olivine, and reaction rims, and, therefore, also possesses meta-
morphic traits.

There remain for comment the Kennack gneisses (pp. 119, et seq.) which Flett and
other workers consider to be the crux of the Lizard problem, though we think that the
serpentine has a claim to that distinction. The Kennack gneisses, however, present
a distinct problem in themselves. Flett has brought forward strong evidence to show
that these gneisses are metamorphosed igneous intrusions in the form of stocks, sills,
dykes, veins, and networks into the original peridotite. If they be pre-serpentinisation
they would necessarily be subjected to the same metamorphosing action as the gabbro
dykes. The dykes will be foliated parallel to their length and the stocks parallel to their
margins. Any peridotite blocks that had been caught up by the invading magma
before the serpentinisation would yield to the serpentinising forces in the same manner
as the large mass. Presence of the serpentine inclusions in the sills or gneiss is not
necessarily evidence that serpentinisation occurred before the intrusions of the gneisses.

The complexity of the Kennack gneisses is due to the mixture of primary basic
rock and primary granite rock. With regard to the " Blocky Gneisses," in which blocks
of basic rock are included in granite gneiss, the fact that the blocks are seldom angular,
but are usually rounded or lenticular, does not show that they were being dissolved by
the granite magma. Where Flett records the evidence of acid magma diffusing into the

* Op. oit.. voL L, p. 79. t Op. oit., vol. II., p. 19.

THE METAMORPHIC ROCKS OF ADELIE LAND. 8TILLWELL. 117

inclusions and yielding an intermediate rock, we may only have a further example of
metamorphic diffusion. In the " Blotched and Streaky Gneisses " the basic material
may be again, as in other cases, the fragments of a disrupted and broken dyke. Such a
possibility is suggested because this type passes into the " Flow Banded Gneisses,"
which may be an injection banding which has preserved its entity throughout the
metamorphism.

The granite gneiss, which occurs within the serpentine area and is intrusive into it,
would be affected on its margin by the serpentinisation in the same manner as the gabbro
dykes and Kennack gneisses. Flett describes (p. 41) how. the older serpentine is often
converted for a short distance from the contact into a soft greenish rock consisting
mainly of talc and tremolite, with sometimes anthophyllite and chlorite. He considers
(p. 142) that these contact phenomena are pneumatolytic changes due to hot vapours
and liquids given off by the cooling gneisses. An analysis is quoted to illustrate the
difference in composition between the tremolite rock and the serpentine. As, however,
tremolite and talc are better known as metamorphic minerals than as pneumatolytic
minerals, it is very probable that Flett is dealing with a metamorphic product rather
than a pneumatolytic one. The observed difference in composition can be readily
explained by metamorphic migration of material.

Summary.

The memoir that is criticised contains an account of the extraordinary number of
rock types met with in the Southern Lizard. The area is one of exceptional
metamorphism, yet it is considered that the study has been approached more from the
standpoint of igneous rocks than from the standpoint of metamorphic rocks. This has
occurred because the metamorphic rock has not been given the same individuality
that is given to igneous and sedimentary rocks. An attempt has been made to state the
metamorphic standpoint and to demonstrate the results obtained thereby.

The Lizard serpentine has been viewed as a metamorphic rock, foliated in part.
We think that the term " fluxion banding " is in most cases a misnomer when applied
to these metamorphic rocks, and that the term " flowing gneiss " is also founded on
misconception. The foliation of the gabbros dykes which penetrate the serpentine
can receive adequate explanation from the combined effect of the external
metamorphosing forces that cause serpentinisation, and the internal pressure and heat
developed by the serpentinisation of the original periodotite. The introduction of the
term " injection foliation " is not necessary, and it does not assist the explanation of
the remarkable dyke foliation.

Descriptions in the area have corresponded to the conceptions of metamorphic
differentiation and metamorphic diffusion. Transition is observed between the
serpentine and the adjacent hornblende schists. Diffusion types are also present in the
so-called " intermingling of acid and basic magma " in the complex Kennack gneisses.

118 AUSTRALASIAN ANTARCTIC EXPEDITION.

Epidosites, that are found in association with the Landewednack schists and with the
green schists and granulites of the old Lizard Head series, can be considered as
metamorphic differentiation products.

At present there is no evidence to discriminate in age between the Landewednack
hornblende schists and the Traboe hornblende schists. The observed differences
can be explained by a varying temperature factor during the metamorphism and
by metamorphic diffusion. Caution is necessary in order that the rocks termed
" gabbro " are not misinterpreted. The flaser gabbro and the gabbro schist are
decidedly metamorphic rocks. It is also possible to consider the " normal gabbro "
as a slightly metamorphosed rock, but if it be maintained that the " normal gabbro "
is a true igneous rock it must be recognised that it is not the normal rock of the area
but a relic gabbro. In the latter case it is a remnant of the original gabbro that has
escaped metamorphism.

The foliation of the dykes and sills of the Kennack gneisses can receive the same
explanation as the foliation of the gabbro dykes. Serpentine inclusions in the sills are
not evidence that serpentinisation occurred before the intrusion of the gneisses. They
are evidence that the gneissic sills are the metamorphosed equivalents of dyke sheets
that invaded the original peridotite. Finally, talc and tremolite are not to be considered
pneumatolytic products and associated with the intrusion of the granite gneiss into the
serpentine. They are metamorphic products, and must receive a metamorphic
explanation.

6- THE VALUE OF CHEMICAL CRITERIA IN IDENTIFYING THE ORIGIN OF METAMORPHIC

ROCKS.

We may perhaps be permitted to add to the discussion on the value of chemical
criteria in determining the origin of metamorphic rocks. Bastin* has quoted authori-
tative opinion on the value of these criteria, and has summarised and discussed the
characteristics of fresh foliated rocks of sedimentary origin. The features indicating
sedimentary origin in a chemical analysis are set out by Bastin thus :

1. Dominance of MgO over CaO is strong evidence.

2. Dominance of K 2 over Na 2 has lesser critical value, but is suggestive.

3. Presence of considerable excess of A1 2 S over and above the 1 to 1 ratio

necessary to satisfy lime and alkalies.

4. High silica may be indicative when supported by other criteria.

When three or all of these relationships hold good the evidence for sedimentary
origin may be regarded as practically conclusive.

With the appearance of Adam's and Barlow's memoir on the Haliburton-Bancroft
area, the value of these criteria seemed to be lessened, inasmuch as Adams and Barlow

" Chemical Composition as a Criterion in Identifying Metamorphosed Sediments," E. S, Bastin, Journ. Geol., vol.
XVII., 1909, p. 445.

THE METAMORPHIC ROCKS OF ADELIE LAND 8TILLWELL. 119

maintained that precisely similar amphibolite rocks could be derived from both altered
limestones and igneous rocks, and that there was no alternative. While we remain
unconvinced in this respect this difficulty is not acknowledged.

Trueman* also discusses chemical criteria as well as the criteria of texture and of
zircon grains. He finds that texture is not successful in determining origin, but he
demonstrates the use of zircon grains, though this method can have only limited
application. Trueman describes the alteration of quartzite into sericite schist
at Waterloo, Wisconsin. The application of the zircon criteria confirms the chemical
work of J. H. Warner, and shows that all gradations exist between the normal quartzite
and the most highly developed sericite schist. The zircon criteria demonstrate that
the sericite schist cannot represent a former argillaceous layer in the quartzite, nor has
the change of composition been affected by the introduction of igneous material from
without. With the development of the sericite schist he finds a simultaneous removal
of silica, and he notices the presence of quartz stringers in some of the bands. The
observation might, perhaps, be stated that the sericite schist with quartz, as a
metamorphic differentiation product, is formed during the metamorphism of a quartzite
under special conditions. Dwelling on the composition of the platy minerals that
develop under the special conditions, he is inclined to attach small importance to
Bastin's criteria, even though they are satisfactory in his particular case. The actual
transfer of material is the difficulty to Trueman, whereas Bastin formulated his criteria
on the supposition that no essential change occurred in chemical composition during
the metamorphism.

Bastinf replies to Trueman's point, and further demonstrates the value of his
criteria in special cases when used as an adjunct to textural and structural evidence.
He then turns to the question of transference of material during metamorphism and
acknowledges that the value of chemical criteria is part of the broader question of the
actual extent to which transfer of material takes place.

Leith and Mead have devoted considerable space to this question in their recent
publication^ titled " Metamorphic Geology." These authors emphasize that the use of a
chemical analysis in this respect depends on two fundamental assumptions, viz., (1) that
the rocks before rock flowage (metamorphism) had a distinctive composition sufficient
to identify them as sediments ; and (2) that there is no essential change in composition
during metamorphism. We think it will be generally admitted, with Bastin, that
sediments do have, in the great majority of cases, a distinctive chemical composition,
and that the crux of the problem lies in the second assumption.

At the outset of their discussion Leith and Mead assume that the criteria set forth
for sedimentary origin by Bastin may be used conversely to prove igneous origin. But

* " The Value of Certain Criteria for the Determination of the Origin of Foliated Crystalline Rocks," J. D. Trueman.
Joura. GeoL, vol. XX., 1912, pp. 229-258, 300-315.

t " Chemical Composition aa a Criterion in Identifying Metamorphosed Sediment*," E. S. Baitin, Journ. Geol., vol.
XI., 1913, p. 193.

J " Metamorphic Geology," C. K. Leith A W. J. Mead. Henry Holt A Co., New York, 1915, p. 226, el teg.

120 AUSTRALASIAN ANTARCTIC EXPEDITION.

this assumption seems to be scarcely warranted from Bastin's paper. Leith and Mead
then attempt to subject the criteria to rigorous proof by application to a number of
analyses of sericite schists, weathered and hydrothermally altered acid and basic igneous
rocks, chlorite schists and hornblende schists. In the tests they fix the critical value
of the A1 2 3 content at 5 per cent, excess over the 1 to 1 ratio with lime and alkalies,
i.e., with more than 5 per cent, excess Leith and Mead consider that the criterion for
sedimentary origin is satisfied, and less than 5 per cent, excess indicates igneous origin.
Yet Bastin has stated that a 5 per cent, excess is only sufficient to cause a suspicion
of sedimentary origin and that a 10 per cent, excess is necessary to make the sedimentary
origin extremely probable. With these interpretations of the meaning of the criteria,
Leith and Mead find, as a result of the indiscriminate application to rock analyses,
without any consideration of other metamorphic characters, that the chemical criteria
have value only when carefully qualified and limited, and that they always fail when
other criteria fail. It needs to be added that the applicability of the criteria to
weathered and hydrothermally altered igneous rocks is not important, because it has
yet to be shown that weathering before the development of foliated structure is not
negligible, as assumed by Bastin.

The problem can be simply stated in terms of metamorphic diffusion and
metamorphic differentiation. Chemical composition may be a very important and
helpful factor in tracing the history of a schist in those cases in which neither
metamorphic diffusion nor metamorphic differentiation has occurred. If Bastin's
criteria are satisfied in these cases, the rock is conclusively sedimentary in origin. If
the analysis is identical with common and definite igneous rock types there will be strong
probability of igneous origin. There will only be doubt where detrital rocks such as
tuffs and arkoses and sediments with approximate igneous composition are possible.

If either metamorphic process is suspected then great caution is necessary. Should
there be the metamorphic differentiation of a single oxide like quartz, the remaining
oxides in the analysis will still bear the same ratios, and the criteria will avail as in the
case of the sericite schist. In general, however, we will be unable to place reliance on
these chemical criteria wherever metamorphic diffusion or differentiation has prevailed.
The composition of the chlorite rock at Cape Denison satisfies the three sedimentary
criteria, the biotite hornblende schist satisfies two of them, while the epidosite does
not satisfy any. The criteria are valueless in the case of metamorphic diffusion and
differentiation types, because such types possess metamorphic individuality alone.
They possess neither the individuality of an igneous rock nor the individuality of a
sedimentary rock. They possess a complex history, and have arisen during the
metamorphism prior to which they did not exist as individuals.

The actual extent of these processes, and with it the extent of the limitation of
chemical criteria, awaits further research. They do seem to be of more frequent
occurrence than hitherto suspected. Yet these processes only have limited range,
and therefore in the complete description of any one area we would normally expect

THE METAMORPHIC ROCKS OF ADELIE LAND 8TILLWELL. 121

to find parts which have been unaffected by the migration. Hence Bastin's criteria
will apply to parts of an area, but which part ? This question will be more readily
answered if it should subsequently be found that metamorphic diffusion and
metamorphic differentiation products are restricted to a certain few rock types.

7. CONCLUSION.

Our reference to other areas of metamorphic rocks may now be concluded by
tabulating the different hypotheses appearing in the geological literature, which are
founded on the transition between two rock types in metamorphic areas.

1. Intermingling of basic magma with acid magma.

2. Differentiation of an intermediate magma into a relatively basic portion

and a relatively acid portion.

3. Local melting or refusion in situ.

4. Gradual assimilation of pre-existing basic sediments by invading granite or

gneiss, producing amphibolite as the final product.

5. Production of amphibolite by the extreme metamorphism of a limestone.

In each case, except the last, the observed transition takes place, as at Cape Denison,
between a granitic gneiss and a basic rock related to amphibolite. In all cases,
as far as can be judged at present, the evidence submitted might be explained on the
hypothesis of metamorphic diffusion.

CHAPTER VI.
THE MACKELLAE ISLETS.

FIELD NOTES*.

The Mackellar Islets are a group situated nearly due north of Cape Denison in
Commonwealth Bay (Plate XXXII.). The smaller members are normally covered with
a thick ice cap built by the frozen spray, which is swept from the surface of open water
by the incessant winds. The largest island is practically ice free, and consists of a low
plain whose highest point is about 40ft. above sea level (Plate XXXIII., fig. 1). The
surface of the islets is, generally speaking, flat, and forms a contrast to the miniature
mountain area of Cape Denison. Where prominences are seen they are well rounded.
No polished areas or striae were found, but the surface is very smooth in places, and
appears to have only been recently roughed up by frost action or other disintegrating
agencies. The surface is everywhere covered with saline matter blown off the sea,
and this may assist the disintegration.

No moraines of truly foreign boulders occur on the islets, but patches of roughly-
rounded boulders of gneiss, similar to the " lower moraines " on the mainland at Cape
Denison, are found. In fact the general appearance is similar to the lower rock belt at
Cape Denison. Detrital gravel patches appear in a few sheltered spots ; their
appearance and situations suggest that they are probably submarine accumulations,
and, if so, they supply some evidence of relative uplift.

The dominant rock is a grey gneiss very similar to that at Cape Denison, but it is
more uniform in character and more granitic. At intervals there are finer-grained
darker patches. Irregular fine-grained black patches and streaks (amphibolites) are
also moderately frequent, but they are not so conspicuous as on the Cape Denison
outcrop. The foliation is sometimes almost horizontal, and at other times nearly
vertical. At the north end the gneiss is particularly massive and granitoid. The
trend of the rock bars through the Mackellar group is in a direction between N.N.W.
and N. by W. Transverse to this dominant structure is a fracturing in a direction
W. by N. (nearly W.N.W.) which has led to the development of cross gullies and ravines.

DESCRIPTION OF ROCK SPECIMENS.

The rock specimens from the Mackellar Islets consist of three specimens of a
granitic gneiss and one of amphibolite.

* The field notes are supplied by Sir Douglas Mawson.

THE METAMORPHIC ROCKS OF ADEL1E LAND. 8TILLWELL. 123

No. 981. This is the specimen of the dark amphibolite, showing massive texture
with fine even-sized grains of hornblende and felspar. In section it shows affinities
to the Cape Denison amphibolites, and its percentage mineral composition is

Hornblende 34-5

Felspar 50-3

Epidote 6-5

Biotite 5-1

Sphene 2-4

Apatite 0-8

Iron ore 0-4

Calcite present

This composition possesses marked differences from the Cape Denison amphibolites.
The proportions of hornblende and felspar are approximately reversed. The composition
is most like that of the epidote biotite schist (No. 153), except that the biotite is replaced
in this rock by hornblende. The high percentage of felspar is reflected in the colour
of the hand specimen, which is not the dense black of the Cape Denison rocks. A
portion of the felspar is saussuritised, and an extinction angle in partly saussuritised
crystals of 40 has been measured from the lamellae, and hence it may be called
labradorite. There is a considerable amount of clear felspar which may bear a trace
of lamellar twinning and which has a refractive index often inseparable from the Canada
balsam. This clear felspar is looked upon as albite. No quartz has been detected
among the clear felspar.

The hornblende is a little paler in colour than in the Cape Denison rocks, but the
bluish tinge is prominent. The edges of the crystals are more ragged and indefinite.
Epidote is found in pleochroic crystals of the same size as the hornblende. It may be
associated with the hornblende in a manner which suggests its derivation from the
hornblende, and it may fringe the biotite crystals. Crystals with definite crystal
boundaries may be set in the biotite plates, and they may contain a brownish nucleus
of allanite.

The biotite appears in relatively large and broken plates, and has a greenish-brown
colour in its darkest position. It contains pleochroic haloes. Sometimes it is fringed
with a rim of opaque iron ore, which is in turn surrounded with epidote. This suggests
that free iron oxide is set free in a reaction between felspar and biotite. Sphene and
apatite are abundant accessories, and occasional plates of calcite are present and are
included in the felspar percentage.

The rock may be called an albite amphibolite, and is placed in the family of albite
amphibolites in the epi division of the Group of the Eclogites and Amphibolites.

124 AUSTRALASIAN ANTARCTIC EXPEDITION.

The separation of the anorthite part of the molecule of the plagioclase, and the
consequent production of the highly sodic felspar, has liberated more lime and probably
accounts for the unusually high epidote percentage. In this respect this rock is different
from the average Cape Denison amphibolite, but agrees with the exceptional case, No.5.
At Cape Denison we find a high percentage of epidote with the high percentages of
biotite and the low percentages of hornblende, whereas in this case, as in No. 5 before,
we get a high percentage of epidote even with a low percentage of biotite. This
difference is no doubt due to slightly different conditions during recrystallisation, which
are reflected both in the composition of the plagioclase and in the epidote percentage.

No. 983. No. 983 is an example of the chief rock type. It looks like a coarse,
massive biotite granite in the hand specimen, showing quartz and felspar and biotite.
The biotite flakes of the normal granite are replaced by aggregates of small biotites.
The pink colour of the felspars is inclined to dominate the colour of the rock, and its
general appearance is different from that of the grey granodiorite gneiss of Cape Denison.

Under the microscope there is little doubt that this rock has been subjected to
metamorphic agencies similar to those interpreted in the Cape Denison rock, and it
must be classed as a gneiss, not as a granite. The large crystals of quartz show strong
cataclasis. Large crystals of felspar, probably orthoclase, have been replaced by
granoblastic aggregates of microcline. A large crystal of orthoclase, which is cloudy
with the development of sericite, encloses areas, with more or less rounded outline, of
perfectly clear microcline which is certainly due to the recrystallisation. The crystals
of plagioclase have not been found with a refractive index above that of basal quartz.
They are interpreted as an oligoclase, being less calcic than in the granodiorite gneiss.
Diablastic structure is often developed in the plagioclase while -rounded and vermicular
pieces of quartz may be set in the felspar. The felspar crystals become more noticeably
cloudy in the crush areas.

The crush areas, produced by the grinding movement, can be recognised between
two large crystals. Mortar structure, however, is not obvious, because recrystallisation
has proceeded in the crush zones and caused the development of comparatively large
granular crystals. The development of these even-sized crystals in the crush zones
and the replacement of large crystals by granulitic crystals may be considered as a
stage in the development of granoblastic structure in a completely recrystallised rock.
Biotite is abundant in these areas and it is often accompanied by muscovite. These
two minerals are nearly always arranged around the contours of the relic minerals.
Sericite, epidote, and muscovite are intergrown with the biotite.

Large crystals of allanite, apatite, and sphene are accessory constituents.

The rock may be called a granite gneiss or an epi orthoclase gneiss. Compared
with the granodiorite gneiss of Cape Denison there seems to be more orthoclase (or its
equivalent) and a less calcic plagioclase.

THE METAMORPHIC ROCKS OF ADELIE LAND. 8TILLWELL. 125

The remaining specimens (Nos. 982, 984) have a finer grain than the preceding,
and with a grey colour in addition they resemble, in outward appearance, the granodiorite
gneiss of Cape Denison. Further, a slight schistose structure can be detected in the
hand specimen. In section andesine, with its refractive index just above basal quartz,
has been detected, and this is a strong point of resemblance to the granodiorite gneiss
of Cape Denison. There is considerably less allanite and apatite than in the preceding
type, No. 983, but pyrite, magnetite, and zircon are present. The mechanical structures
are equally prominent, and possibly muscovite is more abundant, and sericite
correspondingly less than in No. 983. The presence of andesine makes the rock a
granodiorite gneiss rather than a granite gneiss.

CHAPTER VII.
CAPE HUNTER.

Sir Douglas Mawson has supplied the following notes on Cape Hunter, as the result
of his visit on December 22nd, 1913 :

" The rock exposure forming Cape Hunter is quite an imposing sight at close
quarters (Plate XXXIII. , fig. 2). The coastline is steep, and the rocks extend as a
narrow belt, elongated in the direction of foliation. The rock itself is a phyllite, very
uniform in character, but may pass into distinct sericite schists ; narrow bands here and
there are a little talcose. Representatives of similar rocks have been collected from the
moraines at Cape Denison. The foliation and bedding closely correspond, wherever
examined. The foliation is vertical and trends N. 20 W. Jointing along nearly
horizontal planes is prominent. Weathering has developed gullyways at right angles
and across the trend of the rocks.

' The maximum height of the exposure is about 90ft. The top has a rounded
hummocky surface which has been once polished and striated. Where the polish and
striae have been preserved the trend of striae is N. 45 E. to N. 40 E.

" The prevailing schist contains, in some places, a notable amount of iron ore finely
disseminated. In other places talc is found or quartz is prominent, and along some
bands considerable puckering is noticeable. Along the foliation stringers of quartz
are common ; and with the quartz are crystals of hematite, magnetite (?), epidote,
chlorite, garnet (?), etc. Some veins carrying much epidote and a little fluorite cross
the foliation and bedding of the phyllite.

" Compared with Cape Denison there is a noticeable paucity of erratics at Cape
Hunter. A few of these, especially a grey gneiss erratic and a red granite erratic, are
several tons in weight. The following rock types were noted among the erratics :
Red granite, grey granite, both coarse and fine-grained red porphyries, red gneiss and grey
gneiss, garnet gneiss, a gabbroic rock, a dolerite or basalt, red sandstone (one specimen
only). There is a complete absence of representatives of the metamorphic silicated
limestones."

Specimen No. 911 is an example of the Cape Hunter rock, and it is a very fine grained,
highly schistose rock with a bright sheen on the cleavage surface. In section there is a
prominent crystallisation schistosity, and the structure is both finely granoblastic
and blastopelitic.

THE METAMORPH1C ROCKS OF ADELIE LAND. STILLWELL.

127

Relic crystals of quartz, felspar, magnetite, and lenticles of quartz form the pseudo-
porphyroblasts around which the schistose ground mass bends. Some of the larger
crystals of felspar are sericitised, while others are very fine granulitic aggregates of
secondary felspar. The ground mass consists of small flakes of brown biotite, white
muscovite, granular epidote, quartz, clear secondary felspar, prisms of tourmaline,
apatite and zircon, magnetite and pyrite.

The biotite and muscovite are frequently intergrown. Occasionally there are
much larger crystals of muscovite, and these may be bent. All the small biotite flakes
are parallel, so that as the stage is rotated they all occupy the dark position at the same
time. The pleochroism is strong, and consequently the section looks dense in one
position but quite thin in the other position. In the latter position the rarer tourmaline
prisms are in their dark blue position and can readily be picked up. There is a con-
siderable amount of fine granular epidote among the fine material and, like the muscovite,
larger porphyroblasts occasionally appear. Clear secondary felspar has been detected
among the fine quartz but is much less abundant than the quartz. Iron ore is abundant,
and numerous small cubes of pyrite have been seen. Colourless crystals of apatite are
present, and also rounded crystals with high polarisation colours and high refractive
index like zircon.

The rock can be named phyllite.

As far as can be made out the crystalloblastic order is Tourmaline-magnetite,
pyrite-biotite, muscovite-epidote-felspar and quartz.

CHAPTER VIII.
MADIGAN NUNATAK.

The Madigan Nunatak is situated in Lat. 67 8j' and Long. 143 20', about 30 miles
distant from Cape Denison. It lies on a ridge which slopes away to the north, reaching
sea level at Cape Gray, 18^ miles distant. Its appearance is that of a small rock island
rising above the ice plateau at 2,400ft. above sea level, and it forms a small jagged ridge
of rock running north and south. It is 160yds. long and about 50yds. wide in the widest
part, and it rises from the level of the ice sheet at the southern end to a height of about
60ft. at the northern end. Views of the Nunatak are given on Plate XXIV., figs. 1 and 2.

It is composed of gneissic rocks whose foliation strikes approximately north and
south, coincident with the direction of the ridge. There is a steep anticlinal fold at
the southern end (Plate XXVII., fig. 3), pitching slightly to the north. In contrast
to the freshness of the rock exposures on the coast at Cape Denison, Cape Gray, etc.,
there is found considerable surface weathering. The surface is frequently brown and.
iron stained, and the felspars may lose their transparency. Frost action is prominent,
and many of the cracks and joint planes are filled with moderately fine disintegrated
material. There is no sign of recent glaciation and no glacial erratics or ice striae are
found on this area.

Two rock types are found on this area. One is a black massive plagioclase pyroxene
gneiss or pyroxene granulite whose relation to the second type is not obvious in the field.
It was noted that it seemed to form either a band whose trend cut at right angles across
the foliation, or a band that may have been conformable with the anticline. Its
boundary on either side was indefinite or obscured by the angular blocks tumbled about
by the frost action. The second type is the more abundant acid gneiss, containing blue
quartz and hypersthene. In the neighbourhood of the anticline it has a banded
character, but in other parts the gneissic character, though evident, is less prominent.

PLAGIOCLASE PYROXENE GNEISS (PYROXENE GRANULITE).

The fresh specimens of this rock are black and massive with moderately fine and
even granularity. Macroscopically felspar and pyroxene are visible. The weathered
surface is discoloured by brown iron staining. So long as the felspar is sufficiently
fresh to be transparent the dark colour of the pyroxene dominates the colour of the rock.
When, however, the transparency is changed to translucency in the early stages of
weathering, the whiteness of the felspar is noticeable, and the rock assumes a grey colour.

THE METAMORPHIC ROCKS OF ADELIB LAND. STILLWELL. 129

In section the rock (No. 794) has a granoblastic structure, which is modified by
subsequent cataclastic structures. The average absolute grain size is approximately
0-30mm. The mineral composition has been determined by the Rosiwal method to be

Felspar 42-5

Pyroxene 45-5

Hornblende 3-3

Iron ore 84

Biotite 0-3

The rock (Plate IV., fig. 1) is therefore essentially an aggregate of felspar and pyroxene
grains, which are of approximately equal dimensions. Apatite is also present as a
minute accessory.

The greater proportion of the felspar is untwinned. When lamellar twinning is
found it is irregular, patchy, and often bent. There is often undulose extinction, so
that the determination of the felspar by the use of extinction angles is not satisfactory.
The refractive index of most grains is in the neighbourhood of 1551 (nitrobenzol), in
some cases above and in others below. In a few cases the refractive index is below
1-542 (nelkenol). These observations are explained by the presence of two plagioclase
felspars. Sections of the plagioclase with higher refractive index may be found which
possess two good cleavages, and are therefore considered to be normal to (001) and
(010), and an extinction angle of 30 was measured. Hence we consider this plagioclase
to be a calcic andesine. The second plagioclase with the lower refractive index is in
much smaller quantity, and is considered to approach albite in composition. One
fragment with fine lamellation was noticed to have a refractive index less than Canada
balsam. The felspar crystals are often fractured, and show cataclasis. Mortar structure
is common between two felspar crystals.

The pyroxene includes both orthorhombic and monoclinic forms. Hypersthene is
readily detected in the thicker sections by the characteristic pink to green pleochroism.
The form of the grains is granular. The colour of the augite is pale green, and it is on
the whole fresh. It shows cataclasis like the felspar, but not so conspicuously. Strain
polarisation may be found, and some crystals are fractured and show mortar structure.
The crush zones may develop through a crystal. Many of the pyroxene crystals possess
a border of finely pulverised pyroxene produced during the stress action. This crush
border may pass out gradually into crushed felspar, and then an apparent transition
from augite to felspar appears. Very often fine streams of granulated pyroxene tail
out into the felspar and appear as a set of linear inclusions. Ilmenite often forms rims
and borders to the pyroxene crystals, and the appearance suggests in itself that the
iron content of the primary substance had been thrown out during the development of
augite in the first metamorphism. In rare cases the pyroxene is dusty with small
ilmenite inclusions.

Scriee A, Vol. in.. Part 1 I

130 AUSTRALASIAN ANTARCTIC EXPEDITION.

The hornblende is green and has a granular shape. It is clearly developed from
the pyroxene because the passage can be observed. That the hornblende developed
before the crushing is evident from the presence of crushed borders on the hornblende
crystals. Very rarely the normal hornblendisation of the pyroxene is replaced by
glaucophanisation. Small amounts of the blue pleochroic glaucophane have been
detected and are associated with the pulverised augite rather than with the large
crystals. It is interesting to note that such change is recorded by Grubenmann* in the
alteration of the pyroxene of eclogites during transition from one metamorphic zone
to another.

The iron ore percentage is greater than in any of the Cape Denison amphibolites.
The major portion is ilmenite, which is frequently associated with a reddish-brown
mineral, probably rutile. Pyrite is also present. The lustre of the pyrite is bronzy
red, and is suggestive of pyrrhotite. Large ilmenite crystals show crushing and
pulverisation along the borders like the pyroxene. The reddish-brown biotite is scarce,
but is always associated in curious aggregates with ilmenite. It is probable that these
aggregates are produced here, as in other cases, by the interaction of hypersthene and
felspar. The hornblende may be similarly associated with the ilmenite. The biotite
flakes are sometimes bent, and were probably formed before the cataclasis.

The mechanical effects which are typical in the epi zone of metamorphism are a
dominating feature of this rock, and mark the final metamorphic impress. Hence we
name the rock an epi plagioclase pyroxene gneiss.

* Op. cit., vol. II., p. 84.

THE METAMORPHIC ROCKS OF ADELIE LAND 8TILLWELL.

131

Chemical Characters.

The following analysis of specimen No. 794 was made by A. G. Hall, in the
Victorian Geological Survey Laboratory, under the supervision of P. G. W. Bayly.
In the second column is placed an analysis of a hornblende norite from St. Thomas

Mount, Madras, by H. S. Washington*.

I. II.

SiOg 50-62 50-04

AlgO, 11-43 .... 11-65

Fe 2 0s 4-43 .... 2-63

FeO 11-11 .... 15-76

MgO 6-87 .... 5-58

CaO 10-90 7-89

Na 2 1-75 3-08

K 2 0-24 0-89

H *+ ' 62 '"

H 2 0- 0-19 ....

Ti0 2 1-42 1-93

P 2 6 0-08 .... 0-20

SO, nil

Cl tr

MnO 0-28

NiO, CoO 0-03

CoO tr

Li 2 tr

99-97

99-64

Specific Gr 3-076

Group Values.

54-3

2-0

5-1

31-5

7-2

1-0

Projection Values.

1-0

2-7

16-3

The analysis demonstrates the affinity of this rock type to the amphibolites. The
character of the felspar is reflected in the proportion of potash to soda and in the relation
of both of these to the high lime percentage. The iron percentage is a little greater
than in the Cape Denison amphibolites. The silicity is the mean of those of the two
analysed Cape Denison amphibolites

When this analysis of the Antarctic rock is compared with that of the hornblende
norite a basic member of the Charnockite series of India points of great similarity

" The Charnockite Series of Igneous Rooks," H. 8. Washington, Amer. Journ. ScL, vol. XLI., 4th Ser., 1916, p. 323.

132

AUSTRALASIAN ANTARCTIC EXPEDITION.

are noticed. The percentages of Si0 2 and A1 2 S are practically the same. The total
iron is much the same in each case. The relative proportions of MgO and CaO are also
the same, while there is large excess of soda over potash in both cases.

Classification and Origin.

The Ozann group values place the rock in the group of Eclogites and Amphibolites
in Grubenmann's classification. The projection values assign to the rock a position in
the triangular diagram close to the mean group value of Group IV. (fig. 11).

Fig. 11.

I. Mean Position of Group I., the Alkali Felspar Gneisses.
IV. Mean Position of Group IV., the Eclogites and Amphibolites.
797. Hypersthene Alkali Felspar Gneiss, Madigan Nunatak.
754. Garnet Hypersthene Alkali Felspar Gneiss, Aurora Peak.
794. Plagioclase Pyroxene Gneiss, Madigan Nunatak.

We admit the marked traits of the epi zone in the assigned name epi plagioclase
pyroxene gneiss ; but the rock bears a double metamorphic character. Reconstructing
the outlines of the fractured minerals, it can be seen that, before the epi zone imprint
was received, the rock consisted of a granoblastic aggregate of augite, hypersthene,
plagioclase, and ilmenite, with a little hornblende and pyrite and biotite. As such it is
identical with some of the Saxon pyroxene granulites, or the French pyroxene gneisses,
or the Indian norites, whose metamorphic character will be subsequently affirmed
from a comparative study with the basic rocks at Cape Gray and Aurora Peak. As such
it is a member of the plagioclase augite family of the kata division of Group IV. Before
the recrystallisation in the kata metamorphic zone we might judge from the chemical

THE METAMORPHIC ROCKS OF ADELIE LAND. STILLWELL. 133

composition that the rock was either a gabbro or a diabase. In view of the large number
of metamorphosed diabase dykes that are present in this region one would be inclined
to consider that this rock is a diabase dyke which, suffering different metamorphic
conditions, has been converted into a different rock type to the amphibolite at Cape
Denison. Such is in accord with the field observation that the rock seemed to be a band,
and such conclusion will be subsequently supported by correlative argument.

In the kata zone, then, the primary rock suffered its first and thorough
recrystallisation, and the pyroxene and plagioclase and ilmenite were formed. In
the transition from the kata zone to the epi zone we naturally find some evidence of the
passage through the meso zone. The evidence is yielded by the alteration of the
pyroxene into biotite and hornblende. It has been pointed out that the biotite flakes
are often twisted and bent, and the hornblende sometimes granulated ; hence, like the
pyroxene and plagioclase, they are secondary metamorphic relic minerals in the epi zone
metamorphism.

The obscurity in the field concerning the boundary between the epi plagioclase
pyroxene gneiss and the epi hypersthene alkali felspar gneiss may be readily explained
by the presence of a metamorphic diffusion type which would cause a transition from
one type to the other. Near the boundary the lighter coloured constituents become
more prominent, and blue quartz may appear. At the same time the rock assumes a
more schistose character.

Specimens showing the junction between the basic gneiss and the acid gneiss are
present in the collection (No. 795). In these the junction is partly indefinite, and its
position cannot be precisely marked in certain places. Small pieces of the dark rock
are seen to be apparently detached from the parent mass and lie enclosed in the lighter-
coloured rock. These junction specimens were no doubt collected from those parts
where the junction was most obvious in the field.

HYPERSTHENE ALKALI FELSPAR GNEISS.

The second type of gneiss at the Madigan Nunatak is a coarse-grained grayish-white
rock, in which the gneissic structure can be detected. Like the preceding type, it
weathers to a brown colour. Microscopically, one can see blue quartz, felspar, and
smaller amounts of black hypersthene. There is considerable variation in the grain
size of different specimens, and this is especially noticeable with respect to the dark
hypersthene.

In thin section (No. 797) the schistosity is not noticed, but there is abundant
evidence of crushing and cataclasis. Mortar, cataclasic, and diablastic structures are
common. Quartz, orthoclase, and plagioclase form the bulk of the slide. Hypersthene
biotite, and ilmenite are important, though in a lower order of abundance. Zircon,
apatite, and pyrite are accessory minerals.

134 AUSTRALASIAN ANTARCTIC EXPEDITION.

Quartz units show strong undulose extinction, and are sometimes crushed so that
a unit in ordinary light becomes a fine fragmentary aggregate in polarised light. In such
cases a rude schistosity is evident, because many of the fragments are elongated in one
direction. Every large quartz unit is separated from neighbouring crystals by crush zones.

Most of the felspar is orthoclase with perthitic inclusions of albite. The albite
forms lenticular layers in at least two directions in the crystal. In cross section the
albite is rectangular, and in longitudinal sections it appears as thin needles, while in
some sections two sets can be seen crossing at an acute angle. In addition to the
orthoclase and perthite there is a small amount of plagioclase with lamellar twinning
and a comparatively low refractive index. It is considered as an albite oligoclase or
an oligoclase, and it may contain perthitic inclusions of orthcclase. The lamellar
twinning may only appear indefinitely in one corner of the crystal, or the laminae may
be bent and irregularly wedge out. In such cases undulose extinction is present, and
there may be a poorly developed microcline structure. These observations are suggestive
of secondary pressure twinning. However, the plagioclase is definite where the crush
zones cut across the lamellae, for in such cases the lamellae must have existed prior to the
crushing. The felspar shows the crush phenomena even more markedly than the quartz.
The best examples of mortar structure are exhibited in felspar crystals which have
straight fractures. A single crystal may contain one or more fractures, and each fracture
filled with pulverised material. Mortar structure exists between a quartz crystal and
a felspar crystal, but it is always less noticeable between two quartz crystals. Strings
of minute inclusions are common and may extend into neighbouring crystals. Shear
zones of sericite are often present, and become iron-stained during weathering. Sericite
also appears along cleavage planes. Diablastic structure is very common in the crush
areas, and there appear vermicular interlacings of felspars and of quartz and felspar
(Plate III., figs. 1 and 2). Sometimes it is coarse and sometimes it is very fine, but in
many instances it is obviously a secondary structure produced during metamorphism.

The hypersthene is present with its characteristic pleochroism and straight extinction.
Like the quartz and felspar, it has suffered mechanical deformation, and one crystal is
broken with the two pieces separated by a fracture zone. The crystals are sometimes
bordered with a crush rim. It shows considerable alteration to a greenish, fibrous,
serpentinous mineral, with moderately low polarisation colours, usually masked by
the green colour. This mineral is similar to the alteration product of hypersthene
in the Indian charnockites, which is described by Holland as resembling delessite*.
The greenish mineral is a very constant associate of hypersthene in all the acid
hypersthenic gneisses of Adelie Land. It is found in other cases to be intimately mixed
with a pale-green biotite showing brilliant polarisation colours, and this rather suggests
that it is delessite, a chloritic mica. On the other hand, cases have been noticed where
the hypersthene passes through bastite into serpentine, whose appearance is quite similar
to this green mineral. In part brown biotite and ilmenite seem to be developed from
it. The biotite and ilmenite are practically confined to the hypersthene areas, and

* " The Charnockite Series," T. S. Holland, Mem. 28, pt. 2, G.S. India, p. 141.

THE METAMORPHIC ROCKS OF ADELIB LAND STILL WELL.

135

the three minerals are undoubtedly intimately associated. A very small amount of
biotite seems to have developed with the alteration products of the felspar. The biotite
always bears evidence of pressure, and the cleavage flakes are either bent or crumpled,
or else in shreds. The ilmenite is also affected by the stress, and streaks of black ilmenite
dust issue from the ilmenite crystals and traverse the fractured areas. When an ilmenite
crystal partakes in the production of mortar structure the pulverised zone is darkened
by the ilmenite fragments. Apatite and zircon are present in occasional and relatively
large crystals which have been bent and fractured.

Chemical Characters.

The following analysis of No. 797 was made by J. C. Watson in the Victorian
Geological Survey Laboratory under the supervision of P. G. W. Bayly. In the second
column is placed an analysis by H. S. Washington* of acid charnockite from St. Thomas
Mount, Madras.

Si0 2
A1 2 0,

Fe 2 0,
FeO .
MgO
CaO .
Na,0
K 2

72-38
13-39
0-73
1-09
0-67
1-86
2-02
6-57

H 2 O+ 0-44

H 2 -
Ti0 2 .
PA .
SO, ..

a ....

MnO .
Li 2 .

0-12

0-40

0-16

nil

tr.

99-83

Specific Gravity 2-632

II.

77-47
11-00
1-04
2-02
0-43
1-02
2-86
4-14
0-20
0-05
0-26
nil

nil

100-59

Group Values.

Projection Values.

79-8

6-7

1-9

3-0

1-7

11-6

3-3

5-1

Op. oit.. p. 325.

136 AUSTRALASIAN ANTARCTIC EXPEDITION.

The analysis is similar to that of a potash granite and it is considered that the
chemical evidence of igneous origin is very strong in this case. The analysis bears
witness to the high silicity of the rock with corresponding low percentages of iron, lime,
and magnesia. The latter is absorbed in the hypersthene and its alteration products
and is, therefore, some indication of the small quantity of hypersthene in comparison
with the amounts of quartz and felspar. The relative amounts of Ti0 2 and Fe ? 3
confirm the record of ilmenite in the rock. The abundance of alkali felspar is reflected
in the high percentage of alkalies, and the great excess of orthoclase is similarly reflected
in the large excess of potash over soda.

Among the group values the high value of K corresponds with the high silicity,
and the high value of A corresponds with the high alkali percentage. The projection
values are such as to place the rock in the area of Group I. in the triangular projection
(% ID-

Like the associated basic rock, the analysis shows considerable resemblance to the '
quoted analysis of acid charnockite. The silica percentage of the charnockite is nearly
5 per cent, greater ; but this is not important, as Washington has drawn attention to
such a range of variation among the acid charnockites themselves. The relative pro-
portions of ferrous and ferric iron, of magnesia and lime, and of soda and potash are
similar in each analysis.

The Classificatory Position. The group values and projection values bring the rock

into Group I., the group of Alkali Felspar Gneisses. The epi zone metamorphism is

important and is revealed by the cataclastic and mortar structures. Before the epi zone

metamorphism the rock consisted of a granular aggregate of quartz, orthoclase, a little

plagioclase, and small amounts of hypersthene, biotite, and ilmenite with accessory

apatite and zircon. As the schistosity is chiefly marked by the parallel arrangement

of the hypersthene crystals it is obvious that the rock was schistose before the epi zone

imprint. Further, if the inference that the biotite and ilmenite has been formed from

hypersthene and felspar is correct, then such alteration took place before the epi zone

metamorphism, because both biotite and ilmenite show marked mechanical effects.

We must, therefore, recognise two metamorphic phases in the development of the meta-

morphic character of this rock. Parallel with the plagioclase pyroxene gneiss it will

be subsequently affirmed that the primary rock was first recrystallised in the kata zone

of metamorphism. As the rock ascended from the depths of the earth's crust and

became subject to meso zone condition the hypersthene reacted with the felspar and

produced ilmenite and biotite. Possibly here also some of the microperthite was formed

and the diablastic structures produced. These metamorphic results were completed

in the epi zone where the excessive mechanical effects were produced.

The rock may, therefore, be described as an epi hypersthene orthoclase gneiss,
produced by the superposition of epi zone metamorphism upon a kata hypersthene
orthoclase gneiss. The primary equivalent of the latter, judged from the chemical
composition, was probably a granite.

THL METAMORPHIC ROCKS OF ADELIE LAND. STILL WELL. 137

The formation of the anticline* observed at the Madigan Nunatak is probably
associated with the pressure movements which produced the crush structures. The
banding of the anticline is noted by the foliated character of the rock.

Correlation. The two types of gneiss from the Madigan Nunatak bear traces of
similar metamorphic history. Both are rocks with prominent epi zone metamorphism
which has followed kata zone metamorphism. In both cases the passage through
the meso zone was fairly rapid, but it is noticed in the partial alteration of the pyroxene
and the dissociation of the felspar.

The chemical likeness between both rocks and the acid and basic members of the
Indian charnockite series has been pointed out. A discussion on this inter-relation
will be subsequently presented.

The term " Anticline," as here used, is not strictly correct. An anticline is normally marked by the bedding planes
of sedimentary strata. In this case it is marked by the foliation of a gneiss and might be distinguished as a foliation anticline.

CHAPTER IX.
AURORA PEAK.

Aurora Peak (Plate XXVI., fig. 3) is situated in Adelie Land in Lat. 67 24' and Long.
144 12' and is about 50 miles E.S.E. of Cape Denison. It is a solitary peak or nunatak
rising above the snow plain on the west side of Mertz Glacier to a height of 1,750ft.
above sea level. It is distant about 25 miles from the Madigan Nunatak, and similar
rock types appear on both outcrops. It was visited by Madigan's sledging party in
December, 1912, and our information comes from their report and from the examination
of the specimens they brought back.

The similarity in rock types to the Madigan Nunatak is its outstanding feature.
Here again there are two principal rock types, viz., a plagioclase pyroxene gneiss and a
hypersthene alkali felspar gneiss, which are analogous in mineral content with the two
types described at the Madigan Nunatak, differing only in mineral proportion and in
structures. Whereas the rocks at the Madigan Nunatak are remarkable for their crush
structures, the rocks at Aurora Peak are almost devoid of such features. Whereas
the epi zone metamorphism is dominant at Madigan Nunatak, the meso zone meta-
morphism is equally dominant at Aurora Peak. Similar primary rocks have been
metamorphosed in both instances under different physico-chemical conditions.

MESO PLAGIOCLASE PYROXENE GNEISS.

Specimen No. 759 is reported as a black band which cuts across the gneiss. In the
hand specimen it is a fine-grained, dark-coloured rock with a weak schistosity produced
by lenticles of felspar. The average absolute grain size is approximately 0-1 7mm.
Its colour is black when fresh, and it grades up to a light-brown colour as the felspar
becomes cloudy and iron-stained by weathering. The same minerals are present in
the slide as in the analogous rock No. 794 from Madigan Nunatak. Hornblende is
much more abundant, and there is a corresponding decrease in the amount of pyroxene
from which it is derived. The proportion of felspar to ferromagnesia is practically
the same in both cases. Apatite is also more abundant. The following proportions
have been determined by a Rosiwal analysis :

Felspar 44-8

Pyroxene 28-6

Hornblende 15-5

Iron ore 10-1

Biotite 0-3

Apatite 0-7

Thus there is in this rock nearly five times as much hornblende as in No. 794, and
the transformation of pyroxene to green hornblende is correspondingly more obvious

THE METAMORPHIC ROCKS OF ADELIE LAND. STTLL\VELL. 139

in the section. The formation of this hornblende is distinctly a meso zone character.
The small amount of biotite is probably developed by the interaction of pyroxene with
felspar. The felspar is again found in two varieties of plagioclase, and even though
more than half is untwinned it is all believed to be plagioclase*. There is a calcic
andesine or a sodic labradorite, and there is a much smaller quantity of a sodic or
albitic plagioclase which in very exceptional cases possesses a refractive index less than
Canada balsam. Ilmenite is found in indented and irregular grains. In all cases the
pyroxene exerts its crystalline form against the ilmenite, so that in any aggregate of
pyroxene and ilmenite grains the ilmenite is pushed into the interstices between the
pyroxenes. It can therefore be understood, when the ilmenite is crushed and granulated
in subsequent epi zone metamorphism as in No. 794, why the crushed ilmenite should
appear as a border to the pyroxene crystals. Red-brown pleochroic rutile is sometimes
associated with the ilmenite.

The absence of cataclastic structures and the important hornblende percentage
influence the decision that this rock shows meso zone characters rather than kata zone
or epi zone features. Its history will otherwise be the same as the related rock No.
794, and it has been a kata zone metamorphic rock consisting essentially of felspar,
pyroxene, and ilmenite on which a meso zone metamorphic impress has been super-
imposed. It may therefore be called a meso plagioclase pyroxene gneiss (meso pyroxene
granulite), or a hornblende plagioclase pyroxene gneiss (Plate IV., fig. 2).

Had the meso zone metamorphism been complete, all the pyroxene would have
been converted into green hornblende and an amphibolite produced. The chemical
composition must be similar to that of No. 794, and hence both chemical and micro-
scopical characters reveal relation to the amphibolites. As many amphibolites are
altered dykes rocks, this plagioclase pyroxene gneiss, similar to some of the Saxon
pyroxene granulites, is probably a diabase dyke which has suffered metamorphism
under kata zone conditions. The field observation that this rock appears as a band
crossing the gneiss is confirmatory of such an argument.

HYPERSTHENE ALKALI FELSPAR GNEISS.

Specimen No. 754 is the type example of this rock, and it is a coarse-grey gneissic
rock. In some examples the gneissic character becomes more prominent on the
weathered surface. Macroscopically one can see thin lenticles of quartz set in a granular
mosaic of felspar, and pink garnets and black hypersthene are also drawn out in layers
in the direction of the schistosity.

In thin section the rock is granoblastic, with a tendency to a coarse crystallisation
schistosity. Quartz units are built up of interlocking grains. There is little cataclasis.
Felspar consists of untwinned individuals and lamellar twinned individuals. The
former include both orthoclase and clear albite with perthitic inclusions of orthoclase.

The frequent absence of twinning lamelle has been noted by Washington u a peculiarity of the hypersthenic roclu
of India and allied area* ; but we have noticed it in amphibolitea and other metamorphic roclu.

140

AUSTEALASIAN ANTARCTIC EXPEDITION.

The lamellar twinning is generally exceedingly fine, and the small extinction angle
and low refractive index indicates an albitic plagioclase. Sometimes the felspar includes
rounded blebs of quartz, and sometimes there is incipient diablastic structure. A
certain amount of sericite has been produced from the felspar. The hypersthene
is partly altered, and the same greenish, feebly pleochroic mineral (delessite ?) is
present as before. It is also partly altered to enstatite. Brown biotite and ilmenite
are also again associated in a significant manner with the hypersthene. Colourless to
pale-pink garnet is present, and may be associated with the hypersthene. The garnet is
found both in small and large crystals, which are partly idioblastic. Inclusions are not
abundant in the garnet, but biotite, quartz, and ilmenite appear as such. Accessory
grains of monazite or zircon are present.

Chemical Characters.

The following analysis of No. 754 was made by J. C. Watson in the Victorian
Geological Survey Laboratory :

SiO a 6942

A1 2 3 15-03

Fe 2 3 1-66

FeO 2-65

MgO MO

CaO 345

Na 2 4-50

K 2 1-39

H 2 + 0-65

H 2 - 0-07

C0 2 nil

Ti0 2 0-35

P 2 5 tr.

S0 3 nil

Cl tr.

MnO 0-06

I-JzO strong tr.

100-33
Specific Gravity 2-685

75-2

5-7

3-8

Group Values.

5-8

0-2

1-6

Projection Values.

7-4

5-0

7-6

THE METAMORPHIC ROCKS OF ADEL1E LAND- BTILLWELL. 141

This analysis is very similar to that of the granodiorite gneiss of Cape Denison.
There is a little less iron and magnesia in the case, but, except for the relation of one
alkali to the other, there is no important difference. This rock is notable for its excess
soda, though the alkali total is approximately the same in both cases. The analysis
is more similar to the Cape Denison granodiorite gneiss (No. 11) than to the hypersthene
gneiss (No. 797) of the Madigan Nunatak, to which it is closely allied in structure and
mineral composition. Compared with this hypersthene gneiss there is more iron,
magnesia, and lime, corresponding probably with the garnet and the different felspar.
The alkalies furnish the most striking difference. Whereas there is a large excess of
potash and orthoclase in No. 797, there is a large excess of soda and albite in No. 754.

The Classificatory Position.

The group values place the rock among the alkali felspar gneisses of Group I. These
values illustrate the acidity of the rock and its high alkali value. The projection values,
when plotted, give a position not greatly different from that of the mean group value of
Group I., and intermediate between that of the hypersthene gneiss of the Madigan
Nunatak and the granodiorite gneiss of Cape Denison (fig. 10).

Mineralogically this rock differs from No. 797 in the presence of garnet and
the dominance of soda felspar over potash felspar. These differences do not carry
the rock into a different schist group, because the relation between the alkalies does not
enter into the classification. They are both hypersthene alkali felspar gneisses, while
No. 754 is, in addition, garnetiferous. Cataclasis is not important in the Aurora Peak
example, and we thus lose the dominating epi zone character found at Madigan Nunatak.
Assuming for the present that the formation of the hypersthene and garnet belongs
to the kata zone of metamorphism, we can infer that the rock is a kata zone rock. As,
however, we have admitted the meso zone modification of its neighbour, the plagioclase
pyroxene gneiss No. 759, we must consider what evidence of the meso zone conditions
might be found in this rock. It is possible that some of the biotite has been derived
from the reaction of garnet with felspar, or the biotite and ilmenite from the hypersthene
and felspar. There is a significant association of these minerals, but it is not possible
to give a sure interpretation from the study of this specimen alone. A breaking up of
the garnet or the hypersthene might be viewed as a modification due to meso zone
conditions. The abundance of perthite may be looked upon as further evidence.

We may, therefore, describe this rock as a garnet hypersthene alkali felspar gneiss
developed in the kata zone of metamorphism and somewhat modified by the meso zone
of metamorphism.

Without correlative evidence we must depend on the chemical evidence to indicate
the nature of the primary rock. As far as can be determined there is no reason to suspect
metamorphic differentiation or metamorphic diffusion, and the chemical criteria are
valuable. The chemical composition is that of a well-known rock type, viz., a
granodiorite, and this pointe to a primary igneous origin.

142 AUSTRALASIAN ANTARCTIC EXPEDITION.

Comparison of other Specimens with the Type Specimen.

Other specimens from this locality tend to emphasise the subsequent meso zone
impress. Specimen No. 758 possesses more prominent schistosity, partly due to the
presence of prominent quartz lenticles on the weathered surface. In section the example
is noteworthy for its more granoblastic character and its coarse perthite. In addition
there are clusters of coarse brown biotite associated with ilmenite and with odd grains
of hypersthene. The brown biotite with a green transition stage can be found developing
from the hypersthene. Garnet is present, and biotite, without the ilmenite, is developing
from the garnet, probably by reaction with felspar. These observations, therefore,
tend to confirm the meso zone changes reported from the type example.

Two other specimens (Nos. 756, 757) were collected and reported by the sledge
party to be variations in the gneiss, but less plentiful than the type example. These
variations are found to consist of the gneiss with attached portions of metamorphosed
aplitic veins. In each case the boundary is more or less destroyed by metamorphic
diffusion, and consequently they are now all part of the gneiss. Specimen No. 757
consists of a granoblastic mass with weak crystallisation schistosity and well-developed
diablastic structure. Here quartz and cloudy felspar (orthoclase, perthite, and a sodic
plagioclase) form the bulk of the rock. Garnet is present and associated with biotite
with alteration to chlorite. Ilmenite and zircon are common accessories, and a green
spinel, probably hercynite, is present.

Specimen No. 756 is a coarser quartz felspar vein. It is more massive and possesses
more cataclasis. The big crystals have produced in part a mortar structure. The
junction with the gneiss is not noticeable in thin section. In a section across
this junction it is simply noticed that one part of the slide carries garnet and biotite
clusters with a little hypersthene, while these are absent in that part which represents
the original vein.

There remains one other specimen from Aurora Peak. It has been described by
the collector as a specimen illustrating the transition between the dark band (plagioclase
pyroxene gneiss) and the hypersthene alkali felspar gneiss. It is rather a dark-coloured,
banded specimen. Some of the bands are white, others consist of coarse quartz blebs
set in a fine matrix, and others again of very fine black material. The white parts
consist of quartz, felspar, more biotite than usual, occasional garnet and hypersthene
with accessory ilmenite, zircon, and apatite. There has been considerable cataclasis in
which big crystals have frequently assumed a lenticular shape. The crystals may be
surrounded by a granulated zone, and biotite crystals may be set in that zone and tend
to wrap themselves around the crystal. The dark bands appear to be slaty bands,
out of which oval-shaped crystals of secondary quartz have arisen. On closer examina-
tion this is not so. Some of the apparently secondary quartz consist of crushed
granulitic aggregates. Some are relic felspar crystals, while occasionally we find pale
relic crystals of hornblende. Wrapped around them is a fine, dark pleochroic aggregate

THE METAMORPHIC ROCKS OF ADELIE LAND STILL WELL. 143

which is found to consist of minute biotite in which one can detect the incipient appear-
ance of large biotite crystals mixed with some ilmenite dust. In places the dark bands
and the finely granulitic material form a set of parallel bands and sometimes they
penetrate the relic crystals. These dark bands of slaty appearance may possibly be
formed by a continuation of the early stages of the same processes which cause the
biotite to wrap itself around the crystals in the white portions of the rock. If, in addition,
the dark zones represent zones of shear it may become possible to understand why they
should be zones of excessive cataclasis and granulitisation. No large individuals exist
in the crush zone which do not show strain polarisation and are not surrounded by a
zone of granulitised material.

The early examination of this rock gave the impression that it represents the
remains of a recrystallised sediment, and that the dark bands were originally slate.
If this were so, and the field report is correct, it means either that the plagioclase
pyroxene gneiss is a recrystallised sediment a conclusion directly opposed to the study
of the rock or that there is a recrystallised sedimentary gneiss at Aurora Peak which
escaped the observation of the sledging party. The specimen is not a transition type
between the plagioclase pyroxene gneiss and the hypersthene alkali felspar gneiss as it
is reported to be. The only possible explanation is that it represents a shear zone in
which finely powdered biotite and ilmenite dust have dominated the colour. Large
crystals of biotite could form, and it may be a stage of the process in which biotite
crystals wrap themselves around crystals of quartz and felspar in the uncoloured portion
of the rock.

CHAPTEE X.
THE CAPE GRAY PROMONTORY.

DESCRIPTION OF LOCALITIES.*

The promontory terminating northward at Cape Gray is situated between
Commonwealth Bay on the west and Watt Bay on the east. Its seaward edge is a
continuous line of vertical ice cliffs whose monotony is rarely interrupted by rock
exposures. The cliffs often rest on a rocky base and, whenever examined, they consist
of consolidated snow showing distinct lines of stratification.

The promontory is thickly fringed with a large number of rocky islets which form
the Way Archipelago. Some of these islands have a very striking shape. Some are
steeply conical (Plate XXV., fig. 3) and rise out of the water with precipitous faces.
One has its eastern face terminating in an absolutely vertical cliff, while another forms
a sharp, angular wedge (Plate XXV., fig. 1) with its sides rising out of the water at an
angle of 60.

Rock exposures were reached from the mainland in three places, viz., Cape Gray,
Garnet Point, and the Cape Pigeon Rocks, and the descent to them was made possible
by the presence of a steep ramp of ice or snow (Plate XXV., fig. 4). In each case
garnetiferous gneisses are found penetrated by altered basic dykes. As at Cape Denison
the basic rock has been more readily eroded and occupies the gullies and depressions.
It was noted that the islands appear to have the same general character as the rocks
examined, and the subsequent visit to Stillwell Island in the motor launch substantiates
this. Two islands seemed to be composed entirely of the black basic rock, while two
others at the head of Watt Bay are light grey, almost white, in colour, and may consist
of another phase of gneiss.

Compared with Cape Denison there is a noticeable absence of morainic material,
but a few scattered erratics of granite and gneiss are found. Polished surfaces of rock
are frequently noticed on the margin of the exposure, but only one instance of glacial
striae, trending about N.E., is recorded on the Cape Pigeon Rocks.

Cape Gray.

The rock exposure referred to as Cape Gray will, doubtlessly, be an island with
further recession of the ice sheet. It is at present connected with the mainland by a
narrow snow ramp, and a general view is shown on Plate XXIV., fig. 3. The outcrop
is about 250yds. long and 100yds. broad, and it is elongated in an east and west direction.
It is divided in the middle by a transverse gullyway which is occupied by a large basic

* This description embraces the geological field report written conjointly by Laseron and Stillwell.

THE MLTAMORPHIC ROCKS OF ADELIE LAND. STILLWELL. 145

dyke. At the western end of the exposure there are numerous dykes cutting through
the gneiss and have a general trend a little east of north. The dykes repeatedly divide
and unite with one another. They are mostly quite massive, and little evidence of
schistosity is noticeable. The direction of foliation of the gneiss is a little west of north.
In addition to the dykes, veins of quartz and felspar with garnet occur in two sets. A
north and south set are faulted and displaced a few inches by an east and west set. No
ice striae could be found, but the margin of the island, extending back to a height of
20ft. above sea level, consists of well-polished rock. The remainder is rough and jagged
as a result of longer exposure to atmospheric weathering ; the lower zone has probably
been relatively protected by a water or an ice-foot covering.

Garnet Point.

Garnet Point is also approached from the mainland by a steep ice ramp. It is
situated on the north-east portion of the promontory, and is about 10 milea distant
from Cape Gray, and about five miles north of the Cape Pigeon Rocks. It is approxi-
mately the same size as the exposure at Cape Gray. A feature of part of this outcrop is
the presence of abundant aggregates of garnet and mica, up to 2in. broad, which impart
to the rock, even at a distance, a mottled appearance (Plate XXVI., figs. 1 and 2).
The outcrop is divided by a steep transverse gully along which a large black dyke
appears. The marginal zone of polished rock is again noticeable, and a waterworn
pebble of the basic rock was found on a rocky ledge about 20ft. above sea level.

Cape Pigeon Rocks.

The Cape Pigeon Rocks are situated on the east side of the promontory and face
Watt Bay. They are considerably larger than the preceding exposures, and form
two rugged peninsulas which are separated by a narrow sea water channel, and
which terminate seaward in a cliff up to 100ft. in height. A panorama of the northern
peninsula is shown on Plate XXVIII. They are connected with each other on the
landward side by a sloping causeway of ice. The bulk of the rock consists of a
very coarse, grey, garnet felspar gneiss whose foliation trends 20 W. of N. It is
traversed in numerous places by basic dykes which cut across the foliation. Two large
dykes trending a little W. of N. are over 30ft. wide. One outcrops on the northern
peninsula and one on the southern (Plate XXVII., figs. 1 and 4). From one of them
a small dyke is seen branching off at right angles. They dip at a high angle to the west.
Smaller dykes may be only Sin. wide. A large pegmatite vein was noticed. On the
southern portion the rock is excessively contorted, and there are numbers of small dark
amphibolite patches which are elongated and drawn out in the direction of the foliation.

StittuxU Island*.

Stillwell Island is one of the largest members of the Way Archipelago. It is a
steep islet, with poor facilities for landing, and its maximum height is about 120ft.

This information haa been supplied by Sir Douglu Mmwion from hi* diary.
SerioB A, VoL m., Part 1 K

146 AUSTRALASIAN ANTARCTIC EXPEDITION.

(Plate XXIX.). The general outline suggests ice cap erosion, but real smoothing
is only seen up to 30ft. or 40ft. above sea level. At higher levels loose blocks
are scattered about in a manner that indicates no ice sheet has recently passed over the
island. No undisputed erratics are found, though several blocks illustrate a phase of
the local gneiss not observed in situ in the island. Some very large blocks were noted
removed short distances from their original position to situations where gravity could
not possibly place them.

If the snow banks and ice foot were completely melted, the present island would
probably be intersected by one or more sea-water channels. These channels are at the
present time bridged by ice and undermined by the sea, and caverns are produced with
rock walls and ice roofs. These breaks are in an approximate east and west direction,
and remind one of cross-channel structure of the Mackellar Islets. They may
correspond with the sea-water channel that divides the two portions of the Cape Pigeon
Rocks.

The most conspicuous rock is a massive, light-coloured granitoid gneiss, often
carrying abundant dark aggregates of garnet and mica, which are more or less spherical
in shape and from |in. to 2in. in diameter. Varieties of gneiss are also found without
any garnet at all, and the highest part of the island is formed of an acid hypersthene
gneiss. In crossing the islet areas are found consisting of more strongly foliated gneisses,
and the trend of the foliation is a little west and north. Irregular bands of black gneiss,
with dyke form, exist here as at Cape Denison, and some of them are full of fine garnet.

THE GARNET GNEISSES.

In the various outcrops the garnet gneiss exhibits foliation whose general trend
is a few degrees west of north. Both Garnet Point and Stillwell Island are noted for
the large garnet-mica aggregates which are relics of former complete garnet crystals.

Cape Gray.

At Cape Gray there is a rather coarse-grained rock (No. 784) which has a banded
character in the hand specimen. It contains light-coloured bands of coarse felspar
and garnet, appearing through a darker mass containing mica and garnet. The bands
are irregular, being both thick and thin.

In the slide the rock is heteroblastic and the garnet crystals are much larger than
the other constituent minerals. In part the quartz and felspar form granoblastic
aggregates in which cataclasis is absent and diablastic structure is not common. The
fresh character of this quartz and felspar appears in contrast to the finely granulitic
character of the cordierite. The felspar is chiefly orthoclase and perthite. Microcline
and some lamellar twinned sodic plagioclase are also present. The garnet is pink in the
hand specimen and almost colourless in the section. It appears in small crystals as
well as the large individuals, and usually has an irregular outline. There is a tendency

THE METAMORPHIC ROCKS OF ADELIE LAND. ST1LLWELL. 147

to sieve structure, and the most common inclusions are ilmenite and biotite and, to a
lesser extent, blebs of quartz and felspar. Biotite is present, both in large flakes and
very small crystals. It is pleochroic from a reddish brown to a very pale straw. The
small biotite crystals appear abundantly in cordierite. Cordierite is very prominent
with its pleochroic yellow spots, and has the appearance of a fine granulitic aggregate
produced by the crushing of a large crystal (Plate III., fig. 4). In addition to the biotite,
small garnets, ilmenite and sillimanite are frequent inclusions in the cordierite, and the
whole gives the appearance of a hornfels structure. Sillimanite is associated with the
cordierite, both in the form of matted fibrous aggregates and prismatic needles. But
it is not uniformly distributed, being more abundant in some slides than in others.
Monazite is present, and when included in biotite or cordierite is surrounded by strong
pleochroic haloes. Ilmenite is abundant, though more commonly included in the biotite
and cordierite areas. Pyrite is also present.

The garnet and the cordierite provide the dominant characteristic of the rock,
which may be called a garnet cordierite gneiss.

Garnet, Point.

On Garnet Point there are two dominant types of gneiss in which are incorporated
felspar veins bearing abundant garnet. The first type is rather a dark-coloured gneiss
with abundant biotite. The second type is rather light-coloured and carries the large
garnets which give the mottled appearance to the outcrop.

Specimen No. 772, collected from this locality, is a dark-coloured rock with feeble
schistosity in the hand specimen. The abundant glistening biotite is sometimes
aggregated in circular bunches, and felspar and garnet are visible. In thin section the
rock is heteroblastic, and garnet is much more abundant in some sections than in others.
In part it presents a granoblastic aggregate of biotite and plagioclase with some quartz,
but there are, in addition, circular aggregates up to a quarter of an inch in diameter,
consisting wholly of brown biotite. There are also granoblastic areas with grain size
smaller than the average, consisting largely of biotite and quartz ; and there are lenticles
of quartz and felspar in which the diablastic structure may be prominent. The biotite
is the most abundant mineral in all sections and usually has the same reddish-brown
tint as in the previous case. It has a tendency to a parallel arrangement, except in
the circular aggregates. It is remarkable in the possession of numerous and well-
developed pleochroic haloes ; and the nuclei of these haloes are sometimes large, and
seem to be monazite rather than zircon. Apatite inclusions are also present in biotite
but they are not surrounded by pleochroic haloes. The radius of the halo was measured
by a micrometer eyepiece, and found to be 0-040mm. in several cases, thus agreeing
with the ionisation range ThC and furnishing proof of thorium haloes. In several
cases the haloes show the structure, described by Joly*, of an inner dark and an outer
and lighter corona. The pupil of the halo is nearly always a bit fuzzy at the edge,

* " Pleochroio Halow." Joly and Fletcher, Phil. Mag., 1910, p. 630.

148 AUSTKALAS1AN ANTARCTIC EXPEDITION.

and accurate measurement is therefore impossible. Some measurements give the
radius of the pupil as O031mm., corresponding to the ionisation range of RaC, and this
halo is to be considered as a compound thorium radium halo. Other haloes have been
found to be 0-027mm., corresponding best with the range of ThX, while one case was
found in which there was a suggestion of two coronas, and the radius of the pupil was
0-021mm., corresponding with the ionisation range of RaA. There are also small haloes
with radius 0-01 3mm., which Joly accounts for by the slower moving ray of ionium,
radium, or uranium. The structural features are not always very distinct, but the
measurements indicate that haloes exist in the rock which are thorium haloes ; others
are radium haloes ; and others are a mixture of thorium and radium. It is certain
that the thorium haloes predominate. If monazite is the common nucleus in this rock,
we should expect a mixture of thorium and radium in one halo, because monazite may
contain up to 18 per cent, of Th0 2 as well as some radium.

The felspar is usually in clear grains with granular outline. It frequently shows
good sharp twin lamellae and is found to be andesine. In the lenticles, which are com-
paratively free from biotite, the felspar is often more cloudy and shows conspicuous
sieve structure as well as diablastic structure. Some untwinned orthoclase may be
present. Quartz is clear and most abundant in the areas associated with garnet and
biotite. The garnet has a very pale pink colour and is found in part as rounded grains
with corroded outline, and in part as skeletal crystals noticeably associated with quartz
and biotite. The larger garnet grains, which have suffered less alteration, may be
surrounded by a pale greenish mica, distinct from the normal brown biotite. This
pale mica may follow all the cracks that penetrate the garnet crystal, and it may pass
by direct transition into the brown biotite. Pleochroic haloes are equally abundant
in the two types of mica, but they seem to show more often the structure zones in the pale
green type, i.e., they are less often over-exposed. Moreover, the circular zone of the
halo, situated in the pale green mica, is often changed to the brown type of biotite.
Matted fibres of sillimanite may also be present in the quartz biotite areas. It seems
evident that the garnet has reacted with the felspar, and possibly sillimanite, and has
produced biotite and quartz. Such a change is quoted by Grubenmann* as an example
of a zonal change in passing from the conditions of the kata zone of metamorphism
to those of the meso zone of metamorphism. In other examples it will be considered
that sillimanite is not a necessary factor in this reaction, but in this case sillimanite
has been seen associated with the reaction areas.

Cordierite, with its pleochroic yellow spots, is also associated with the same areas
of relic garnet and sillimanite. Within the granoblastic area of biotite and plagioclase
coarse crystals of a colourless mineral may be found. It has a moderately high refractive
index and oblique extinction in a section showing cleavage. Sections with imperfect
cleavage are normal to a bisectrix, and the mineral is negative. There appears to be a
simple twin whose two halves show a marked change of colour in parallel polarised light
without difference in extinction. These characters cause the identification of cyanite.

* Grubenmann, op. cit., vol. I., p. 52.

THE METAMORPHIC ROCKS OF ADELIE LAND STILLWELL. 149

This identification has been confirmed by the preparation of more sections, in which
we learn that the alteration of garnet to quartz and biotite is not the complete story
of the change. Granoblastic areas are found which consist of cyanite and the pale green
mica which is developed from the garnet. It seems, therefore, that the normal reaction,
which produces quartz and biotite, may be replaced by one which produces cyanite
and biotite. In the latter case there has been an excess of AJ 2 S present, and possibly
corundum has been involved. Ilmenite is not as abundant as in the Cape Gray gneiss,
and monazite and apatite occur as accessories. The rock may be called a cyanite
biotite gneiss produced from a garnet cordierite gneiss.

Specimen No. 770, obtained from the same locality, is similar to the preceding,
though cyanite is not found in it. The hand specimen consists of the biotite gneiss
with a piece of felspar garnet vein attached. In the slide the vein consists of colourless
areas of orthoclase, perthite, soda plagioclase with abundant myrmikite and its diablastic
structure. Occasional areas of ilmenite (with its alteration product leucoxene) are also
associated with the biotite. As hypersthene has been found in similar veins in a similar
locality (Stillwell Island) it is not at all impossible that these may represent the decom-
position of hypersthene. In addition to the colourless areas there are large garnet
areas in the hand specimen of the vein, with which biotite is associated. The biotite
fills up the cracks and surrounds detached pieces of garnet, while the outline remains
that of a large crystal. In the slide of this rock the relic areas containing sillimanite,
garnet, and cordierite are more prominent than in No. 772. The sillimanite is found in
coarse prismatic needles as well as in fibres, and is occasionally in parallel position with
the biotite. There are also the aggregates of biotite and quartz which have certainly
developed in the same way as No. 772. Sometimes the normal brown biotite is replaced
by a much paler mica crowded with opaque magnetite dust. Some reaction has caused
the separation of the iron content of biotite as magnetite. Associated with the biotite
are numerous needles and grains of a yellow-brown mineral with high refractive index
and double refraction, and with a tendency to be opaque. It is frequently included in
biotite and is never surrounded by pleochroic haloes, and is considered to be a variety
of epidote. Aggregates of muscovite are occasionally found with the biotite, while some
of the biotite flakes are bent, crushed, and broken.

Specimen No. 777 is an example of the second type of gneiss from Garnet Point,
and contains the large porphyroblastic garnets. The hand specimen is massive, and
shows felspar and quartz, as well as the pink garnet. In the section the porphyroblastic
garnets are found in skeleton form and penetrated by quartz and biotite. These may
appear as inclusions in the garnet or else along the cracks and edges developed by inter-
action with the felspar. The biotite, as before, contains the pleochroic haloes, and is
here again found to develop through a yellowish-green micaceous mineral. Areas of
aggregated biotite and quartz with ilmenite are present, as in the preceding examples.
Apart from the garnet areas, the rock consists of a granoblastic aggregate of cloudy
plagioclase, orthoclase, microperthite, and quartz. The felspar is cloudy, partly

150 AUSTRALASIAN ANTARCTIC EXPEDITION.

through serialisation and partly through saussuritisation. Scattered through these
felspathic areas are small and large crystals of monazite, while in local patches there
are numerous small crystals of corundum. The corundum is more or less rounded in
cross sections, while longitudinal sections are long prisms with evidence of basal cleavage.
The rock may be called a garnet felspar gneiss.

The Junction of the Cyanite Biotite Gneiss with Amphibolite. Specimens showing the
junction of this type of gneiss with the amphibolite are in the collection. Macroscopi-
cally there is a short and rapid transition from the gneiss to the amphibolite. The
line of junction is straight, and by no means indented as it would be if the gneiss had
been eaten away by the invading amphibolite. The dykes are relatively small and the
transition can not readily be explained by assimilation.

Under the microscope there is perfect crystalline continuity across the junction,
and hornblende appears and increases in quantity with the diminution of biotite and
garnet (Plate III., figs. 5 and 6). The cyanite and quartz seem to travel further into
the amphibolite than the biotite and garnet.

The gneiss in the specimens is similar to the biotite cyanite gneiss No. 772, except
that the garnet is more abundant and the character of the cyanite is different. The
cyanite possesses here pronounced lamellar twinning, and its polarisation colours reach
the lower part of the second order colours. Quartz in the section never shows a higher
polarisation colour than a very pale yellowish white, and hence the double refraction
of this cyanite must reach at least 0-019. The highest recorded value for cyanite
is 0-016. Yet it seems necessary to associate this mineral with cyanite. It
is colourless, with a cleavage parallel to the twin lamellae. Cross sections, which show
indistinct twinning, show two cleavages, both of which are oblique to the direction
of the twin lamellae. Crystal outlines are completely absent and it appears in irregular
plates with marked sieve structure. The abundant inclusions consist of biotite, garnet,
quartz, felspar, and ilmenite. Its refractive index is about the same as the other cyanite.
Extinction angles have been measured up to 28 from the lamellae, but they are often
less. In the region of hornblende it is found to pass by direct transition into hornblende.
Part of an individual crystal may be green hornblende, and part the colourless cyanite,
and, further, the cleavage continues indiscriminately through the green and colourless
portion. The hornblende usually has a smaller angle of extinction, measured from
the trace of the lamellae. The regular arrangement of hornblende with the cyanite
in the hornblendic part of the section is in contrast with the irregular inclusions of
biotite in the cyanite on the biotite part of the slide.

The cyanite extends some distance out into the amphibolite, and some of it may
be found in most sections of the amphibolite. Unfortunately, the number of specimens
is limited, and the collection is too incomplete for us to be able to deal fully with this
case of migration. Nevertheless, it is certain that the amphibolite dyke intruded the

THE METAMORPHIC ROCKS OF ADELIE LAND. 8TILLWELL. 151

sediments, now represented by biotite cyanite gneiss, before the recrystallisation, and
that the junction between the two has been rendered indefinite by the recrystallisation.
It also seems probable that a migration of molecules has taken place across the original
boundary during the metamorphism, and the position of the original junction is marked
by the mixed rock. It also seems probable that certain simple minerals like quartz
and cyanite are able to migrate further than the more complex garnet or biotite or
hornblende. We may refer to this as another instance of metamorphic diffusion, and
it would be very interesting to see how far this cyanite could be traced into the
amphibolites. This we are, unfortunately, unable to do from the material in our
collection. The specimens are small, and from them we can only determine that cyanite
is found in amphibolite at least an inch from the apparent contact. All the specimens
of this dyke show fragments of the attached gneiss.

Cape Pigeon Rocks.

Garnet gneisses are recorded in the field notes from this locality, but there are no
specimens of it in the collection. The garnets are noted as being particularly abundant
in part. A different phase of the gneiss with large porphyritic crystals of felspar is also
recorded. Hypersthenic gneisses were collected from this locality and will be dealt
with later.

Stittwett Island.

A garnet gneiss (No. 917), similar to the garnet felspar gneiss (No. 777) collected
from Garnet Point, is obtained from Stillwell Island. The large garnet-mica aggregates
are again a feature in the gneiss on the island. The mica associated with the garnet is,
in most sections, the pale-greenish variety, from which the brown biotite is only feebly
developed ; but in one example (No. 939), collected a little below the summit of the
island, the brown biotite completely replaces the greenish variety. The green biotite
forms the marginal fringe to nearly every fragment of garnet, and there can be no doubt
that the aggregates were originally complete crystals of garnet. Pleochroic haloes are
abundant in the green mica, and we notice, again, that the alteration caused by the
radio-active particle has caused transition to brown biotite. In such cases the brown
biotite emphasises the halo area, and an example has been noticed where only the inner
ring of a halo with structure is marked by the brown biotite.

This specimen shows in part more evidence of cataclasis than the Garnet Point
example. Granulation of the quartz, which has developed with the mica from the
garnet and mortar structure, are present, though not in any marked degree. Large
garnets (No. 9176) occasionally show cataclasis and are then represented by a granular
aggregate. The granulated garnet, like the granulated quartz, may be drawn out in a
linear manner in the direction of schistosity. The plagioclase is very cloudy, and the
diablastic and sieve structures are prominent. Large orthoclase and perthite may
be traversed by lines of sericite. Occasionally we find areas of cordierite with its

152

AUSTRALASIAN ANTARCTIC EXPEDITION.

pleochroic spots associated with granular garnet and biotite ; and sometimes the small
crystals of corundum appear as in the related example. Ilmenite, pyrite, and epidote
have been noted. The rock may be called, like No. 777, a garnet felspar gneiss.

Chemical Characters.

Analyses have been made of the garnet cordierite gneiss from Cape Gray and of
the cyanite biotite gneiss from Garnet Point. The analyst is A. G. Hall, Victorian
Geological Survey Laboratory.

I. II.

Si0 2 60-93 55-39

A1 2 3 18-09 18-36

Fe 2 3 1-88 .... 1-76

FeO 5-55 6-81

MgO 4-54 4-74

CaO 0-90 .... 2-79

Na 2 1-78 .... 3-36

K 2 3-89 .... 3-74

H 2 + 1-15 1-46

H 2 - 0-14 0-13

Ti0 2 1-07 .... 0-86

P 2 5 tr 0-14

SO S nil .... nil

Cl tr tr.

MnO 0-14 0-24

NiO, CoO 0-02 .... 0-03

CoO tr tr.

LizO strong tr tr.

Total 100-08

Sp. Gr 2-752

Group Values.

I.
II.

68-2
62-5

4-6
6-2

1-1
3-3

14-3
15-9

6-0
2-6

99-81

2-804

1-5
1-0

Projection Values.

4-6
4-9

1-1
2-6

14-3
12-5

I. Specimen No. 781-Garnet Cordierite Gneiss, Cape Gray, Adelie Land
II. Specimen No. 772-Cyanite Biotite Gneiss, Garnet Point, Adelie Land.

THE METAMORPHIC ROCKS OF ADELIE LAND.-STILLWELL.

153

The outstanding feature of these analyses is the high percentage of alumina, a
considerable excess over the 1 to 1 ratio necessary to satisfy the lime and alkalies. In
addition there is, in both cases, a dominance of MgO over CaO, and of K 2 over Na20.
Bastin's criteria of sedimentary origin* are therefore satisfied.

The total alkali percentage of No. 772 is high, and finds mineral expression in the
abundance of biotite and of felspar. The felspar is less important in No. 784, and there
is a greater dominance of K 2 over Na 8 than in No. 772. The great dominance of
alkalies over CaO in each case is an important factor when considered with the silica
percentage. The low CaO percentage of No. 784 means that there can be little lime
in the abundant garnet and probably, also, little CaO in the plagioclase. In No. 772,
where the plagioclase is more important, there is three times as much CaO as in No. 784,
but there is still the large excess of MgO.

The high alumina percentage is reflected in the Group Value T, and the values
for T are again indicative of sedimentary origin. T is lower in No. 772 because the
higher alkali percentage absorbs more A1 2 8 . The value F, which expresses the
ferromagnesian constituents, is very high in both cases. As a consequence the projection
values are dominated by the excessive value of f. No. 784 occupies a position (fig. 12)

8 C

Fig. 12.

II. Mean Position of Group II., the Aluminium Silicate Gneisses.

III. Mean Position of Group III., the Plagioclase Gneisses.

784. Garnet Cordierite Gneiss, Cape Gray.

772. Cyan te Biotite Gneiss, Garnet Point.

- Chemical Composition a* a Criterion in Identifying Metamorphoaed Sediment*," E. S. Baitin, Joum. Geol., 1909
Tol. 17, p. 445.

154 AUSTKALASIAN ANTARCTIC EXPEDITION.

on the triangular projection close to the mean position of Group II., the group of the
aluminium silicate gneisses. No. 772 also lies in the same area, but in a direction tending
towards the mean position of the plagioclase gneisses.

Classification.

The gfoup values, considered collectively, place both rocks in the group of the
aluminium silicate gneisses. In all cases the values fall within the assigned limits of
this group, except the value of T in No. 772, which is just below the lower limit (3-0).
The other types from this region, viz., Nos. 770, 777 from Garnet Point, No. 917 from
Stillwell Island, No. 785 from the Cape Pigeon Rocks, probably belong to the same
group, though certainty is unattainable without a chemical analysis.

The dominating garnet and cordierite with sillimanite in No. 784 means that the
recrystallisation of the sediment took place under conditions of very high temperature
and great hydrostatic pressure, i.e., under the conditions of the kata zone of
metamorphism. No. 784 is therefore placed in the kata division of the aluminium
silicate gneisses. In the rocks Nos. 772, 770 from Garnet Point it has been noticed
that the garnet and felspar with, perhaps, sillimanite have been replaced by biotite
and quartz. This change is considered to occur in transition from the kata zone to the
meso zone. When the meso zone metamorphic conditions are dominant, the areas of
sillimanite, cordierite, and garnet become areas of secondary relics and indicate the
double phase of metamorphism of the original sediment.

No. 777, the second type from Garnet Point, containing enormous garnets, also
shows trace of the meso zone conditions. The large garnets show considerable alteration
to biotite and quartz, and some of the felspar is dissociated into perthite and myrmikite.
The small development of sericite and saussurite brings in an epi zone element. Hence,
though the abundant large garnet rocks at Garnet Point are indicative of kata zone
metamorphism, there is also the impress of meso zone conditions, which is sufficient
to place these examples in the meso division of the aluminium silicate gneisses.

The garnet felspar gneiss, No. 917, from Stillwell Island, is like the garnet felspar
gneiss from Garnet Point, and shows evidence of meso zone conditions. In this case
cataclasis is present, and we find a portion of the quartz and garnet granulated. The
felspar has become more cloudy, owing to further sericitisation and saussuritisation.
Hence, while the evidence of meso zone conditions is greatest, there appear the initial
stages of epi zone metamorphism.

THE ACID HYPERSTHENIC GNEISSES OF STILLWELL ISLAND AND THE CAPE PIGEON

ROCKS.

In addition to the garnet felspar gneisses a second type, related to the acid
hypersthene gneisses of Madigan Nunatak and Aurora Peak, was discovered on Stillwell
Island and the Cape Pigeon Rocks. In one case on Stillwell Island this type of gneiss
appears in dyke form. The same form of occurrence is strongly suspected at the Cape
Pigeon Rocks, and similar rocks can be remembered at Garnet Point, though no specimens
are in the collection.

THE MKTAMORPHIC ROCKS OF ADEL1E LAND. 8T1LLWELL. 155

Stittwett Island.

No. 949 is an example of this rock type, and was collected from a fine-grained band
several inches wide, which crossed the gneiss irregularly near the summit of the island.
This specimen is a rather dark-coloured rock with a vitreous lustre. Quartz and felspar
are visible with a lens, and specks of pyrite are sprinkled unevenly through it.

The section consists of an even-sized granoblastic aggregate of quartz and felspar,
through which grains of pyroxene and its associated biotite and ilmenite, pyrite and
apatite are scattered (Plate III., fig. 3). The average absolute grain size is approximately
0-20mm. There is a general absence of crystal boundaries, and the manner in which
blebs of quartz are set in the felspar is clearly metamorphic. The felspar is very clear
and unaltered, and includes orthoclase and plagioclase. Lamellar twinned individuals
have a refractive index often above basal quartz. The large extinction angle, measured
from the trace of the lamellae, is 20, and hence the felspar is andesine. The pyroxene
is largely hypersthene, and only a few grains do not possess straight extinction*. The
pale-pink to pale-green pleochroism is very marked. The same green serpentinous
alteration product which appears associated with the hypersthene in the Madigan
Nunatak and Aurora Peak gneisses is found in this rock. In part biotite and ilmenite
are developed in its alteration. The biotite may be mixed with the green alteration
product, but probably the green mineral has developed after the biotite, as the latter
may be associated with perfectly fresh hypersthene. Whenever the green mineral
appears the alteration is more advanced. The association of the ilmenite with the
biotite is fairly constant. Apatite, zircon, and pyrite are accessory minerals. The
rock may be called a hypersthene felspar gneiss.

This rock has not suffered the subsequent crushing that is evident in the Madigan
Nunatak gneiss, and the absence of garnet makes it different from the Aurora Peak
gneiss. It therefore possesses, without any modification, the characters of the Indian
charnockitet. A rough determination of its specific gravity gave the value 2-67, which
is the same as that for normal charnockite, and greater than the specific gravity of the
Madigan Nunatak gneiss, and just a little less than that of the Aurora Peak gneiss. Its
composition would not be very different from that of the Aurora Peak rock, and would,
therefore, possess the igneous characteristics which are in agreement with the dyke
form of its occurrence.

No. 979. Another example of gneiss, related to the preceding charnockite-like
rock, is No. 979, which was collected about 150yds. from the boat moorings at Stillwell
Island. In the field it was noticed to be unusually free from garnet.

* Grain* of pyroxene with apparently oblique extinction hare been shown by Washington (Amer. Journ. Sci., vol. XLI.,
4th ST., 1916, p. 323) to possess the optical character of hypersthene. This effect is ascribed to the development of a cleavage
other than the usual prismatic cleavage.

t The charnockite series will be discussed later. A special rock name is desirable for the acid hypersthenic gneisses,
yet Holland has definitely asked that the name chamockito should not be used for extra-Indian rocks. Still, if it be acknow-
ledged that the Indian charnockite series does not consist merely of phenomenal igneous rocks but of definite metamorphic
types, then it may be suggested that, with Holland's permission, " charnockite " should supply the need.

156 AUSTEALABIAN ANTARCTIC EXPEDITION.

In the hand specimen it is a coarse brownish-coloured rock, showing quartz, felspar,
and hypersthene, and a little biotite. In section it is a coarse granoblastic aggregate.
The quartz shows a little granulitisation. The felspar consists of both twinned and
untwinned varieties. Peg structure and diablastic structure are common in the felspar.
The development of the diablastic structure by the dissociation of plagioclase crystals
in situ is very plain in some instances. Lamellar twinned felspar has again a refractive
index occasionally higher than that of the basal quartz, and it is probably andesine.

A brown biotite is the most abundant ferromagnesian mineral. It may be
associated with ilmenite, and it is found in the hypersthenic areas. The hypersthene
in the section is largely decomposed to the serpentinous greenish alteration product.
Biotite is intimately associated with the serpentine and seems to develop from it in a
pale-green form. Sometimes the hypersthene loses its iron content, becomes colourless,
assumes lower polarisation colours, and changes into enstatite. In places a very pale-
green biotite is associated with the hypersthene, and this seems again to be an
intermediate stage between brown biotite and the green delessite. Occasionally large
crystals of apatite and zircon are present. Grains of ilmenite and fragmentary garnet
are near the hypersthene. The garnet may form fragmentary rims around the biotite,
and occasionally the biotite is grouped in radial sprays.

In a second slide of No. 979, cut from the opposite end of the specimen, the sprays
of radial biotite are more prominent surrounding the ilmenite nuclei (Plate V., fig. 3).
Some of the biotite flakes are associated with fan-shaped felspar vermiculse, as in Plate
V., fig. 4. The association of quartz with the biotite sprays is also noticeable, especially
in the aggregates of small basal biotites and quartz, which are in some cases cross sections
of biotite sprays. This slide also contains a plate of basal biotite which is surrounded
by a rim of later biotite straws set in quartz, which in turn has a thin coating of iron
ore (Plate VI., fig. 4). The garnet is more abundant in this slide and may be detected
as fragmentary rims around ilmenite as well as biotite. These features will be found
to be better developed in the next example, No. 947.

In this case we have also noticed large crystals of untwinned plagioclase with
inclusions of orthoclase distributed in the same manner as the schiller inclusions in
olivine in peridotite from the Isle of Rum*. The appearance is not unlike a graphic
structure, but it is distinct from the vermicular intergrowths in the same slide. The
inclusions of orthoclase have a considerably lower refractive index than the untwinned
plagioclase (andesine), and the larger pieces contain minute fusiform inclusions of a
felspar with higher refractive index. The crystal plate of plagioclase possesses cleavage
which extends in places across the orthoclase inclusions, and the extinction angle,
measured from the cleavage, is 4. When the stage is rotated in the opposite direction
an extinction angle of 3 is measured in the orthoclase inclusions. The complete
recrystallisation of this rock, together with the abundant metamorphic felspar
intergrowths, indicates that this graphic-like structure is also of metamorphic origin,
and possibly connected with diffusion phenomena.

* " Natural History of Igneous Rocks," A. Barker, p. 258.

THE METAMORPHIC ROCKS OF ADELIE LAND. 8TILLVVELL. 157

The rock may be described as a meso-hypersthene felspar gneiss. We may conclude
that the hypersthene is less in quantity than in No. 949, because more of it has reacted
with the felspar to produce biotite and ilmenite. It is certain that the type is more
closely related to No. 949 than to No. 917, the garnet felspar gneiss. It has only been
reported as indefinite bands associated with the garnet gneisses, which are considered
to be sedimentary in origin, but it possesses undoubted affinities to igneous rocks and
to No. 949, which occurs as a dyke. It is quite likely, therefore, to be the metamorphosed
equivalent of an acid dyke whose identity has been wholly or partially lost by the
operation of metamorphic diffusion.

No. 947. Another example of the hypersthenic gneiss was obtained from the
summit of the island. It is a coarse rock similar in outward appearance to No. 979.
The brownish colouration of the rock is a little more prominent, and, at the same time,
more like the brownish coloured rocks of Madigan Nunatak. While related to the
preceding example, No. 979, it differs from it in possessing much more garnet, more
pyroxene, less biotite, and very little quartz. Yet a rough determination of the specific
gravity gave 2-74 in the. first case (No. 979) and 2-76 in the second (No. 947).

The place of the quartz in No. 979 is taken by areas of untwinned felspar (orthoclase)
with cryptoperthitic inclusions. Augite, as well as hypersthene, is present. The pink
garnet appears in two ways : it may appear first as large crystals with felspar, biotite,
and ilmenite inclusions, or it may appear as granular garnet surrounding ilmenite, biotite,
and hypersthene. The garnet rims around the ilmenite and biotite clearly follow closely
all the irregularities in shape of the ilmenite and biotite nuclei (Plate V., figs. 5 and 6).
Biotite and ilmenite are often associated with the pyroxene, and have no doubt been
formed from it in part in the familiar reaction with felspar. The biotite may partly
enclose the pyroxene and it may fill cracks and indentations in the pyroxene crystals.
Further, as the garnet rims around the pyroxene may be in part separated from the
pyroxene by biotite, we can fairly safely conclude that an explanation of the garnet
around the biotite will provide an explanation of the garnet rim around the pyroxene.
Again, we can find crystals of ilmenite symmetrically enclosed by a biotite zone, and,
if this biotite zone were replaced by a garnet zone, we should get the garnet rims
around ilmenite as are observed. Consequently, an explanation of the garnet-biotite
reaction will also provide an explanation of the garnet zone around ilmenite. This
conclusion is supported by the discovery of an aggregate in which an ilmenite crystal is
surrounded by biotite, and this in its turn is practically surrounded by garnet, while
pyroxene crystals jut against it in part.

It is therefore clear that biotite is on one side of the equation and garnet on the
other. Sometimes outside the garnet rim a change in character of the plagioclase is
quite evident. Hence the plagioclase may be considered to take part in the reaction
and to supply a lime molecule which may enter the garnet. The following equation
shows that for average compositions of biotite and garnet the change is chemically

158 AUSTRALASIAN ANTARCTIC LXPEDiTlON.

possible, and at the same time an explanation why the quartz of the related type (No.
979) is replaced by orthoclase in this example (No. 947)*.

(KH) 2 (MgFe) 2 A1 2 (Si0 4 ) s + CaAl, Si 2 8 + 4Si 2
Biotite Anorthite Quartz

^ 2(MgFe)0,CaO, A1 2 3 , 3Si0 2 + 2KA1 Si s 8
Garnet Orthoclase

The related rock (No. 979), in which there is very little garnet but much quartz
and plagioclase, can then be explained as a rock type in which the biotite side of this
equation is expressed. No. 947, in which there is considerable garnet and orthoclase
and practically no quartz, may be looked upon as a rock type in which the garnet side
of this equation is expressed.

We have interpreted the reaction of the hypersthene with felspar in these
hypersthenic rocks to biotite and ilmenite as associated with a change of kata zone
conditions to meso zone conditions. If, after these changes, the biotite reacts with
quartz and felspar to produce garnet and orthoclase, we should, on the same reasoning,
interpret the cause as a reversal to the kata zone conditions. For this we cannot imagine
any variation in the depths of the earth's crust, because there is no similar evidence
in any other variety of gneiss on Stillwell Island. We can only imagine that the
temperature and pressure have been increased locally, possibly by neighbouring chemical
reactions which have liberated heat and caused expansion of volume. If this were so
the area subjected to the reverse conditions would be highly localised.

The specimen No. 947 is an irregularly shaped piece about Sin. long and, roughly,
l^in. square in section. A second section was cut from the opposite end of the specimen,
distant Sin. from the first section. In this section no garnet is found, but large pieces
of quartz. Hypersthene is present, again showing some alteration to ilmenite and
biotite, but none of the biotite is rimmed with garnet. This result was surprising,
and a third section was cut from the middle of the specimen, half-way between the two
previous sections. In this middle section some garnet is found, but less than in the
first section. It again borders biotite in the same remarkable manner. There is some
quartz in the section. We have also noticed in this section a large crystal of pale-green
mica with included ilmenite. This green mica is evidently an intermediate stage between
delessite and biotite, but it is not possible to say in which way the reaction is going.
Hence the supposition of the highly localised distribution of the garnet rims seemed
to be confirmed.

The rock may be described as a hypersthene alkali felspar gneiss, in which the
hypersthene has first partly changed to ilmenite and biotite. This change has been
followed in localised portions of the rock by a partial reaction of biotite with quartz
forming garnet and orthoclase.

Similar conclusions may be formed about the primary igneous origin of No. 947,
as in the case of No. 979.

* The (KH) molecule is reckoned as K 2 for simplicity. In the analyses of some biotites the K 2 is in great excessover
the H ; O, and this is assumed to be the case here.

T11K MKTA.MORl'HIC ROCKS OF ADEL1E LAND STILLWELL 159

Cape Pigeon Rocks.

The hypersthene gneiss from the Cape Pigeon Rocks possesses many of the
peculiarities noted in the preceding rocks from Stillwell Island. The specimen, though
a little larger than No. 947, is no more than 3^in. long, and reveals the same remarkable
variation in mineral content. Four sections cut from different portions of the specimen
have been necessary to understand the character of the rock. These will be dealt with
separately in order to again illustrate this variation. A rough determination of the
specific gravity of the specimen No. 785 gave the value 2-75, and, therefore, its total
composition is likely to be very similar to the composition of No. 947 or No. 979.

No. 785 (1). No. 785 (1) was the first slide cut and examined from the specimen
from the Cape Pigeon Rocks. In it there are only scattered fragments of garnet which
has been largely replaced by biotite and quartz. A crystal of ilmenite often occupies
the central position of the biotite aggregates as before. The larger biotites are some-
times bent or crushed, but they often open out into radial sprays set in quartz, which
again enter into the fan-like myrmikoidal intergrowths of felspar. In other cases we get
aggregates of small biotites with small quartz crystals. Pleochroic haloes are still
common. The felspar is often cloudy and in part there is a good deal of sericite. In
part the orthoclase is transformed into microcline. The plagioclase has. in most cases,
a refractive index less than quartz and a small extinction angle, and is probably an
oligoclase andesine. Some of it includes the common blebs of rounded quartz, and it
frequently presents a diablastic structure. Along the junction of two felspar crystals
we may find one of them bordered with a diablastic zone. In one instance where the
diablastic structure has developed in a corner of a crystal, the twin lamellae can be traced
from the unaltered part through the diablastic area. Some of the plagioclase is
saussuritised, and epidote is found both in sporadic grains and in the finely granular
form with the saussurite. Chlorite is more abundant in this slide than in the others.
Apatite appears in fairly large crystals, and pyrite and zircon are also accessory. No
hypersthene is present ; and on this description alone the rock would have to be named
a biotite felspar gneiss.

No. 785 (2). The slide No. 785 (2) is cut from the opposite end of the specimen,
distant 3jin. In general, there is less chlorite, epidote, saussurite, or sericite than in
the preceding slide. The garnet rims are well developed, and these, with the presence
of hypersthene, indicate the relation of the rock to the hypersthenic gneisses of Stillwell
Island. In the hand specimen there is nothing to indicate this variation. The igneous
origin of this rock type is further evidenced by the large, well-defined crystals of apatite
and zircon.

The hypersthene possesses a beautiful and intense pleochroism from pink to green.
The depth of the pleochroism in hypersthene is usually associated with the iron per-
centage ; but when one recalls the pleochroism of the titaniferous augites, it seems
probable that the deeply pleochroic nature of these hypersthenes may be partly due to

160 AUSTKALASIAN ANTARCTIC EXPEDITION.

the Ti0 2 content. The hypersthene may contain ilmenite inclusions, which are situated
either irregularly or in planes. Occasionally the hypersthene loses its colour and
pleochroism, assumes the lower polarisation of enstatite and changes into enstatite.
Sometimes it is partially replaced by a platy brown mineral with the deep red brown
colour of biotite but with very low polarisation colours. This brown mineral is an
iron-stained serpentine.

Biotite is again abundant and appears in large platy crystals, in close aggregates
of smaller crystals surrounding ilmenite, and as small crystals set in quartz. The fan-
shaped biotite sprays may appear in the zone around ilmenite or with the biotite plates.
Some of the biotite flakes are crushed and bent. Sometimes the biotite plates possess
a dark border in which the integrity of the plate is broken. A slight perforated appear-
ance develops and the dark colour is due to the separation of minute crystals of iron ore.
It is an alteration which is either associated with the crush phenomena or else with the
reaction which produces the biotite sprays. A further state is noticed where the biotite
has completely lost its brown colour, though still surrounded by a fragmentary garnet
rim. It has assumed a pale greenish colour and is dotted with small magnetites (or
ilmenites) but still retains its bright polarisation colours. Residual patches of brown
biotite may remain in the pale biotite, and as chlorite is present in the slide this may be
interpreted as the passage of biotite into chlorite.

A feature of this slide is the presence of garnet rims similar in nature to those in
No. 947. Apart from the rims, garnet only occasionally appears in moderate sized
crystals. The garnet rims may surround biotite and hypersthene, and are usually
composed of small, idioblastic crystals. The rims have not been observed around ilmenite
as in No. 947, but an excellent example is found of an ilmenite nucleus, surrounded by
biotite, which in turn is surrounded by a garnet rim (Plate VI., fig. 1). We also find
the hypersthene surrounded by biotite and this in turn by garnet (Plate V., fig. 2). A
thin layer of orthoclase may exist between the garnet and the biotite, but it is often
absent. The garnet may come into direct contact with the hypersthene, and may
even penetrate the hypersthene in seams. As biotite is often intimately mixed with
the hypersthene it is possible, in many cases, to still explain the presence of the garnet
in the hypersthene by a biotite-plagioclase-quartz reaction as in No. 947 ; but the
examination of the fourth slide of this specimen proves that this explanation is
inadequate in certain cases. The biotite-plagioclase-quartz reaction still explains the
garnet rims on the biotite, but all the garnet is not so formed.

We find here, also, that the garnet rims surround aggregates of biotite and quartz
(Plate V., fig. 1). Some of these have a definite rectangular outline and others may be
irregular or approximately hexagonal or octagonal. The aggregates are very often
without ilmenite, but they may enclose fragments of enstatite. The definite outline
indicates that they formerly surrounded a single crystal, and that they existed before
the biotite-quartz aggregate. The presence of a portion of an enstatite crystal suggests
that the original mineral was a pyroxene. As, in addition, we may find the biotite-
quartz aggregate containing an ilmenite nucleus and scattered fragments of garnet,

THE METAMORPHIC ROCKS OF ADELIK LAND. STILL WELL. 161

extending as a circular bight into the side of a hypersthene crystal (Plate IV., fig. 3),
there can be little doubt that the hypersthene has taken part in the formation of this
aggregate.

The radial arrangement of the small biotite crystals is very noteworthy, and it is
so constantly associated with the intergrowth of felspars (Plates IV., fig. 5 ; V., fig. 4).
The intergrowth has normally a fan-shaped arrangement and frequently branches from
a biotite flake (Plate XVI., fig. 6), and there can be little doubt that there is a genetic
connection. The biotite rosettes often surround an ilmenite nucleus, and, while there
is a similarity with an ordinary zone, significance must be attached to the different
structure. Though it is difficult to offer definite proof, the whole arrangement suggests
the reversal of the biotite-garnet reaction described in No. 947. If we imagine first
the formation of the garnet zone around ilmenite as in No. 947 (Plate V., figs. 5 and 6),
and then a reversal of the metamorphic conditions to those on the biotite side of the
reaction, whereby the garnet disappears and the biotite reappears, we might get the
rosetted biotite zones. Such hypothesis provides an intelligible account of the connection
between the biotite sprays and the associated myrmikoidal felspar. The evidence of
this reversal includes the presence of the garnet fragments in the quartz-biotite zone
(Plate IV., fig. 3) which lies between a hypersthene crystal and a large ilmenite surrounded
by biotite rosettes. Secondly, a break in the garnet zone around biotite (Plate IV.,
fig. 3) is found, and the break is marked by a biotite spray which opens out into a
myrmikoidal fan which is only visible between crossed nicols, and, therefore, not seen
in the photograph. The sprays are also poorly developed on the outside of this garnet
rim. The incompleteness of the garnet rim around ilmenite and biotite (Plate VI.,
fig. 1) may be explained in the same manner.

The felspar in the slide consists of orthoclase, plagioclase (oligoclase andesine), and
the myrmikoidal intergrowths. Blebs of quartz may be set in the plagioclase which may
be rimmed with the intergrowths (Plate IV., fig. 5). A case has been noticed where the
plagioclase is separated from orthoclase, containing abundant minute fusiform inclusions
with higher refractive index, by a zone of intergrowtha. Large crystals of quartz are
irregularly distributed through the slide, in addition to the fine quartz in the biotite
aggregates and in the felspar. Pyrite and zircon are accessory.

No. 785 (4). No. 785 (4) is a second slice cut from the same end of the specimen
as No. 783 (2). In many respects this slide is similar to the preceding, but there is a
little less hypersthene and garnet. The large crystals of apatite are still prominent,
and we have now noticed that the garnet rims may extend on to a crystal of
apatite.

The garnet rims extend around the biotite-quartz aggregates in a manner noted
in the previous slide. Now the garnet, in addition to the rims, may extend as seams
through the aggregate in the same way that has already been seen in the hypersthene.
Further, the relic hypersthene in the same aggregates leaves no doubt whatever that the

Serial A, VoL ra., Part 1 L

162 AUSTRALASIAN ANTAECTIC EXPEDITION.

biotite and quartz can be produced in a reaction in which the hypersthene has taken part.
For average compositions of these minerals the reaction may be expressed chemically
as follows :

(KH) 2 0, 2(FeMg)0, A1 2 3 , 3Si0 2 + 5Si0 2 ^ 2J(MgFe)0, Si0 2 i + K 2 0, A1 2 3 , 6Si0 2 .
Biotite Quartz Hypersthene Orthoclase

Excess iron may separate out as iron ore, and K 2 is assumed again to largely dominate
over H 2 in the biotite. This reaction has undoubtedly followed the production of
the garnet from biotite or pyroxene, and the orthoclase which accompanies the formation
of the garnet may react again with the hypersthene.

In one case a biotite crystal, partly crushed, extends into a hypersthene aggregate
(Plate IV., fig. 4). Part of the biotite has a garnet rim, and one corner of the biotite
area has a perforated appearance with the development of quartz, and is in intimate
relation with the biotite. There is only occasional chloritisation of the biotite and
serpentinisation of the hypersthene.

No. 785 (3). The slide No. 785 (3) is cut from the middle of the specimen. The
same general features can be recognised here ; but there is less biotite and more hyper-
sthene, most of which is considerably altered to serpentine.

A new feature appears in this slide in a large aggregate of hypersthene and altered
hypersthene in which the outlines of the crystals are marked by thin garnet borders
(Plate VI., fig. 2). The garnet also penetrates some of the crystals in thin irregular
seams. Similar seams have already been noticed in fresh hypersthene and in the garnet-
rimmed areas of biotite and quartz. In this aggregate the original hypersthene crystals
have assumed a pale-green colour, are slightly pleochroic, and are in part finely fibrous.
The least altered still have the polarisation colours of hypersthene, but in many cases
the mottled colours of serpentine appear. Strong pleochroic haloes appear in the ser-
pentine. The alteration takes place here through bastite to serpentine.

During the serpentinisation a considerable amount of magnetite (or ilmenite) has
separated out ; and this separation is well illustrated in a crystal of partially altered
hypersthene, which is apart from the aggregate. The centre of this crystal is still the
unaltered pleochroic hypersthene ; but its low polarisation colours indicate that its
iron content is small, and that it is passing over into enstatite. The outer portions
have changed to clear enstatite or to serpentine, which is brownish in part ; but along
the fringe of the crystal there are numerous, small, opaque crystals formed from the
liberated iron.

In another crystal the hypersthene has completely changed to enstatite in which
serpentinisation has freely developed along the cracks and fractures in a manner common
in olivine. This enstatite crystal is seamed irregularly with planes of colourless garnet,

THE METAMORPHIC ROCKS OF ADKLIE LAND. STILLWELL. 163

and it is certain that the development of the garnet took place prior to the serpentini-
sation. It probably occurred before the development of the enstatite, because we have
previously noted garnet seams in deeply pleochroic hypersthene.

Sometimes these altered crystals of hypersthene contain inclusions of biotite, or
of ilmenite surrounded by biotite sprays with associated quartz. But biotite crystals
are mostly confined to the margin of the serpentine hypersthene aggregate. These
biotites may have the normal garnet rim, produced, no doubt, in the same manner as
before by a biotite-plagioclase reaction. Such reaction may explain the garnet fringe
around the edge of the aggregate, but it will not reasonably explain the rims and seams
in the inner part of the aggregate. The garnet has also been derived in some other
manner.

It seems necessary to account for the garnet without any reaction at all, and to
assume that the garnet is derived directly from the hypersthene. Holland* has reported
the decomposition of augite into garnet and felspar ; but there is no reason to suppose
that this instance cannot be explained as has been done in the garnet plagioclase
pyroxene gneiss (No. 953) of Stillwell Island, in which augite has reacted with labradorite
to produce garnet with andesine and quartz. In the same publication Holland refers
to the description by Brauns in 1888 of the formation of a lime iron garnet in a palaeopi-
crite by the alteration of augite in which the chemical analysis indicated a removal of
A1 2 S . L. Hezner mentions the record of a pseudomorph of garnet after augite by
Pelikan.f These instances, however, are probably not parallel with the present instance.

Van Hise quotes the change of pyrope into enstatite, spinel, and quartz. J It is
not unlikely that this reaction is reversible, with suitable conditions, and pyrope may be
derived from enstatite, provided the suitable amounts of A1 2 3 are available. If this
is so then the type of reaction may be indicated thus

SMgSiOg + A1 2 8 <- Mg 8 A1 2 Si, 12

The hypersthene that enters into the reaction.very probably contains some alumina.
According to Dana, hypersthene may contain as much as 10 per cent, of A1 2 8 , and a
Victorian example in a titaniferous dacite was found by Richards to contain 4 per cent.
It can be conceived that the A1 2 8 content of the hypersthene may provide the alumina
in the above reaction, because the amount of serpentinised hypersthene is much greater
than the amount of garnet formed. The iron content of the hypersthene may separate
out as iron oxide as in the formation of enstatite or enter the garnet molecule. Any
content of lime in the hypersthene would also enter the garnet molecule.

As a result of the examination of these four slides we think the most comprehensive
name is hypersthene biotite felspar gneiss.

T. H. Holland. " Origin and Growth of Garnet*," Rec. G.S.I., XXIX., p. 20.
t Op. cit., L. Hezner, p. 67.
J " Treatise on Metamorphism," p. 304.

" On the Separation and Analysis of Minerals in the Dacite of Mount Dandenong, Victoria," H. C. Richards, Proo
Roy. Soc. Vic., vol. XXI., n.s., p. 533.

164 AUSTRALASIAN ANTARCTIC EXPEDITION.

THE CRYSTALLOBLASTIC ORDER.

It is found difficult in some cases to satisfactorily name the crystalloblastic order.
This order contains a list of minerals which have arisen more or less simultaneously
during its recrystallisation. If all the minerals in the rock have not formed at the
same period, then they cannot be placed in a single crystalloblastic order, and it
frequently happens that a rock carries traces of two metamorphic phases each
characterised by certain minerals. Sometimes two minerals in a rock may not come
in contact, and their relative position in the order cannot be fixed. In some cases, as
at Madigan Nunatak, the contacts may be wholly or partially replaced by areas of
pulverised material.

In the garnet cordierite gneiss, garnet and sillimanite crystals have not been observed
in contact, and are bracketed in the crystalloblastic, which appears to be as follows :
Garnet, sillimanite, ilmenite, biotite, felspar, cordierite, quartz.

In the cyanite felspar gneiss from Garnet Point the cyanite exerts its form against
the biotite and must, therefore, be placed above the biotite in the sequence. In this
case the cyanite cannot be compared with the garnet or cordierite, as these are looked
upon as secondary relics from the kata zone metamorphism.

In the hypersthene alkali felspar gneiss, No. 949 (charnockite), the observed order
is Apatite, ilmenite ; hypersthene ; biotite ; felspar ; quartz.

In the less acid, hypersthene felspar gneisses, containing garnet, the garnet is sub-
sequent to the formation of most of the biotite, and is, therefore, omitted from the
sequence, which appears to be Apatite, ilmenite ; pyroxene ; biotite ; felspar ; quartz.

SUMMARY.

Garnet gneisses are obtained from Cape Gray, Garnet Point, and the Cape Pigeon
Rocks three rocky outcrops on the present shore line that are accessible to a sledging
party on the mainland. A fourth locality is Stillwell Island, distant nearly two miles
from the shore line, and was visited by the ship's boat.

The gneisses may be summarised thus

Cape Gray Garnet cordierite gneiss.

Garnet Point Cyanite biotite gneiss.

Garnet felspar gneiss.
Stillwell Island Garnet felspar gneiss.

Hypersthene felspar gneisses.
Cape Pigeon Rocks Garnet gneiss.

Hypersthene biotite felspar gneiss.

The garnet felspar gneisses on Garnet Point and Stillwell Island are light coloured
and mottled by large aggregates of garnet and biotite, which are more or less spherical
in shape and up to 2in. in diameter. These aggregates represent original and complete

THE MBTAMORPHIC ROCKS OF ADEL1E LAND. STILLWELL. 165

crystals of garnet. Cordierite and sillimanite are most prominent at Cape Gray, but
are found at the other localities, where they are interpreted as secondary relics. With
the recession of garnet and cordierite, biotite with quartz and felspar become prominent.
Biotite and quartz have been produced by the reaction of garnet and felspar. The
biotite so produced is a pale green variety which develops later into the normal brown
biotite. The alteration from the green to the brown colour may be effected by the
radio-active rays which produce the pleochroic haloes in the biotite.

The junction between the cyanite biotite gneiss and an amphibolite dyke has been
described at Garnet Point. Near the junction there is considerable garnet and cyanite
in the gneiss. The cyanite is not normal and shows prominent lamellar twinning, but
its double refraction is estimated to be about 0-019 a value higher than recorded values
for cyanite. There is perfect crystalline continuity across the junction, which is only
indefinitely marked by the gradual appearance of hornblende in the section. The
cyanite and quartz travel further into the amphibolite than the garnet or biotite, and
the cyanite has been noticed in the amphibolite at a distance of lin. from the apparent
junction. The cyanite may be intergrown with the hornblende in the amphibolite.
It is not considered possible to explain these features by assimilation of the sedimentary
rock by the igneous rock prior to the metamorphism. The characters of the complex
of sediment and dyke are solely due to the recrystallisation, during which it is supposed
that a limited migration of material occurred across the pre-existing junction, tending
to efface it. It is viewed as another example of metamorphic diffusion.

The chemical composition, as well as the mineral composition, shows that these
gneisses are sedimentary in origin.

The gneisses in each case are placed among the Aluminium Silicate Gneisses in
Grubenmann's classification of the crystalline schists. The rock at Cape Gray, the
most northerly outcrop, is placed in the family of the cordierite gneiss in the kata
division. Kata zone metamorphism is found in each of the other outcrops but is
modified first by meso zone metamorphism and, later, by additional traces of epi zone
metamorphism. At Madigan Nunatak, situated on the ridge which terminates in
Cape Gray and 18 miles due south of it, we have already described the rocks as examples
of kata zone metamorphism modified by strong epi zone features. We now find that
of these four localities, the nearest in point of distance from the Madigan Nunatak is
the locality in which traces of epi zone metamorphism have been described as super-
imposed upon kata zone metamorphism. Cape Gray, the furthest in point of distance
from the Madigan Nunatak, possesses the least modified kata zone metamorphism.
The intermediate localities possess kata zone metamorphism modified by meso zone
metamorphism to a degree sufficient to place the rocks in meso division of the schist
group. In the latter case the specific families of cyanite gneiss and meso garnet gneiss
are represented.

In addition to the garnet felspar gneiss on Stillwell Island, acid hypersthenic gneisses
occur. The only specimen of gneiss, apart from the altered dykes, collected from the

166 AUSTRALASIAN ANTAKCTIC EXPEDITION.

Cape Pigeon Rocks, is also a hypersthene gneiss. It is also probable that similar rocks
occur at Garnet Point. These rocks are related to the acid hypersthenic gneisses of
Madigan Nunatak and Aurora Peak. One example from the summit of the island is
found in dyke form, and is no doubt of igneous origin. This rock is a granulitic
aggregate of quartz, orthoclase, plagioclase, and hypersthene. Biotite and ilmenite are
developed by the reaction of the hypersthene with the felspar. The hypersthene also
changes to a greenish serpentinous mineral, as in the rocks at Madigan Nunatak and
Aurora Peak. The rock is identical in kind with the normal charnockite of the Indian
charnockite series.

Two other examples of hypersthene alkali felspar gneisses are described from Still-
well Island and are distinguished from the first by a higher specific gravity and by the
coarse-grained character in the hand specimen. In one of these there is considerable
quartz, but only fragmentary garnet, and the brown biotite is found developing through
the stage of pale-green biotite. The hypersthene may lose its iron content and change
into enstatite. In the second considerable garnet is found in part, but very little quartz.
The quartz in the first is replaced by orthoclase in the second. The garnet appears
not only as large crystals but also as granular zones surrounding ilmenite, biotite, and
hypersthene. As zones of biotite may surround ilmenite and hypersthene, these garnet
rims may be explained in each case by a reaction between biotite and plagioclase and
quartz, producing garnet and orthoclase. This reaction is found to be highly localised,
being absent from a second section cut at the other end of the specimen, and distant
Sin. from the first section. A third section cut from the middle of the same specimen
shows some garnet. The garnet-forming conditions are, therefore, very limited.

Remarkable variation of a similar kind is found in the specimen of hypersthenic
gneiss from the Cape Pigeon Rocks. The specimen was not more than 3jin. long before
the slicing, and four sections have been studied. In the first of these there is very little
garnet and no hypersthene. At the other end of the specimen there is considerable
hypersthene and the garnet rims are equally developed as in the preceding example
from Stillwell Island. Here, in addition to the ilmenite, biotite, and hypersthene nuclei,
we find the garnet rims enclosing curious areas of small biotites and quartz. These areas
are looked upon as formed by the reaction of hypersthene with orthoclase. In another
case the garnet penetrates the hypersthene crystals in the form of thin seams. In the
fourth there is a curious aggregate of hypersthene and altered hypersthene in which the
outlines of the crystals are marked by garnet borders. The garnet rims are in most cases
explained, as at Stillwell Island, by a reaction between biotite, plagioclase and quartz ;
but in the hypersthene aggregate, it is supposed that the hypersthene, containing some
A1 2 S , changes in part into garnet.

A very noticeable feature in this type of rock is the presence of biotite both in the
form of platy crystals and in fan-shaped sprays. The fan-shaped sprays of biotite are
constantly associated with an intergrowth of felspars, and a genetic connection is assumed
between them. It is considered likely that these biotite fans are produced from garnet
by a reversal of the biotite-plagioclase-quartz reaction.

THE METAMORPHIC ROCKS OF ADELIE LAND. STILLWELL.

167

The specific gravity of these hypersthenic gneisses of varying content ranges between
2-74 and 2-76, and are comparable with the intermediate members of the Indian
charnockite series. They resemble the intermediate charnockites in the irregular distribu-
tion of the ferromagnesian silicates and in the prominence of the felspar intergrowths
and inclusions. The Antarctic specimens differ from the Indian rocks in the possession
of the well-defined garnet rims.

The primary igneous origin of these intermediate types is determined by analogy
with the normal charnockite-like rocks, though the occurrence at the Cape Pigeon Rocks
is probably that of an original dyke.

CHAPTER XL
THE CAPE GRAY METAMORPHOSED DYKE SERIES.

In each exposure on the Cape Gray Promontory basic dykes are found traversing
the garnet gneisses. Photographs were obtained from the localities visited by the
sledging party from the mainland, and these show the obvious dyke characters (Plate
XXVI., fig. 4 ; Plate XXVII., figs. 1, 2, 4).

Each locality will be dealt with separately, as each possesses different metamorphic

features.

Cape Gray.

At Cape Gray a perfect network of dykes is visible on the bare rock floor. A
diagrammatic sketch of this network is given in Fig. 13. The dykes branched and
junctioned frequently and small tongues could be seen running from the dyke channel
out into the gneiss (Plate XXVII., fig. 2). In places they enclose large fragments of
gneiss. So perfectly preserved is the network, we immediately assumed that the dyke
series would be much younger than the development of the gneiss. Examination,

Fig. 13.

DIAGRAMMATIC SKETCH OF THE RELATION OP THE GARNET CORDIERITE
GNEISS TO THE PLAGIOCLASE PYROXENE GNEISS AT THE WEST END

or CAPE GRAY.
The shaded area represents the garnet cordierite gneiss.

THE METAMORPHIC ROCK8 OF ADELIE LAND. 8T1LLWELL. 169

however, shows that foliation can be traced in part of the dykes. Most of the dyke
specimens look like dense fine-grained basalt, except that a vitreous appearance is
more noticeable on the fractured surface. Microscopic examination shows that they
are not normal igneous rocks, and we find a definite metamorphism of a varying kind.

No. 773. No. 773 is an example of the massive rock. The structure is finely
granoblastic and relic structures can be seen. The outlines of the felspar laths of the
primary dolerite are plainly visible (Plate VI., fig. 5), and are marked by lines of granular
augite producing a blastophitic structure. Diablastic structure is produced by an
intergrowth of augite and felspar. The mineral composition is as follows :

Felspar 40-5

Pyroxene 45-3

Hornblende 3-4

Ilmenite 6-6

Biotite 4-2

The original felspar laths are replaced by a granoblastic aggregate of clear secondary
felspar, which becomes evident in polarised light (Plate VI., fig. 6). The average absolute
grain size of the aggregates is 0-05mm. In some places relic felspar is found ; it is always
dusty with minute inclusions, and, therefore, appears in contrast to the clear secondary
felspar. The relic felspar is both simple twinned and lamellar twinned, and an extinction
angle of 29 has been measured from the trace of the lamellae. This felspar is, therefore,
labradorite. The relic felspar is usually surrounded by a zone of granular clear felspar,
which may contain vermicular grains of pyroxene. The refractive index of the
secondary felspar is less than that of the labradorite, but the difference is not great as it
is not noticeable under low power objectives. The maximum extinction angle obtained
in pieces showing twinning is 18, and hence we consider it to be an andesine. The
pyroxene forms 45-3 per cent, of the rock and includes hypersthene and augite and relic
augite. The relic augite is dusty through numerous minute inclusions of ilmenite.
These inclusions are more or less regularly arranged and may be called schiller inclusions.
Plates of this dusty augite have been found which have been ophitically inlaid with
laths of relic felspar, now represented by strings of secondary felspar. The recrystaliisa-
tion of the primary dusty augite has produced a granular aggregate of clear secondary
pyroxene, while the minute dusty inclusions have coalesced and now form a number
of small ilmenite crystals. The clear granular pyroxene sometimes forms a zone around
the dusty augite, but it may appear as a parasitic aggregate enclosed within the primary
dusty plate. Some of this granular pyroxene is certainly hypersthene with its pink
to green pleochroism and its straight extinction ; but it is impossible to determine its
proportions to the secondary augite.

The development of the secondary augite and hypersthene from the primary augite
means that the secondary augite will be more aluminous than the primary augite. A
high value for alumina in this secondary augite provides a point of resemblance to the

170 AUSTRALASIAN ANTARCTIC EXPEDITION.

omphacite in the Otz Valley* eclogites. The development of the secondary pyroxene
involves a change in the double refraction and many augite crystals show uneven polarisa-
tion colours. If the primary augite is showing blues and greenish blues of the second
order, the recrystallised augite may show the higher greens ; but, on the other hand,
the degree of colour is often lowered to the reddish purples and violet at the top of the
first order. In the latter, grains of hypersthene have been seen as a nucleus. No
corresponding change in the extinction angles of the augites has been noted.

A second type of alteration of augite that can be traced in this rock is the passage
into green pleochroic hornblende. The 3-4 per cent, hornblende in this rock has developed
in this way. The hornblende grains possess the same average size as the secondary
augite or felspar, but their distribution is not uniform. It is found in sporadic patches
which sometimes seem to indicate the outline of a prismatic crystal of pyroxene. The
4-2 per cent, of secondary brown biotite is distributed more uniformly throughout the
rock, and some of it may be reliu. Ilmenite abounds in small crystals and as minute
inclusions in the augite. It is definitely recognised as some of the larger crystals show
alteration to greyish leucoxene. Occasional grains of pyrite are also present.

The metamorphic character of this rock certainly dominates the igneous character,
and, therefore, we call the rock a plagioclase pyroxene gneiss.

No. 766. This example is a modified variety of No. 773, and shows distinct schis-
tosity. The secondary felspar and the secondary pyroxene are arranged approximately
in layers, producing a crystallisation schistosity. It is possible that the original rock
had a coarser grain size than the original rock of No. 773, because the plates of relic
augite are much larger in this example. The average grain size of the recrystallised
individuals is about twice as large. Relic dusty felspar is still present, and is surrounded
by a granulitic mass of clear felspar. In one case sharp lamellar twinning is present, and
the lamellae have extinction angles of 36 and 37, again indicating labradorite. The
recrystallisation of the primary pyroxene is more diagrammatic than in No. 773. Plates
of primary augite may form the nucleus of beautiful granoblastic zones of clear secondary
pyroxene whose growth in the direction of the schistosity may produce long tails (Plate
VII., fig. 1). Many of the grains can be identified as hypersthene ; in many cases
where the primary augite is completely replaced, the layer of granular pyroxene may
enclose areas, sometimes circular, of fine vermicoidal pyroxene set in an aggregate of
felspar producing a diablastic structure. At other times the vermicular pyroxene
forms a fringe around the outline of a primary pyroxene in the same manner as is more
prominently exhibited in No. 951 from Stillwell Island. This diablastic pyroxene has
character different from the granular pyroxene, and is not unlike the intermediate
stage m the formation of garnet which is seen in the Stillwell Island rocks. The
suggestion, therefore, is that these are incipient garnet areas.

' The Percentage of Alumina in the Omphacite in the Otz Valley Eclogite is 10-91 per cent. " Beitrag zur Kenntnis
der Eklogite und Amphibolite," Laura Hezner. Wein, 1903, p. 10.

THE METAMORPH1U ROCKS OF ADBLIE LAND. STILLWELL. 171

Apart from the schistose character the important difference from No. 773 is the
large percentage of hornblende. The grain size of the hornblende tends to be a little
larger than that of the secondary pyroxene. The green hornblende is found developing
directly from the relic platy augite and from the secondary pyroxene. Hornblende
grains are found indiscriminately associated with the granular pyroxene, and they
border plates of primary augite and arise parasitically within them. Further, horn-
blende with small crystals of ilmenite, completely replaces the pyroxene in some of the
crystalline layers of the rock. Biotite is much less important in this rock and shows a
very strong tendency to be confined to the pyroxene areas.

The percentage mineral composition of this section is

Felspar 33-6

Pyroxene 24-4

Hornblende 39-5

Ilmenite 2-3

Biotite 0-2

Apatite Present

This composition cannot be directly compared with No. 773, because it is a schistose
rock, and because the section has been cut at about 30 to the plane of schistosity instead
of 90. The difference in the ratio of the felspar to the ferromagnesian cannot be con-
sidered to demonstrate a change in chemical composition. The expression is, however,
useful to demonstrate the degree of hornblendisation. There is far too much pyroxene
for the rock to be considered an amphibolite. It is a transition type between the plagio-
clase pyroxene gneiss (No. 773) and an amphibolite, and should, therefore, be called a
hornblende plagioclase pyroxene gneiss.

Stittwett Island.

The basic rocks observed by Sir Douglas Mawson on Stillwell Island were considered
by him in the field to be altered dykes. " Irregular bands of black rock," he says in
his diary, " exist as at Cape Denison : some of these are not much altered, others are
full of fine garnet. The black bands usually extend long distances, and all have the
appearance of original dykes. A vertical section of one, exposed in a cliff face, showed
that it dipped regularly at 45 to the west." The island contains some of the most
interesting members of this dyke series, and remarkable stages of incipient alteration
are found.

No. 951. This example has a coarser grain than most members of the dyke series.
The gram size is sufficient to suggest a primary gabbro, because there is little alteration.
On the other hand the rock may be a completely recrystallised example, and the " little
alteration " may be the incipient development of a second metamorphic phase. The
latter interpretation is rather supported by the granulitic texture and the recrystallised
character of the surrounding rocks. In section the rock consists of felspar, augite.

ite, ilmenite, and apatite.

17 2 AUSTKALASIAN ANTAECTIC EXPEDITION.

The bulk of the felspar is in clear transparent crystals in which the twinning is some-
times indefinite and irregular. A maximum extinction angle of 28 has been measured,
and the felspar is, therefore, interpreted as labradorite. Both augite and hypersthene
are present and are free from the dusty inclusions of ilmenite. Brown biotite is present
both in large platy crystals and in small secondary crystals. The ilmenite is abundant.

The large pyroxene and biotite crystals, as well as the ilmenite, are almost invariably
bordered by a zone which follows the outline of the crystal, no matter how irregular
and ragged it may be (Plate VII., fig. 5). The zone may be described as a diablastic
intergrowth of vermicular pyroxene and felspar ; but the pyroxene is different from
the normal pyroxene, and the felspar is not the relic labradorite. The vermicoidal
pyroxene has a lighter colour than the normal pyroxene, suggesting that part of the
iron may have separated out to form ilmenite. The felspar is a more sodic felspar
and its development from the calcic felspar is very noticeable. Sometimes small
secondary biotites are seen in these diabiastic fringes, as well as small ilmenites. Ilmenite
crystals, large or small, are always associated with the pyroxenic parts of the slide
though there is no direct evidence to show here, as in No. 773, that they form during
the recrystallisation by the coalescence of minute inclusions in the primary pyroxene.
The same diablastic fringe is also found surrounding large ilmenite and biotite crystals ;
but it does not accompany these with the same regularity as it accompanies the pyroxene.
Ilmenite crystals may be found with a rim of pyroxene, and if this rim should pass into
the vermicoidal type we should get the ilmenite crystal surrounded by the diablastic
zone in the way we have often observed in this section. Some of the iron ore has the
appearance of pyrrhotite.

There can be no doubt that this diablastic zone is a product of a reaction between
labradorite and pyroxene, or between labradorite and biotite. Stages may be observed
between augites surrounded by a thin rim and small augites surrounded by a thick zone.
In the latter the remaining augite is mouse eaten and has nearly disappeared.

The formation of biotite in the vermicoidal zone must be associated first with a supply
of K 2 from the felspar, and secondly with the temperature factor during metamorphism.
The temperature factor must be high to permit the formation of secondary pyroxene,
and it may have been, in the first stage of metamorphism, too high for biotite. Biotite
may have been formed only after a lowering of the temperature, and the appearance
of the biotite is quite in agreement with the suggestion that the biotite is subsequent
to the initial formation of the rim. A study of the phenomena in No. 935 shows that this
reaction is the initial stage in the formation of garnet.

The rock may be described as a plagioclase pyroxene gneiss, which shows the
incipient stages of garnet formation.

No. 942. No. 942 is another example which still retains normal igneous structures.
It occurs in dyke-like bands, up to 10ft. wide, crossing the garnet gneiss. It is a much

THK MKTAMOKI'HIC ROCKS OF ADELIE LAND.-ST1LLWELL. 173

finer grained type than No. 951, and, in the hand specimen, might be taken for a
slightly altered dolerite, because there seem to be fine-grained portions, representing
the unaltered dolerite, surrounded by more coarsely crystalline rock, representing the
altered part. No part, however, is found to be unaltered in section.

In this section we find that the outlines of primary felspar and pyroxene of the
dolerite have quite disappeared. The former crystals are now replaced by a finely
diablastic aggregate of pyroxene and felspar. The individuals in the aggregates are
more granular in contrast to the vermicoidal appearance in the preceding. The
aggregates may contain small garnets and biotites with numerous small crystals of
ilmenite ; a little quartz has been detected and is probably associated with the formation
of garnet. We call these aggregates diablastic, because we consider them to be pro-
duced in the decrystallisation, or breaking down, of the primary pyroxene and labra-
dorite which results partly in the secondary pyroxene and a more sodic plagioclase.
This decrystallisation is followed by a recrystallisation, and we find here and there
granulitic aggregates of secondary pyroxene, including both orthorhombic and mono-
clinic forms, identical in kind with those produced in the Cape Gray rocks.

The recrystallisation or the building of large crystals from smaller ones seems to
have taken place under conditions in this case which have favored the formation of green
hornblende and biotite. Hornblende and biotite possess an average grain size much
greater than the pyroxene, and are both much more abundant than in No. 951. The
large crystals of hornblende and biotite are frequently aggregated in clusters, just as if
each cluster were a metamorphic differentiation centre of hornblende or biotite. Horn-
blende and biotite are frequently intergrown, indicating that they have formed at the
same time. Often the hornblende clusters have a linear trend, and sometimes they are
circular, enclosing areas of the diablastic felspar and pyroxene (Plate VII., fig. 6). In
doing this, they provide the initial stages of the growth of the phenomena to be described
in No. 953.

Hypersthene is again noted among the pyroxene, and apatite and odd grains of calcite
are present.

The metamorphic character of this example dominates the igneous, and most of
the rock has suffered complete decrystallisation. It may be called a hornblende
plagioclase pyroxene gneiss.

No. 952. In some instances the hornblendisation noted in the preceding has
proceeded to such an extent that a normal amphibolite has formed. No. 952 is an
example of this type, obtained from among the basic plagioclase augite rocks of Stillwell
Island.

In the hand specimen this rock is similar to the fine grained, massive varieties at
Cape Denison. In section, it consists chiefly of hornblende and felspar (labradorite-
andesine), with small amounts of ilmenite, biotite, and garnet. Sphene, calcite, and
apatite have been detected.

174 AUSTRALASIAN ANTARCTIC EXPEDITION.

A small portion of the hornblende is very pale in colour, and has the appearance
of uralite rather than that of the normal green hornblende. The uralite has green spots
of normal hornblende, and the cleavage passes indiscriminately through both. The
uralite is evidently passing into hornblende, or vice versa. Sometimes there are bluish
glaucophane borders on the hornblende crystals.

The biotite is intergrown with the hornblende as before. There are occasional
small blebs of garnet usually set in the felspar. Similar in outline and situation are
occasional small blebs of calcite, and these are probably the remains of former garnet
from which the A1 2 3 and the Si0 2 have been withdrawn, and the excess lime has been
converted by carbonation into calcite. Small dusty ilmenite areas have been found,
and suggest that some of the ilmenite has formed by the aggregation of this dust.

The rock is a true amphibolite, which has formed under meso zone conditions,
and its presence is noteworthy among a large number of recrystallised basic rocks in which
garnet and pyroxene predominate.

No. 935. This specimen was obtained from a broad band about 20ft. wide with ill-
defined boundaries. The rock is dark coloured and massive, with a vitreous lustre,
but without any suggestion of schistose texture in the hand specimen. The doleritic
character is suggested by the presence of felspar laths and large black augites, which can
be seen with the naked eye. With the aid of a pocket lens small garnets are found to
be numerous.

In thin section we find abundant garnet, pyroxene, hornblende, ilmenite, and felspar.
Pyrite and apatite are also found. The mineral proportions in slide No. 935 (2) have
been determined as follows :

Felspar 25-4

Pyroxene 21-3

Hornblende 25-3

Garnet 15-9

Ilmenite 5-7

Biotite 6-0

Apatite 0-4

The transformation of augite and its reactions with the felspar are very plain.
There are large plates of augite crowded with the minute dusty ilmenite inclusions
which we know are a relic of the pre-existing dolerite. We can trace the following
changes in this primary augite :

(1) There are parasitic clumps of small, interlocking, granular pyroxene crystals
which are clear and have been formed in the recrystallisation of the dusty pyroxene.
The primary schiller inclusions have been thrown out, and have coalesced to form large
ilmenite crystals. This is the same change as was observed in No. 773 from Cape Gray.
Partial aggregations of the minute ilmenite dust are often seen (Plate VII., fig. 2).

THE METAMORPHIC ROCKS OF ADELIE LAND. STILLWELL. 175

(2) There are seams and patches of granular green hornblende in cracks and fractures
of the relic pyroxene, and among the aggregates of secondary granular pyroxene. It
appears that the hornblende has formed directly from both the relic dusty augite and
from the secondary pyroxene. A large plate of relic augite may enclose parasitically
a granular aggregate of green hornblende crystals in which an ilmenite crystal may be
set as a nucleus, formed, as before, by the coalescence of the primary ilmenite inclusions.
Sometimes larger hornblende crystals have grown out of the aggregates of small granular
hornblende.

(3) There is frequently a considerable amount of small secondary biotite associated
with the granular hornblende so intimately that there can be no doubt they have arisen
at the same time as the hornblende. Its formation depends on the chemical supply
of potash and water.

(4) The large relic augites, which may be replaced by secondary pyroxene or
hornblende, are bordered by a zone of small garnets which may be partly idioblastic
(Plate VIII., figs. 4 and 5). Between the edge of the pyroxene crystal and the garnet
rim there is usually a thin zone of clear felspar (or quartz). The manner in which the
garnet rim follows the outline of the relic pyroxene can be easily seen when the section
is observed with a low power lens. The character of the felspar in the neighbourhood
of the garnet undergoes an obvious change when observed in polarised light (Plate
VIII., fig. 6). The formation of the garnet absorbs lime from the primary labradorite,
and we may find a labradorite crystal zoned with a rim of more sodic felspar. By analogy
with phenomena in metallic alloys, the manner of extraction of the anorthite from
the solid solution of plagioclase is strongly suggestive of solid diffusion. The reaction
that has taken place is one that has been quoted by Grubenmann,*and may be written
in this case

Augite Labradorite Garnet Andesine

CaMgSi 2 + 2CaAl 2 Si 2 8 , NaAlSi 8 8 ^ Ca 2 MgAl2Si,0 12 + CaAljSiA, NaAlSi 3 8

Quartz
+ Si0 2

There is no evidence to lead us to ascribe these compositions to augite, felspar, and
garnet, but, by doing so, we can more readily understand how the garnet is formed
and the more sodic plagioclase produced. More augite may combine with another
anorthite molecule of the andesine, and a still more sodic plagioclase produced. The
separation of the quartz has been definitely no^ed in a second section No. 935 (2) from
the same specimen, and that it does appear with garnet is abundantly evident in
No. '953.

Sometimes where a blastophitic structure can be recognised in No. 935 (2), and a
relic labradorite crystal crosses a plate of dusty augite, we may find no garnet border.

Along the edge of the felspar there are numerous small rounded inclusions like incipient

Die KruUllinen Schiefer, vol. I., p. 34."

176 AUSTKALASIAN ANTARCTIC EXPEDITION.

garnets, and the part of felspar near the contact with augite is more sodic than the
central portion. The relic augite is bordered by the secondary hornblende and this
may have prevented the interaction along this junction.

The proportions of the minerals garnet, biotite, and hornblende vary in different
slides with the varying amounts of recrystallisation.

If these varying alterations had not occurred the rock would consist chiefly of
augite, labradorite, and ilmenite, and, perhaps, some biotite. The primary ophitic
structure has been detected, and there can be no doubt at all that the rock is a meta-
morphosed dolerite. Yet it has been reported as a band with ill-defined boundaries.
Such ill-defined borders are probably to be explained by some such process as meta-
morphic diffusion. The formation of the garnet in this case is most certainly not due
to any absorption of any sedimentary material, as has been suggested by Cole for the
origin of certain garnet amphibolites in Ireland.*

The rock may be called a garnet plagioclase pyroxene gneiss.

No. 953. Several specimens of a garnet amphibolite have been collected from
Stillwell Island, where a dark basic band becomes definitely banded. Specimen No.
953 is a moderately coarse-grained rock with noticeable schistosity showing pink garnet,
black hornblende, and biotite.

In section, the rock is remarkable for its percentage of hornblende, garnet, and
quartz ; and a casual study would suggest that the rock possesses a composition different
from that of the more obvious dyke rocks. The chemical analysis shows that this is
not so. The percentage mineral composition has been determined as follows :

Felspar and quartz 24-1

Pyroxene 2-4

Hornblende 38-7

Garnet 19-7

Ilmenite 6-7

Biotite 7-9

Apatite and sphene 0-5

The green hornblende is the most abundant constituent, and at times seems to be
wrapped round a garnet crystal in the manner suggested in No. 942. The grains are
much embayed and sometimes poikiloblastic. The pink garnet is often crowded with
small inclusions. The grains are mostly rounded, but exert their form against the felspar
and quartz, and tend to do so against the hornblende. While quartz forms the bulk
of the colourless constituents, a garnet crystal is always set in a felspar base an
association which clearly has genetic meaning. A twinned felspar has been found to
give an extinction angle of 33 measured from the lamellae, and to possess a refractive

" On the Growth of Crystals in the Contact Zone of Granite and Amphibolite," G. A. J. Cole, Proe. Roy. Irish Acad.,
vol. 25, sect. B, 1905, p. 117.

THE METAMORPHIC ROCKS OF ADELIE LAND.-STILL\VELL. 177

index above quartz. This is labradorite, but I think the bulk of the felspar has a refrac-
tive index below quartz and a small extinction angle. There is a little pyroxene present,
but there are still good examples of the pyroxene-felspar vermicoidal intergrowths.
These intergrowths may extend as a bite into a garnet crystal, but it may be interpreted
either as a breaking down of the garnet or as a patch of unformed garnet (Plate VIII.,
fig. 1). It may form a zone around ilmenite crystals in the manner suggested in No. 935 ;
and as the ilmenite crystal may be embedded in garnet, the pyroxene-felspar interlacing
may form an annulus between the ilmenite and the garnet (Plate VIII., figs. 2 and 3).
The pyroxene " fingers " are often radial, both to an ilmenite nucleus and to a garnet
nucleus, and then a " centric structure " is formed. The hornblende has developed
from the pyroxene, and we sometimes find the normal pyroxene " fingers " of the inter-
growth converted into spokes of hornblende ; more rarely we find spokes of biotite.
Sometimes we find the intergrowth embedded in a crystal of hornblende.

A large individual of ilmenite is often a network rather than a compact mass, and this
is due to the imperfect coalescence of the small primary ilmenite crystals. Occasionally
the ilmenite network is set in a pyroxene base, and this is clear evidence that it is due
to the aggregation of minute inclusions in the primary augite, in the same way as was
observed in several sections. In all cases here this pyroxene-felspar intergrowth, which
we have included as a diablastic structure, may be explained on the hypothesis of the
interaction between pyroxene and anorthite to produce garnet and quartz. This reaction
is, doubtless, reversible. It has proceeded in the direction of the garnet in this example ;
but there is no reason why it should not proceed in the reverse direction in certain
examples in which garnet is said to be disappearing.* The abundant quartz accom-
panying the abundant garnet is clear evidence that Si0 2 is separated in the reaction.

The size of the garnet crystals is large in comparison with that of the garnet crystals
which form the garnet rims in No. 935. This can be readily explained as being due to
the growth of larger garnets at the expense of smaller crystals, a phenomenon which
has been exemplified by the hornblende in the Cape Denison series, and which will
subsequently be exemplified by the pyroxene in this series.

The rock is described as a garnet amphibolite. The felspar of the normal
amphibolite is here partly replaced by garnet.

Cape Pigeon Rocks.

Several dykes of basic rock exist on this locality. The obvious nature of the dykes
is recorded in photographs. Two large dykes (Plate XXVII., figs. 1 and 4) cut obliquely
across the foliation and are upwards of 30ft. wide. There are numerous smaller ones
as well (Plate XXVI., fig. 4), and some are only Sin. wide.

No. 767. Specimen No. 767 is an example of the large dyke. It is a dark, fine-
grained rock in which a faint schistosity may be detected. The schistosity is recognisable
in the slide, and there has been complete recrystallisation of the primary dolerite.

* " Untorauchungen die AltkrutaUiniwhen Sohiefergwtoine," Lehmann, Bonn, 1884, Tafel XXIV., fig. .
Series A, VoL ni.. Part 1 M

178 AUSTKALASIAN ANTAECTIC EXPEDITION.

The Rosiwal analyses of two slides gave the following results :

I. II.

Felspar : 45-3 .... 38-2

Pyroxene 25-8 .... 26-7

Hornblende 19-1 .... 24-1

Garnet 3-2 .... 2-9

Iron ore 4-3 4-9

Biotite 2-1 .... 2-9

Sphene, apatite 0-2 .... 0-3

The first of these is cut parallel to the schistosity, and its higher felspar percentage
is due to the fact that the schistosity is marked by strings of felspar in the hand specimen.
The second slide is cut in a haphazard direction, and the measurement is made to
determine the variation in the garnet percentage. This variation proves to be less than
anticipated.

The rock has a finely granoblastic structure. The felspar consists of water-clear
grains which sometimes show diablastic structure. The pyroxene, which occupies
one*quarter of the rock volume, includes plates of relic, dusty augite ; but it mostly
forms small granular crystals of augite and hypersthene, aggregated in areas which
originally represent large primary augite crystals. The clear recrystallised augite has
a pale green colour as before, and is practically free from the ilmenite inclusions. The
more pleochroic hypersthene is again present among the recrystallised pyroxene. The
percentage of green hornblende is not much less than that of the pyroxene, and indicates
the prominent degree of hornblendisation of the pyroxene. The garnet appears in small
pink crystals and is usually set in felspar areas ; this can be taken as evidence that it
has formed in the same way as in the basic rocks of Stillwell Island. The brown biotite
is usually associated with the pyroxene and hornblende areas.

An interesting feature in this rock is the presence of a shear line which cuts across
the schistosity. This line is marked chiefly by a decolouration of the hornblende and
by broken strings of pyrite. The hornblende may assume a pale green colour, and,
if the bright polarisation are absent, it may look like chlorite. Sometimes the shear line
may cut straight a crystal of green hornblende and then there appears a belt of colourless
hornblende in the green crystal, and this belt is even more noticeable in polarised light.
Sometimes there is a pale green mineral with high polarisation colours in the shear
zone, and as it has straight extinction it is looked upon as a pale biotite. In addition,
there is a very fine granular aggregate of highly polarising mineral, which is possibly
talc. The felspar becomes saussuritised and, in general, there is a fuzziness in the
neighbourhood of the line. Conditions along a shear plane would correspond in some
measure with the conditions of the epi zone of metamorphism ; and the pale hornblende,
the chlorite, the talc, and the saussurite are, in general, looked upon as epi zone
products.

THE METAMORPH1C ROCKS OF ADELIE LAND. 8TILLWELL. 179

The rock may be called a hornblende plagioclase pyroxene gneiss. It is similar
to the plagioclase pyroxene gneisses of Cape Gray, and in its garnet content it shows
affinities with the garnet plagioclase pyroxene gneiss (No. 935) and with the garnet
amphibolite (No. 953).

No. 782. Specimen No. 782 was collected from one of the narrower dykes at the
Cape Pigeon Rocks. It is a dark, fine-grained rock with abundant glistening hornblende.

In section, the rock is found to be quite different in general appearance from No.
767, a fact which is eloquently expressed by the following mineral composition :

Hornblende 49-0

Felspar 31-8

Pyroxene 7-6

Iron ore 7-5

Biotite 3-7

Apatite 0-4

The increased amount of hornblende and the decreased amount of pyroxene is the
most important difference ; and it is now noticed that the mineral composition approxi-
mates to that of the Cape Denison amphibolites. If all the pyroxene had disappeared
the proportion of hornblende to felspar would be the same as in some members of that
series.

The green hornblende is thus the most abundant mineral in this slide. The horn-
blende crystals, together with the more rare crystals of brown biotite, show a more
or less parallel arrangement, indicating the schistose nature of the rock. Very rarely
a colourless hornblende is intergrown with the green hornblende, similar to part of that
seen in the shear zone in No. 767. Both hypersthene and augite can be found among
the relic pyroxene distributed in patchy areas throughout the slide. It is often in
fragmentary form, and the fragments which are set in felspar can be determined by
polarised light to have been parts of a large crystal showing poikiloblastic structure.
Part of the relic pyroxene is altered to a greenish-brown micaceous product. The
felspar is again perfectly clear and ilmenite is abundant as usual. Pyrite is present.

The presence of the pyroxene makes the relation of this specimen to the hornblende
plagioclase pyroxene gneiss No. 767 obvious, and the primary types must have been very
similar. The differences are due to varying conditions during metamorphism. The
pyroxene felspar areas also suggest a likeness to the type No. 942 from Stillwell Island,
in which hornblende is not so abundant but the pyroxene areas more prominent. The
rock may be called an augite amphibolite.

No. 771. A closely related type to No. 782 is No. 771. This specimen has a much
finer grain and is less schistose.

It consists of a fine granoblastic mass of hornblende and felspar, with insignificant
amounts of biotite and ilmenite, but the latter may be surrounded by sphene. There

180 AUSTRALASIAN ANTARCTIC EXPEDITION.

are occasional large crystals of saussuritised felspar and neither pyroxene nor garnet
is present. The rock is a typical amphibolite.

A shear zone, developed subsequently to the formation of the hornblende, can be
detected in this rock as in No. 767. Without the microscope the shear plane looks like
a thin vein running through the slide. Under the microscope it is again marked by a
line of decolourised hornblende, saussuritised felspar, and some fine, highly polarising
aggregates. The broken strings of pyrite do not appear in this case, but specks of
this mineral are found in this zone.

These specimens of amphibolite, Nos. 782 and 771, were collected from the narrow
dykes on the Cape Pigeon Rocks, whereas the very broad dyke produces a hornblende
plagioclase pyroxene gneiss. We have insufficient data to determine whether this is
generally the case. It may be so, and it is quite possible that thin dyke sheets may tend
to become shear planes during the compression of a composite rock body, in which case
the thin dykes may be subjected to metamorphic conditions of the meso or epi zone
rather than those of the kata zone.

No. 786. Specimen No. 786 was collected as an amphibolite associated with the
gneiss. It did not appear in the field as a definite dyke-like band. It is much coarser
grained than the other amphibolites, and felspar and hornblende are plainly visible
in the hand specimen.

Under the microscope, however, it is found to be similar in kind to the altered
dyke rocks. The same type of green hornblende is again the most abundant mineral
and its development from the pale green pyroxene is apparent. The hornblende some-
times contains inclusions of sphene. Both quartz and felspar make up the colourless
components of the rock. There is a considerable amount of quartz which does not
show cataclasis or undulose extinction. Part of the felspar is aaussuritised and part
is quite clear. Labradorite has been recognised, but as some pieces of felspar have a
lower refractive index than basal quartz, there is some andesine or oligoclase as well.
Fragments of garnet are occasionally set in the felspar areas. Ilmenite, sphene, and
apatite are accessory minerals.

The rock may be described as an augite amphibolite.

We are inclined to think that this rock is related to the dyke bands at the Cape
Pigeon Rocks, in the same manner that the coarse-grained amphibolites (No. 9) at Cape
Denison are related to the corresponding amphibolite dykes. This example differs
from the coarse amphibolites of Cape Denison in the possession of augite and garnet ;
but in a like manner the altered dykes at the Cape Pigeon Rocks differ from the Cape
Denison series in the possession of augite and garnet.

The history of the coarse-grained patches at Cape Denison is considered to be
probably associated with great stress which has rendered former dyke channels dis-
continuous. It is interesting to note that the area near No. 786 at the Cape Pigeon
Rocks has suffered intense crumpling (Plate XXV., fig. 2).

THE METAMORPHIC ROCKS OF ADBL1E LAND. 8TILLWELL. 181

Garnet Point.

Among the specimens of altered dyke rock from this locality two varieties have been
collected. The extraordinary features along the junction with the cyanite biotite
gneiss of one type of amphibolite (No. 769 or 781) have already been mentioned. In
this case a mineral which has been referred to as cyanite appears in the cyanite biotite
gneiss, and can be traced across the junction to a distance of at least lin. away from it.
At this distance it is less abundant than in the cyanite biotite gneiss. It is found in all
the specimens of amphibolites from this locality.

In this amphibolite hornblende is the most important constituent ; but of nearly
equal importance are the circular areas of diablastic felspar and pyroxene (Plate VII.,
fig. 4). These areas are similar in outline to some of the felspar-pyroxene areas in No.
942, or to the felspar-garnet areas in No. 953 from Stillwell Island. This rock, like
No. 953, also possesses a noticeable amount of quartz ; but nowhere do we find the
pyroxene fragments set in quartz. Occasionally the fibres of pyroxene are set radially
in the felspar. The pyroxene in the aggregates may be altered to hornblende or to a
cloudy fibrous mineral. Very often, when it can be determined, the pyroxene has straight
extinction. The aggregates may be dotted with ilmenites and small biotites, while
in polarised light they nearly always show a little scapolite, arising out of a fibrous mass.
Sometimes a mineral with high refractive index and low bluish polarisation colours
can be detected and suggests a zoisite. The felspar in the aggregates may be clear
and possess very fine, irregular, twin lamellae. The low refractive index, high polarisa-
tion colours, and straight extinction of the scapolite can always be observed, but its
determination is rendered more certain by the observation of uniaxial character and
negative sign in the second amphibolite from this area.

Gajnet is present, but in most slides protracted search is required to find the small
pieces of garnet that may be set in the felspar. A portion of one slide, however, contains
considerable garnet. This garnet is very ragged in outline and contains inclusions of
felspar, biotite, ilmenite, and a colourless, brightly polarising mineral, probably scapolite.
The decomposition of garnet into pyroxene cannot be observed in these rocks in the
manner recorded in the Saxon area.

The relation of the pyroxene-felspar areas to the garnet is difficult to determine
in this section. We know they are connected by our study of other sections, and in
this instance the aggregate is occasionally replaced by garnet. Scapolite is observed to
be included in the garnet and in the pyroxene areas, but it does not seem possible to
say whether the pyroxene felspar has been developed from the garnet or vice versa.

The rock may be described as an amphibolite which is related both to the garnet
amphibolites and to the hornblende plagioclase pyroxene gneisses.

No. 799. The second type of amphibolite from Garnet Point is distinguished
from the preceding by a complete absence of the diablastic areas of pyroxene and felspar
and associated minerals.

182 AUSTBALASIAN ANTARCTIC EXPEDITION.

The rock is a little more coarsely crystalline than most of the examples from Cape
Gray and the Cape Pigeon Rocks and the granoblastic structure is again prominent.
The mineral proportions may be indicated by the following :

Hornblende 57-7

Felspar and quartz 35-3

Pyroxene 2-3

Garnet 1-2

Iron ore 1*2

Biotite 0-3

Apatite 0-3

Sphene 0-1

Residue, including scapolite and talc 1-6

The green hornblende is again the most abundant constituent and it is occasionally
fringed with a little blue glaucophane. At other times an irregular brown tinge is
noticeable in some crystals. The felspar is just as clear as the quartz from which it is
difficult to distinguish in ordinary light, because their refractive indices are nearly
the same. The felspar is an andesine, and the quartz may be set as rounded blebs
in the hornblende as well as in the felspar, producing a poikiloblastic structure. There
is, however, much more felspar than quartz.

The 2-3 per cent, of pyroxene is localised in one part of the slide, where it is nearly
as abundant as the hornblende. The clear portion has a very pale green colour, but
some of it is turbid and dense. Hypersthene is present because a large number of grains
show straight extinction. Monoclinic pyroxene is also present because an extinction
angle of 37 has been measured. The garnet is again invariably set in felspar (Plate
VII., fig. 3). Some crystals are very small and fragmentary but yet perfectly clear
and unaltered. Bigger individuals, granular in outline, appear in the larger areas of
felspar.

There are areas in this slide which seem to be analogous to the shear planes that
are recorded in Nos. 767 and 771 at the Cape Pigeon Rocks. The bulk of these areas
are included in the 1-6 per cent, residue in the percentage mineral composition. In
this case these areas have no linear trend except that the pyrite in part seems to occupy
a definite plane ; but the analogy is found in their mineral content. The areas are
noted for an abundance of fuzzy material which has high polarisation colours and which
may be finely granular talc. Equally prominent with this talc is a colourless hornblende
which may be bordered with blue glaucophane. The colourless hornblende is often
feebly pleochroic and sometimes contains patches of normal green hornblende. Some-
times the plates and fibres are bent or broken, and if an extinction angle of 30 can be
measured it is interpreted as a colourless pyroxene. There are more prominent areas
of scapolite associated with these shear areas along the edge of the slide.

THE MET AMORPHIC ROCKS OF ADEL1E LAND .-STILLS-ELL.

183

Though the garnet percentage is small it is distinctive and the rock may be called
a garnet amphibolite. This name indicates its relation to the garnet amphibolite of
Stillwell Island (No. 953), though the garnet percentage of the latter is many times
greater.

CHEMICAL CHARACTERS OP THE CAPE GRAY DYKE SERIES.

The following chemical analyses of four members of this series of rocks have been
made by Messrs. P. G. W. Bayly and J. C. Watson, in the Victorian Geological Survey
Laboratory :

II.

III.

IV.

Si0 2

47-74

49-91

49-99

48-06

A1.0,

15-10

13-02

13-84

14-19

Fe,O.

2-47

2-84

1-97

1-95

FeO

12-43

13-70

13-18

15-66

MjjO .

6-85

4-74

6-01

5-29

CaO . .

9-41

9-28

9-72

9-24

Na 2

2-09

2-03

1-94

0-71

K.O .

O61

0-83

0-79

1-29

"2"

H,0 +..

0-73

0-87

1-40

1-21

H,0 -.

0-19

0-12

0-05

0-13

"2"

co z .

n.d.

tr.

n.d.

Ti0 2 . . .

1-83

2-39

1-78

2-54

P.O.

O30

0-20

tr.

0-28

SO,

nil

n.d.

tr.

str. tr

n.d.

MnO

0-24

0-12

0-07

n.d.

NiO, CoO . .

0-02

0-01

n.d.

CoO

n.d.

LiO,

tr.

str. tr.

tr.

n.d.

Total.

100-01

100-06

100-75

100-55

Sp. Gr.

3-0988

3-1283

3-0974

3-2457

Group Values.

Projection Values.

1 52-9

2-5

7-0 28-1

3-8

0-9

1-3

3-7

15-0

II 56-1

2-7

5-5 27-5

5-2

1-0

1-5

3-1

15-4

III 54-8

2-6

6-1

27-7

5-0

1-0

1-4

3-4

15-2

IV 54-4

1-6

7-4

27-6

3-3

1-0

0-9

4-0

15-1

I. No .773 Plagioclase Pyroxene Gneiss. Cape Gray, Adelie Land

II. No. 767. Hornblende Plagioclase Pyroxene Gneiss. Cape Pigeon Rocks, Adelie Land

III. No. 799. Garnet Amphibolite. Garnet Point, Adelie Land.

IV. No. 953. Garnet Amphibolite. Stillwell Island, Adelie Land.

184 AUSTKALASIAN ANTARCTIC EXPEDITION.

There is a very strong family likeness in the chemical composition of the dyke
rocks from these four localities. Each analysis has the general characters of a basic
igneous rock and closely resembles that of the Cape Denison amphibolite (No. 629).
The minor differences can readily be explained as primary variations in the compositions
of the dykes at the separate localities.

It may be recalled, however, that the mineral compositions of these four rocks
show great variation, and range from 45-3 per cent, pyroxene in the Cape Gray rock
to 2-3 per cent, pyroxene in the Garnet Point rock ; from 57-7 per cent, hornblende
in the Garnet Point rock to 3-4 per cent, hornblende in the Cape Gray rock ; from 19-7
per cent, garnet in the Stillwell Island rock to 1-2 per cent, garnet in the Garnet Point
rock, and from 6-7 per cent, ilmenite in the Stillwell Island rock to 1-2 per cent, in the
Garnet Point rock. These varying mineral combinations are independent of the
chemical composition and are interpreted as due to varying metamorphic conditions.
The similarity of the four analyses provides an argument for the general constancy of
chemical composition during metamorphism.

The specific gravities of these rocks are all higher than the specific gravity (3-030)
of the Cape Denison amphibolite, No. 629, which is considered to be a product of more
superficial conditions. These higher specific gravities agree with the general deep
seated metamorphism of Cape Gray Promontory. The garnet amphibolite from Stillwell
Island has a value distinctly greater than the others, and this high value can be ascribed
to the same cause as the production of garnet.

The general family likeness is reflected in the table of Osann group values and
projection values. These group values place each rock in the group of eclogites and
amphibolites. When the projection values are plotted they produce a cluster of dots
around the mean projection value of this group (fig. 14).

The production of secondary pyroxene requires a high temperature factor, and the
production of garnet requires a high uniform pressure factor during the recrystallisation.
Rocks which contain these two minerals can confidently be classed as kata zone products.
The plagioclase pyroxene gneiss of Cape Gray is a kata zone rock, though only the
incipient forms of garnet are found. Hornblendisation of the pyroxene is looked upon
by Grubenmann as a meso zone characteristic, and, therefore, the hornblende plagioclase
pyroxene gneiss from the Cape Pigeon rocks represents a transition stage between the
kata zone type and the meso zone type. Other dyke rocks described from the Cape
Pigeon rocks are distinctly meso zone types.

The garnet amphibolites from Garnet Point and Stillwell Island are members of
the garnet amphibolite family which Grubenmann places in the Meso division. The
development of both garnet and hornblende from the pyroxene and felspar has been
described from the same rock, but it cannot be considered to be proved that the garnet-
forming conditions are the same as the hornblende-forming conditions. An increase of

THE METAMORPHIC ROCKS OF ADELIE LAND STILLWELL.

185

pressure without alteration in temperature may produce garnet, while a decrease of
temperature without alteration in pressure may produce hornblende. The hornblende-
forming conditions may follow the garnet-forming conditions. We think this is
indicated by rocks like No. 942, which is similar to the garnet amphibolite in structure,
but the garnet is replaced by pyroxene felspar areas. If, then, we place the garnet
amphibolites among the meso zone rocks, it must be borne in mind that the same meso
conditions do not produce both garnet and hornblende. The garnet amphibolites
are not kata zone rocks and their family characteristics are too definite to allow them
to be considered as transition types between the kata types and the meso types.

a c

Fig. 14.

IV. Mean Position of Group IV., the Eclogites and Amphibolites
773. Plagioclase Pyroxene Gneiss, Cape Gray.
799. Garnet Amphibolite, Garnet Point.

767. Hornblende Plagioclase Pyroxene Gneiss, Cape Pigeon Rocks.
794. Plagioclase Pyroxene Gneiss, Madigan Nunatak.
953. Garnet Amphibolite, Stillwell Island.

The CrystaUoblastic Order.

The crystalloblastic order for the plagioclase pyroxene gneisses appears to be
Pyroxene, hornblende ; biotite ; ilmenite ; felspar. If garnet appears, as in No. 935,
garnet is placed above the hornblende.

In the garnet amphibolite the order is Garnet, hornblende, biotite, ilmenite,
felspar, quartz.

186 AUSTKALASIAN ANTAKCTIC EXPEDITION.

SUMMARY.

In all exposures on the Cape Gray Promontory basic gneisses are found associated
with the garnet gneisses. With one exception, these basic gneisses are found in dyke
form which is so definite that the field examination convinced the observers of the
igneous origin. Microscopic and chemical examination have confirmed this observation
and interesting mineralogical changes have been traced.

At Cape Gray the outlines of primary felspar laths and augite crystals can be
determined and a blastophitic structure is found. The primary felspars are now repre-
sented by aggregates of interlocking, clear, secondary felspar. The primary augite,
recognisable in all cases by the presence of minute ilmenite inclusions, becomes trans-
formed into clear, granular, secondary augite and hypersthene, with associated ilmenite.
A varying amount of hornblendisation of the pyroxene occurs. The basic rocks at
Cape Gray bear evidence of kata zone metamorphism like the surrounding cordierite
garnet gneiss.

At Stillwell Island massive types occur and further changes are traced. A coarsely
crystalline rock (No. 951), which is probably a completely recrystallised dolerite, consists
of granular crystals of clear pyroxene and clear felspar. It shows the incipient changes
of modification in a rim of diablastic intergrowth of pyroxene and felspar which surrounds
crystals of pyroxene, biotite, and ilmenite. This rim is looked upon as the incipient
stage of reaction between pyroxene or biotite and felspar ; which produces garnet and
quartz or garnet and orthoclase.

This reaction is advanced in another example (No. 935), and a well developed rim
of garnet can be traced around pyroxene areas. The aggregation of the small garnets
which form the garnet rim may produce the larger garnet crystals of the garnet
amphibolite. This origin explains why the garnet crystals are always set in a felspar
base a constant association which must have genetic meaning. If hornblendisation
of the remaining pyroxene occurs, we get the garnet amphibolite, of which No. 953 is
an example. In some cases the hornblendisation of the pyroxene occurs and a normal
amphibolite, No. 952, is found.

At the Cape Pigeon Eocks the large dyke is found to be a hornblende plagioclase
pyroxene gneiss. Garnet is present and illustrates the relation with some of the Stillwell
Island gneisses. Hornblendisation is prominent but not sufficient to mask the relation
of the gneiss to the plagioclase pyroxene gneisses. The narrower dykes on this area
show a much greater degree of hornblendisation than the large dyke. The percentage
of pyroxene decreases from 26-8 per cent, to 7-6 per cent, in No. 782, and is zero
in others. These last are amphibolites. A coarser amphibolite comes from this area
which did not maintain the dyke form in the field. It is a rock which is clearly related
to the dyke rocks, and the relation is considered to be the same as that between the
coarse-grained amphibolite patches at Cape Denison (No. 9 type) and the well-defined
dyke bands.

THE METAMORPHIC ROCKS OF ADELIE LAND. STILLWELL. 187

The presence of definite shear planes has been noted in two examples from this
locality. The shear plane may look like a thin vein in the hand specimen, and under
the microscope is marked by decrystallisation. The hornblende and biotite may become
very pale and even decolourised. The presence of finely granular talc is indicated,
and pyrite is distributed linearly along the shear plane.

At Garnet Island two types of amphibolite have been described. One type is noted
for the abundant circular areas of pyroxene diablastically set in felspar. The pyroxene
is sometimes altered to hornblende and sometimes to a fibrous product. Scapolite
is frequently discovered in these areas. Rarely these areas are replaced by ragged
garnets ; but no definite evidence can be gathered to show that the garnets break up
into pyroxene and felspar.

The second type of amphibolite at Garnet Point carries a small garnet percentage
evenly distributed through the rock. Again the garnet is always set in a felspar base,
and a relation to the garnet amphibolite of Still well Island is indicated. Features,
similar to those in the shear zones in the Cape Pigeon Rocks, are also found, but no
definite linear direction is obvious in them. Decolourised hornblende and talc are
prominent in these areas, and the colourless hornblende may be fringed with blue
glaucophane. Sometimes the crystals of colourless hornblende are bent and broken
and linear pyrite may be found. Scapolite is here associated with these areas.

The various types may be summarised thus

No.

Cape Gray 773 Plagioclase Pyroxene Gneiss.

766 Hornblende Plagioclase Pyroxene Gneiss.

Stillwell Island 951 Plagioclase Pyroxene Gneiss.

942 Hornblende Plagioclase Pyroxene Gneiss.

935 Garnet Plagioclase Pyroxene Gneiss.

953 Garnet Amphibolite.

952 Amphibolite.

Cape Pigeon Rocks .... 767 Hornblende Plagioclase Pyroxene Gneiss.

782 Augite Amphibolite.

771 Amphibolite.

786 Augite Amphibolite (without dyke form).

Garnet Point 781 Amphibolite.

799 Garnet Amphibolite.

Four chemical analyses of this rock series are given. These show a strong family
likeness and possess the general characters of basic igneous rocks. The Osann group
values place them quantitatively in the group of eclogites and amphibolites. The
plagioclase pyroxene gneisses are placed in the kata division of this group and the
amphibolites belong to the meso division. The hornblende plagioclase pyroxene gneisses
are transition members between the two divisions.

188 AUSTKALASIAN ANTARCTIC EXPEDITION.

The general similarity in field characters and in composition at the four localities
permit the assumption that the altered dykes in each area are part of one intrusive
series. Differences in all cases can be ascribed to varying metamorphic conditions.

There is no direct evidence to correlate this intrusive dyke series with the Cape
Denison metamorphosed dyke series ; but all differences can again be explained by
varying metamorphic conditions. The dominating factor among the metamorphic
conditions at Cape Denison is strong stress, whereas the general metamorphic conditions
in the Cape Gray series involve high uniform pressure and high temperature with only
subordinate stress. The strong stress at Cape Denison has destroyed all those finer
features of dyke form which have been preserved at Cape Gray. The differing mineral
suites are considered to be a direct reflection of the different conditions during
recry stallisation .

CORRELATION.

The basic pyroxenic gneisses are found in many areas of the crystalline schists.
In the classical area of the Saxon pyroxene granulites there are examples to which
members of the Cape Gray series are analogous. The fine grained plagioclase pyroxene
gneiss (No. 773) is similar to the pyroxene granulite from America near Penig.* This
Saxon type, however, is described as schistose, whereas the Cape Gray rock is massive
and relic dolerite structures are recognisable. There is an analogy between the garnet
amphibolite from Stillwell Island (No. 953) and the pyroxene granulite from Bahnstation,
Wittgensdorf ; f but the hornblende in the former is replaced by pyroxene in the latter.
The manner in which the garnets are set in a colourless base in the Antarctic rocks is
a phenomenon that also appears in the Saxon rocks illustrated on Table XXIII., figs.
3, 5, and 6, of Lehmann's memoir. There is also a likeness between the pyroxene
granulite from Chemnitzbiege, by Mohsdorf,J and the amphibolite No. 769 from Garnet
Point. In No. 769 there is little garnet, but the structures are similar in both. In
this Saxon example Lehmann considers that the separation of the pyroxene occurs
at the expense of the garnet, whereas Holland has found the reverse to be true in some
of the Indian rocks. Other observers have formed similar conclusions to both Lehmann
and Holland. In our observations we have been led to suspect evidence for Lehmann's
position and we obtained definite proof in favor of Holland's position. The explanation
probably lies in the fact that the reaction which involves both pyroxene and garnet
is a reversible one. The direction in which the reaction goes is determined by the external
conditions.

The same class of rock has been described among the pyroxenic and hornblendic
gneisses by Lacroix, in India and other places. || The pegmatoidal pyroxene, set in
oligoclase and quartz, that is figured by Lacroix (p. 179) is similar to some of the structures
described as diablastic, e.g., No. 942.

* " Entstehung der Altkrystallinischen Schiefergesteine," J. Lehmann, Bonn, 1884, Tafel XXIII., fig. 2.

t Tafel XXIII., fig. 3.

t Tafel XXIV., fig. 5.

" Origin and Growth of Garnets," T. H. Holland, Rec. G.S.I., vol. XXIX., p. 20.

|| " Gneissose Rocks of Salem and Ceylon," Lacroix, trans, by Mallet, Rec. G.S.I., XXIV., p. 155.

THE METAMORPHIC ROCKS OF ADELIE LAND STILLWELL. 189

Similar rocks have been described by Holland as norites among the charnockite
series of India.* The augite norite and the hornblende augite norite present analogies
to the plagioclase pyroxene gneisses. Yet they are, perhaps, more comparable with the
basic gneisses at Madigan Nunatak and Aurora Peak ; but the latter rocks have a direct
relation to the Cape Gray dykes.

In his description of plagioclase pyroxene rocks from Parasnath and the I jri Valley,
from the Madras Presidency and Bengal, Holland f describes the original augite as
darkened, almost blackened, by minute rods and plates forming an ordinary example
of schillerisation. The hornblende which is derived from the augite is free of such
inclusions. This augite reads precisely similar to the primary dusty augite that has
been described from Cape Gray and Stillwell Island. The development of secondary
pyroxene has not occurred in the Indian rock as in the Antarctic.

In some of the pyroxene granulites or basic charnockites from the neighbourhood
of Salem, Holland J describes a corona of garnet around the hypersthene. This seems
to be similar to the corona around the pyroxene in the garnet plagioclase pyroxene
gneiss (No. 935) from Stillwell Island. In this publication a sketch is given of hyper-
sthene with a corona of spongy garnet. This spongy garnet appears to be similar to
what we have referred to as diablastic pyroxene, or vermicular pyroxene, which is very
well developed in some of the Stillwell Island rocks. This material is sometimes isotropic,
sometimes with very low polarisation colours, but sometimes it shows the brighter
polarisation colours of pyroxene. Possibly it is not constant in composition and
represents some intermediate form between garnet and pyroxene. The separation of
quartz in this garnet-pyroxene reaction is noticed in the Indian rocks as well as in the
Stillwell Island rocks.

Similar pyroxene gneisses have been described in many parts of the world, in Canada,
in Scotland, in Madagascar, from the moraines in South Victoria Land, and in many
other places.

" Charnockite Series," T. H. Holland, Mem. G.S.I., XXVIII., pt. 2, p. 166.

t " Origin and Growth of Garnets," T. H. Holland, Reo. G.S.I., vol. XXIX., p. 20.

J " Geology of the Neighborhood of Salem," T. H. Holland, Mem. G.S.I. 30, p. 12.

CHAPTER XII.

1. RELATION BETWEEN THE ROCKS AT CAPE GRAY, MADIGAN
NUNATAK, AND AURORA PEAK.

We have shown in our descriptions that the two rock types found at the Madigan
Nunatak correspond closely with the two chief types at Aurora Peak. If we subtract
the epi zone metamorphism from the Madigan Nunatak rocks, and the meso zone
metamorphism from the Aurora Peak rocks we get, in both cases, kata zone meta-
morphic types. The basic rocks of the two localities then become identical and the acid
rocks are analogous ; but they all agree in possessing a granulitic structure and the
mineral hypersthene.

The basic rock at Aurora Peak is reported as a dyke cutting across the foliation,
but, though the basic rock at Madigan Nunatak appeared to form a band, nothing
definite could be observed in the field. Still, from general considerations, it has been
considered to be probably a metamorphosed basic igneous rock. The likeness to the
Aurora Peak rock renders this more probable, but it receives striking confirmation
by comparison with the undoubted dyke at Cape Gray.

We place here, side by side, the mineral proportions, determined by the Rosiwal
method, of No. 773 (the plagioclase pyroxene gneiss from Cape Gray), of No. 794 (the
plagioclase pyroxene gneiss from the Madigan Nunatak), and of No. 759 (the hornblende
plagioclase pyroxene gneiss from Aurora Peak). We add, for the sake of comparison, the
mineral proportions of a hornblende norite from St. Thomas Mount, Madras, determined
by Washington.*

No. 773.

No. 794.

No. 759.

Hornblende
Norite, Madras.

Felspar

40-5

42-5

44-8

40-8

Pyroxene

45-3

45-5

28-6

31-0

Hornblende

3-4

3-3

15-5

19-6

Iron Ores

6-6

8-4

10-1

8-6

Biotite

4-2

0-3

Apatite

0-7

Approx. average absolute grain size . . .

O05mm.

0-30mm.

O17mm.

The likeness of the proportions of felspar and ferromagnesian minerals in all four
cases is obvious. The pyroxene of No. 773 is partly secondary and partly primary ;
but the secondary pyroxene is very similar in type to the pyroxene of Nos. 794 and 795.

* " The Charnockite Series of Igneous Rocks," H. S. Washington, Amer. Journ. Sei., Vol. XLI., 4th Ser., 1916, p. 323.

THE METAMORPHIC ROCKS OF ADELIE LAND. STILLWELL. 191

Orthorhombic and monoclinic forms are present in each case. The primary augite in
No. 773 contains abundant minute inclusions of ilraenite, and these were neglected in
the count of iron ore. There is, therefore, reason for the smallest iron ore percentage
in No. 773.

Thus the proportions of felspar and pyroxene in the plagioclase pyroxene gneiss
at Cape Gray with average absolute grain size approximately 0-05 mm., whose primary
dyke origin is beyond all possible doubt, are practically identical with the proportions
in the epi plagioclase pyroxene gneiss (pyroxene granulite) at Madigan Nunatak with
average absolute grain size approximately 0-30mm. The minerals are in each case
the same, and there is no important difference in the chemical composition, and they are
not far removed from one another in the Ozann triangular projection. The chief
difference between these two rocks is the grain size. The fine grained type, No. 773,
possesses the relic structure and is most like the primary dolerite. Hence we can
conclude that the metamorphic conditions at Madigan Nunatak were longer continued
and caused certain crystals to enlarge themselves at the expense of other crystals, thus
producing fewer and larger crystals. The fine-grained facies has been replaced by a
coarse-grained facies.

We have the direct evidence that the felspar and pyroxene of a coarse-grained
dolerite at Cape Gray have been replaced by a fine granoblastic aggregate. Secondary
enlargement of these fine grains can proceed till we get an aggregate many times coarser.
If the pyroxene in the aggregate be then partially converted into hornblende we get
the meso plagioclase pyroxene gneiss at Aurora Peak which is identical with the rock
called hornblende norite from St. Thomas Mount, Madras. If the hornblendisation
of the pyroxene be completed, an amphibolite, comparable to those at Cape Denison
would result. If the conditions of hornblendisation are replaced by those of the epi zone
of metamorphism a certain amount of granulation appears, and we find some crystals
are fractured, some are crushed, and some have granulated selvages as in the Madigan
Nunatak example. The direct transition stages between dolerite and amphibolite
have not been directly traceable in these areas as it has been done in other parts of the
world.

The two-phase metamorphism of these basic rocks at the nunataks is considered
to rest on direct and sure evidence. Therefore it is not reasonable to doubt the inter-
pretation of a two-phase metamorphism of the acid hypersthene gneisses which are
associated with and intruded by the basic rocks. Subtracting respectively the effects
of epi and meso zone metamorphism, we find a family likeness which is exhibited by
the granoblastic structure and the presence of hypersthene. We attribute this family
likeness to kata zone metamorphism.

As we have shown that the hypersthene is a metamorphic product in the plagioclase
pyroxene gneiss, it is little assumption to assert that the hypersthene is also a meta-
morphic mineral in the acid hypersthene felspar gneisses which have suffered similar

192 AUSTRALASIAN ANTARCTIC EXPEDITION.

conditions of recrystallisation. Indeed, the definite parallel arrangement of the hyper-
sthene crystals would be difficult to explain on any other hypothesis. In many
metamorphosed granites and similar rocks the ferromagnesian percentage is expressed
in a content of biotite or chlorite. If the temperature of the metamorphism should
exceed that at which biotite is capable of holding its water of combination, what will
be the product? Pyroxenes are high temperature minerals and it would not be
unreasonable, on a priori grounds, to expect biotite to be replaced by pyroxene under
such circumstances. If there were sufficient lime in the rock we might equally well
expect garnet, provided the pressure factor is suitable. Biotite has been proved to be
an alteration product of hypersthene in the zone of the gneissic dacites at Belgrave,
Victoria.* Hence the metamorphism of a granite under conditions of high temperature,
high uniform pressure, and weak stress, might produce a hypersthene alkali felspar
gneiss. Apart from such considerations, however, or any correlative evidence, the
study of the rock relations has made it clear that hypersthene is a metamorphic mineral
in the basic rocks and, as these and the acid hypersthenic gneisses have suffered complete
recrystallisation under similar conditions, it is unreasonable to deny the metamorphic
character of the hypersthene in the acid gneisses.

To this complete recrystallisation under similar conditions we must assign the
family likeness of the Madigan Nunatak and the Aurora Peak rocks, in spite of chemical
differences. On the other hand we must ascribe the differences between the gneisses
at Aurora Peak and at Cape Denison to dissimilar conditions of recrystallisation, in spite
of marked chemical likeness. The Cape Denison granodiorite gneiss is almost a product
of epi zone conditions.

We think that it is impossible to deny the metamorphic character of these acid
hypersthenic rocks, and yet we find that they correspond very closely with the descrip-
tion of the hypersthenic rocks of Peninsular India, which have been called the Charnockite
Series. We have already demonstrated the similarity in chemical composition of the
acid and basic types at the Madigan Nunatak to members of the Charnockite Series ;
and we have just found that the mineral composition of the hornblende plagioclase
pyroxene gneiss at Aurora Peak is very close to that of a hornblende norite, a basic
charnockite from the type locality, St. Thomas Mount, Madras. But the charnockites
have been considered by Holland f to be igneous rocks which have consolidated under
phenomenal conditions. Can, then, these Antarctic metamorphic rocks be strictly
compared with igneous rocks, or is it possible that the charnockites are really
metamorphic rocks ? It behoves us to critically examine the evidence.

* " Gneisses and Dacites of the Dandenong District," E. W. Skeats, Q.J.G.S., vol. LXVL, 1910, pp. 450-469.
t " The Charnockite Series," T. S. Holland, Mem. G.S. India, XXVIII., pt. 2. Numbers in brackets in the following
refer to pages in this publication.

THE METAMORPHIC ROCKS OF ADEL1E LAND.-STILLWELL. 193

2. THE CHARNOCKITE SERIES.

The charnockites are a group of crystalline rocks which appear among the Archtean
gneisses in Southern India. The distinguishing features of the unaltered members
of the series (p. 125) are the constant even-grained granulitic structure and the constant
presence of the mineral hypersthene, while garnet uniformly appears in the gneissose
forms, just as in the Antarctic rocks. The chief types range from acid charnockite
and leptynite (granulite) to basic norites and ultrabasic pyroxenites and hornblendites.
They are a series which possesses resemblances to the Saxon pyroxene granulites and
the French pyroxene gneisses, and to other ancient pyroxenic eruptives ; but they are
determined from evidence within the series itself to be of igneous origin. At times they
are acknowledged to have suffered some alteration, but the igneous character is held
to be dominant.

The pyroxene granulites and the pyroxene gneisses are looked upon as crystalline
schists, and hence it must be considered possible, apart from Antarctic evidence, that
the charnockites are similarly so. Besides, we are told (p. 195) that the charnockites
are quite old enough to be affected in the same way as the Dharwar system of crystalline
schists.

The charnockite series groups together acid and basic rocks in a way that is known
to be (p. 154) contrary to the usual practice of petrographical (igneous rock) classification,
and Holland (pp. 131, 210) is quite aware that metamorphism tends to reduce points
of difference between rocks of diverse origin and to produce similarity. Holland is also
aware that the granulitic structure (p. 154) with general absence of idiomorphism favours
a metamorphic origin, but he believes that a granulitic structure may result from dis-
turbance of the magma during the process of consolidation. This belief rests on the
observation that dykes of pyroxenite cut the norites, and the igneous origin of the
pyroxenite cannot, therefore, be doubted. But this observation does not preclude
the possibility that the pyroxenite dyke and the norite have suffered similar meta-
morphic conditions during which the dyke characters have been preserved in the same
manner as at Cape Gray. Such grouping, then, of the charnockite series indicates
the outstanding nature of the genetic relationship between the various members.

Now we find (p. 125) that nearly all varieties possess a linear arrangement of the
constituent minerals, i.e., a foliation. In the case of rocks of St. Thomas Mount and
Pallavaram the direction is constant between N.N.E. and S.S.W. Holland insists
that this foliation is not a metamorphic feature, because it may occur (p. 125) in rocks
with a complete absence of all signs of crushing, and because (p. 137) the most delicate
interlocking structures may be preserved, and all signs of dynamo metamorphism are
wanting. Hence Holland concludes that this disposition of the minerals occurred
before consolidation. It seems to us that the degree of dynamo metamorphism is here
determined wholly by the amount of induced mechanical structures, e.g., granulitisation,
mylonisation, etc., and that this is an instance that would justify Wienschenk's criticism

Serin A, VoL m., Part 1 N

194 AUSTKALASIAN ANTARCTIC EXPEDITION.

of the term dynamo metamorphism and his introduction of the terms piezo contact
metamorphism, etc. It must now be considered a fundamental fact that mechanical
structures are often absent in thoroughly recrystallised rocks. We have elsewhere
argued that the gneissic foliation in granites and similar rocks is always a metamorphic
feature in that it involves a rearrangement of crystals after, not before, consolidation.
We reached this conclusion because stress is considered to be an essential factor in its
production. Holland acknowledges the action of stress in saying (p. 125) that " the
crystals are arranged with their long axes at right angles to the direction of maximum
pressure." But when he adds " before consolidation " he is implying the action of a
stress through a liquid which is impossible. Holland's interpretation of the foliation
and banding has produced the chief difficulty in the determination of the charnockites
as metamorphic rocks.

If this difficulty be removed there is no barrier to the interpretation of the
charnockites, like the Antarctic rocks, as a suite of igneous rocks that have suffered
complete recrystallisation under the conditions of the kata zone of metamorphism.
The demonstration of the primary igneous character by the form and structure of the
great massif, by the existence of dykes and apophyses proceeding from the main mass
and by the presence of included fragments of foreign rocks, is unaffected by the
acknowledgement of kata zone metamorphism. Nor is the chemical and mineralogical
evidence affected. Eecrystallisation under kata zone conditions (high uniform pressure,
dominating largely over stress and high temperature) does not destroy the dominating
igneous structures, e.g., the dyke structures at Cape Gray are so wonderfully clear that,
foliation being almost absent, no metamorphism was suspected in the field. Observation
of destroyed dyke structures are found restricted to areas where strong stress is evident.
It seems to us that a metamorphic history applied in general to the Charnockite series,
together with a double metamorphic history in those cases where mechanical structures
are evident, satisfies the recorded evidence including the supposed igneous abnormalities.

The inclusion of rounded blebs of quartz in the felspar (quartz de corrosion) is a
metamorphic structure analogous to the case of rounded blebs of quartz or felspar in
the hornblende. The dissociation of felspar into microperthitic intergrowths is frequently
associated with metamorphism, and it is also to be noted that the hypersthene is
associated with a similar greenish alteration product in both the Indian and Antarctic
rocks.

We are also inclined to read further evidence of recrystallisation after the primary
consolidation of the associated rocks in the frequent observation of crystalline continuity
across junctions. Sections cut across the junction of quartz felspar veins (p. 145) which
cut across the charnockite and which bear the same family likeness, show no abrupt
junction but a line of interlocking crystals. The same crystalline continuity is present
across the junction of quartz veins in the norite (p. 157), across the junction of pyroxenite
and norite, and across the junction of tongues of charnockite (p. 226) which ramify
the biotite gneiss, while there is a gradual passage from charnockite to garnetiferous

THE METAMORPHIC ROCKS OF ADELIE LAND. STILLWELL. 195

leptynite (granulite). Such crystalline continuity is, therefore, quite independent of
the nature of the rocks, and is most likely an impressed metamorphic character. In
specimens showing a junction from the Cape Gray Promontory there is a gradual
microscopic transition from one rock to the other.

In describing the garnetiferous leptynite as a pressure altered form of charnockite
(p. 142), Holland assumes that the set of metamorphic conditions which produced the
granulated selvages has also produced the garnet. He argues that the temperature
of crushing must be high. It can be pointed out that such is unnecessary as the
acquisition of mechanical structures may be subsequent to the development of the
garnet.

The typical augite norite from St. Thomas Mount (p. 156) is very similar to the
plagioclase pyroxene gneiss at the Madigan Nunatak minus the mechanical structures,
but the hornblende plagioclase pyroxene gneiss from Aurora Peak seems identical with
the more common form of hornblende augite norite. If this comparison is correct we
would expect that the hornblende is derived at the expense of the pyroxene, but I have
not noticed this to be stated in references to the norites. Amphibolisation of pyroxene
in the pyroxenites, an ultrablastic form of the charnockites, is described (p. 169), and
it is noted (p. 170) that, whenever hornblende becomes a prominent constituent of the
pyroxenites, the norites have also a conspicuous amount of hornblende. This amphiboli-
sation is, therefore, a character which has probably developed by the impress of certain
external conditions which have affected both the norite and pyroxenite together, i.e.,
it is a recrystallisation subsequent to the formation of both pyroxenite and norite, and
hence a metamorphic character. If the whole series of charnockites be metamorphic
this development of hornblende may belong to a second metamorphic phase. Similar
evidence of metamorphism is reported from the Nilgiris. It is stated (p. 121) that
where the charnockite series is garnetiferous, the coarse-grained segregation or con-
temporaneous veins composed of quartz, felspar, and hypersthene also include garnet.
Like the hornblende the distribution of the garnet is not controlled by igneous
boundaries.

It is to be noted that, in using the term norite for these rocks, Holland is quite
aware (p. 153) that they possess different appearance from olivine norite and augite
norite which can appear as normal dyke rocks. If the metamorphic character be upheld
this fact should not permit the retention of the name " norite." The norites, rich in
garnet, should be acknowledged as metamorphic rocks, because Holland considers
that the garnet, with its numerous characteristic inclusions, is formed from the
pyroxene.

Where we find a description (p. 186) of pegmatoidal pyroxene plagioclase rocks in
the Nilgiri mass we may read further metamorphic evidence. These rocks occur in
lenticular masses which appear in trains along the same band of charnockite. The
lenticular shape is looked upon as the result of the pinching of a once continuous band

196 AUSTRALASIAN ANTARCTIC EXPEDITION.

of charnockite, and in view of the Cape Denison phenomena this is quite consistent with
metaniorphic action. The pegmatoidal structure described between the felspar and
pyroxene in these rocks reads like a known metaniorphic structure, and the amphibolisa-
tion of the pyroxene certainly is.

Finally, it may be pointed out that the so-called abnormal igneous features (p. 244)
find ready explanation on this metamorphic hypothesis. Whereas the granulitic
structure and the almost complete absencg of pronounced porphyritic crystals are
remarkable for large masses of igneous rock, they are normal features in large masses
which have been thoroughly recrystallised under kata zone conditions. The frequent
presence of garnet cannot be adequately explained by an igneous hypothesis supple-
mented by subsequent mechanical deformation, and the parallel arrangement of the
constituent minerals requires the action of stress after consolidation. If the pressure
during recrystallisation were wholly the hydrostatic type no foliation would result, but
such cannot be generally expected. The small amount of foliation denotes a weak
stress.

An acknowledged metamorphic character can readily admit a variety like biotite
pyroxene gneiss referred to by Fermor,* and considered by Holland to be probably
an abnormal member of the charnockite aeries.

3. THE " INFRAPLUTONIC ZONE" HYPOTHESIS.

Could the conclusion concerning the metamorphic nature of the charnockite series
be avoided if we accept Fermor's conception of an infraplutonic zone ? In this hypo-
thesis Fermor postulates the existence of a shell in the earth's crust, situated below the
depth at which plutonic rocks consolidate, and characterised by garnets. The shell
must lie at considerable depths, and the temperature and pressure are very high. It
extends round the earth, and the whole of it is a potential magma.f

Fermor's theory arises from the study of an area of garnetiferous and manganiferous
rocks which have been named the " Kodurite Series." In certain members of this series
Fermor discovered that a calculation of the specific gravities of the mode and norm
of the spandite rock (Ca-Mn garnet), and of kodurite (orthoclase, Mn garnet and
apatite), showed that the spandite rock occupied 20 per cent., and the kodurite 10 per
cent., smaller volume than its norm. Such indicates that the conditions favourable
to the formation of garnet rocks are those of high pressure. If high pressure conditions
have prevailed in the case of the kodurites, Fermor expects to find garnets in the various
rock series associated with the kodurites. As this is so, he then suggests that eclogite
must be a high pressure form of gabbro. It may be pointed out that eclogite has already

* " Manganese Deposits of India," L. L. Fermor, pt. II., p. 245, Mem. G.S.I., 37.

t " Preliminary Note on Garnet as a Geological Barometer and on an Infraplutonic Zone in the Earth's Crust " L. L.
Fermor, Rec. G.S.I. XLIII, pt. I., 1913.

THE METAMORPHIC ROCKS OF ADELIE LAND. 8TILLWELL. 197

been considered as a high-temperature and high-pressure alteration form of gabbro,*
and further, that it is a product of the kata zone of metamorphism in which Grubenmann
has recorded garnet as a common mineral. f

The hypothesis has only been put forward, so far, in a short preliminary paper
in a general manner. The absence of detail and of references makes it difficult to arrive
at a just estimate of its worth. The shell is supposed to be normally solid, and only
becomes liquid on release of pressure ; yet the constituents in this zone are, in a general
way, spoken of as " crystallising out." The phrase, " crystallising out," is generally
used in reference to magmas and solutions, but it may be applied to solid solutions.
Fermor does not indicate that he is referring to solid solutioas.

In the infraplutonic zone the formation of garnet occurs in those reactions and
rearrangements which are accompanied by reduction of volume and absorption of heat ;
and he points out that the formation of garnet from other minerals, such as pyroxenes,
olivines, and iron ores, is always accompanied by decrease in volume. But he does not
offer any explanation as to how the pyroxene, olivine, and iron ore happen to reach
the infraplutonic zone. If we are to suppose that some pre-existing rock is buried
by earth movements to such a depth that it reaches the infraplutonic zone, we are
merely imposing a set of metamorphic conditions upon the rock. In this case the infra-
plutonic zone is indistinguishable from a metamorphic zone defined by the particular
set of conditions. If melting should follow the exit of the rock from the infraplutonic
zone the rock will assume the characters of a normal eruptive rock. If melting does
not occur with release of pressure, and it has not occurred in the garnet rocks that I have
studied, a normal metamorphic rock would appear to result.

The difference between the infraplutonic zone and the kata zone of metamorphism
has not been considered by Fermor. If there is a similarity we point out that the infra-
plutonic conception involves the worst feature of the conception of metamorphic zones
viz., that of depth. A metamorphic zone can only be adequately defined by a set of
physico-chemical conditions, and not by varying depths in the earth's crust.

When a plutonic rock forms from its magma we date its existence as a unit from the
time of its consolidation. The petrologist must, at present, be content to leave open
the questions concerning the origin of the magmas, because " cosmogony can afford
no firm foundation for a priori reasoning."f Perhaps we should do the same with the
products of the infraplutonic zone. Perhaps an infraplutonic zone product is meant
to be analogous to a plutonic zone product, differing only in the conditions of temperature
and pressure under which it forms. If this is so the infraplutonic conception does not
offer any explanation at all of the large number of garnet rocks which have formed
by recrystallisation in the solid state, and whose former igneous or sedimentary origin
can be traced.

* " Ein Beitng zur Kerntnu der Eclogite und Amphibolite mit beaonderer benicksichtigung der Vorkommniane dw
Mittleren Otrtalea," Laura Hezner, published by Alfred Holder, Wien, 1903.
f Grubenmann, op. cit., vol. II., p. 83.
J " Natural Hutory of Igneous Rock," A. Barker, p. 4.

198 AUSTRALASIAN ANTARCTIC EXPEDITION.

Contrary to Fermor's supposition we find on the close examination of a given
garnetiferous area that some rocks may contain garnet while others do not. In all
localities on the Cape Gray Promontory most rocks have been described with garnet,
but in each case there are types without garnet. In some cases we have described the
incipient stages of garnet formation. Further, we have found that the garnet-forming
conditions are highly localised in the garnet hypersthene felspar gneisses of Stillwell
Island and the Cape Pigeon Rocks. They may be present at one end of a specimen
Sin. long, while totally absent at the other end. In these the evidence is quite definite
that the garnet has formed by reactions between existing minerals without fusion,
and the metamorphic nature is undoubted. Such pronounced variation is scarcely
compatible with the infraplutonic zone hypothesis.

The infraplutonic hypothesis has been advanced also without consideration of
those instances in which garnets are known to be products of magmas. It is well known
that some garnets, like melanite, appear in alkaline volcanic rocks, and we also found
garnet at Cape Denison appearing both in a large crystal, Gin. in diameter, and in graphic
intergrowth with quartz in the same pegmatite associated with the granodiorite gneiss
in which garnets are absent. We do not doubt at present, therefore, that some garnets
may form directly from solution.

Until Fermor discusses this mode of formation of garnet and distinguishes the
infraplutonic zone from the kata zone of metamorphism and points out its relative
advantages, a geologist will be unable to use his conception. For the present we must
conclude that it does not give a reasonable account of the origin of the charnockite
series.

4. THE KODURITE SERIES.

We have turned to the short accounts available of the kodurite series to discover
whether this remarkable series, which is responsible for the infraplutonic hypothesis,
is considered to possess any metamorphic traits. At the commencement we notice
that Fermor* assumes that certain gneissose granites gain their foliated character
because they were intruded at the time of the folding of the Dharwar series, while other
granites, which were intruded subsequent to this series, have only a banded structure
due to flow. We, therefore, anticipate an incomplete appreciation of metamorphic
individuality, while it has already been pointed out by Crossf that further proof is required
before geologists can be expected to accept the view of igneous origin.

In this publication we find that the kodurite series is stated (p. 244) to be part of
an ancient group of rocks which include the charnockite series, a gneissose granite, calc
gneisses, and the khondalite series (metamorphosed sediments). As we consider these

* " Manganese Ore Deposits of India," L. L. Fermor, Mem. G.S.I., XXXVII., pt. 2, 237.

t " Problems of Petrographic Classification suggested by the Kodurite Series of India," W. Cross, Journ. Geol., XXII.,
1914, p. 794.

THE METAMORPHIC ROCKS OF ADELIE LAND. 8TILLWELL. 199

rocks possess metamorphic character*, and as they all outcrop in the form of parallel
bands, it is probable that the kodurite series, which also possesses the banded arrange-
ment, likewise possesses such characters.

Fermor admits the likeness to metamorphosed manganiferous sediments in some
cases, but he is influenced by the belief that there is little or no evidence that the kodurite
series has suffered much by earth movements. There remains, however, a suspicion
that the " suffering from earth movements " means nothing more than the presence
of mechanical crush structures.

Even if we admit, as in the case of the charnockites, the evidence of igneous origin,
there remains the possibility of a metamorphic character superimposed upon the igneous
character ; and, if this were so, the rocks must be classed as metamorphic, because
it is quite impracticable, as argued by Crook,* to classify altered rocks according to their
original condition. We notice (p. 250) that the kodurite family likeness extends over
types varying from acid to ultra-basic, i.e., there is a family characteristic, as in the
charnockites, which is independent of chemical composition, and which we interpret
as a metamorphic character. One of the supports of the igneous hypothesis is the
assumed magmatic differentiation, which depends on the observation that the more
basic rocks (p. 254), such as spandite rock, occur as large patches or streaks, surrounded
by zones of less basic composition, such as kodurite, in a general matrix of quartz felspar
or felspar rock. In a general report for 1914 it is stated f that Fermor is inclined to
replace this interpretation with the supposition of assimilation on a large scale, and to
consider that a granite magma has bodily dissolved entire manganese ore deposits.
Either of these hypotheses will be difficult to substantiate, but the discussion will not
be complete without a consideration of metamorphic differentiation or metamorphic
diffusion.

The full and detailed account of the petrology of these rocks is not yet published,
but the information available suggests that there has not been full consideration of
possible metamorphic characters.^

"The Genetic Classification of Rocks," T. Crook, Min. Mag., XVII., 1914, p. 70.

t " General Report of the Geological Survey of India for the Year 1915," C. S. Middlemiss, Roc. O.S.I., vol. XLV., pt. 2,
p. 103.

J In addition to the two reference* quoted there is " The Systematic Position of the Kodurite Series, especially with
reference to the Quantitative Classification," L. L. Fermor, Rec. G.S.I., XLII., pt. 3, p. 208.

CHAPTER XIII.

THE GENERAL PROBLEM OF TRANSFERENCE OF MATERIAL DURING

METAMORPHISM.

There is no fundamental difference between the processes which have been termed
metamorphic diffusion and metamorphic differentiation. The two terms have been
introduced for convenience. Metamorphic diffusion merely involves a migration of
material in the solid rock during the recrystallisation, and has been studied along pre-
existing junctions. Metamorphic differentiation requires, in addition to a migration,
a segregation of migrated molecules. Metamorphic diffusion of some constituents is
necessary to bring about metamorphic differentiation, and in this way a metamorphic
differentiation product is also a metamorphic diffusion product ; but the converse is
not true. Both processes are probably governed by the same fundamental principles,
and both involve the wide problem of the general transfer of material during meta-
morphism. The principle asserted in this problem is one that has not been accorded
general acceptance chiefly on account of the paucity of sure evidence and the possibility
of alternative hypotheses.

EVIDENCE OF MIGRATION IN GEOLOGICAL LITERATURE.

Chemical changes in total composition during metamorphism has been recently
argued by Leith and Mead,* and each chemical change requires a migration. These
authors look upon amphibolite as an end product of the metamorphism of marble.
Though we see reason to question this view, we think that the phenomena described
by Adams and Barlow f at the junction of a crystalline limestone and an amphibolite
provide sound evidence of a transfer of material during metamorphism. Further
examples are quoted by Leith and Mead. W. S. Bayley has provided the record in the
Menominee District of Michigan that " at some places the dolomites at their contact with
their overlying iron formation have been entirely changed from their original condition
and are now represented by talc and serpentine."! Such a change would involve an
important change of composition. The transformation of quartzite into sericite schist
is a change that has been substantiated by Truemann, but, so long as the authors ascribe
the removal to the agency of solutions, it cannot be considered as evidence of the
molecular transfer of material in the solid state during dynamo metamorphism. Strong
evidence has been put forward by G. H. Williams, who has traced a gabbro into a

* " Metamorphic Studies," C. K. Leith and W. J. Mead, Journ. Geol., vol. XXIII., 1915, p. 602.
f Op. cit., Adams and Barlow, p. 87.

J " Menominee Iron Bearing District of Michigan," W. S. Bayley. Mon. 46, U.S. Geol. Survey, 1904, p. 221.
" The Greenstone Schist Areas of the Menominee and Marquette Regions of Michigan." Bull. 62, U.S. Geol. Surv.,
1890, p. 76.

THE METAMORPHIC ROCKS OF ADELIE LAND. 8TILLWELL. 201

green-stone schist (sericite chlorite schist), and the alteration in composition is proved
by chemical analyses. He correlates the chemical changes between fresh gabbro,
8 aussuritised gabbro, and greenstone schist with varying amounts of chloritisation,
sericitisation, and carbonation.

Further evidence still is claimed in the theory advanced by Leith and Mead that there
is a convergence under metamorphic conditions towards definite mineralogical types,
chiefly hornblende, chlorite, or mica. If such a conception is warranted it does provide
extra evidence. The unpublished text book referred to in their paper is now available,*
but no detailed petrological study is forthcoming of the instances in which these types
are produced. The theory depends largely on the comparison of groups of analyses of
sericite schists, chlorite schists, and hornblende schists, which have been divorced from
all field associations and from questions of origin. The theory at present remains a
mere speculation until some knowledge is gained of the fate of that part of the chemical
composition of the rock which must be rejected in order to produce hornblende, mica,
or chlorite rocks. The theory further neglects to take account of those types of meta-
morphic products which are produced under metamorphic conditions incompatible
with the existence of chlorite, hornblende, or mica, e.g., pyroxene gneisses or any rock
that is formed under the conditions of Grubenmann's kata zone where the normal
structure is massive not schistose.

It may be suggested that the " convergence " theory is an incomplete expression
of the theory of metamorphic differentiation, because it has been shown that hornblende
rock, chlorite rock, and biotite-hornblende rock are formed by metamorphic differentia-
tion. If it be subsequently found that metamorphic differentiation can only occur under
those metamorphic conditions which tend to produce hornblende, chlorite, or mica,
then there will be a closer agreement between the theories than is anticipated. As,
however, we have instanced a large garnet crystal as a minute differentiation we must
at least suppose that differentiation can occur under garnet-forming conditions which
are normally those of the kata zone.

Some evidence, therefore, has been culled from the geological literature by Leith
and Mead to establish the theory that some transference of material occurs during
metamorphism. On the other hand a wealth of evidence can be obtained to prove
that in many instances no important change has occurred. This, indeed, is the basic
principle upon which Grubenmann has built his classification of schists. Hence the
degree and range of the transference must necessarily be limited. No more than this
is implied in the theory of metamorphic diffusion.

If transference of material occurs across a pre-existing junction of two rock types
the junction will be replaced by a transition which we call a metamorphic diffusion
product. Many examples of destroyed junctions have been described between igneous
and sedimentary rocks, but in most cases it has been ascribed to assimilation of the

" MeUmorphic Geology," C. K. Leith and W. J. Mewl. New York. 1016.

202 AUSTRALASIAN ANTARCTIC EXPEDITION.

sedimentary rock by the igneous. The examples have been looked upon as igneous
phenomena, not metamorphic. Cole, for instance, describes a granite which, he claims,
has invaded and partially absorbed amphibolite. ' The various stages of absorption can
be traced with the unaided eye. Lumps of amphibolite seem to swim in the gneiss and
fade off into it, as if melting before our eyes. The gneiss becomes enriched with streaks
of basic matter in which biotite begins to predominate over amphibole."* The same
attitude has been maintained by Cole in his recent address to the British Association for
the Advancement of Science,t and a list of references is quoted where similar passages
between metamorphosed igneous and sedimentary rocks have been recorded in other
areas. As far as it can be verified, all the instances could be given a metamorphic
interpretation, in which case they become examples of metamorphic diffusion and pro-
vide evidence of the transfer of material during metamorphism. We cannot accept
Cole's statement that amphibolite is the " final term of various metamorphic series."
It is a statement in which considerable migration during metamorphism is tacitly
assumed, and it is an unwarranted assumption when we reflect that Cole is endeavour-
ing to apply the theory of contact metamorphism on the evidence of transition.

A further example may be quoted from the Broken Hill region. Mawson states J
that where the metamorphic conditions are intense a passage of materials may take place
from the intrusive rock into the intruded rock. He instances a gradation between
quartzite and quartz felspar rocks east of Mookaie Hill. Mawson, however, is inclined
to regard these transitions as part of a pneumatolytic effect.

An igneous magma may, of course, completely alter the character of the invaded
sediments, but there still remain to be found the examples where a junction has been
destroyed and replaced by transition types except where powerful stresses are evident.
Even with the play of stresses such junctions frequently survive, e.g., 12 miles north of
Casterton, Victoria, the junction of a foliated granite, presumably of Archaean age,
with the invaded sedimentary schists, is still perfectly sharp and transition types are
apparently absent. The fact that the bulk of the dyke rock has preserved its apparent
sharp junction with the grey gneiss is evidence that only special conditions will permit
metamorphic diffusion. Intrusions of igneous rock into igneous, with the junctions
subsequently modified by metamorphic diffusion, will necessarily be more difficult to
demonstrate than igneous intrusions into sedimentary rock.

Summarising, we may assert that geological literature provides some evidence
for a molecular transfer during processes of recrystallisation. It has been found a difficult
matter to demonstrate, but the interpretation of the Cape Denison phenomena, with the
aid of solid diffusion, is probably applicable to a number of rock junctions partially
destroyed during metamorphism. If this is so a large number of unrecognised examples
involving migration is indicated.

* " The Intrusive Gneiss of Tirerrill and Drumahair," G. A. J. Cole, Proc. E.I. Acad., vol. XXIV. B, pt. 4, p. 361.

t G. A. J. Cole, Pres. Add Sect. C. Brit. Ass. Adv. Sci., Manchester, 1915, p. 5.

% " Geological Investigations in the Broken Hill Area," D. Mawson, Mem. Roy. Soc. S.A., vol. II., pt. 4, p. 237.

THE METAMORPHIC ROCK8 OF ADEL1E LAND. KTILLWELL. 203

THE PROCESS OF MIGRATION.

General remarks on the problem of transference of material during metamorphism
may be offered under the heads

1. Solution.

2. Solid Diffusion.

3. Force of Crystallisation.

The remarks are confined to instances of metamorphism unaccompanied by the
addition of material from any foreign source. The rocks during metamorphism have
remained throughout as a solid, self-contained mass. There has been no essential change
in chemical composition, and metamorphism is conceived to be merely a recrystallisation
caused by super-imposed physico-chemical conditions. We also do not consider instances
in which partial fusion is assumed, because we have not met with any examples in our
field study.

1. Solution.

Solution as a transferring agent is always the factor to which there is ready appeal,
and it is not easy to understand how the decrystallisation of a rock can proceed without
it. There is always a measurable quantity of water ( H 2 0) in rocks, and this is
present in capillaries and hollow spaces ; in addition there is the chemically combined
water ( -f H 2 0) which is freed at higher temperatures, and may become an active solvent.
Magmatic water may also be included when finding its way to the surface for the first
time. Thus there is always a minute amount of solvent present.

There is no need to enter into the discussion of conditions that assist or prevent
solution. But it must be admitted here, that, provided that the metamorphic conditions
are such that water can circulate, solution in some places, with corresponding deposition
in other places, can be an agent in the transference of material in rocks which do not lose
their rigidity. This factor is, perhaps, more noticeable in the epi zone of metamorphism
which is more particularly characterised by the hydrous silicates. The presence of
minute fractures is possible in this zone, and these cracks provide more ready channels
of water percolation. In the production of amphibolites by the metamorphism of
dolerites under epi conditions, we occasionally find microscopic veins of scapolite,
lawsonite, calcite, and epidote, whose origin has been traced to large crystals of
saussuritised felspar. In such cases solution has been a direct agent in the transference
of material.

Microscopical solution and deposition is implied in the principle called Riecke's
principle by Becke and Grubenmann. This principle is discussed by Johnston and
Niggli,* who point out that it depends on the thermodynamical fact that unequal pressure,
acting only on the solid phase, increases its vapour pressure and its solubility (in any

" General Principles Underlying Metamorphic Processes," Journ. Geol. XXL, p. 603.

204 AU6TEALASIAN ANTARCTIC EXPEDITION.

particular solvent) and lowers its melting point. If a stress is applied to a rock, solution
tends to occur at the points immediately under the stress with simultaneous deposition
in planes at right angles. If solution occurs on the ends of a prismatic crystal, for
example, and deposition occurs in the plane at right angles, the prism will be flattened
first into a lenticular aggregate and finally into broad flakes. By molecular displace-
ments in this way the rock can yield to a stress as if it were plastic, and the development
of a schistose structure can be pictured with the ordinary conception of solution. If
there appear new mineral forms, such as the platy minerals like biotite, which can be
stable against the stress, a completely recrystallised rock may arise.

2. Solid Diffusion.

In many cases, however, microscopic solution cannot be so readily pictured aa
the means of transport, as for example, in the development of large porphyroblastic
crystals in metamorphic rocks which have only been subjected to weak stress, when
the pressure is mainly hydrostatic. The growth in such cases may be uniform in all
directions. When a corona of garnet crystals develops during the recrystallisation of
dolerites, by the- reaction between felspar and augite, the garnet crystals must occupy
space formerly occupied by felspar and augite ; and this space is not provided by
solution as in metasomatic replacement. As the garnet crystal grows it draws its supplies
from adjacent regions, and in this there is a molecular transfer of material and, at the
same time, a minute metamorphic differentiation. The larger the crystal grows the
greater must be the distance over which it draws its supplies, and this distance must
be appreciable when the garnet becomes over an inch in width. The well known
secondary enlargements of hornblende, augite, plagioclase, and orthoclase which have
been likened by Holland* to the growth of garnet, imply similar transfer. So also does
the growth of larger crystals at the expense of the smaller crystals, with its consequent
increase in grain size a fact that is well known among the crystalline schists. f

A full discussion of the problem of solid diffusion is given by Desch in a report to
the British Association for the Advancement of Science. J It is pointed out that the
devitrification of glasses involves molecular diffusion in solids. As glassy rocks of
Palaeozoic age are unknown it becomes evident that solid diffiision is a process that has
been operative in geological time. It is shown that diffusion in metals has been estab-
lished beyond doubt. The cementation and decarburisation of iron, the segregation
and recrystallisation of constituents in solid metallic alloys are processes involving
true solid diffusion.

In steels the iron carbide separating from solid solution is at first in a state of ultra-
microscopic division (troostite). On reheating and on different conditions of cooling

* " Origin and Growth of Garnets," T. S. Holland, Kec. G.S.I., XXIX., p. 26.
f Op. cit., Grubenmann, vol. 1, pp. 39, 78.

J " Roport on Diffusion in Solids," C. H. Desch, Brit. A.A.S., Dundee, 1912, p. 348.

Troostite, Sorbite, and Peartite are not definite compounds but aggregates of ferrite (Fe) and cementite (Fe 4 C) with
different structures.

THE METAMORPHIC ROCKS OF ADELIE LAND. -STILL WELL. 205

finely granular sorbite and laminated pearlite are successively obtained. Further
heating causes the pearlitic laminae to contract, producing beaded forms and ultimately
the carbide segregates into coarse masses.

The segregation of this constituent, iron carbide, in steels seems to be in some
measure analogous to the segregation of chlorite in the amphibolite rocks at Cape Denison.
Though the amphibolite cannot be looked upon as a solid solution, yet in each case there
is the segregation of a single constituent within a solid mass. In both cases it is an
adjustment of physical equilibrium, not chemical equilibrium. It is, of course, true
that the segregated mass of chlorite is many times larger than the segregated mass of
carbide in steel, but the former appears in the rocks of the greatest geological antiquity,
while the latter is developed in a few hours in the laboratory. In the other cases of
metamorphic differentiation where more than one constituent has been segregated,
the analogy is still suggestive of a mental picture whereby the process can go on.

" In alloys," says the report, " which contain two solid solutions in equilibrium
with one another, such as the <* ft alloys of copper and zinc, the structure becomes
coarser when the alloys are heated at a temperature at which diffusion takes place.
The increase in size of crystals is equally pronounced when only a single constituent
a pure metal or a solid solution is present. The growth of the ferrite grains in soft
steel at 700-720 is extremely rapid, and the process can be conveniently watched in
other instances. It is always the larger crystals that absorb the smaller."

' Whether the metal in which such recrystallisation takes place is homogeneous
or heterogeneous, diffusion must occur in order that rearrangement may come about.
The effect has been explained by the principle that small crystals have a greater
solubility than large, so that if small and large crystals of the same substance are both
in the presence of a solvent, solution and redeposition tend to go on until only crystals
above a certain size are present. This has been verified for the case of calcium sulphate
in water. A thermo-dynamical explanation has also been given of the principle that
the bounding surface between a crystal and its saturated solution tends to become a
minimum, so that equilibrium is only finally reached when all the small crystals have
united to form a single crystal."

" The principle of differing solubility is rejected as an explanation by G. Tammann,
who assumes that the surface tension, which is less than the forces producing rigidity in
a crystal at the ordinary temperature, may become much more considerable with increase
of temperature. When the surface tension exceeds the opposing forces, two crystals
unite as two drops of fluid do. The hypothesis is applied to explain the recrystallisation
of strained minerals."

This increase in grain size in metallic alloys seems to be closely analogous to the
increase in grain size in certain metamorphic rocks. If solid diffusion is an operating
factor in one case, it must occur in the other. At Cape Gray the first metamorphic
product derived from a dolerite has a very fine granulitic structure, with average absolute

206 AUSTRALASIAN ANTAECTIC EXPEDITION.

grain size of 0-05mm., while localised areas of slightly coarser grains may be found. This
fine-grained product has been shown to be practically identical in kind with the much
coarser-grained plagioclase pyroxene gneiss at Madigan Nunatak, with average absolute
grain size of 0-30mm. The differences in the grain size (T5mm. and 0-22mm.) of the
amphibolites (No. 9 and No. 629), at Cape Denison, illustrate the same phenomena.
In these instances all the constituents of the rock have been uniformly enlarged. There
are also cases where certain constituents are enlarged at a relatively greater rate than
other constituents and heteroblastic rocks result. It is believed that the garnets in
the garnet amphibolite (No. 953) have formed by the aggregation of smaller crystals,
arid they are uniformly larger than the average crystal of hornblende in the same rock.
The large magnetite crystals in the amphibolites, Nos. 143 and 637, at Cape Denison,
are also enlarged to a greater degree than neighbouring minerals, and consequently
appear as porphyroblasts. An even more striking example is found in the localised
variation in the size of the felspar grains that have developed in the metamorphic contact
products, the biotite felspar gneisses, at Cape Denison. In three specimens from one
locality the average width of the felspar crystal may vary from 0-90mm. in one case, to
8mm. in the second, and 27mm. in the third. Similar examples are known in other areas
of crystalling schists and in glacier ice, and Grubenmann considers that increase in
crystal size corresponds with a striving towards a condition of smallest surface tension,
because the sum of free surface is decreased.

A little evidence of solid diffusion through crystalline solids and artificial crystals
is given in Desch's report. But it is stated that experiments are lacking to prove the
occurrence of diffusion in minerals, even under favourable conditions, though indirect
evidence points to its possibility. The most favourable observations according to him
are those of schiller inclusions, as of magnetite in the olivine from peridotite, Isle of
Rum ; and in the hypersthene from norite, Labrador ; and also rutile in certain felspar
and pyroxenes. These examples are phenomena connected with igneous rocks. If
solid diffusion has been operative in rocks it is more likely to be traced in the rock
bodies that have been completely recrystaliised, without fusion, than in igneous rocks
which have merely suffered slight changes on cooling after solidification. We, there-
fore, proceed to summarise favourable instances from our study of metamorphic rocks.

1 . The presence of large porphyroblasts which have grown during the metamorphism
and illustrated by

(a) Large garnets, 5cm. broad, in the garnet gneisses at Stillwell Island and

Garnet Point.

(b) Large felspars (oligoclase-andesine), 27mm. long, at Cape Denison, in

certain metamorphic contact products, the biotite felspar gneisses.

amphibolite found on the moraines at Cape Denison.

(d) Large magnetite crystals in the amphibolites Nos. 143 and 637, at Cape

Denison.

THE METAMORPHIC ROCKS OF ADELIE LAND . STILLWELL. 207

2. The variation in the crystal grain size in the same type of metamorphic rock,
illustrated by

(a) Plagioclase pyroxene gneisses at Cape Gray, Madigan Nunatak, and Aurora

Peak, with average absolute grain sizes 0-05mm., 0-30mm., 0-17mm.
respectively.

(b) Coarse and fine amphibolites (No. 9 and No. 629), at Cape Denison, with

average absolute grain sizes l-5mm. and 0-22mm. respectively.

Denison, in which the absolute size of the felspar may vary from 0-30mm.
up to 27mm.

3. The existence of rocks which have been described as metamorphic differentiation
products

(a) Chlorite rock.

(6) Epidosite.

(d) Bands and lenses of pure hornblende or felspar, associated with the coarse-

grained amphibolites.

(e) The nodular zones of magnetite, sphene, and felspar in the amphibolite

(No. 143).

4. The existence of the rocks which have been described as metamorphic diffusion
products

(a) The indefinite junction between some amphibolites and the granodiorite

gneiss at Cape Denison, with the formation of hornblende and biotite
gneisses.

(b) The apparently sharp junction between amphibolite and granodiorite

gneiss illustrated in the specimen No. 372, collected from the moraines
at Cape Denison, in which the amphibolite and gneiss are separated
by a zone of biotite felspar gneiss 1cm. wide.

at Cape Denison.

(d) The change of the thin threads of original basic dyke into biotite felspar

gneisses at Cape Denison.

(e) The observed passage between cyanite garnet gneiss and amphibolite at

Garnet Point.

5. In the recrystallisation of the primary augite of the basic rocks in the Cape Gray
dyke series the segregation of the ilmenite involves the transmission of material through
solid crystalline material.

208 AUSTRALASIAN ANTARCTIC EXPEDITION.

6. The highly localised nature of the reactions which result in the formation of
garnet, etc., in the garnet hypersthene felspar gneisses at the Cape Pigeon Rocks and
Stillwell Island, indicate that the reactions have occurred in solid rock and the molecular
supply is only provided from a very limited range.

7. The manner of change in the isomorphous mixture of plagioclase, from labradorite
to andesine or oligoclase, during the reaction of labradorite with augite in the production
of garnet in the garnet plagioclase pyroxene gneiss (No. 935) from Stillwell Island,
indicates diffusion in the solid state.

8. The development of felspar intergrowths in metamorphic rocks, similar in
character to some of the micrographic intergrowths, probably indicates solid diffusion.
The intergrowths are especially abundant in the garnet hypersthene felspar gneisses
at Cape Pigeon Rocks and Stillwell Island. These intergrowths bear some analogy to
certain pearlitic intergrowths which have developed with conditions permitting solid
diffusion. Such pearlitic intergrowths of iron carbide in phosphoferrite (solid solution
of iron phosphide in iron) were produced by Stead* during the cementation of an ingot
of iron containing 2 per cent, phosphorus and a little carbon. During this process a
small iron core was not penetrated by the external carbon, and it gave up even the small
percentage of carbon that it already had, and fan-shaped pearlitic intergrowths were
produced along the junction of the altered and unaltered portions.

Though we can thus secure reason to suppose that solid diffusion can occur in
crystalline minerals, we cannot find in solid diffusion the whole cause of the phenomena
mentioned. In some of these cases the metamorphic diffusion products are highly
schistose. This means that solid diffusion has not merely resulted in indiscriminate
mixing of molecules, in the manner tending to reduce heterogeneous systems to homo-
geneous rocks. The operation of microscopic solution and deposition under Riecke's
principle, already mentioned, helps us to picture the development of the schistosity ;
but there is something else involved in the production of some of the structures.

3. Force of Crystallisation.

In the production of the metamorphic differentiation products, such as the chlorite
clot, we picture solid diffusion providing the means of molecular supply ; but there is
still some reason why all the chlorite molecules should converge to one centre. In the
formation of the zoned nodules of felspar, sphene, and magnetite in the amphibolite
No. 143, there must be some reason other than Riecke's principle why there should be
such an orderly arrangement. There must be some driving force which results first
in an attraction of the magnetite molecules to produce the magnetite nucleus, and
secondly in an attraction of the felspar molecules to give the felspar zone. Contempo-
raneously there is an outward diffusion of the ferromagnesian minerals. The production

" The Segregatory and Migatory Habit of Solids in Alloys and in Steels below Critical Points," J. E. Stead, Journ.
Soo. Chem. Ind., 1903, p. 340.

THE METAMORPHIC ROCKS OF ADELIE LAND. 8TILLWELL. 209

of the sphene rim can be pictured, if necessary, as produced by reaction of the Ti0 2
content of the magnetite with the felspar. Mere differences in rates of solid diffusion
of the different constituents will not explain why felspar and magnetite have travelled
in one direction while the ferro-magnesian goes in the other. The driving force wliich
controls the direction of migration of the chlorite or of magnetite molecules we express
in the term " Force of crystallisation." Stead, in the above-mentioned study on
the segregatory habits of solids in alloys, recognised this driving force in the term
" Crystallic attraction."

The force of crystallisation, or the power that a crystal has of drawing to itself
molecules of its own kind, has been demonstrated by Becker and Day* to be, in
supersaturated solutions, commensurate with the crushing strength of the rock. It is
the same force which, varying in different minerals, produces the crystalloblastic order.
It is, therefore, an important. factor, and can be pictured as a directive agency in the
recrystallisation of rocks.

" linear Force of Growing Crystals," Becker and Day, Proc. Wash. Acad. Sci., vol. VII., 1905, pp. 283-288.

Series A, Vol. in.. Part 1 O

DESCRIPTION OF PLATES.

PLATE I.

Fig. 1. Amphibolite, No. 629, Cape Denison, showing mainly hornblende and felspar.
Sphene with included magnetite can be seen near the centre. Mag. 35 diam.

Fig. 2. Biotite amphibolite, No. 412, Cape Denison, showing biotite and hornblende
in nearly equal proportions. Mag. 35 diam.

Fig. 3. Epidote biotite schist, No. 153, Cape Denison, showing biotite, epidote,
and felspar. Sphene with included magnetite can be seen in the centre, with a crystal
of hornblende a little to the left of it. Mag. 35 diam.

Fig. 4. Lawsonite amphibolite, No. 635, Cape Denison. Hornblende is the most
abundant mineral and the felspar is cloudy with saussuritisation. Biotite and lawsonite
are intergrown and the lobate outline of the latter can be distinguished. Mag. 35 diam.

Fig. 5. Amphibolite, No. 628-6, with a vein of lawsonite, Cape Denison. The
walls of the vein are lined with epidote. Mag. 35 diam.

Fig. 6. Lawsonite amphibolite, No. 720, Cape Denison. Hornblende is the most
abundant constituent and the felspar is cloudy. The lawsonite is colourless, with good
cleavage, and in large crystals, and it may be seen bending against the hornblende and
extending across the field. Mag. 35 diam.

PLATE II.

Fig. 1. The junction of a meta-xenolith of gneiss with the amphibolite, Cape
Denison. The xenolith, the colourless portion, consists of a granular aggregate of quartz
and felspar. Mag. 35 diam.

Fig. 2. The same field as the preceding, in polarised light, and the granulitic aggre-
gate of quartz and felspar forming the xenolith is apparent. Mag. 35 diam.

Fig. 3. A large relic crystal of quartz in a meta-xenolith of gneiss. The early stages
of the granulitisation can be seen at the extremity of the large crystal. X nicols. Mag.
45 diam.

Fig. 4. The junction of a meta-xenolith of saussurite with the amphibolite. The
boundary of the primary felspar has been preserved in this example. Large crystals
of epidote are set in the saussuritic aggregate. Mag. 35 diam.

THE METAMORPH1C ROCKS OF ADELIE LAND. 8TILLWELL. 211

Fig. 5. Epidosite, No. 415, Cape Denison, showing epidote and felspar. The
dark mineral in the upper half is sphene, and a crystal of hornblende is situated in the
left centre. It occurs as a clot in the amphibolito, No. 628. Mag. 35 diam.

Fig. 6. Biotite hornblende schist, No. 4, Cape Denisop, showing long prisms of
hornblende intergrown with biotite. This rock occurs as a clot in the amphibolite
bands. Mag. 35 diam.

PLATE III.

Fig. 1. Diablastic structure, or a secondary intergrovvth of quartz and felspar
in the hypersthene alkali felspar gneiss at Madigan Nunatak. X nicols. Mag. 35 diam.

Fig. 2. A similar, but much finer, diablastic intergrowth in the same slide as Fig. 1.
Mortar structure is seen and the intergrowth has developed in the crush area. X nicols.
Mag. 35 diam.

Fig. 3. Hypersthene alkali felspar gneiss, No. 949, Stillwell Island. It is a rock
similar to " charnockite " and occurs in dyke form. Mag. 35 diam.

Fig. 4. Garnet cordierite gneiss, Cape Gray, showing the finely granulitic character
of the cordierite. The large clear crystal at the top is quartz. X nicols. Mag. 35 diam.

Fig. 5. This and the following illustrate a slide cut across the junction of the cyanite
biotite gneiss with the amphibolite at Garnet Point, No. 781. This figure illustrates
the hornblende part of the slide. The dark mineral is hornblende, and the colourless
mineral with which it is intergrown is cyanite. Mag. 35 diam.

Fig. 6. The biotite part of the slide, No. 781. The dark mineral is biotite, and the
mineral with high refractive index is garnet. The colourless portion consists of cyanite
and quartz chiefly. Mag. 35 diam.

PLATE IV.

Fig. 1. Plagioclase pyroxene gneiss, No. 794, Madigan Nunatak. Plagioclase,
pyroxene, and ilmenite with a very little hornblende are visible. The narrow granulated
selvages around the pyroxene crystals can be detected in part. The rock has practically
the same percentage mineral composition as No. 773 (Plate VI., fig. 5). Mag. 35 diam.

Fig. 2. Hornblende plagioclase pyroxene gneiss, No. 759, Aurora Peak. The
darker hornblende is readily distinguished from the pyroxene on the one hand and from
the ilmenite on the other. The rock has practically the same mineral composition as
No. 794, Fig. 1, except that part of the pyroxene is replaced by hornblende. Mag. 35
diam.

Fig. 3. Garnet hypersthene felspar gneiss, No. 785 (2), Cape Pigeon Rocks. A
crystal of ilmenite is surrounded by a zone of biotite and quartz. A little ilmenite and
garnet are scattered through the zone which extends as a bight into a large hypersthene
crystal on the left and top of the figure. Mag. 45 diam.

212 AUSTRALASIAN ANTAECTIC EXPEDITION.

Fig. 4. Garnet hypersthene felspar gneiss, No. 785 (3), Cape Pigeon Rocks, showing
an association of biotite and hypersthene. A garnet fringe is present between the biotite
and felspar. The biotite has in part developed a perforated appearance, and a portion
of the flake is bent. Mag. 35 diam.

Fig. 5. Garnet hypersthene felspar gneiss, No. 785 (2), Cape Pigeon Rocks, showing
the intergrowths of felspars and quartz in polarised light. Mag. 45 diam.

Fig. 6. The same slide as Fig. 5, showing the relation of myrmikoidal felspar inter-
growth with the biotite. The lower right consists of biotite sprays and a thin lath of
biotite is seen as a dark line from which the felspar vermiculse radiate. Other inter-
growths appear as the stage is rotated. A crystal of microcline is situated in the top
left hand corner. X nicols. Mag. 65 diam.

PLATE V.

Fig. 1 . Garnet rims surrounding aggregates of biotite and quartz in the hypersthene
felspar gneiss, No. 785 (2), Cape Pigeon Rocks. A fragment of hypersthene remains
on one side of the aggregate. Mag. 45 diam.

Fig. 2. A crystal of hypersthene surrounded by biotite and an outer garnet rim
in the same slide as Fig. 1. Mag. 45 diam.

Fig. 3. Sprays of biotite issuing from an ilmenite nucleus in the hypersthene felspar
gneiss, No. 979 (2), Stillwell Island. Mag. 35 diam.

Fig. 4. The same field as Fig. 3 in polarised light showing the manner in which some
of the sprays open out into felspar intergrowths. Mag. 35 diam.

Fig. 5. Hypersthene alkali felspar gneiss, No. 947, Stillwell Island. The slide
is cut from the garnetiferous portion of the specimen. Biotite, with a rim of garnet,
is seen in the upper portion. Garnet, with inclusions of ilmenite, is a little lower on
the right, and pyroxene appears on the lower left. Mag. 35 diam.

Fig. 6. The garnet rims around ilmenite in the same slide as Fig. 5. One crystal
of ilmenite has been partly torn out of the slide in the grinding. Mag. 35 diam.

PLATE VI.

Fig. 1. A crystal of ilmenite surrounded by biotite and an outer garnet rim in the
hypersthene felspar gneiss from the Cape Pigeon Rocks, No. 785 (2). Mag. 45 diam.

Fig. 2. The serpentine-hypersthene aggregate in the hypersthene felspar gneiss,
No. 785 (3), from the Cape Pigeon Rocks. The crystals of hypersthene are partly altered
to serpentine and are outlined by seams of garnet. A pleochroic halo is seen in the
lower right. Mag. 35 diarn.

THE METAMORPHIC ROCKS OF ADELIE LAND 8T1LL\VELL. 213

Fig. 3. The garnet rim around the biotite is broken by the development of a spray
of secondary biotite which opens out into the usual vermicular intergrowths of felspar,
but these are not visible in ordinary light. A certain amount of the secondary biotite
also appears on the outside of the garnet rim. Hypersthene felspar gneiss, No. 785 (2),
Cape Pigeon Rocks. Mag. 35 diam.

Fig. 4. A flake of basal biotite is surrounded by a rim of younger biotite and quartz.
Hypersthene felspar gneiss, No. 979 (2), Stillwell Island. Mag. 45 diam.

Fig. 5. Plagioclase pyroxene gneiss, No. 773, Cape Gray, showing chiefly a fine-
grained aggregate of pyroxene and felspar with scattered ilmenite. The outlines of
the felspar laths of the primary dolerite can be plainly detected. The rock has the same
percentage mineral composition as No. 794 (Plate IV., fig. 1). Mag. 35 diam.

Fig. 6. The same field as Fig. 5 in polarised light. The primary felspar laths are
converted into granulitic aggregates of secondary felspar. Mag. 35 diam.

PLATE VII.

Fig. 1 . Hornblende plagioclase pyroxene gneiss, No. 766, Cape Gray. The schistose
character is seen, and the dark hornblende is distinct from the paler pyroxene. A plate
of relic pyroxene is seen partly altered to the fine granulitic aggregate of secondary
pyroxene and hornblende. Mag. 35 diam.

Fig. 2. Garnet plagioclase pyroxene gneiss, No. 935, Stillwell Island. The field
is occupied by a large crystal of relic pyroxene which is darkened by numerous minute
crystals of ilmenite. On the left and extreme right the relic pyroxene is replaced by
granulitic aggregates of clear secondary pyroxene with the partial coalescence of the
minute ilmenites. The relic pyroxene is also partly replaced by granulitic aggregates
of hornblende which surround an ilmenite nucleus. Mag. 45 diam.

Fig. 3. Garnet amphibolite, No. 799, Garnet Point. Note the manner in which
the garnet crystal is set in a felspar base. Mag. 35 diam.

Fig. 4. Amphibolite, No. 781, Garnet Point. This is the amphibolite which
junctions the cyanite biotite gneiss in Plate III., figs. 1 and 2. Mag. 35 diam.

Fig. 5. Plagioclase pyroxene gneiss, No. 951, Stillwell Island. The crystals of
biotite and pyroxene are surrounded by a diablastic intergrowth of pyroxene and felspar.
This pyroxene is partly granular and partly vermicular, and is interpreted as the first
stage in the production of garnet. Mag. 35 diam.

Fig. 6. Hornblende plagioclase pyroxene gneiss, No. 942, Stillwell Island. The
centre is a roughly circular area of diablastic pyroxene and felspar which is partly
surrounded by hornblende and biotite. In the lower part of the photograph the larger
granular pyroxene is seen. Mag. 35 diam.

214 AUSTKALASIAN ANTARCTIC EXPEDITION.

PLATE VIII.

Fig. 1. Garnet amphibolite, No. 953, Stillwell Island. The large garnet is pene-
trated by an area of vermicular pyroxene and felspar. It appears as if a portion of the
garnet crystal has broken up into the pyroxene and felspar, but this is not the general
case. Mag. 35 diam.

Fig. 2. Garnet amphibolite, No. 953, Stillwell Island. The pyroxene vermiculse
are set radially to an ilmenite nucleus and are partially enclosed by the garnet. The
crystals of ilmenite have been set in secondary pyroxene which has partly reacted with
the felspar to produce garnet while the residue now appears as vermiculse. Mag. 35 diam.

Fig. 3. Garnet amphibolite, No. 953, Stillwell Island. The same phenomena as in
Fig. 2 appears enclosed within a garnet crystal, but it has the same origin. This example
only differs from Fig. 1 in the presence of the ilmenite, and hence it is not necessary
to assume the decomposition of the garnet in Fig. 1 . The large garnets are due to growth
at the expense of the smaller garnet crystals, and, like all crystals in recrystallised rocks,
may include all other constituents. The fact of inclusion has little significance. Mag.
35 diam.

Fig. 4. Garnet plagioclase pyroxene gneiss, No. 935, Stillwell Island. The centre
of the field is occupied by a granulitic mass of secondary pyroxene, and this is surrounded
by a garnet rim, produced by the interaction between pyroxene and felspar. Mag. 35
diam.

Fig. 5. Another field in the same slide as Fig. 4. The nucleus of granular pyroxene
is much smaller. Mag. 35 diam.

Fig. 6. Almost the same field as in Fig. 5 in polarised light. The garnet is black
and the granulitic character of the pyroxene is noticeable. The manner of the change
in the composition of the felspar, concurrent with the formation of the garnet, is seen.
The outer rim of the felspar crystal, together with the " graphic " inclusions, have a
smaller angle of extinction than the bulk of the crystal. Mag. 45 diam.

PLATE IX.

Figs. 1, 2, and 3. These are different examples of the composite type of meta-
morphosed xenoliths composed of saussurite and hornblende, from Cape Denison. The
saussurite marks the junction with the enclosing amphibolite. Fig. 2 shows a remark-
ably angular xenolith. The crystal boundaries of the primary felspar in the primary
aggregate can be distinguished in Fig. 3. The specimens in Figs. 2 and 3 were not found
in situ though very close to the actual outcrop of the xenoliths. Similar specimens
were found in situ and are in the collection, but the pebbles make excellent diagrams.

Fig. 4. This is also a diagrammatic specimen, obtained from the " lower moraine,"
at Cape Denison, close to the in situ occurrence. It is an amphibolite dotted with pieces
of saussurite of varying sizes.

THE METAMORPHIC ROCKS OF ADELIE LAND.-8TILLWELL. 215

Fig. 5. Specimen of augite amphibolite found on the moraines at Cape Denison.
It contains a large porphyroblast of anorthite (Ab, An,,). These crystals are tinted
dark green and are perfectly clear and are a product of the recrystallisation.

Fig. 6. Specimen of biotite amphibolite, No. 143, collected from an enclosed basic
patch in the gneiss at Cape Denison. It contains small nodular crystals of magnetite
surrounded by a felspar zone, which passes out into normal amphibolite. Each nodule
is looked upon as a metamorphic differentiation centre.

PLATE X.

Fig. 1. Metamorphosed gneissic xenolith embedded in amphibolite. The xenoliths
are drawn out into an elongated oval form in the direction of the schistosity. Collected
in situ in the amphibolite band, No. 629, at Cape Denison.

Fig. 2. Side view of the specimen in Fig. 1. In section the angular shape of these
gneissic xenoliths can be seen, and the schistosity of the rock passes through them
irrespective of its outline.

Fig. 3. This is a more massive specimen than that in Figs. 1 or 2. Note the
irregular outline of the metamorphosed gneissic xenoliths. The outline of at least
two of these has been rendered indefinite during the recrystallisation. Collected in situ,
Cape Denison.

Fig. 4. The reverse view of the specimen in Fig. 3. The sharp angular outline
of a xenolith can be seen near the bottom left hand corner. In this specimen there are
xenoliths both with sharp boundaries and without sharp boundaries. This rapid
variation renders it unlikely that the xenoliths were partially absorbed by the dyke
magma before its primary consolidation.

Fig. 5. A remarkably angular xenolith consisting of almost pure saussurite. The
boundary is perfectly sharp and definite, except for a very small length in the bottom
left hand corner. Collected in situ, Cape Denison.

Fig. 6. A saussurite xenolith in which the re-entrant angle of the primary felspar
twin is retained. Collected in situ, Cape Denison.

PLATE XI.

Fig. 1. Composite gneiss from Cape Denison in which threads of quartz felspar
veins laminate the darker gneiss. The original boundaries are now indefinite.

Fig. 2. Specimen of granodiorite gneiss with an excessively contorted vein, Cape
Denison.

216 AUSTRALASIAN ANTARCTIC EXPEDITION.

Fig. 3. Specimen of hornblende gneiss which has been fractured by frost. It was
collected from the moraines at Cape Denison, and the two halves of the boulder were
found lying within a few feet of one another. The fracture plane does not correspond
with the direction of the schistosity.

Fig. 4. Specimen of granodiorite gneiss showing curved foliation planes. Collected
in situ from Cape Denison.

Fig. 5. Specimen of granodiorite gneiss showing a banded character, Cape Denison.

Fig. 6. Specimen of aplitic gneiss showing a large crystal of allanite. Collected
in situ, Cape Denison.

PLATE XII.

Figs. 1, 2, 3, are photographs of biotite felspar gneiss, Nos. 144, 146-1, 146-2, of
similar composition. Collected from within a few feet of one another at Cape Denison.
The porphyritic character is a metamorphic variation, and has nothing to do with the
original character of the rocks, because these rocks are considered to be metamorphic
hybrids.

Fig. 4. Specimen showing a thread of grey gneiss between two portions of dark
biotite felspar gneiss, No. 145. The boundary on either side of the grey gneiss is indefinite
as a result of metamorphic diffusion.

Fig. 5. Specimen No. 372, collected from the moraines at Cape Denison. The
apparent sharp junction between the dark amphibolite and the grey granodiorite gneiss
is seen. Actually there is a transition and a zone of biotite felspar gneiss separates the
two.

Fig. 6. Specimen of hornblende gneiss from the moraines at Cape Denison. It
contains a dark band of amphibolite and shows the normal sharp junction of the
amphibolite bands and the granodiorite gneiss at Cape Denison.

PLATE XIII.

Fig. 1. External, weathered surface of beach rock (No. 702), Cape Denison. X 4.

Fig. 2. Cut surface of ditto, showing the cavernous and detrital nature of the
rock. X 4.

Fig. 3. Thin section of the rock, showing angular sand-grains and fine calcareous
and detrital cement. X21.

Fig. 4. Part of thin section passing through the coral fragment. X21.

PLATE XIV.

Fig. 1. The junction of the rocky cliffs at Cape Denison and the ice cliffs of
Commonwealth Bay at " Land's End."

THE METAMORPHIC ROCKS OF ADELIE LAND. 8T 1 1,1. \\KLL. 217

Fig. 2. The lower part of the glacier wall at John o' Groats, where it rests on a rocky
floor below sea level, and where it is darkened by inclusions of rock detritus. The curved
lines are big conchoidal fractures which are probably connected with the presence of a
steep rock wall immediately on the right of it. Pancake ice in the foreground.

PLATE XV.

Fig. 1. The ridges are crowned with numerous small peaks and possess the character
of a miniature mountain range. Cape Denison.

Fig. 2. The narrowest valley at Cape Denison. The rough surface has been
caused by frost action.

PLATE XVI.

Fig. 1. A band of epidote biotite schist (No. 153) which has been more resistant
to weathering than the surrounding granodiorite gneiss. The reverse is usually the case
at Cape Denison.

Fig. 2. A steep wall of granodiorite gneiss at Cape Denison with a black amphibolite
band at the base.

PLATE XVII.

Fig. 1. The northern end of Lake II. which is nearly frozen over. The furrowed
and encrusting character of the lake ice is due to freezing during agitation by the winds.
Skua gulls are bathing on the edge of the water.

Fig. 2. The broad valley in which the hut was situated at Cape Denison. The
photograph was taken while low surface drift was sweeping down the valley. The
surface drift produces the haziness over portions of the rocks. The feet of the figure
on the right are invisible for the same reason.

PLATE XVIII.

Fig. 1. Highly polished rock which is characteristic of the peripheral area below
the 40ft. contour level at Cape Denison.

Fig. 2. A glacial pavement. A portion of a block about 9ft. square with well marked
parallel striae trending N. 32 E. at Cape Denison.

PLATE XIX.

Fig. 1. The " wave-sorted moraine " or " lower moraine " at Cape Denison, showing
a collection of large rounded boulders.

Fig. 2. A large boulder of silicated limestone found on the moraine at Cape Denison.
It has probably not been carried far. A parallel set of ice striae can be seen on the

218 AUSTRALASIAN ANTARCTIC EXPEDITION.

boulder trending across the schistosity. The parallelism of these striae indicate that they
were probably received before the rock was plucked out of its in situ position by the
onward travel of the glacier.

PLATE XX.

Fig. 1. View across the most easterly valley on Cape Denison. Lake V. can be
seen on the floor of the valley.

Fig. 2. View of Lake IV., looking south towards the glacier slopes. Beyond the
lake a moraine bar is visible. Still further on the upper limit of the discoloured ice is
marked by the upper limit of the white snow.

Fig. 3. Highly contorted granodiorite gneiss at Cape Denison.

Fig. 4. Bock surface disturbed by frost action. It occupies a position where the
drainage of the thaw water is retarded.

PLATE XXI.
Fig. 1. Jointing in the granodiorite gneiss, Cape Denison.

Fig. 2. View looking down the glacier slopes towards Lake IV. The moraine bank
is more prominent than in Plate XX., fig. 2. The north bank of the lake in the distance
can be seen to be thickly covered with morainic material, shown in greater detail in
Plate XXIII., fig. 1.

PLATE XXII.

Fig. 1. View showing three parallel amphibolite bands at Cape Denison. The
place is situated above the 40ft. contour level, and the surface is very rough compared
with that in Fig. 2.

Fig. 2. View illustrating the surface below the 40ft. contour level. Note the
smoothed and polished appearance. Two dark basic schlieren can be seen.

PLATE XXIII.

Fig. 1. Glacial detritus thickly strewn along the rocky bank of a small glacial lake.
Fig. 2. An erratic on the moraine at Cape Denison.

PLATE XXIV.

Fig. 1. The Madigan Nunatak from the south-east.
Fig. 2. The Madigan Nunatak from the south-west.

Fig. 3. Cape Gray from the edge of the barrier ice cliffs, looking north-west. The
gully-way which divides the island is visible. It is formed by a large metamorphosed
basic; dyke.

THE MBTAMORl'HIC ROCKS OF ADELIE LAND. 8TILLWELL. 219

Fig. 4. View from the barrier ice cliffs near Cape Gray, showing various members
of the Way Archipelago. A large amount of heavy floe ice is visible. This ice had
iiccumulated between the islands of the Way Archipelago and broken out before the
arrival of the sledge party on December 16th, 1912.

PLATE XXV.

Fig. 1. View from the eastern side of the Cape Gray Promontory, looking north-
east out of Watt Bay. Four individuals of the Way Archipelago can be seen, including
a very curious, wedge-shaped island.

Fig. 2. Crumpled gneiss at the Cape Pigeon Rocks.

Fig. 3. View from Garnet Point, looking north-east out of Watt Bay. A steeply
conical member of the Way Archipelago can be seen. Penguins are inspecting a mitten
in the foreground.

Fig. 4. The steep descent to Garnet Point. An ice ramp enabled the sledge party
to descend from the top of the barrier cliff down to the rock exposure.

PLATE XXVI.

Figs. 1 and 2. These are two views showing the large aggregates of garnet and biotite
in the garnet felspar gneiss at Garnet Point. The ice axe, in Fig. 1, is 36in. long, lO^in.
wide at the pick end, and the handle is l|in. in diameter.

Fig. 3. Aurora Peak.

Fig. 4. A close view of a narrow recrystallised basic dyke at the Cape Pigeon
Rocks.

PLATE XXVII.

Fig. 1 . The large recrystallised basic dyke cutting the garnet gneiss on the northern
part of the Cape Pigeon Rocks. An offshoot can be seen in the photograph. Specimen
No. 767 was collected from the large dyke.

Fig. 2. A recrystallised basic dyke at Cape Gray. It cuts the garnet cordierite
gneiss and a fine stringer can be seen branching out into the gneiss on the right hand side.

Fig. 3. The foliation anticline at the southern end of the Madigan Nunatak.

Fig. 4. The large recrystallised basic dyke on the southern half of the Cape Pigeon
Rocks. The view is taken from the northern part.

PLATE XXVIII.
Panorama of the northern half of the Cape Pigeon Rocks.

220 AUSTEALASIAN ANTARCTIC EXPEDITION.

PLATE XXIX.
Stillwell Island, one of the largest members of the Way Archipelago.

PLATE XXX.

Fig. 1. A panorama looking across Cape Denison. On the left and in the distance
are the rising slopes of the inland ice. The moraine is in the foreground.

Fig. 2. A panorama of the sea front looking eastward from Cape Denison. A
stretch of waterworn boulders is seen on the right, which are part of the beach deposits
which are referred to as the "lower moraines." The plateau slopes are visible to a
height of about 1,500ft.

PLATE XXXI.

Fig. 1. A panoramic view looking south from near the hut. In the distance are
the slopes of the inland ice sheet. In the foreground is the terminal moraine. Between
the rocks and the figure is a zone of ice impregnated with detritus which causes rapid
thawing on calm summer days.

Fig. 2. A panoramic view looking north towards the sea. In the middle of the
picture is Round Lake.

PLATE XXXII.
The Mackellar Islets viewed from an elevation of 800ft. on the mainland.

PLATE XXXIII.

Fig. 1. A large island of the Mackellar group, showing its planated surface.
Colonies of Adelie penguins are distributed over it, and the rocks in the foreground are
encrusted with salt.

Fig. 2. Cape Hunter, composed of phyllites with vertical cleavage planes.

PLATE XXXIV.
Locality map of Adelie Land.

PLATE XXXV.
Locality map of Cape Denison.

GENERAL INDEX.

PAGE.

Adelie Land 7, 9, 93

Adams, F. D 81, 82, 98, 114, 118, 200

alkali felspar gneiss 90

alkali felspar gneisses, group of 44, 88, 90, 136, 141

aluminium silicate gneisses, group of 44, 153, 154

aluminium oxide rocks, group of 44

amphibole gabbro schist 24

amphibole para-gabbro 24

amphibolite 9, 29, 32, 35, 36, 39, 67, 93, 102, 123, 150, 174, 180, 184, 187

amphibolite, augite 179, 180, 185

amphibolite, biotite 25, 29

amphibolite, lawsonite 25, 34

amphibolite, meaning of term 24

amphibolite, quartz 25

amphibolite. quartz biotite 35

amphibolite series, Cape Denison 10, 23, 131

amphibolite series, Cape Denison, chemical characters 41

amphibolite series, Cape Denison, classificatory position 45

amphibolite series, Cape Denison, field characters 25, 79

amphibolite series, Cape Denison, mineral composition 28

amphibolite series, Cape Denison, origin 55

amphibolite series, Cape Dcnison, petrographical characters 27

amphibolite schist 25, 36

amphibolite schist, biotite 33

amphibolite schist, lawsonite 37, 39

aplite gneiss 9, 89

assimilation 105

augit* amphibolito 179, 180, 185

Aurora Peak 7, 9, 10, 13, 138, 142, 155, 166, 190, 195, 207

Azimuth Hill 10, 26

Bahnstation 188

Bancroft 81, 98, 114, 118

Barlow, A. E 81, 98, 114, 118, 200

Barna 74

Bastin, E. 8 118, 119, 153

Bayly, P. G. W 8, 41, 131, 135, 183

Bayley, W. 8 200

Becke, F. . 108, 203

222 AUSTRALASIAN ANTARCTIC EXPEDITION.

PAGE

Becker, G. F 209

Beinn Lair 55

Belgica 93

Belgrave 192

Bendigo 56

Bickerton, F 8

biotite amphibolite 25, 29

biotite amphibolite, quartz 35

biotite amphibolite schist 33

biotite felspar gneiss 12, 61, 69, 75, 159, 207

biotite hornblende schist 12, 58, 207

biotite schist, epidote 25, 29, 31, 41, 47

Bonney, T. G 107

Brauns, D 163

Cape de la Motte 7

Cape Denison 7, 8, 9, 12, 15, 16, 21, 22, 48, 55, 71, 72, 76, 79, 84, 94, 105, 106, 114, 120, 122,

126, 128, 144, 171, 192, 206, 207

Cape Denison Physiography 15

Cape Denison amphibolite series 10, 23, 95, 123, 131, 179, 180, 188

Cape Denison amphibolite series, chemical characters 41

Cape Denison amphibolite series, classificatory position 45

Cape Denison amphibolite series, field characters 25

Cape Denison amphibolite series, mineral composition 28

Cape Denison amphibolite series, origin 55

Cape Denison amphibolite series, petrographical characters 27

Cape Gray. . 7, 9, 10, 13, 72, 94, 95, 128, 144, 146, 152, 153, 154, 164, 168, 183, 184, 186, 189, 190, 207, 208

Cape Hunter 7, 9, 10, 20, 126

Cape Pigeon Rocks 7, 9, 10, 13, 14, 28, 144, 145, 151, 154, 159, 164, 177, 183, 186, 198, 208

Cape Wrath 95

Cardiff 103

Casterton 202

Challenger 93

Chamberlin and Salisbury 18

Chapman, F 21, 22

charnockite 13, 134, 135, 136, 155, 166, 193

charnockite series i 14, 131, 134, 137, 155, 190, 192, 198

Chemnitzbiege 188

chloromelanite rocks, group of 44

chlorite rock 12, 58, 62, 63, 64, 207

Cima d'Asta 71

Clark, F. W 56

classification of crystalline schists 43

Clough, C. T 55, 94

Coker, E. G 81

Cole, G. A. J 74, 82, 93, 105, 176, 202

Commonwealth Bay 7, 15, 18, 122, 144

THE METAMORPHIC ROCKS OF ADELIE LAND.-STILLWELL. 223

PAGE.

composite gneiss 73

consolidated beach sand 21

cordierite gneiss, garnet 9, 10, 146, 152, 153, 164, 168

Coverack 115, 116

Crook, T 23, 108, 199

Cross, Whitman 198

Crousa Downs 115

crystalline schist 23

crystalloblastic order 40, 127, 164, 185

cyanite biotite gneiss 9, 149, 152, 153, 164

Dana, E. 8 163

Day, A. L 209

Daylesford 80

Depot Island 96

Desch, C. H 71, 204

diablastic structure 25

diorite gneiss 25

eclogites and amphibolites, group of 44

Elsden, J. V 71

epidosite 12, 58, 62, 63, 64, 115, 207

epidot biotite schist 25, 29, 31, 41, 47

Fannich Mountains 97

felspar gneiss, biotite 12, 61, 69, 75, 159, 207

felspar gneiss, garnet 9, 150, 164

felspar gneiss, hypersthene 155, 157, 164

Fcnner, C. N 103

Fermor, L. L 14, 196, 198, 199

Ferrar, H. T 93

Flett, J. 8 48, 55, 64, 106

fluxion gneiss 110

foliation anticline 137

force of crystallisation 207

frost action 17

Frosterus 102

Gairloch 55, 95, 101

Galway, County 74

Garbh Allt 43

garnet amphibolitc 13, 177, 183, 185, 187

garnet cordierite gneiss 9, 10, 14, 146, 152, 153, 164, 168

garnet felspar gneiss 9, 1 50, 164

garnet hypersthene alkali felspar gneiss 9, 141

224 AUSTEALASIAN ANTARCTIC EXPEDITION.

PAGE.

garnet plagioclase pyroxene gneiss 176, 187

Garnet Point 7, 9, 14, 144, 145, 146, 154, 164, 166, 181, 183, 184, 187, 207

Geikie, A 110

glacial lakes 16

glacial plucking 16

glacial valleys 16

glacier action 15

Glencaloie Lodge 43

Glenelg 97

gneiss 23

gneissic type of meta-xenolith 51

granite gneiss 9, 124

granodiorite gneiss 9, 18, 25, 84, 87, 125, 141, 192

Great Mackellar Island 8, 9, 122

Greenly, E 71

Grubenmann, U 10, 23, 24, 40, 43, 47, 48, 94, 97, 106, 108, 109, 130, 148, 185, 175,

184, 197, 201, 203, 204, 206

Gruinard 95, 96

Gunn, W 94

Haliburton 81, 98, 114, 118

Hall, A. G 41, 86, 131, 152

Barker, A 156, 197

Hart, T. G 80

Herman, H

Hesket 88

Hezner, Laura 24, 163, 170, 197

Highlands, New Jersey 103

Highlands, N.W. Scotland 55, 94

Hill, J. B 55, 106

Hinxman, L. W 94

Holland, T. S 134, 163, 188, 189, 192, 193, 204

hornblende gneiss 12, 69

hornblende plagioclase pyroxene gneiss 9, 139, 171, 173, 179, 181, 183, 185, 187

hornblende schist, biotite 12, 58, 207

Hurley, F 8

hypersthene alkali felspar gneiss 9, 132, 133, 138, 139, 158, 164

hypersthene felspar gneiss 155, 157, 164

Iddings, J. P 64

Ijri valley 189

Ilfracombe 21

Indian charnockite series 14, 131, 134, 137, 155, 190, 192, 193, 198

infraplutonic zone 14, 196

injection banding 110

injection foliation 110

iron oxide rocks, group of 44

THE METAMORPH1C ROCKS OF ADELIE LAND.-BTILLWKLL. 225

PAGE.

jadeite rocks, group of 44

Johnston and Niggli 23, 108, 109, 203

Joly, J 147

Kasshabop Lake 101

Kennack gneisses 106, 111, 116, 118

King Edward VII. Land 93

kodurite series 196, 198

kristallinen schiefer 23

Kukri Hills 93

Kylesku 81, 96

Lacroix 188

Lahee, F. H 40

lakes, glacial 16

Lande wednack 114, 1 18

Laseron, C. F 8, 144

Lawson, A. C 86

lawsonite amphibolite schist 37, 39

Lehmann, J 177, 188

Leith, C. K 40, 109, 119, 120, 200, 201

Lewisian gneiss 94

lime silicate rocks, group of 44

lime soda felspar gneisses, group of 44

Lizard 55, 64, 106, 117

Lochalsh 97

Loch Broom 81

Loch Carron 95

Loch Laxford 95, 96

Loch Maree 55, 95, 97, 101

Loewinson- Leasing 24

" lower " moraines 18

Macedon 86

Mackellar Islets 7, 16, 18, 21, 122, 146

Madigan, C. T 8, 138

Madigan Nunatak 7, 9, 10, 13, 17, 18, 128, 138, 141, 154, 155, 157, 164, 165, 166, 189, 190, 193, 207

magnesium silicate schists, group of 44, 59, 64

Man of War gneisses 114

marmorites, group of 44

Mawson, Sir Douglas 8, 16, 20, 21, 93, 145, 171, 202

M-a.1. W. J 40, 119, 120, 200. 201

Meall Mheinnidh 55

Menominee 200

Mertz glacier 8, 138

Serie* A, VoL m.. Part 1 P

226 AUSTEALAS1AN ANTARCTIC EXPEDITION.

PAGE.

metamorphic differentiation 12, 58, 62, 76, 91, 94, 115, 119, 200

metamorphic diffusion 12, 71, 72, 94, 114, 120, 121, 133, 150, 200

metamorphosed dyke series of Cape Denison 10, 23, 95, 188

metamorphosed dyke series of Cape Denison, chemical characters 41

metamorphosed dyke series of Cape Denison, field characters 25, 79

metamorphosed dyke series of Cape Denison, mineral composition 28

metamorphosed dyke series of Cape Denison, origin 55

metamorphosed dyke series of Cape Denison, petrographical characters 27

metamorphosed dyke series of Cape Gray Promontory 13, 72, 168, 183

metamorphic rock, meaning of term 23

metamorphosed xenoliths (meta-xenoliths) 12, 48, 56

metamorphosed xenoliths (meta-xenoliths), gneissic type 51

metamorphosed xenoliths (meta-xenoliths), origin of 53

metamorphosed xenoliths (meta-xenoliths), saussuritic type 48

metamorphosed xenoliths (meta-xenoliths), significance of 54

Methuen 101

micrographic structure 25

micropegmatitic structure 25

Mid Argyll 97

Middlemiss, C. S 199

Miliolina 22

Mohsdorf 188

Mookaie Hill 202

moraines 19

morainic mud 22

New Jersey Highlands 103

Niggli, Johnston and 23, 108, 109, 203

Nilgiris 193

nomenclature 23

norite 13, 24, 131, 190, 193

North- West Highlands of Scotland 55, 94

Ormsby ' 102

Otz Valley 170

Ozann group values 44, 45, 59, 63, 87, 131, 135, 140, 152, 183

Ozann triangular projection 45, 46, 60, 88, 132, 153, 185

Pallavaram 193

para-diorite 24

para-granodiorite 25

Parasnath 189

Peach, B. N 94

pearlite 204

Pelikan . 163

THE METAMORPHIC ROCKS OF ADELIE LAND 8T1LL\VELL. 227

PAGE.

188

9, 127

physiography of Cape Denison 15

Pine Lake 103

plagioclase gneisses, group of . 44, 88

plagiodase pyroxene gneiss 24, 128, 132, 168, 170, 172, 183, 185, 187

plagioclase pyroxene gneiss, garnet 176, 187

plagioclase pyroxene gneiss, hornblende 9, 139, 171, 173, 179, 181, 183, 185, 187

plucking, glacial 18

Point Sleat 95

Pol Cornick 114

Powers, S 55

projection values, Ozann 44, 46, 59, 63, 87, 131, 135, 140, 152, 183

pyroxene alkali felspar gneiss 24

pyroxene gneiss 24, 132, 193

pyroxene gneiss, plagioclase 24, 128, 132, 168, 170, 172, 183, 185, 187

pyroxene granulite 13, 24, 93, 128, 132, 193

quartz amphibolite 25

quartz biotite amphibolite 35, 39

quartzite rocks, group of 44

Queen Mary Land 93

Rainy Lake 87

Richards, H. C 163

Riecke's principle 208

Roberts- Austen 71

rock flowage 108

Rosenbusch 24

Roeiwal method 10, 28

Ross 95

Boss Sea 20, 93

Rudha Caol 97

Rum, Isle of 156

Salem 189

Salisbury, Chamberlin and 18

saussurite 48

saussuritic type of meta-xenolith 48

Saxon pyroxene granulite 13, 188

ichiefer, kristallinen 23

schist 23

Scotland, North- West Highlands . 55, 94

Scott 20, 93

Sederholm J. J 23

serpentine 55, 117

228 AUSTRALASIAN ANTARCTIC EXPEDITION.

PAGE.

Shackleton 20, 93

shore ice 18

Skeats, E. W 8, 192

Skye 95, 97

solid difiusion 71, 203, 204

solution, transference by 203

sorbite 204

South Victoria Land 93, 96, 189

Stead, J. E 208, 209

Stillwell Island 7, 9, 10, 13, 14, 28, 144, 145, 151, 154, 155, 159, 164, 166,

170, 171, 184, 186, 189, 198, 206, 208

Strath Bromm 97

St. Thomas Mount 131, 135, 190, 193

subrotunda, Miliolina 22

Sutherland 95

Tagil River 24

Teall, J. J. H 55, 94, 96

thaw water 18, 19

Traboe schists 114, 118

troostite 204

Traeman, J. D 119

valleys, glacial 16, 18

origin 18

Van Hise, C. R 23, 40, 57, 83, 108, 110, 112, 163

Victoria 80, 192, 202

Walkom, A. B 93

Warner, J. H 119

Washington, H. S 43, 131, 135, 136, 139, 155, 190

water action 17

Waterloo 119

Watson, J. C 59, 63, 86, 135, 140, 183

Watt Bay 7, 144, 145

Way Archipelago 7, 144, 145

Weinschenk 48, 106, 108, 193

West Antarctica 93

Williams, G. H 200

wind action 17, 18, 19

Wittgensdorf ; 188

Wright, F. E 109

xenolith 27, 48, 55

metamorphosed 12, 48, 56

Zirkel 24

NUMBER INDEX OF ROCK SPECIMENS.

BOCK

NUMB KB

NAME or ROCK.

LOCALITY.

4 biotitc hornblende schist

5 amphibolite

9 amphibolite

10 biotite felspar gneiss

ICU aplite gneiss

1 1 granodiorito gneiss

13 hornblende gneiss

60 alkali felspar gneiss

143 j biotite amphibolite

143A granodiorite gneiss

144 biotite felspar gneiss

145 ' biotite felspar gneiss

146 biotite felspar gneiss

150 aplite gneiss

153 epidote biotite schist

154 alkali felspar gneiss

160 i granitic gneiss with band of amphibolite

212 augite amphibolite

372 junction of amphibolite and grano liorite gneiss

411 biotite amphibolite

412 . biotite amphibolite schist

415 epidosite

424 i biotite felspar gneiss

628 meta-xenoliths in amphibolite

629 amphibolite

630 quartz biotite amphibolite

630A biotite felspar gneiss

631 amphibolite

634 amphibolite schist

634A lawsonite amphibolite schist

635 lawsonite amphibolite schist

637 amphibolic

640 chlorite schist

720 lawsonite mica amphibolite

754 hypersthene alkali felspar gneiss

756 alkali felspar gneiss

757 hypersthene alkali felspar gneiss

758 hypersthene alkali felspar gneiss

759 hornblende plagioclase pyroxene gneiss

Cape Denison

from moraine, Cape Denison..
from moraine, Cape Denison..
from moraine, Cape Denison..

Cape Denison

Aurora Peak

Aurora Peak .

90
30,41

48
35,41

139
142
142
142
138

230

AUSTRALASIAN ANTARCTIC EXPEDITION.

ROCK
NUMBER.

NAME or ROCK.

LOCALITY.

PAGE.

766

hornblende plagioclase pyroxene gneiss . . .

Cape Gray ... .

170

767

garnet plagioclase pyroxene gneiss

Cape Pigeon Rocks

177, 183

769

junction ol amphibolite and cyanite biotite gneiss

Garnet Point . .

150

770

biotite gneiss

Garnet Point

149

771

amphibolite ...

Cape Pigeon Rocks

179

772

cvanite biotite gneiss

Garnet Point

147, 152

773

plagioclase pyroxene gneiss

Cape Gray ...

169. 183

777

garnet felspar gneiss

Garnet Point

150

781

junction of amphibolite and cyanite biotite gneiss

Garnet Point

150

782

augite amphibolite

Cape Pigeon Rocks

177

784

garnet cordierite gneiss

Cape Grav . .

146 152

785

hypcrsthene biotite felspar gneiss . .

Cape Pigeon Rocks

159

786

augite amphibolite

Cape Pigeon Rocks

180

794

plagioclase pyroxene gneiss

Madigan Nunatak .

128

797

hypersthene alkali felspar gneiss

Madigan Nunatak .

133

799

garnet amphibolite

Garnet Point .

181 183

911

phyllite . .

Cape Hunter ....

127

917

garnet felspar gneiss

Stillwell Island

151

928

sphene biotite felspar gneiss

Cape Denison

935

garnet plagioclase pyroxene gneiss

Stillwell Island

173

939

garnet felspar gneiss .

Stillwell Island

151

942

hornblende plagioclase pyroxene gneiss

Stillwell Island .

172

947

hypersthene alkali felspar gneiss

Stillwell Island . .

151

949

hypersthene felspar gneiss .

Stillwell Island

155

951

plagioclase pyroxene gneiss .

Stillwell Island

171

952

amphibolite

Stillwell Island

173

953

garnet amphibolite

Stillwell Island

173 183

979

hypersthene felspar gneiss ...

Stillwell Island

176

981

albite amphibolite

Great Mackellar Island

123

982

granitic gneiss

Great Mackellar Island

124

983

granite gneiss

Great Mackellar Island

152

984

granitic gneiss

Great Mackellar Island

125

PLATE I.

Fig. l.

SlUlm-U.

Fig. 2.

.s-liV/irr/7.

Fig. 3.

Fig. 4.

.I,//,. '

Fig. 5

Fig. 6.

: .''.

PLATE II.

J i *t

Fig. 1.

.;*.

Fig. 2.

Fig. 3

StUlwrU.

Fig. 6.

PLATE III.

Fig. 2.

Fig. 3.

Fig. 4.

Fig. 5.

Fig. 6.

Fig. 1.

Fig. 2.

UlillifeU.

Fig. 3.

tilillurll.

Fig. 4.

StiUlcrtl.

Fig. -.-

,-7.7/iB-ff.

Fig. 6.

I'l.ATK V.

Fig. 1.

.ftillirfll.

Fig. 2.

SHOwB.

Fig. 4.

v D '/.

Fig. 5.

Fig. 6.

PLATE VI.

Fig. l.

Stillurll.

;, *-

Fig. 3.

Fig. 4.

MBMB.

Fig. 5.

Fig. 6.

I'l.ATK VII.

Fig. 1.

Fin. a

Fig. 2.

Fig. 4.

Fig. 5.

PLATE VIII.

Fig. l.

Fig. 3

Fig. 2.

StUlmll.

Fig. 5.

lT ft *' ' '

iii. u.

PIRATE IX.

Fig. 1.

Fig. 3.

Fig. 4.

Still tcr II.

1-51 T-

Fig. 5.

Fig. 6.

Stillmll.

PLATE X.

Fig. l.

Slillmll.

Fig. 2.

Stilluvll.

Fig. 3.

HtillHYll.

Fig. 4.

Stillmll.

Fig. 5.

Fig. 6.

I'l.ATK XI.

Fig. 1.

SEV

Fig. 2.

Slillvrll.

Fig. 3.

ShllHYll.

Fig. 4.

.SI .//XT//.

Fig. 5.

Fig. 6.

Stillitrll.

PLATE XII.

Fig. 1.

fitillirrll.

Fig. -2.

Slillin-ll.

Fig. 3.

Stillicfll.

Fig. 4.

.Sli//uvU.

Fig. 5.

Fig- G.

filUltrrll.

I'LATK XIII.

PLATE XIV.

Fig. 1

PLATE XV.

Fig. l.

// H rley.

Fig. 2.

Hurley.

PLATE XVI.

Fig. l.

Hurley,

Fig. 2.

Hurley.

PLATE XVII

Fig. l.

Hurlty.

1'LATK XVIII.

Fig. 1.

Hurley.

Fig. 2

i. ' i

PLATE XIX.

Fig. 1.

Hurley.

PLATE XX.

PLATE XXI.

Fig. 1.

Hurley.

Fig. 2.

limit]/.

PLATE XXII.

Fig. l.

Hurtev.

Fig. 2.

PLATE XXIII.

Fig. 1.

Hurls if.

Fig. 2.

PLATE XXIV.

PLATE XXV.

PLATE XXVI.

Fig. 1

Fig. 2.

Fig. 3.

U / I

Fig. 4.

; M

(I I <

I'LATK XXVII.

PLATE XXVIII.

PLATK XXIX.

PLATE XXX.

I'l.ATK \\.\l.

PLATE XXXII.

PL ATK XXXIII.

Fig. 1.

Hurley.

Fig. 2.

I'l.ATE XXXIV.

PLATE XXXV.

AUSTRALASIAN ANTARCTIC EXPEDITION

s . 1911-14. ;

UNDER THE LEADERSHIP OP SIR DOUGLAS AAWSON, KT.. D.Sc, B.C.

SCIENTIFIC REPORTS.

SERIES A.

VOL. III.

GEOLOGY.

PART

THE METANIORPHIC LIMESTONES

COMMONWEALTH BAY, ADELIE LAND.

C. E. TILLEY, B.Sc.

WITH TWO PLATES.

PRICE: ONE SHILLING AND SIXPENCE.

Printed bv Alfred Jin Kent, Action Gownnwnt Printtr, PblUip-tlrMt .Sydney. 1933-

JED JULY, 1923.

SERIES A REPORTS.

HON. EDITOR: PROF. SIR DOUGLAS MAWSON, Ex., D.Sc., B.E., University of Adelaide,

VOL. PRICE.

S. d.

I. GEOGRAPHY AND PHYSIOGRAPHY. (In preparation.)

II. OCEANOGRAPHY.

PART 1. SEA-FLOOR DEPOSITS FROM SOUNDINGS

By FREDERICK CHAPMAN, Ass. Linn. Soc. (Lend.), F.R M.S., &c., National Museum, Melb. 060

III GEOLOGY. (Adelie Land and King George Land.)

PART 1. THE METAMORPHIC ROCKS OF ADELIE LAND

By F. L. STILLWELL, D.Sc., Aust. Antarc. Exped. Staff 220

2. THE METAMORPHIC LIMESTONES OF COMMONWEALTH BAY, ADELIE

LAND. By C. E. TILLEY, B.Sc 016

3. THE DOLERITES OF KING GEORGE LAND AND ADELIE LAND.

By W. R. BROWNE, D.Sc., Lecturer, Geological Department, Sydney University

4. AMPHIBOLITES AND RELATED ROCKS FROM THE MORAINES, CAPE
DENISON, ADELIE LAND.

By F. L. STTLLWELL, D.Sc., Aust. Antarc. Exped. Staff

IV. GEOLOGY. (Will deal principally with Queen Mary Land.) (In preparation.)
PART 1. THE ADELIE LAND METEORITE.

By P. G. W. BAYLY, F.I.C., and F. L. STILLWELL, D.Sc. 1 C

V. GEOLOGY (Macquarie Island). (In preparation.)
VI. GLACIOLOGY. (In preparation.)

AUSTRALASIAN ANTARCTIC EXPEDITION

1911-14.

UNDER THE LEADERSHIP OP SIR DOUGLAS AAWSON. KT.. D.Sc., B.E

SCIENTIFIC REPORTS.

SERIES A.

VOL. Ill

GEOLOGY.

PART II:

THE METAMORPHIC LIMESTONES

COMMONWEALTH BAY, ADELIE LAND.

C E. TILLEY. B.Sc.

WITH TWO PLATES.

PRICE: ONE SHILLING AND SIXPENCE.

Printer! by Alfred ]*met Kent, Acthw Govrnmel Pristtr. PhilUp-tit ,Sr<l"

ISSUED JULY. 1923.

THE METAMORPHIC LIMESTONES OF COMMON-
WEALTH BAY, ADELIE LAND.

('. E. TILLEY, B.Sc.

WITH TWO PLATES.

PA1IK.

CONTENTS.

I. Introduction 231

1 1 . Petrography :

1. Forsterite- Marbles 232

2. Tremolite-Marhles 234

3. Diopside-Tremolite- Marbles 235

4. Pyroxene-Garnet-Marbles 230

."). Pvroxene-Epidote-Marbles 237

<i. Epidote-Marbles 238

7. Carbonate-free Calc-silicate Rocks 239

III. Review of the Metamorphic Processes involved 240

IV. Paragenesis 242

V. Explanation of Plates ... --Ml

I. iXTHonrrnnx.

Tin; group of rooks which are described below represent a collection of erratics made
from tin- moraines at Cape Denison, Adelie Land. With reference to their occurrence,
! ><>uglas Mawson states (*) :

"None of them were met in situ; but tin- evidence of their occurrence in
ciTtJiin moraines (their distribution), and the evidence of the rock types which do
actually outcrop in situ on that coast of Adelie Land, point to the calc-silicate
M-ries being located under the ice Cap to the south-west of Cape Denison, probably
in the depression at the head of Commonwealth Bay, between C'ape Denison ;uul
Cape Hunter. This is rendered more likely, since this calc-silieat series would

1 Penonil communication.

232 AUSTRALASIAN ANTARCTIC EXPEDITION.

erode more rapidly than the tougher schists and gneisses which appear in the
visible outcrops thereabouts. There is no doubt that the rocks of this series
represent phases of alteration of the same sedimentary series. The alteration has
been effected, it would appear, by the intrusions of extensive granitic magmas
now appearing as gneiss typically developed at Cape Denison and the Mackellar
Islets."

The carbonate sediments from which these rocks have been derived were
characterised to a greater or less degree by the presence of detrital material, which in.
the process of metamorphism has reacted with the carbonate minerals. This group of
rocks, however, with two exceptions, still possesses a content of free carbonate mineral.
Its quantitative amount is dependent in some cases on the degree of metamorphism of
the rocks concerned, and in others on the quantity of foreign material present in the
original sediment, capable of chemical reaction with calcite or dolomite. Amongst
these rocks, there is no example which suggests any extensive addition of material
from magmatic sources, other than purely volatile constituents.

According to their mineralogic content, these rocks may be divided into the
following classes :

(1) Forsterite-Marbles.

(2) Tremolite-Marbles.

(3) Diopside-Tremolite-Marbles.

(4) Pyroxene-Garnet-Marbles.

(5) Pyroxene-Epidote-Marbles.

(6) Epidote-Marbles.

(7) Carbonate-free Calc-silicate Rocks.

The description of these various classes can now proceed seriatim.

II. PETROGRAPHY.

(1) FORSTERITE-MARBLES.

The rocks characterised by the presence of magnesian olivine, comprise the
following : Nos. 135, 137, 307, 318, 392, 395, 402, 653, 992, 993, 994.

As a class, they are medium-grained, white to grey rocks. The majority are
characterised by the presence of yellowish-green pseudomorphs of serpentine after
forsterite, and these project on weathered surfaces. In No. 395, serpentine veins the
rock in two parallel bands, in the centres of which narrow venules of chrysotile asbestos
are developed. Nos. 135 and 137 are characterised by the presence of orange-coloured
crystals with vitreous lustre, and these on examination prove to be chondrodite. In
No. 135, this mineral is largely developed along a plane surface. Flakes of light-
coloured mica are sparingly distributed in a number of these rocks.

MKTA.MolM'lllr I.IMKsToNKs TIU.KY

It will In- sufficient to describe the nature <if the minerals uf these rocks as a
whole, points of particular interest shown l>y any of them being refererd to in the
course of this description.

The constituent mineral are dolomitf. cnlcitf. forxh-rite. <-/t<nlr<nlit<-.
hornblende, dm/mnlr. pUogoptte, .-//*'/////'. and

The minerals which characterise the class are dolomite, calcite and forsterite.
the remainder forming a subordinate group, which, while often abundant, yet rank as
accessory constituents to the class.

Dolom ite. While the ordinary method of differentiation between this mineral
and calcite is the staining method of Lemberg, yet the criteria which have been found
to operate in metamorphosed dolomites of other regions are often of use here. These
iiii hide the different types of twinnning shown by. dolomite and calcite respectively,
and the degree of turbidity.

As earlier noted (') the dolomite twinning on the 0221 plane is sufficiently
distinctive in favourable sections, and the turbid character of the calcite also affords a

further means of discrimination. This poly synthetic twinning in dolomite suggests
that a secondary twinning along a glide plane is involved.

Calcite is quite frequently twinned in these rocks and the turbidity often observed
is due to the presence of minute inclusions which are probably carbonaceous.

Forsterite. The usual habit is in rounded grains or prismatic crystals, in which
the trace of the 010 cleavage parallel to the elongation is imperfectly developed. In
rocks Nos. 137 and 318 the olivine is almost completely free from decomposition, but
in the remainder all stages of serpentine development are revealed. Usually this is a
colourless type, but in some cases it appears of a pale yellowish green tint. In No. 992
the serpentine is accompanied by granules of secondary magnetite. Where developed
as an inclusion in the serpentine, the carbonate is usually calcite.

Chondrodite. Orange yellow crystals of a member of the humite group ol
minerals are developed in rocks Nos. 135 and 137. In No. 135 they are partly
arranged along a plane through the rock, as if indicating the passage of fluorine bearing
vapours along a bedding plane or other .surface < I interruption.

In thin section the mineral is noticeably pleochroic, varying from golden
yellow (X). to colourless (V, Z). Twinning is faintly developed in some sections.
The extinction measured from the 001 cleavage, the plane of lamellar twinning,
corresponds to that of chondrodite in the section available, reading 20 degrees.
Clinohumite is thus excluded.

There is no trace of intergrowth with forsterite, and serpent inisation proceeds in
the same manner as in normal olivine.

1 Oe.il. Mag., Vol. Ivii. 1020. p. 453.

234 AUSTRALASIAN ANTARCTIC EXPEDITION.

Spinel. The spinel of these rocks is usually the colourless magnesia spinel
developed in rounded grains or more rarely subidioblastic with octahedral outline. It
occurs isolated in the carbonate minerals or may be intimately associated with
forsterite, occasionally enclosed in the outline of the latter.

In No. 992, the spinel has the green colour of the pleonaste type, and is there
associated with magnetite which is developed peripherally and along cracks. It is
usually quite free from alteration, but in some examples has developed a peripheral
ring of colourless serpentine.

Hornblende. A colourless amphibole is sometimes abundantly developed in
these rocks. This is especially the case in rocks No. 318. In this rock, the horn-
blende is developed as subidioblastic crystals enclosing grains of forsterite, and also
as narrow corona-like rims to the same mineral. The mineral shows the typical
amphibole cleavages, and cross-sections show the emergence of an optic axis, with the
optic axial plane bisecting the obtuse intercleavage angle. These grains are optically
positive, and there can be no doubt that the mineral is edenite, and not the normal
tremolite. Moreover the extinction angle exceeds the value for this latter type.
Tremolite is, however, not absent from these rocks, but is sparingly developed.

Diopside is present as an accessory constitutent in clear colourless grains, with
prismatic habit. In No. 395 it is present with tremolite fringing a band of serpentine.
There is no definite evidence however to suggest that the serpentine has a pyroxenic
derivation. The bands in this rock, with their accompanying chrysotile venules are
essentially of forsterite derivation.

Pfdogopite. The colourless mica which is a frequent member of these rocks is
a type with very small optic axial angle, approaching uniaxiality, being the magnesia-
rich phlogopite variety common to metamorphosed dolomites.

(2) TREMOLITE-MARBLES.

This class includes the following rocks : Nos. 306, 306a, 355, 406, 673, and 707.
The distinct habit of crystalline schists is given to these rocks by the abundant develop-
ment of fibrous amphibole. They are grey to green-grey rocks in which the amphibole
is present in light-green fibres often with a parallel orientation. This, however, is not
a constant feature, the porphyroblasts of tremolite being developed as in No. 406 in
diverse orientations.

A radiate arrangement appears in the lighter-coloured rock No. 707. On
weathered faces the more resistant amphiboles usually project from the general surface.

The constituent minerals are dolomite, calcite, tremolite, pUogopite, biotite,
(chlorite), magnetite, and apatite.

Dolomite is again revealed by the characteristic type of twinning, and both it
and the calcite are universally twinned on a polysynthetic scale.

MKTA.Moltl'lllr U.MKSToNKS TIU.KY.

ilr is a general constituent in idioblastic prisms, with positive elongaticm.
negative sign and large optic axial angle. It is colourless in thin section. l)iit macro-
scopirally may develop a pale-green tint indicative of tlie presence of (lie actinolite
molecule.

/'///(/"/*//< is cohtnrless to pale yellow -green, and slightly plrochroir ;md always
with a very small optic axial angle.

Brown hiotilf is |. resent in No. ;}.3,>, and shows peripheral alteration to a light-green
chlorite. The phlogopite molecule is probably present in solid solution in its constitution.
It is associated with highly refracting and birefringent prisms of rutile. which are of
secondary origin, and result from its degradation. The remaining constituents of
these rocks call for no special remark.

(3) DlorslDK-TliKMol.lTK HoCKS.

Two locks are strictly included within this class. Nos. :><:> and 657. No. <i~. I
may be considered here for while tremolite is absent, calcite and diopside are the chief
oonstituentB.

No. :{n:{ is a coarse-grained grey rock with large crystals of diopside measuring
up to 1 inch or more in length. It is associated with light-green fibrous amphibole.
I 'ink-coloured calcite is sparsely distributed, and quartz can also be recognised macro-
scopically.

NO. (i.")7 is a finer grained rock in which on weathered surfaces, the new-formed
silicates project. As such, can be recognised biotite. mica and green diopside.

No. <).">! is a flesh-coloured crystalline marble containing porphyroblasts of light-
green diopside.

The constituent minerals of these rocks are m/r/'/r. ilio/isidt'. lirnntlil,-. actinnlitc.
luntiti'. i>ltH/iiH-ltiNr. Ilium!,'. \i n-<>,>. itfHiiiic. and secondary white mint. Dolomite has not
been recognised in thin sections by the 1 winning, nor in those sections examined by
staining. In No. 651 all carbonate can be decomposed in dilute hydrochloric acid.

The colourless />i/r<>irtn' <>| No. :{o:{ shows a lamellar twinning, and parting planes
parallel to the 001 and 100 faces. The extinction angle / A c is 41. In No. 657 the
pvroxene is a colon rles.-, to pale green type. The colourless fri'imilitc of No. :5<>:! i>
fibrous parallel to the c axis and is wedged between larger crystals of diopside. The
amphibole of No. 657 is distinctly coloured and pleochroic in light-green tints. Tremolite
in this KM k is also a constituent of pseudomorphous assemblages of white mica and
xoisite. which in some cases can be shown to take the place of plagioclase felspar.

I'luf/iiii-lns, is developed in rounded grains, sometime^ .\itli twinning lamella'
after the albite and pericline laws. The more b;isic types are the less stable, giving
rise to pseiidomorphs in which mica and x.oisite are important constituents. There
are rounded grains whose refractive index approximates to that of canada-balsam.
and appear to be near oligocla>e in composition.

236 AUSTEALAS1AN ANTARCTIC EXPEDITION.

Biotite is present in Hakes showing a brown to pale-yellow pleochroism.

A crystal of a mineral showing parallel intergrowth of several individuals with
good crystal outline, is developed in this rock. It is pleochroic in bluish-green to yellow-
green with brown tints, is optically negative, and an optic, axial plane perpendicular
to the elongation, the sign of which is negative. The pleochroic tints are those of soda
amphiboles, and the extinction is oblique (14 degrees). The position of the optic axial
plane cannot be the symmetry plane, however, if prismatic habit obtains. The only
metamorphic soda amphibole in which the optic plane is perpendicular to the 010 face
is crossite. In the absence of further sections, no definite determination is possible.

(4) PYROXENE-GARNET-MARBLES.

Under this class come the rocks Nos. 658, 730, and 849. They are essentially
types which would be grouped under the term " Kalksilikatgneise " by German and
Scandinavian petrographers.

No. 658 is a dark-coloured banded rock, consisting of layers of pink calcite in
which silicate minerals are sparingly distributed, alternating with bands rich in greenish-
black pyroxene, white felspar, and thin layers of brown garnet.

No. 730 is also a banded calc-silicate-gneiss, showing layers of flesh-coloured
calcite containing greenish-black pyroxene, and porphyroblasts of garnet intimately
associated with epidote. Such a band is separated from finer-grained bands rich in
felspar and pyroxene, by narrow layers of pyroxene and epidote.

No. 849 shows large grains of dark-brown garnet, pyroxene, and yellowish-
green epidote in addition to calcite and colourless silicates.

These banded rocks with varying mineralogies 1 composition attest the changing
composition of the original carbonate sediment during its deposition.

The constituent minerals of these rocks are calcite, pyroxene, garnet, epiflote,
scapolite, microcline, titanite, actinolite, quartz, apatite, magnetite and plof/iodase.

The pyroxene of No. 658 is green in colour and noticeably pleochroic : X ;
yellow-green, Y and Z = sea-green. The value of the extinction Z A c > 48 degrees,
and this figure when considered with the colour of the mineral in thin section indicates
that the hedenbergite molecule must dominate, but that a content of sesquioxides is
also present. The pyroxenes of Nos. 730 and 849 are also green in thin section, and
have a high extinction-angle corresponding to a content of hedenbergite. The garnet
of these rocks has a clove-brown colour very similar, to that of the associated titanite.
It is not idioblastic, and is often moulded on the green pyroxene. It is always isotropic.
In No. 658 it is associated with green pyroxene, and may enclose the other minerals.
In No. 730 it usually forms a core to the epidote-garnet aggregates, the central garnet
being enclosed by a narrow corona of pleochroic epidote,

MKTA.MiMM'im U\IKsToM-> MUM 1'HT

From il> associations tin- -:arnct \\ould I.e expected t. consist largely i.t the
Milantc and andradite molecule.-. Tin- rcfringence of the garnet us measured
Irom lia<_'meni- in No. s! is somewhat greater than 1-78.

A'/)-W<.// i^ xenoblastic . In Nd. 7:10 it is essentially a corona mineral, ami i>
found as sin-li BUITOUnd ing garnet, pyroxene. scapolite plagioclasc. and calcite. Arouiui
{iirnet it is often closely associated with quart/ in vermicular Intel-growth. In these
Intel-growths caleite is often included, and they correspond to the svmplcktitic structures
-'derholm. 1 The epidote shows a well-developed basal cleavage, and is plc<thi<>n-
in yellow-green tints.

I'niaxial grains of scapolite of negative sign and elongation is an abundant
tituent of No. 658. A typical pilaster texture is often given by this constituent.

The liirefringence approximates 0-03. A secondary white mica with positive elongation
common alteration product. In No. 730 scapolite is absent and plagioclase take-,

its place. Rock No. 849 shows pseudomorphs after (?) scapolite, in which white mica

and /.oisite are the prime constituents. The birefringence of the unaltered mineral

of these rocks indicates that the meionite member is predominant.

M it-rtH-liiif grains with noticeable cross twinning are present in association with
pyroxene, and give a pflaster texture to bands which are rich in this felspar.

In No. 730 plofjioclase is mueh decomposed, yielding aggregates of white mica.
These grains are often surrounded by a narrow rim of epidote. Some iw id plagioclasr
oligoclase is interspersed among potassic felspar grains.

(Jitiirt- is a variable constituent, but is present in all three rocks. In No. 068
a number of vermicular intergrowths between scapolite and quart/ are revealed (cf.
symplektites of Sederholm).

.\riiiitiliii- is developing at the expense of pyroxene in No. S4!i.

Tiliimt, i, a constant member of the>e rocks. It may be associated with garnet,
but there is never any indication that it is secondarily derived from this mineral.

(.->) I'YKOXK.M: Ki'ii)<iTK-M.\Kni

Kock.- NOB. :5S. :HO. M\. W(\. and HUH are included within this da>s. They .ue
medium-grained rocks, rich in caleite (with the exception of No. llliS). On weathered

'.ices, pyroxene and epidote project, the latter with the ch.irac tei isti. vellow-j
colouration. The epidote may be developed in narrow bands through the rock.
\b-taiiM.rphic tel>par and quart/ likewise project from the general >mface of the rock.

The constituent minerals are m/r/V . / ,/,(!. *\,,<e. itcapnhtf.

l>l<i | ijiuirt-.. ami in-1 ami itf .

The !,,,,, I these mirks is of a hedenbergitic tyj.e in > and H\(t. with

ie pleochroism, a- in the rocks of class (4). In NV I 1'^ Uu pyrOMM in thin
H less strongly coloured, but the extinction / r is not less than 17 to 48,

1 Sederholm Bull. Conim. Ceol. Finl \ ^. I'JIfl. p. 'H.

238 AUSTRALASIAN ANTARCTIC EXPEDITION.

so that the presence of sesquioxides is indicated. As a whole the pyroxenes may be
regarded as hedenbergitic rather than diopsidic. A parting parallel to the 001 face
is often developed (cf. No. 1168).

Epidote is a very characteristic mineral of these rocks, present in irregular grains'
and again as narrow rims to other minerals. A basal cleavage is usually strongly
manifest. The extinction in sections parallel to the optic axial plane is approximately
30 from the 001 cleavage. The optic sign is negative, with the optic plane perpendicular
to the cleavage. It is strongly pleochroic in yellowish -green tints, and a high content
of Fe 2 3 is indicated by a high birefringence of 0-04 (as in No. 38). Intergrowths
with quartz are not uncommon.

A potash felspar with microdine twinning is present in variable amount, and
calls for no special remark.

Scapolite is present in Nos. 38 and 310, but is absent in the remaining rocks.
It has the high birefringence of the scapolites of the pyroxene garnet rocks. In No. 38
it may be surrounded by a narrow shell of basic plagioclase. In many cases this cannot
be regarded as a secondary development of scapolite from plagioclase, as the boundaries
are quite sharp. Rather it has the appearance of primary growth, and suggests that
plagioclase appeared in place of scapolite when the supply of essential mineralisers for
the formation of the latter mineral became exhausted.

Bluish-green actinolite is present in No. 996, but is absent in the other rocks.
It forms independent crystals in the rock.

Quartz is sparingly distributed in these rocks, and very characteristically in
association with epidote. Its presence here suggests that it is a by-product in the
synthesis of epidote.

Two single grains of pleochroic' tourmaline are observed in No. 996. The
pleochroism scheme is 0>E with =- dark green and E - reddish -brown.

(6) EPIDOTE-MARBLES.

The rocks of this class are Nos. 132, 305, 308, 384, 386, 534, 654, 676, 995, 997,
and 1163. As a class they are fine-grained rocks, which on weathered surfaces show
an abundant development of yellow-green epidote. Some of them are banded, due
to streaks and lenses of this mineral, and irregular segregations can often be observed.

The constituent minerals are calcite, epidote, actinolite, microcline, plat/itx-liisc.
titanite, biotite, quartz, chlorite, apatite and magnetite.

Calcite shows abundant lamellar twinning, and the grain -size is very variable.
It is especially fine-grained in rocks Nos. 305, 386, 534, and 1163. In some examples
distorted twin lamellae and mortarisation of the larger calcite grains (as in No. 997)
bear witness to the shearing movements which have affected the rocks.

MKTA.MdlM'HIC LIMKSTONKs TILKKY.

is present in yellowish -green crystals, which are often idioblastic. with
a dominant basal cleavage. It is not infrequently disposed in layers through the rock,
associated with <|iiartx and felspar. This is strikingly seen in No. 132, where the felspar
is largely microdine. The epidote is always a variety rich in ferric oxide, n
indicated by its birefringence and pleochroism. It is only accessonly present in No
In Ni>. lit;:! the mineral is undergoing change to a chloritic |>rodnct with anomalous
interference tints.

iii pleochroic plates of blue-green to yellow-green tint is often present
in association with the epidote.

M icnicl/ni'. plagioclase, and quart/, are the colourless silica containing minerals.
The potash felspar with ink-recline twinning is closely associated with epidote and
quart/. The twinning is, however, not a constant feature, and in the absence oi
cleavage and twinning, it is sometimes difficult to discriminate from plagioclase.

The /tlngiodase present in these rocks is usually of an acid type. It may pos-
both pericline and albite twinning lamella?. The felspar of No. 995, for example, is
an optically positive type, whose refractive index is greater than that of canada-balsam.
but less than that of quartz. It corresponds to albite-oligoclase. Secondary
paragonitic (?) flakes are distributed through the grains. In the other rocks the
plagioclase very generally shows a refractive index less than that of quartz. The place
of the more basic felspars is evidently taken by the abundant epidote.

<><>artz is quite common as a constituent of the ground mass with calcite. and
may sometimes vein the rock as in No. 997.

In the bands of epidote in No. 132, the associated quartz shows rows of pores
along cracks, oriented at right-angles to the long axes of the bands.

(Jreen lnOtite is present in No. 132, associated with epidote. and a single grain of
highly pleochroic 1-mnntdine is also to be observed.

(7.) (.'ARHOXATK-KRKK CAI.C-SII.H '.\TK HOCKS.

Two rocks which cannot strictly be defined in the classes already instituted
complete the suite of rocks. Both these rocks are completely free from primary
carbonate minerals. No. 1153 is a pinkish-white rock with a band of calc-silicate
minerals, partlv diopside and partly a dark fibrous amphibole.

I'nder the microscope, the calc-silicate band shows a very pale tremolitic
actinolite. on the outer edge of which colourless diopside is present. This band is
Mic.-eeded by a band <>! saussuritic material, in which secondary white mica is plentifully
distributed. Quartz is abundant in streaks and lenses. It shows the effect
shenrinv stress in the undulose extinction and mortar texture.

No. l-js is of greater interest. The hand specimen is a greenish -grey rock, in
which small porphyroblasts of felspar can be seen. In thin section the constituent
minerals are observed to be felspar, epidote, and clinox.oisite MS porpliyroblasts. whilst

240 AUSTRALASIAN ANTARCTIC EXPEDITION.

the ground mass is constituted of tremolite, zoisite (clinozoisite), quartz, felspar,
titanite, and apatite. Epidote is present in idioblasts measuring up to 1-5 mm. in
length. It shows a light yellow tint, but is seldom noticeably pleochroic. A basal
cleavage is strongly developed. The double refraction does not exceed 0-027. The
optic axial plane is perpendicular to the elongation and the cleavage, and the optic
sign is negative. Twinning on the 100 plane is rarely developed, and is seen in sections
cut parallel to the optic axial plane. The extinction is 3 to 4 degrees from the plane
100. From the basal cleavage the angle Z A a is 30.

In sections nearly perpendicular to the optic axes, the interference tints are
often anomalous. A few of the porphyroblasts are optically positive. These are
clinozoisite. It is clear that the porphyroblasts are types comparatively poor in
Fe. 2 3 , and grade in composition from clinozoisite (optically positive) with increasing
iron content to the more birefrigent epidote (negative).

The porphyroblasts of felspar are dominantly plagioclase. They often show
both albite and pericline twin lamellae. In composition they grade from albite-
oligoclase to oligoclase. Both optically positive and negative varieties are present,
and the refractive index is usually equal to or greater than that of canada-balsam.
The refrigence is, however, always less than that of quartz. Potash felspar is sparingly
represented, optically negative and has a refringence less that that of canada-balsam.

The ground mass is constituted of quartz and alkali felspar, in which are set
fibres of colourless tremolite, prisms of zoisite, and less frequently highly refringent
grains of titanite and apatite. The porphyroblasts of felspar are remarkably free
from the constituents of the ground mass.

III. REVIEW OF THE METAMORPH1C PROCESSES INVOLVED.

The mineralogical variation of the suite of rocks described in the foregoing is
clearly resultant of two independent factors, viz. :

(a) The range of chemical composition of the original sediment,

(b) The grade of metamorphism which the several classes have experienced.

Accepting throughout that an additive metamorphism has been inoperative,
the rocks of classes (1), (2) and (3) are examples of metamorphosed dolomites. Classes
(4) and (5) represent types in which a gradation to a more calcareous sediment is
apparent, whilst in class (6) the metamorphosed equivalent of calcareous limestones
are revealed. A varying grade of metamorphism is apparent.

The forsterite marbles bear witness to a metamorphism of essentially thermal
type. The silica and alumina of detrital origin have reacted with the carbonate

MBTAMORPHIC LLMESTo.NKS TILLEY. 241

minerals, bringing about a partial dedoloniitisation. The presence of fluorine is
established by the occurrence of chondrodite. The halogen is doubtless derived from
the igneous magma to which the metamorphism of these rocks is due.

In the second class of marbles, however, whilst dolomite is present with the
calcite, the dominant silicate mineral is neither forsterite nor diopside, but tremolite.
The habit of crystalline schists which these rocks possess suggests that in their
metamorphism the element of shearing stress was dominant. Under these conditions
the antistress mineral, forsterite, does not develop.

In the diopside-tremolite-marbles, dedoloniitisation has been complete, and the
residual carbonate mineral is calcite. With the disappearance of dolomite, the entry
of magnesia-free silicate is permitted, and in accordance with this, these rocks develop
subordinate plagioclase. The composition and texture of the rocks of this class allows
them to be regarded as contact types.

In class (4) the pyroxene-garnet-marbles, an abundance of detrital constituents
has given rise to a varied assemblage of metamorphic minerals. Microcline arises
from detrital sericite, the excess alumina being absorbed in the accompanying silicates,
and scapolite develops in place of plagioclase, when the necessary volatile elements are
present.

The pyroxene, garnet and epidote, show by their optical properties that a not
inconsiderable amount of iron is present in their molecules. The source of this iron
oxide, is doubtless ferrous carbonate or hydrated iron oxide (limonite), in the original
sediment. The hedenbergitic character of the pyroxene shows that portion of the
magnesia is replaced by FeO.

The coloured garnet must be regarded as a grossular andradite solid solution.
This is confirmed by its paragenesis, and optically by the refringence. A possible
type of reaction which has led to the development of this mineral may be represented
as follows :

6CaC0 3 + H 4 Al 2 Si 2 () 9 + Fe 2 a + 4Si0 2 = 2Ca 8 AlFeSi 3 O u + 6C0 2 + 2H a O.

Epidote must be regarded as the youngest of the silicates. The presence of
the epidote-quartz symplektites around the garnet suggests that the development of
epidote in these cases is attributable to a degradation of the garnet molecule in a later
stage of decreasing metamorphism,

3Ca :i (AlFe) 2 Si 3 12 + H 2 = 2HCa 2 (AlFe) :s Si :! O l3 + 3SiO 2 + 5CaO.,

the lime with carbon dioxide giving the associated calcite. A development from the
anorthite member of the plagioclase has also played a part.

In the rocks of class (5) epidote is an abundant constituent, and must be given a
place as a primary constituent, derived either from a reaction of carbonates and ferric
oxide, with kaolin or sericite, in the latter case being accompanied by microcline. The
rocks of this class resemble those of class (4) with the distinction that garnet is not
represented and its place is taken by epidote.

242

AUSTRALASIAN ANTARCTIC EXPEDITION.

The rocks of class (6) resemble the products of the outer zone of contact
metamorphic regions. Pyroxenes are absent, and they are represented by actinolite.
Epidote is the most abundant coloured silicate. The rocks are characterised by thei r
fine grain, and both structurally and texturally they bear witness to a comparatively
low grade of metamorphism. This metamorphism has been sufficient however to
give rise to metamorphic felspar in association with epidote. Much of the acid
plagioclase, however, must be considered as being originally developed in the rocks, for
its mineralogical association with calcite or with quartz points to this conclusion. The
degree of metamorphism has allowed recrystallisation, probably with increase of grain
size.

In the carbonate-free rock, No. 128. the epidote and tremolite reactions have
proceeded to the exhaustion of the carbonate mineral.

IV. PARAGENESIS.

In the following table, the occurrence of the various minerals in the suite of
rocks is shown in their paragenetic relationships :

Calcite.

Dolomite,

Forsterite

Chondrod

Edenite.

Tremolite

Pyroxene

1
&

Scapolite.

^
c

o
o

'.
H

Calcite

-)-

-|_

_)_

-j-

_^-

_)_

_|_

_l_

_|_

Dolomite

-j.

_|_

_)_

_|_

_j_

_|_

_j_

_|_

Spinel

-)_

_|_

_j_

Forsterite ... ' ..

_l_

_|_

_)_

_j_

Chondrodite

_)_

_j_

_l_

Edenite

_)_

_l_

_(_

_j_

Tremolite

_)_

_j_

_l_

_j_

_[_

_j_

_l_

_j_

_|_

Pyroxene ...
Microcline

Plagioclase

Scapolite

Garnet

-j-

-)-

Epidote
Quartz

The part played by the lime-alumina silicates is clearly shown in the paragenetic
relations of such minerals as plagioclase, scapolite, garnet and epidote. In dolomitic
limestones, they enter only after dolomite and the pure magnesian silicates and
aluminate (spinel) have disappeared. Microcline appears in a similar fashion.
Detrital sericite is first converted into the magnesian-mica, phlogopite, which is
characteristically a member of the forsterite marble group of rocks. This mineral
reacts later with silica and calcite giving pyroxene and microcline.

\IKTA.Moi: I'll lr I.IMKSKiNKS TII.I.KV. 243

In these rocks pure maguesiuiu-iiluminiiuu silicates do not occur. Of these
latter. conlierite and pyrope garnet are the types. Cordierite is unkno\vn amongst
metamorphosed carbonate rocks, and is unstalile in tin- presence of inonoclinic
pyroxene, (Joldsrhmidt (') (Considers th.it, its absence mav lie attributed to a reaction
of the form -

.M.L'.Al t Si,0 18 +2CaMgSL0 = 4.MgSi(>,+ 2<'aAl .Si.O., + Sin.

yielding enstatite or hypersthene, as in the limestone contact zones of the < 'hristiania
region. A rhomliic pyroxene is, however, wholly wanting in these Antarctic rocks, as
in the product of other areas of regional metamorphism. In this respect the meta-
morphism in the Christ iania region may he regarded as the ideal thermal type
unaccompanied by any notable shearing stress (-).

In the rocks now studied, the alumina of potential cordierite is represented in
plagioclase, and the magnesia in monoclinic pyroxene.( ) According to Eskola (*)
garnets containing more than 75 molecules per cent, of pyrope are unknown, and even
these are confined to eclogitic rocks, and unknown amongst metamorphosed dolomites.
In this respect pyrope and grossularite differ widely in their paragenesis.

Throughout this suit of rocks quartz is stable in the presence of calcite. Even
in the carbonate-free calc-silicate rock (No. 128) there is no development of wollastonite.
The application of Xernst's heat theorem to the equilibrium

+ SiO. < 'aSiO, + CO

gives a PT curve which rises with the pressure. Kor aluminous and ferromagnesian
lime silicates the corresponding curves, as Uoldsehmidt () has pointed out, run at less
elevated temperatures.

Whilst these more complex silicates (epidote, tremolite, pyroxene, Arc.) were
abundantly produced, the development of wollastonite the high-water mark of
metamorphism in calcareous sediments was not attained.

1 V \! lloliU. Imiiilt. \'iri,-ii-.k. Skrift.-r. IHII. X.. 1. p. I3S. : A. Miiikn. l'r,-. A.hl..^ Geo. S.H .. UHs. p I. \\VIII

* C!.'.,|. Ma.. vol. Ivii. l!l-'ll. |.. 4!Mi. ' I'. Kxk..lu. \,,rsk. (Jrol. Ti.lMkr. \1. n. 172-X

V. \l. lHiMdiniilt. Vi.li-n-k Skiifln. l!Hl'. X... L'2. p. li.

244 AUSTRALASIAN ANTARCTIC EXPEDITION.

V. EXPLANATION OF PLATES.
PLATE XXXVI.

Fig. 1. Gale-silicate marble. Photo of the natural surface developed by weathering
under Antarctic conditions of wind and drifting snow. The softer carbonates
are removed, leaving the harder silicates standing prominently in relief.
Concomitant with the metamorphism of the rock, are evidenced contortion
and brecciation of the harder bands. Two-thirds natural size.

Fig. 2.Calc-silicate-Gneiss (Garnet-pyroxene-Marble.) Specimen No. 658. The
photograph shows alternating bands of calcite, green pyroxene and felspar,
and brown garnet.

The continuous black bands are seams of grossular-andradite, and the
remaining black lenses and streaks are hedenbergitic pyroxene associated
with white felspar. About natural size.

PLATE XXXVII.

Fig. 1. For sterile-Marble. Specimen No. 318. The microphotograph shows calcite,
dolomite, forsterite, spinel, and the aluminous amphibole, edenite.

The amphibole is typically developed as coronae around the forsterite
grains. Isolated grains with amphibole cleavage are present in the lower
half of the photograph. The clear carbonate grains immediately above are
dolomite. Magn. 40 diameters.

Fig. 2. Garnet-pyroxene-marUe. Specimen No. 658. The constituents .shown are
garnet, green pyroxene, scapolite, microcline, a symplektitic intergrowth of
scapolite and quartz, and grains of titanite, apatite and calcite. Magn. 40
diameters.

Fig. 3. Garnet-pyroxene-Marble. Specimen No. 730. The constituents are garnet,
epidote, plagioclase, green pyroxene, calcite, and quartz. The garnet is
surrounded by a shell of epidote, intergrown with quartz and some calcite.
The large plate of plagioclase is filled with a sericitic decomposition product,
and is itself bordered by a perimorph of epidote. Magn. 40 diameters.

Fig. 4. Carbonate-free cole-silicate rock. Specimen No. 128. The constituents are
clinozoisite, epidote, plagioclase, microcline, tremolite and quartz. The
porphyroblasts of epidote and felspar are set in a ground mass of tremolite,
zoisite, quartz and felspar. The felspar porphyroblasts are remarkably free
from the groundmass constituents. Magn. 40 diameters.

Sydney: Alfred James Kent, Acting Government Printer 1023.

AUSTRALASIAN ANTARCTIC EXPEDITION.

SERIES A. VOL. III. PLATE XX>VI.

Fig. 1

Fig. 2

AITRALASIAN ANTARCTIC EXPEDITION

SERIES A VOL. III. PLATE XXXVII.

.
A JKK\MB

Fig. 1.

F'g. 2.

' ~\ ^

tJ M

> mm

* -

Fig. 4/

1st

AUSTRALASIAN ANTARCTIC EXPEDITION
^ 1911 14.

UNDER THE LEADERSHIP OP SIR DOUGLAS J\AWSON, KT., D.Sc, B.E

SCIENTIFIC REPORTS.
SERIES A.

VOL. III.

GEOLOGY.

APR 13 19^

PART III :

THE DOLERITES

KING GEORGE LAND AND ADELIE LAND.

W. R. BROWNE. D.Sc

WITH TWO PLATES.

PRICE: ONE SHILLING AND SIXPENCE.

: Alfred Jama Kent. Government Printer IMS.

ISSKD AUGUST. 1923.

SERIES A REPORTS.

HON. EDITOR : PROP. SIR DOUGLAS MAWSON, KT., D.So., B.E.. University of Adelaide.

VOL.

I. GEOGRAPHY AND PHYSIOGRAPHY. (In preparation.)

II. OCEANOGRAPHY.

PART 1. SEA-FLOOR DEPOSITS FROM SOUNDINGS

By FREDERICK CHAPMAN, Ass. Idftn. Soc. (Lond.), F.R.M.S., .fee., National Museum, Melb.

Ill GEOLOGY. (Adelie. Land and King George Land.}

PART 1. THE METAMORPHIC ROCKS OF ADELIE LAND

By F. L. STILL WELL, D.Sc., Aust. Antarc. Exped. Staff 2

2. THE METAMORPHIC LIMESTONES OF COMMONWEALTH BAY, ADELIE

LAND. By C. E. TILLEY, B.Sc ... 016

3. THE DOLERITES OF KING GEORGE LAND AND ADELIE LAND.

By W. R. BROWNE, D.Sc., Lecturer, Geological Department, Sydney University 1 6

4. AMPHIEOLITES AND RELATED ROCKS FROM THE MORAINES, CAPE
DENISON, ADELIE LAND.

By F. L. STII.LWELL, D.Sc., Aust, Antarc. Exped. Staff 020

IV. GEOLOGY. (Witt deal principally with Queen Mary Land.) (In preparation.)
PART 1. THE ADELIE LAND METEORITE

By P. G. W. BAYLY, F.I.C., and F. L. STILLWELL, D.?c. 1

V. GEOLOGY (Macqitarit Island). (In preparation.)
VI, GLACIOLOGY. (In preparation.)

AUSTRALASIAN ANTARCTIC EXPEDITION

1911-14.

UNDER THE LEADERSHIP OP SIR DOUGLAS AVAWSON, KT., D.Sc., B.E

SCIENTIFIC REPORTS.

SERIES A.

VOL. III.
GEOLOGY.

PART III :

THE DOLERITES

KING GEORGE LAND AND ADELIE LAND.

W. R. BROWNE, D.Sc

WITH TWO PLATES.

PRICE: ONE SHILLING AND SIXPENCE.

. Alfred June* Ktnt, liuvcrnmrnt Printer 1J.

ISSUD AUGUST. 1923.

' A

THE DOLERITES

KING GEORGE LAND AND ADELIE LAND,

r,v
W. II. BROWSE, D.Sc.

(Lecturer in Geology, University of Sydney.)
WITH TWO PLATKS.

CONTENTS.

PAGE

I. The Dolerites of the Horn Bluff, King George Land.-

1. Petrography 246

2. Order of Consolidation 250

3. Discussion of the Petrological Features 250

' 4. Chemical Composition and Relationships 252

II. Dolerite Erratics from the Moraines, Cape Denison, Adelie Land ... 252

III. Explanation of Plates 258

Tin-: most important section of this collection of basic igneous rocks consists of three
specimens of dolerite from the great sill at the Horn Bluff, King George Land. Other
dolerite types related to that of the Horn Bluff sill were collected as erratics along the
coasl of Adelie Land, notably at Cape Denison, where the main collection of the

Kxpnlition \\iis made.

At the Horn Bluff the great sill, intrusive into a sandstone formation, rises in
organ-pipe formation as gigantic cliffs upwards of five hundred feet in height. The
specimens from this important locality were collected by C. T. Madigan's party.

246 AUSTRALASIAN ANTARCTIC EXPEDITION.

I. THE DOLERITES OF THE HORN BLUFF.
1. PETROGRAPHY.

The rocks present certain variations in texture, but consist essentially of the
same minerals, plagioclase, pyroxene, and iron ore, with a mesoetasis of micropegmatite.
It will be convenient to preface short descriptions of each of these rocks with an account
of the characteristics of the principal mineral constituents.

Plagioclase is as a rule fresh, but occasionally showing slight alteration to sericite.
The habit is usually stout prismatic, but in the finest-grained rock (No. 732B) the
prisms are more elongated. Zoning is marked, especially in the two coarse rocks
(Nos. 732A and 733), and twinning is present on the albite and occasionally on the
Carlsbad law, pericline twinning being visible only in the felspars of the coarsest phase
(No. 733). The limits of chemical composition are bytownite (Ab 20 An 80 ) to andesine
(Ab 55 An 45 ) in the case of No. 732A, and Ab 29 An 71 to Ab 45 An 45 in No. 733. The felspar
of No. 732B shows less strong zoning and is about Ab 35 An 65 .

The pyroxene is almost entirely monoclinic. It is very pale-grey in colour,
sometimes changing peripherally to green or brownish, and without noticeable pleo-
chroism. Some of it is apparently ordinary augite or diopside, but much is of the
magnesium-rich variety, known as enstatite-augite or magnesium-diopside, characterised
by an optic axial angle which is variable, but always much smaller than that of ordinary
augite, approaching in the limit. Extinction angles up to 43 have been obse.ved,
but it seems as though extinction and double refraction both decline with the optic
axial angle. This is well seen in No. 732A, where two pyroxene individuals often occur
in parallel intergrowth, that with the smaller optic axial angle being distinguished
by a slightly lighter colour in ordinary light.

In addition to the usual prismatic cleavage, there is often in both pyroxenes an
extremely fine basal or salite striation, which, when combined with simple or repeated
twinning parallel to (100), gives rise to herring-bone structure. This basal striation
is not at all constant in its occurrence, often not extending all the way across a crystal,
and being developed in irregular patches. It is often brought into prominence by
local yellowish alteration of the pyroxene, with decline in the birefringence. For the
most part, nothing definite was observed tending to show that this basal parting
betokens twinning, as has been sometimes suggested, (cf. Iddings, Rock Minerals, 1906,
p. 305), although in one case there was something that looked like simultaneous ex-
tinction of alternate lamellae.

A similar very fine striation is also sometimes observed parallel to (100), and
this is likewise often incomplete, and accompanied by alteration. With the develop-
ment of these salite and diallagic striations, there is often local obliteration of the ordinary
prismatic cleavage.

Twinning, when present, is usually of the normal type, parallel to (100), but
occasional cruciform twins on (101) are observed,

THE DOLER1TES OF KIN(J KEORGE LAND AND ADELIE LAND BROWNE. 247

An almost universal characteristic of the enstatite-augite is its undulose extinc-
tion, which is sometimes zonal in character, but generally quite irregular. This is a
feature which has often been observed in the pyroxenes of the quartz-dolerites, but has not
yet been satisfactorily explained. Benson remarks on something analogous in connec"
tion with the rhombic pyroxenes in the dolerites of South Victoria Land 1 , comparing
it with the similar appearance presented by anorthoclase, and ascribing it tentatively
to submicroscopio twinning. It will be remembered that Brogger and Harker 2 believe
anorthoclase to be of the nature of a cryptoperthite ; and it seems possible that this
peculiar form of augite, containing a large proportion of the enstatite molecule, may
be a similar cryptoperthitic intergrowth of ihombic and monoclinic pyroxene, a view
developed by Elsden 3 .

It may be that the ruonoclinic pyroxene at the temperature of crystallization
was able to hold in solid solution a certain proportion of enstatite, more than it could
retain at a lower temperature. On cooling, therefore, the excess of enstatite was
expelled and took up a cryptoperthitic relation towards the augite, and the volume
change, if any, involved in this process possibly set up a state of strain and produced the
undulose extinction. Alternatively if the mineral is still a true homogeneous mix-
crystal, a state of internal strain may perhaps have been set up owing to a tendency
on cooling for the mineral to change its system of crystallization.

The habit of the pyroxene is peculiar ; there is a suppression of the prism faces
and a flattening of the crystal parallel to the (100) pinacoid, so that cross-sections are
rectangles with the length about four times the breadth. The prisms are perhaps
rather more elongated than usual parallel to the c axis, and this characteristic is very
marked in specimen No. 733, in the description of which it will be more particularly
referred to.

The mesostasis which is present in all three of the rocks is usually of such fine
grain as to render the exact determination of its constituents a matter of much difficulty.
Under low magnifications it appears colourless but slightly clouded, and sometimes
almost isolates the plagioclase laths projecting into it. Viewed with higher magni-
fications it takes in most places the form of long slender rods of plagioclase arranged
in different patterns on a background of another felspar with the optical characters of
orthoclase. Tin- rods may be aggregated in groups, each consisting of three or four
])ar;illel rods, crossing each other at various angles; or sheaf-like and plumose aggre-
gates, occasionally sprouting from the end of a plagioclase may suggest the effect
of frost on a window-pane ; or again, the rods may form a confused interlacing network.
These phenomena are similar to those described by Osaun for the Tasmanian dolerite 4
mid n\- Benson for that of Sm.th Yi( toria Land.

1 Report of British Antarctic Expedition, 1907-9, (ieology (vol. ii), page 15 1.
1 Harker : Natural History of the Igneoua Roclu, page 246.
The St. David'* Head Rock Seriea. Q.J.O.8., vol. Ixiv, 1908, pag.
0ann: Central, fur Mill., 1907, pp. 701-11.

248 AUSTRALASIAN ANTARCTIC EXPEDITION.

Elsewhere the intergrowth consists of tiny squares and rectangles and gnomons
of felspar in optical continuity with the outer margins of the plagioclase laths, forming
a pattern on a larger surface of orthoclase. Other patches of the mesostasis consist
of graphically intergrown quartz and ortboclase, the former being recognised by the
familiar triangular forms of its skeleton crystals, while again, extensions of the plagio-
clase laths may be graphically intergrown with quartz. This latter mineral is also
found as independent interstitial grains quite evidently primary in many cases, though
at times possibly secondary. On the whole, however, quartz plays a very minor part
in the mesostasis.

What appears to be minutely granular pale chlorite functions occasionally as
an interstitial filling in the two finest grained rocks.

Iron ores, including apparently both magnetite and ilmenite, are not very
abundant. They occur mostly embedded either in the peripheral parts of the pyroxenes
or else in the mesostasis.

Tiny apatite needles are pretty abundant in the mesostasis of all the rocks, but
practically nowhere else, so that the total proportion of apatite in the rocks is insig-
nificant.

Of the two specimens labelled No. 732, the finer-grained, which we may call
No. 732B, was collected " 6 feet above the contact of the volcanic rock and the sand-
stone." The hand specimen shows very marked prismatic jointing on a. small scale.
The main constituents of the rock are, in volume percentages :

Plagioclase 41

Pyroxene 44

Iron Ore 1

Altered Olivine (?) 3

Mesostasis ... ... ... ... 11

The plagioclase is mostly in thin to stout laths averaging perhaps -3 mm. ; a few
larger individuals apparently slightly more basic than the rest are aggregated in
glomeroporphyritic fashion. Enstatite-augite is probably the dominant pyroxene, but
an interesting feature is the presence of a very little rhombic pyroxene, very pale-
coloured and non-pleochroic, but having negative birefringence. It generally occurs
completely enclosed in augite, with which it is in parallel orientation. A very few
minute flakes of biotite were detected attached to grains of pyroxene.

A peculiar feature of this rock is the presence of scattered patches of a substance,
green to brownish in colour, pleochroic, with a fibrous structure and strong birefringence.
The patches are streaked with carbonates and crossed with cracks containing magnetite
dust. These characters suggest aggregates of talc, or perhaps iddingsite, after olivine,
but this is by no means certain, especially as enstatite or hypersthene might alter in
similar fashion. The outlines of the aggregates do not help much in the determination
of the original mineral.

THi: Dol.KlMTKs <)K Kl\.; GEORGE I. VXD VXD ADK.L1K LAND I'.HOWNK. 249
The pyroxene is evenly distributed through the rock in typical ophitic relation-

ship ti. the felspar. (See hate XXXYI1I. Fig. 1.)

The specimen Nn. 732A was taken from the debii* at the foot of the Organ Pipes,
and is eoarser than No. 732B. It is a greyish-black rock in hand-specimen, with
occasional pyroxene prisms visible up to 3 mm. in length. The principal constituents
are. in volume percentages :

Plagioclase ............ 54

Pyroxene ............... 33

Mesostasis ............ 11

Iron Ores ............ 2

This constitution corresponds pretty closely with the norm of the rock, given below.
and indicates that the rock has a normative mode.

I'lagioclase is in stumpy prisms up to -5 mm., though a few larger individuals
attain a length of 1-5 mm. For the most part the prisms only dent the outlines of the
pyroxenes, which are often subidiomorphic, so that typical ophitic fabric is absent.
Parallel intergrowths of pyroxenes with different optic axial angles are frequent. There
is an absence of rhombic pyroxene, biotite, and the green and brown pseudomorphs so
characteristic of No. 732B. {See Plate XXXVIII, Fig. 2.)

The third of this series (No. 733) is coarser in grain than the other two, and differs
much from them in other respects. It is a brownish-coloured rock with an average
grainsize of about 4 or 5 mm. The dominant characteristic is the presence of many
flashing cleavage-planes of pyroxene, slightly curved at times and exhibiting quite a
bronzy lustre.

The minerals present are :

Plagioclase ............ 33 per cent, by volume.

Pyroxene ............ 27 ,,

Iron Ore ............ 2 ,,

...... ...... 38 ,,

The plagioclase forms thick prisms varying in length up to about li mm. and giving
almost >< plan 1 sect ions at right angles to (001) and (010). There are also smaller prisms.
some of microlitic dimensions. The felspars often show magmatic resorption and
embayment of their edges just like the phenocrysts of a porphyritic rock, and inclusions
of the mesostasis are fairly frequent. Indeed the mesostasis is so abundant that it
ives rise in places to what Iddings would have called porphyritic intersertal fabric,
the pyroxene and felspar playing the part of abundant phenocrysts.

The pyroxene is all monoclinic. and. judging by the si/.e of the optic axial angle
where it can \>e observed, is mostly, at all events, enstatite-augite. The salite
striation is extremely common, and to a less extent the diallage striation, both accom-
panied by a brownish decomposition product which gives rise to the bronzy lustre

250 AUSTRALASIAN ANTARCTIC EXPEDITION.

observed in hand-specimen, while herring-bone structure is very characteristic. The
alteration associated with the striation is often so arranged as to produce a kind of
hour-glass appearance in the sections. A curious feature of these augites is the frequen^
elongation parallel to the c axis, one section parallel to (010) measuring more than
6 mm. in length by about -7 mm. Some of the crystals show a marked bending into
a circular arc, with undulose extinction. This bending causes a tapering in some of
the sections, which, combined with the central pinacoidal twinning-line and the herring-
bone structure, produces the appearance of a goose-quill, an effect heightened by the
curiously serrated edges of the crystals. This serrated effect is possibly a result of the
bending of the crystals during their growth, as suggested by Iddings 1 for the curved
felspars of the Obsidian Cliff rhyolite. Much of the augite is graphically intergrown
with the plagioclase.

The iron ore is mostly ilmenite, in characteristic plates and skeletal forms ; there
is also much ilmenite and magnetite in tiny granules scattered through the mesostaeis.

The rock is much stained with brownish-green and brown decomposition products,
which penetrate the pyroxene peripherally and along cracks, and also discolour large
portions of the mesostasis. (See Plate XXXVIII, Figs. 3, 4, and 5.)

2. ORDER OP CONSOLIDATION.

In the finest-grained rock the felspar has very evidently crystallized first, followed
by the augite. In No. 732A this is not so evident ; the fabric is subophitic, but the
augite is often nearly idiomorphic and only marginally indented by the felspars, as
though the former had been the first to start crystallizing and had continued during
part of the period of crystallization of the felspar. In No. 733 the graphic intergrowth
of felspar and pyroxene indicates much simultaneous crystallization of these two
minerals.

It is noteworthy that much of the iron ore has crystallized late, and that the
apatite is largely concentrated in the mesostasis. This is often the case in basic rocks 2 ,
and is in harmony with the common experience that the last or pegmatitic products
of consolidation of basic magmas are often rich in apatite and iron ore.

Teall, dealing with the Whin Sill dolerite 3 , has remarked on the fact that the
coarse-grained phases of that rock exhibit graphic or perthitic intergrowths of their
constituents. A somewhat similar statement might be made about these Horn Bluff
rocks.

3. DISCUSSION OF THE PETROGRAPHICAL FEATURES.

These three rocks are evidently to be placed among the quartz-dolerites, although
quartz does not enter very conspicuously into the constitution of the mesostasis. There
is an almost complete absence of biotite and hornblende, which are common minor
constituents of the typical hunnediabas and kongadiabas.

1 Iddings : Igneous Rocks, vol. i, p. 226.

* cf. for example Elsden, Q.J.G.S., vol. Ixiv, 1908, p. 289.

' Q.J.O.S., vol. xl, 1884, pp. 640-57.

I UK DOLERITES OF KING GEORGE LAND AND ADELIE LAND- UROWNE. 251

It is instructive to compare the mineralogical constitutions of the three rocks
as given roughly by Rosiwal measurements :

No. N... No.

732B. 732A. 733.

Felspar 41 54

Pyroxene n .r? 27

.Mesostasis ... II 11 38

Iron Ore 1 2

Altered Olivine (?) 3

The rocks are placed in order of increasing average grainsize, and it will be noticed

(1) That the ratio of light to dark constituents increases with grainsize, and

(2) That the proportion of mesostasis has notably increased in the coarsest
phase.

It should be remarked that accurate figures for the mesostasis in the other two
rocks were very difficult to obtain, and that a mental note was made, before results had
been calculated, that the mesostasis of No. 732B had probably been over-estimated,
and that of No. 732A under-estimated at the expense of felspar. Hence there is pro-
bably an actual progressive increase in the amount of mesostasis present with the
increase of grainsize.

The texture of the mesostasis is likewise somewhat coarser in No. 733 than in
the other two rocks, a fact which is in harmony with the observations of Holland 1 ,
Benson 2 , and others on the quartz-felspar mesostasis of quartz-dolerites.

It is unfortunately not known from what part of the sill two of these specimens
originally came. That labelled No. 732B, from 6 feet above the base of the intrusion,
and exhibiting prismatic jointing, probably represents a marginal facies of the rock.
No. 732A, l>eing a bit coarser in grainsize, is probably from a more interior portion of

the .si 11.

It would be interesting to know the precise mode of occurrence of the relatively
coarse-grained type No. 733. This rock is strikingly similar in texture and constitution
to a coarse phase of the Tasmanian dolerite from the Domain, Hobart, similarity
i -.\ tending even to the curvature of the flattened and elongated pyroxene prisms. This
rock, according to Professor Sir Edgeworth David 8 , occurs as schlieren or pegmatitic
segregation veins in the normal fine-grained dolerite. In a specimen in the collection
of the Geological Museum of the University of Sydney, the pyroxene individuals
at times exceed an inch in length, and the mesostasis forms a large proportion of th e
rock.

> Holland, Q.J.G.S.. vol. liii, 1807, p. 408. ' Bwuon, of. cit.. p. 155. ' Verbal communication.

780-2 I'.

252 AUSTRALASIAN ANTARCTIC EXPEDITION.

A search of some of the literature of the quartz-dolerites reveals the fact that
the peculiar habit and curvature of the pyroxene has often been noted. It is true
there is no mention of it, for example, in Marker's description of the Carrock Fell
intrusion 1 , or Elsden's account of the St. David Head Rock series 2 , or in J. V. Lewis's
report on the diabases of New Jersey, U.S.A. 3 .

But Teall, examining the dolerite of the Whin Sill 4 , noted a coarse-grained
variety occurring " only where the rock attains a very considerable development," in
which " crystals of pyroxene measuring an inch in length are not uncommon." The
crystals are flattened parallel to (100), and the cleavage planes are " bent and
undulating," pinacoidal twinning is present, also a fine basal striation, absent when the
rock is fresh but present in the most altered specimens. The coarser rock occur a
apparently in irregular veins, the junctions of which with the fine-grained dolerite are
" remarkably abrupt."

A. H. Phillips 5 noted in the trap of Rocky Hill, New Jersey, a coarse-grained
phase whose relation to the other phases is not mentioned, except that it is regarded as
occupying a central position in the intrusion. The pyroxene crystals " constantly
increase in length as we pass in from the border " of the intrusion; in the very coarsest
varieties of the rock they measure " often an inch and occasionally 2 inches in length."
A schistose arrangement was sometimes seen, and slight curvature of the long axis was
observed, but this is evidently attributed by the author to the pressure which gave
rise to the schisosity.

The internal characters of the pyroxenes are similar to those described for the
Adelie Land rock, but the basal striation is attributed to polysynthetic twinning parallel
to (001). Undulatory extinction was observed of a zonal nature.

The " plumose diabase " occurring as a phase of the trap sheet of Holyoke,
Massachusetts 6 , appears to be a particularly coarse-grained variant of the basalt of
which the sheet is composed. The pyroxenes of this rock, up to 4 inches in length,
are curved, twinned on the pinacoid, and basally striated ; and in addition the vertical
sections show notched or serrated edges due to " the development of unit faces." These
coarse-grained patches occur as lenticular masses or sctdieren in the trap.

The similarity on many points of the Adelie Land rock with those just mentioned
is very clear, and it seems as though the characteristics emphasised, especially those of
the pyroxene, are peculiar to certain pegmatitic phases of the dolerites, so that the
rock No. 733 may with some degree of confidence be assigned to this category.

4. CHEMICAL COMPOSITION AND RELATIONSHIPS.

To indicate the chemical characters of the Horn Bluff dolerite a chemical analysis,
for which I am indebted to Mr. G. D. Osborne, B.Sc., and Miss M. L. Graham, B.A.,

1 Q.J.G.S., vol. 1, 1894, pp. 311-336. 2 Elsden, op. cit. * Annual Report of State Geologist of New Jersey, 1907.
Teall, op. cit., p. 643. '" A.J.S., 4th series, vol. viii, 1899, pp. 267-285. " Emerson, Bull. Geol. Soc. Amer.,
vol. 1904, pp. 91-130.

THE DOLERITES OF KING GEORGE LAND AND ADEL1E LAND- BROWNE. 253

the rock No. 732A, the results of which are given in column J l>do\v, the
analyses of three other quartz-dolerites from Antarctica and one from Tasmania being
added for comparison :

II.

in.

IV.

SiO,

53-06

64-17

64-16

53-2<i

52-49

A1.0, ...

16-95

14-90

15-08

15-64

16-44

Fe,0,

0-79

1-09

0-79

0-24

2-60

FeO

6-69

7-74

8-08

7-44

5-30

MgO

6-91

10-66

7-14

8-64

6-18

CaO

11-56

8-79

10-57

12-08

11-71

NajO

2-05

1-26

1-60

1-25

2-06

K,0

d-07

0-54

1-11

0-58

1-09

H,0 + ...

0-49

059

0-36

0-41

1-42

H,0 - ...

0-43

0-17

0-20

0-35

0-15

CO,

Trace

Trace.

TiO,

0-65

0-64

0-70

0-62

P,0 S

Trace.

0-04

Trace.

MnO

0-07

0-15

0-14

0-11

Trace.

100-61

100-65

99-93

100-74

100-06

Analvtto

Otborne and

? Walkom and

1'rior.

Dittrich.

Graham.

Borrows.

Burrow*.

I. Horn Bluff Dolerite No. 732A;

TI and III. Erratics from Cape Royds, South Victoria Land. Geology (Vol. II),
British Antarctic Expedition, 1907-9, p. 157.

IV. Erratic from Knob Head Moraine, South Victoria Land. National Antarctic
Expedition, 1901-4, Natural History, Vol. I., Geology, p. 137.

V. Enstatite-augite-bearing Diabase from Launceston, Tasmania. Osann, Central, fur
Min. 1907, pp. 701-711.

An inspection of these analyses shows that of all the South Victoria Land rocks
N<>. Ill is most closely related to the Horn Bluff rock, and that there is a distinct
chemical resemblance between them, but that a very much closer and very remarkable
agreement exists between the Horn Bluff dolerite and the Tasmanian diabase. The only
points of apparent difference are in the higher Fe 2 3 and H 8 of the Tasmanian rock,
which are possibly to be attributed to surface alteration. The norms of the three rocks
in question emphasise the closeness of the relationships:

III.

Quartz

l-'.Mi

6-18

4-50

Orthoclase

6-12

6-67

Albite

17-2'.t

13-62

17-29

Vnorthit' 1

33-92

30-86

32-20

Diopside

19-01

17-50

2083

Magnetite

18-93
1-16

22-54
1-16

12-06
3-71

llmrnitP
Water

1-22
0-92

i :',:
OM

1-22
1-57

264 AUSTRALASIAN ANTARCTIC EXPEDITION.

This is particularly true in the case of Nos. I and V. In the latter as compared
with the former there is an increase in normative quartz and magnetite, which is offset
by a decrease in hypersthene. This is due to the oxidation of the FeO into Fe 2 3 ,
which causes more of the FeO to be required for normative magnetite, decreasing the
amount available for diopside and hypersthene, and consequently liberating SiO a for
normative quartz.

It is interesting to note, in passing, the possible effect of slight weathering on
the norm, and perhaps too on the magmatic designation of a rock. It is quite conceivable
that by the oxidation of the FeO, and the consequent liberation of normative Si0 2 , a rock
which actually contains modal olivine may show normative quartz. Further, the
diversion of this Si0 2 from the femic to the salic portion of the norm would disturb th e
relative proportions of salic and femic constituents and of normative quartz and felspar'
putting the rock into a more salic class and a more quaric order.

In the present instance the disturbance has not been sufficient to change the
magmatic position of V relatively to I, and both rocks belong to III. 5. 4. 3 Auvergnose.
No. Ill, is placed in IH.4.4.3.

CONCLUSION.

The wonderful similarity between the quartz-dolerites of South Victoria Land
and those of Tasmania has been pointed out by Benson 1 and Thomson 2 . The present
writer has had the opportunity of examining some of the Tasmanian rocks microscopical^
and of comparing them with the Adelie Land rocks, and the resemblance is certainly
remarkable. Chemical and mineralogical investigation indeed show that the rocks
at present under discussion are very closely related to those of Tasmania, and that they
are, beyond reasonable doubt, co-magmatic with the dolerites of South Victoria Land
encountered by the Scott and Shackleton Expeditions.

II. DOLERITE ERRATICS FROM THE MORAINES, CAPE DENISON.

The collections made from the moraines near the Winter Quarters at Common-
wealth Bay have yielded a number of specimens that have features in common with
the dolerites described above, and should be grouped with them.

Specimen No. 837, of which No. 838 is a duplicate, appears in hand-specimen
as a dark, fairly compact rock, somewhat pitted on the weathered surface, and showing
occasional small phenocrysts of felspar. Under the microscope the rock is seen to be
a typical fine-grained dolerite, wherein the pyroxene is predominant over the felspar,
and resembling No. 732B in many respects. The felspar is present in two generations,

1 Op. cit, ' Jour. & Proc. Eoy. Soc. N.S.W., vol. xliii, 1911, p. 312.

IHK IMH.KRITES OF KING GEORGE LAND AND ADELIE LAND BROWNE. 255

the earlier being in thick tabular crystals, up to 3 mm. in length, and zoned (Ab gfi An 75
to Ab tt An M ). Schiller inclusions are arranged along the pinacoidal cleavage-planes
The later plagioclase laths, averaging "> nun., have a composition about Ab^ An^
(Plato XXXIX Ki.n. 2).

The pyroxene is all monoclinic, pale-grey to pale purplish-grey, and is, in part
at least, enstatite-augite ; it is ophitic towards the felspar. Olivine is represented
by abundant small individuals pseudomorphed by brown iddingsite (?), but a few aggre-
gates are found representing original olivine nodules now changed to a pale-green strongly
birefringent substance, suggestive of talc, and sometimes accompanied by carbonates.
The marginal passage of this into a brown-coloured substance suggests that the material
doubtfully referred to above as iddingsite may really result from the staining of this
green talc-like mineral by iron (Plate XXXIX, Fig. 1).

The iron ore is skeletal ilmenite with a little magnetite ; it is usually moulded
in felspar but enclosed in pyroxene. A little pyrites has made its way along cracks in the
felspar.

There is a small proportion of rnesostasis in the rock, consisting of a greyish
mineral that appears to be orthoclase, sometimes in parallel intergrowth with what is
probably another felspar, and often crowded with magnetite granules and tiny apatite (?)
needles. Very occasionally quartz fills the little interstices between the felspar laths.

The association of olivine and quartz in the same rock, though not unknown, is
rather exceptional. Certainly the olivine is now completely altered, but the identity
of the original mineral is beyond doubt. The quartz for the most part has the appearance
of being primary, acting as an interstitial filling in which are embedded apatite needles
and the ends of plagioclase prisms. A possible explanation is that the olivine grains
and nodules represent early intratelluric crystallizations from the dolerite magma, which
had sunk to the bottom of the magma-reservoir and were caught up in the still liquid
portion of the magma at the time of its injection.

Specimen No. 456 is a medium-grained rock in which felspar (Ab 35 An 63 ) and
>xene are present in about equal proportions, and ilmenite of semi-skeletal habit is
unusually abundant. The pyroxene is, in part at least, enstatite-augite, showing the
salite striation at times and altering into chlorite enclosing tiny sphene granules, and
into brownish uralite, the felspar showing sericitic and calcitic alteration. For the most
put the pyroxene is moulded on the felspar, but occasionally the reverse relation holds.
rtz occurs interstitially in very small amount and never in pegrnatitic intergrowth,
and a very little apatite and biotite are also noticed. A feature of the rock is the presence
of irregular interstitial patches of finer grain than the main body of the rock, consisting,
ir as can be made out, largely of plagioclase. with subordinate augite largely altered
1 - 1 uralite. Ilmenite and magnetite are fairly plentiful in close association with augite ;
also elongated, tiny needles of what appear to be apatite, often with parallel arrange-
Little flakes of biotite are not uncommon, and there is a mesostasis consisting

256 AUSTRALASIAN ANTARCTIC EXPEDITION.

apparently of a felspar (? orthoclase) crowded with confused radiating aggregates of
very tiny brown rods of some intermediate substance, possibly rutile'. The last phase
of consolidation of the rock has been the introduction of pyrites, often along cracks
in the felspar.

A very fine-grained type is represented by Specimens Nos. 449 and 457, which
are really duplicates. The rock is hard, compact, and aphanitic, with a subconchoidal
fracture in places, and showing tiny dendritic patches of pyrites on joint planes. It is
holocrystalline, intergranular in fabric, and of notably uniform grainsize.

Felspar (acid labradorite) is arranged in little bundles of parallel laths up to
3 mm. long, the bundles being oriented in all directions. The pyroxene is very plentiful
in tiny light-green prisms and granules. It is predominantly if not solely monoclinic,
but though the presence of enstatite-augite is suspected the small size prevents con-
clusive proof. Quartz is in fairly abundant interstitial patches, and an interstitial
felspar of low R.I., probably orthoclase, also occurs, but never in intergrowth with
quartz.

Magnetite is quite plentiful, and minute shreds of biotite and needles of apatite
are fairly common. Infrequent vesicles are filled with chlorite. (Plate XXXIX, Fig. 3.)

A different type of rock is No. 459, which is close-grained and slightly vesicular
in hand-specimen. It contains two generations of felspar prisms, both about Ab 35 An 65 ,
somewhat zoned, and often showing only Carlsbad twinning. Enstatite-augite, the
dominant mineral, is moulded on the felspar, but typical ophitic fabric is not developed,
the pyroxene being in small grains. Iron ores include both ilmenite and magnetite,
and a little pyrites has been introduced subsequently to consolidation. Frequent
patches of a fox-red pleochroic mineral (? iddingsite) may represent pseudomorphs
after olivine, and a good deal of greenish and brownish chloritic material is present,
sometimes forming cores to the felspar crystals. (Plate XXXIX, Fig. 4.) The abundant
small interstitial spaces of the rock are filled with material of a light brownish-green
colour, thickly charged with microlitic magnetite. This has a feeble polarization
and may represent the devitrification of a glassy base.

All these rocks possess certain characteristics which link them with each other
and with the quartz-dolerites, the most important being the presence of enstatite-augite
and of quartz and orthoclase. Textural and mineralogical variations are such as might
be expected in a differentiation series, and such indeed as have been described as resulting
from the crystallization of quartz-dolerite magmas.

It is doubtful whether the rock labelled No. 208 should be grouped along with
the dolerites just described. It is a dominantly felspathic rock, with prisms of labradorite
averaging about 1-75 mm. in length, another felspar, probably orthoclase, being present
in minor amount. Pyroxene is for the most part moulded on the plagioclase, and plays

1 cf. J. V. Lewis, Annual Report of State Geologist of New Jersey, 1907, p. 117.

IFIE DOLERITES OF KING OBOKUK I,\NI> \ND ADKME LAND- BROWN K. 257

an interstitial role. This pyroxene is entirely iiuiiuicliiuc. and has a violet, or rather
rose, colour, with distinct pleochroisiu. betokening the presence of titanium. Biotite
i> ulinndant. likewise ilmenite in semi-skeletal forms and rods. Apatite is scarce, and
t here are a few irregular patches of pyrites.

The rock is considerably altered, the felspar being spangled with sericite flakes
as well as with calcite and chlorite ; in some places it appears to be altering to a zeolite.

Augite is much changed to a fibrous uralite, pleochroic as follows :
X = Pale yellowish-green.
Y = Green.
Z = Bluish -green.

With Y>Z>X.

With the uralite change, there has been separation of iron oxide. Chlorite is
another alteration product of the pyroxene. Biotite is irregularly bleached. Some
of this mineral is probably primary, but much appears to be an alteration product of
the augite. The section of an augite crystal is often covered over with tiny scraps of
biotite, many of which are optically continuous and simulate the appearance of a graphic
intergrowth with the augite. Some of the biotite and uralite contain pleochroic haloes
surrounding tiny indeterminate colourless minerals. A number of irregular patches
of yellow-green chlorite are seen, sometimes associated with little granules of carbonates
and of secondary sphene ; these may represent the ultimate alteration products of
augite.

The rock may be termed an essexitic dolerite. (Plate XXXIX, Fig. 6.)

258 AUSTRALASIAN ANTARCTIC EXPEDITION.

EXPLANATION OF PLATES.

All photographs have been taken in ordinary light unless when otherwise stated.

PLATE XXXVIII.

Fig. 1. Fine-grained quartz-dolerite from Horn Bluff (No. 732B). Note typical
ophitic fabric. Patches of mesostasis are to bo seen at the centre of the
picture and elsewhere, x 27.

Fig. 2. Quartz-dolerite from Horn Bluff (No. 732A). The felspars, which act as a
matrix, may be seen indenting the periphery of the pyroxene. The
augite crystal on the left is cut parallel to (100), and shows indistinct basal
striation. The mesostasis in which the felspars are set is well shown.
x 17.

Fig. 3. Part of the pegmatitic quartz-dolerite (No. 733) from Horn Bluff. To the
right of the vertical diameter is part of a long pyroxene individual twinned
on (100) and showing serrated edges. The dark longitudinal band on it
represents an alteration area in which the exceedingly fine basal striation
is developed. The rectangular sections coming out horizontally from the
top of this crystal are cross -sections of augite. At the right-hand side
of the long pyroxene some granophyric mesostasis may be seen. Note
the variation in the size of the plagioclase. x 17.

Fig. 4. Basic plagioclase, cut normal to (010) and (001), in graphic intergrowth with
augite, in No. 733. Chlorite-stained granophyric mesostasis at top and
sides, x 17.

Fig. 5. No. 733. Most of the field is occupied by mesostatic material, some in the
position of extinction, consisting largely of felspar with rod-like or acicular
development. To the left this is in contact with a twinned plagioclase
crystal whose broad lamellae are extinguished. Crossed nicols. x 59.

PLATE XXXIX.

Fig. 1. No. 837. Showing part of a nodule composed of olivine grains with some
associated felspar and pyrites. The olivine has been altered to talc (?),
the colour of which changes from very pale-green on the left to a
greenish-brown on the right side of the nodule, x 17.

Fig. 2. No. 838. Showing part of a glomeroporphyritic aggregate of plagioclase.
Crossed nicols. x 17.

Fig. 3. No. 449. Containing plagioclase, with granular augite and magnetite. A
few small white patches represent interstitial quartz, x 27.

Fig. 4. No. 459. Showing on the left part of a glomeroporphyritic aggregate of
plagioclase crystals whose central portions have been replaced by chlorite.
x 27.

Fig. 5. Essexitic dolerite (No. 208). There may be seen augite, plagioclase, ilmenite
(sometimes in rods) and chlorite (light irregular patches at top and right-
hand side), x 17.

Sydney: Alfred James Kent, Government Printer 123.

AUSTRALASIAN ANTARCTIC EXPEDITION

SERIES A VOL 111. PLATE XXXVIII.

Fig. 1.

Fig. 2.

Fig. 3.

Fig. 4.

AUSTRALASIAN ANTARCTIC EXPEDITION.

SERIES A. VOL. III. PLATE XXXIX

Fig. 1.

Fig. 2.

&&&
^%

Fig. 3.

Fig. 4.

Fig. 5. . ' '.

. .

.' .

,. .-..

AUSTRALASIAN ANTARCTIC EXPEDITION
~ 1911-14.

UNDER THE LEADERSHIP OP SIR DOUGLAS nAWSON, KT . D.Sc., B.E

SCIENTIFIC REPORTS.

""""""SERIES A.

VOL. 111. APR 131927

GEOLOGY.

PART IV:

AHPHIBOLITES AND RELATED ROCKS

FROM

THE MORAINES, CAPE DENISON, ADELIE LAND.

F. L. 5TILLWELL. D.Sc

WITH TWO PLATES.

PRICE: TWO SHILLINGS.

AJhuI Ja>~! Kcct. G<mrmMM Pnrnwr. PkMlp-MfW< Sftuff i

ISSUED AUGUST, 1923.

SERIES A REPORTS.

HON. EDITOR : PBOF. SIR DOUGLAS MAWSON, KT., D.Sc., B.E., O.B.E., F.R.S.

VOL. PEICB.

s. d.
I. GEOGRAPHY AND PHYSIOGRAPHY. (In preparation.)

II. OCEANOGRAPHY.

PART 1. SEA-FLOOR DEPOSITS FROM SOUNDINGS

By FREDERICK CHAPMAN, Ass. Linn. Soc. (Lond.), F.R.M.S., &c., National Museum, Melb. 060

III GEOLOGY. (Addie Land and King George Land.)

PART 1. THE METAMORPHIC ROCKS OF ADELIE LAND -

By F. L. STIELWELI,,: D.Sc., Aust. Antarc. Exped. Staff 2 2 O!

2. THE METAMORPHIC LIMESTONES OF COMMONWEALTH BAY, ADELIE

LAXD. By C. E. TILLEY, B.Sc ...................... 016

3. THE DOLERITES OF KING GEORGE LAND AND ADELIE LAND.

By W. R. BROWXE, D.Sc., Lecturer, Geological Department, Sydney University 016

4. AMPHIBOLITES AND RELATED ROCKS FROM THE MORAINES, CAPE
DENISON, ADELIE LAND.

By F. L. STILL WELL, D.Sc., Aust. Antarc. Exped. Staff 020

IV. GEOLOGY. (Witt deal principally with Queen Mary Land.) (In preparation.)
PART 1. THE ADELIE LAND METEORITE.

By P. G. W. BAYLY, F.I.C., and F. L. STILLWELL, D.Sc. 1 6

V. GEOLOGY (Maoquarie Island). (In preparation.)
VI. GLACIOLOGY. (In preparation.)

AUSTRALASIAN ANTARCTIC EXPEDITION

1911-14.

UNDER THE LEADERSHIP OP SIR DOUGLAS AAWSON, KT., D.Sc, D.C

SCIENTIFIC REPORTS,

SERIES A.

VOL. III.

GEOLOGY.

PART IV:

AMPHIBOLITES AND RELATED ROCKS

FROM

THE MORAINES, CAPE DENISON, ADELIE LAND.

F. L. STILLWELL, D.Sc

WITH TWO PLATES.

PRICE: TWO SHILLINGS.

ftiatti It Altnd Janet feat, GOWBIMBI f>nMr, Pbillip-ttcMt ,ST***T t**t-

ISSUED AUGUST, 1923

AMPHIBOLITES AND RELATED ROCKS FROM
THE MORAINES, CAPE DENISON.

F. L. BTILLWELL, D.Sc.

WITH TWO PLATKS.

CONTENTS.

PAGE.

I. Introduction ........................... 261

II. Group IV*. Amphibolites and Eclogits ............... 263

1. Kata Division ........................... 263

llnrntblende-Plagiodase-Pijroxene-fTneiss.
No. 937, with relic structure of original dolrrit?.
Nos. 902, 067. Erratics from Cape Hunter.

Garnet- Plagioclase-Pyroxene-Gneits.

No. 693, related to eclogitcg.
Garntt-Plagioclase-Biotite-Gneisg.

No. 227, pyroxene largely replaced by biotite.

2. Meso Division .................... . ...... 264

Pyroxene-amphibolites.

Noe. 380, 351, 212. No. 212 contains large porphyroblasta of Ab, AD,.
No. 547, with felspar replaced by quartz.

Amphibolites.

Nos. 946, 956, with " sieve " structure.
No. 865, with porphyroblasto of felepar.

Biot ite- A mphibolite.

No. 597, with greyish colour in hand specimen, and a little lawsonite.

3. Epi Division ........................... 266

( 'hlnrite-A mphibolites.

No. 985, with saussurltised felspar.

20, with the separation of blue glancophane from green hornblende.
No. 965, an amphibolite with similar features and " sieve " structure.
No. 247, with porphyroblasts of felspar.

it h more pronounced cataclaois and mortar stnicture than No. 247.

No. 961, with remarkable mortar structure.
No. 2.13, hornblende extensively chloritised.

Kpiilntf-( 'liliirite-schist.

'10, n-n-inbles the epidote-biotite-chist. No. 1.53, found in i(u, except that bintite if replaced by

'hl'irit*.

Lawsm\itf -Chlorite- A mph iboHte-xrli < - / .

No. 936, contains pyroxene and an excellent developmpnt of lawsonite.

* The olMdBomtion adopted is that proposed by U. Grubenmann in " Die Kristallinen Bchiefer. ' Berlin, 1910.

-'A

260 AUSTRALASIAN ANTARCTIC EXPEDITION.

riot,
4. Gabbro Schists and Gneisses 268

These differ from the other amphibolites in possessing an apparent coarse grain-size due to their
derivation from coarsely-crystalline gabbros.
No. 915, a garnet-biotite-amphibolite.

No. 990, similar to No. 915 in which brown biotite is replaced by green chlorite.
No. 339, garnet-plagioolase-gabbro-gneisg.

No. 593, the gabbro structure is so well pronounced that the rock is sirailiar to a hornblende -gabbro
No. 367, with augen structure and the augen consist of aggregates of calcic andesine.
No. 969, with augen of saussurite.
No. 317, saussurite-gabbro-gneiss.

No. 728, gaussurite-gabbro-gneiss with traces of the outlines of the original felspar.
No. 861, saussurite-gabbro-gneiss with well-preserved form of original felspar.
No. 515, saussurite-gabhro-gneiss with branching veins of lawsonitc and chloritisation of the hornblende.

III. Group V. Magnesium Silicate Gneisses 271

1. Kata Division 271

No. 587, pyroxene rock.

No. 963, hornblende-pyroxene rock.

2. Meso Division 271

No. 594, garnet-hornblende-schist.

No. 513, anthophyllite-schist, with abundant green hornblende and areas of talc.

No. 516, anthophyllite-schist with green hornblende and without areas of talc.

No. 109, anthophyllite-schist with plagioclase and without green hornblende.

No. 548, hornblende-fels, consisting wholly of hornblende.

No. 950, biotite-hornblende-schist.

No. 941, hornblende-fels, with granulitic hornblende developed from lar^c platy hornblende.

No. 209, hornblende-schist with areas of talc.

No. 246, aotinolite-schist.

3. Epi Division 274

No. 229, chlorite-hornblende-schist.

No. 888, hornblende-schist with large crystals of chlorite.

No. 229, ohlorite-hornblende-schist.

No. 916, epidote-chlorite-schist.

No. 931, epidote-hornblende-schist.

IV. Group III. Plagioclase Gneisses 275

Tliese rocks are more acid than the amphibolites and contain less hornblende or its equivalent.

1. Meso Division 275

Nos. Ifi9, 170, hornblende-plagioclase-gneiss with schliercn of amphibolite and pyroxene.
No. 378, hornblende-felspar-gneiss, containing microcline as well as plagioclaso.
No. 259, hornblende-plagioclase-gneiss, with subordinate biotite.
No. 240, Biotite-plagioclase-gneiss, with subordinate hornblende.

No. 956, Biotite-hornblende-felspar-gneiss. The hornblende resembles actionolite in form nd colour,
and biotite is more reddish than usual.

No. 51, hornblende-quartz-plagioclase-schist, with hornblende porphyroblasts embedded in a fine grami-
litic ground mass of quartz and felspar.

No. 924, hornblende-plagioclase-schist, resembles a porphyroide, but the base is a mass of densely-packed
needles of hornblende.

AMI'HIIUiUTKS AND RELATED !!<>< K- >TI l.l.\\ KM.. 201

-'. Mj.i Division 277

_'"i7, (pid.itv-fi'lHpar-sclmt.
N l.'iT. :>22, .piilote-felspar-sehi-.t with broader banding than No. 257, and containing a little quutt?

72fl, epidote-fe|gpar-Hchil with very broad hands of felspar and epidotr.
NM. 398, epidote-felspar rork with large porphyroblaU of microcline.
N.I. 954, chlorite-raicriK-lme-gncisH. belonging to the group of the alkali-felspar gnei*
No. 590, epidote-fclsp.u -. lu.i with pink purphyrobliuits or augen of febtpar.
No. 910, epidote-frlmir breccia, in which i'|>idnt>- nccum in a finely -granular form.

.'._'. i'pi<l.iUj-fel.par-gneis, brerri:<ti>cl l,y vein* of quartz and felspar and finely-granular cpidote.

591, epidote-fclspar-Kneiss in whii-h tin- tine granular epidote occurs in wavy band* producing a helicnl
structure.

No. 589, epidotc-fcUpar-gneiiU) in win. h >-pi<l..t< i> m.>n- abundant than No. 591.

V'. ">!!>. epidote i-hlorite-HchiHt containinu more chlorite and less felspar than No. 589.

No. 255, an epidocita of Group IX which results when there is a deficiency of both chlorite and felspar.

V. Description of Plates 280

I. INTRODUCTION.

I'.KAT variety of metamorphic rocks were collected from the moraines at Cape
Di-nison, in Adelie Land, and the following notes give some account of the amphibolites
;in<l related rocks. Three specimens, which are specifically mentioned, are fragments
nl initii s found at Cape Hunter, nine miles west of Cape Denison. The basis of the
grouping of these rocks is Grubenmann's chemical classification, with its zonal
subdivisions, and this section includes Groups IV, V, and III, the Amphibolites and
Kclogites, the Magnesium Silicate Gneisses, and the Plagioclase Gneisses. Each group
is subdivided into three divisions (based on mineral composition and on structure)
rjilli'd the Kata, Meso, and Epi divisions. In the Kata division of these groups the
pyroxene is in excess of the hornblende; in the Meso division the hornblende i>
in excess of the pyroxene ; and in the Epi division there is a pronounced development
of saussurite, lawsonite, epidote or chlorite, and also of mechanical structures. These
groups are defined by their chemical composition, and the use of the classification is
therefore restricted by the absence of complete chemical analyses. The chemical
(imposition is, however, reflected in the mineral composition, from which the divisions
nl '-;ich group have been in most cases recognised.

The term amphibolitr is used with the same meaning adopted in the memoir
mi the metamorphio rocks occurring in situ at Cape Denison. 1 It may be defined as
u completely recrystallised rock of basic igneous chemical composition, whose mineral
content is essentially hornblende and plagioclase. The latter mineral can be replaced
wholly or partly by other minerals, such as zoisite, epidote, garnet, or scapolite. When
tin- amphibolite possesses a strongly-foliated character it has been called an
(iin/ihibolite-schist. In many cases the schistose structure is accentuated by the partial
replacement of hornblende by biotite or chlorite, and the term ainphibolit
becomes equivalent to the terms biotite, chlorite or tnica-amphibolite.

' " MUmorphic IWLi of .Hrlie Land." F. L SUIlwrll. S< imtitk RporU, A.A.E , Vol. iii, pt. 1, pp. 24, 26.

262

AUSTRALASIAN ANTARCTIC EXPEDITION.

This usage does not conform with the definition issued by the Joint Committee of
the Geological Society of London and of the Mineralogical Society in the recent report
on British Petrographic Nomenclature 1 . This Committee retain the term amphibolite
for unfoliated or slightly -folia ted metamorphic rocks of doubtful or other than igneous
origin. It is composed essentially of hornblende and felspar, often containing various
accessories, such as epidote and garnet. They distinguish hornblende-schist from
amphibolite by the possession of a foliated texture.

On this basis completely recrystallised rocks, which consist of hornblende and
felspar, and which possess a slightly-foliated structure, are excluded from the family
of amphibolites when they are derived from igneous rocks. It would appear to be the
intention of the Committee to include such cases among the epidiorites which are defined
as unfoliated basic igneous rocks in which the augite is completely altered to
hornblende. They seem to have overlooked the fact that two types of altered basic
hornblende rocks, occurring in the form of a dyke or sill, can be recognised, (1) those
in which the ferro-magnesian has suffered recrystallisation, and (2) those in which
both the ferro-magnesian and felspar have suffered recrystallisation. These two
types may be encountered in one area, as at Broken Hill (New South Wales), where
they are readily distinguishable from one another in the field, in the hand specimen
and under the microscope. The term epidiorite covers the first type, and it is clearly
undesirable to extend it to the second type, which is identical with amphibolite.

A typical amphibolite will contain 60 to 70 per cent, of hornblende and about
25 to 30 per cent, of felspar. Other metamorphic rocks occur which contain nearly
all hornblende and practically no felspar, and it is necessary for precise description
to distinguish these from the typical amphibolite. It is also desirable to distinguish
between the foliated amphibolite and the hornblende schist with felspar in excess of
the hornblende. These metamorphic types, in which hornblende is an important
constituent, may be tabulated as follows :

Mineral composition.

With massive structure.

With foliated structure.

Mainly hornblende

Hornblende in excess of felspar...
Felspar in excess of hornblende

Hornblende fels

Amphibolite

Hornblende-plagioclase-gneiss

Amphibole or hornblende-schist.
Amphibolite- schist or mica -amphibolite
Hornblende-plagioclase-schist.

The massive types pass by gradual transitions into the foliated types. It is
obviously difficult to restrict the term hornblende-schist to foliated amphibolites, as
recommended by the Joint Committee. It is more logical to use the term for the pure
amphibole types as has been done by Grubenmann 2 . The term amphibole-schist
includes rocks which contain more than one variety of amphibole, but those which
contain anthophyllite are referred to as anthophyllite schists in the following pages.

" Report on British Petrographic Nomenclature." Min. Mag., Vol. xU, No. 82, pp. 137-147, 1921.
" Die Kristallinen Schiefer." U. Grubenmann. Berlin, 1910, p. 216.

AMPHIBOL1TI> \ND RELATED ROCKS-ST1U.U KLL. 263

II. GROUP IV. THE AMPHIBOLITKS AND ECLOOITKS.

At Cape Denison the variety of types occurring in situ is considerably increased
by a study of specimens collected from the moraines. Tin- main occurrences in xitn
i (insist essentially of felspar and hornblende, with or without biotite. They have been
looked upon as basic igneous dykes which have recrystallised under conditions varying
from tlmse of Grubenmann's Meso zone to those of the Epi zone. The moraine types
include representatives of the Epi, Meso, and Kata divisions, and some are similar to
members of the Cape Gray series, occurring '2~> miles to the east.

KATA DIVISION.

The distinguishing feature of the Kata division of this group is the abundance
of pyroxene. A representative is specimen No. 937, a hornblende-plagioclase-pyroxene-
gneiss, which possesses similar features to the plagioclase-pyroxene-gneiss, No. 773,
from Cape Gray, and to the plagioclase-pyroxene-gneiss, No. 935, from Stillwell Island 1 .
Like Nos. 773 and 935, it possesses large relic crystals of augite, which have recrystallised
partly as clear granular pyroxene and partly as hornblende. Like No. 773, it possesses
traces of the original felspar laths of the original dolerite. Its felspar is a clear basic
labradorite, appearing partly as lath-shaped crystals, but mostly as coarse granular
crystals of the same average grain size as the hornblende and pyroxene. It contains a
few flakes of biotite and a few disseminated and minute grains of ilmenite.

Related to this type are two erratics (Nos. 962 and 967), from Cape Hunter,
9 miles west of Cape Denison. These are hornblende-plagioclase-pyroxene-gneisses, in
which the pyroxene is in excess of the hornblende. Garnet is absent, and the pyroxene
includes both augite and hypersthene. Basic felspar is an important constituent,
and, in addition, the rocks carry a little biotite and accessory apatite and ilmenite.

Numerous pink garnets are present in No. 595, a type closely related to the
garnet-plagioclase-pyroxene-gneiss, No. 935, from Stillwell Island. Some of the
garnets tend to form a zone between the pyroxene aggregates and the plagioclase. The
felspar occurs as large crystals of basic labradorite and as granulitic aggregates of a less
calcic felspar, probably andesine. Borne of the larger crystals contain cloudy alteration
products. Quartz is present, and has probably developed with the formation of garnet
by the interaction of felspar and pyroxene. There is a considerable amount of
hornblende, derived from the alteration of the pyroxene, which is associated with
numerous large grains of ilmenite, probably derived from the same source. This
garnet-bearing type is related to the eclogite family, but the clear omphacite of
the typical eclogite is absent.

Specimen Xo. 227 is a dark, glistening schistose rock, showing abundant pink
garnet and black biotite in the hand specimen. It also possesses a thin lenticle of
segregated quartz along the schistosity, which may have developed with the formation

1 Op. tit.. VV . 1W. 17*.

264 AUSTEALASIAN ANTAECTIO EXPEDITION.

of the garnet. In section, the rock has a coarsely-crystalline schistose structure, and
the brown biotite is the most abundant mineral along some of the schistose bands.
The plagioclase is mostly andesine, and varies towards labradorite. It is mostly
clear, and only occasional grains show alteration. Pyroxene occurs in large crystals
along one schistose band, and in smaller crystals in other parts of the rock. It
includes hypersthene, augite, and fibrous diallage, from which ilmenite has
separated out. Only occasional crystals of green hornblende are present. Garnet
occurs in large pink crystals up to 6 mm. in diameter, with numerous inclusions, and
also in small idioblastic crystals. A little quartz is present, and there are large
accessory crystals of apatite and zircon. The rock may be called a garnet-biotite-
plagioclase-gneiss. It differs from members of the eclogite family in the large develop-
ment of biotite in place of pyroxene.

MESO DIVISION.

The conditions of the Meso zone metamorphism are considered to become more
important in the formation of the augite-amphibolites. Several examples, including
Nos. 380, 351, and 212, exist in the collection; and the percentage of hornblende in
these is in large excess of the percentage of pyroxene. They resemble the pyroxene
amphibolites described from the Cape Pigeon Rocks. 1 Of these specimens No. 212
is remarkable in possessing large, dark -green porphyroblasts of a very calcic felspar.
Some of these crystals are over an inch in width, and possess well-defined crystal
boundaries. The texture of the rock is massive, and the structure is typically
granoblastic and porphyroblastic. The mineral composition of the rock, excluding
the porphyroblasts and the accessory minerals apatite and ilmenite, is :

Felspar 31-4 per cent.

Hornblende 64-8 ,,

Pyroxene ..3-8 ,,

The proportion of felspar to the ferromagnesian in this case is practically the
same as in the Cape Denison amphibolites. The felspar is quite clear and unaltered.
Most grains show lamellar twinning and are highly calcic. The hornblende is green
and fresh, without the tinge of blue colour noticeable in many of the amphibolites
with Epi zone features. The pyroxene includes both augite and hypersthene.

The porphyroblasts possess crystal outlines, and, like the felspar in the base,
are perfectly clear and free from traces of decomposition. They show complex
twinning, and, in addition to simple twinning, there are at least two sets of lamellar
twinning. They contain small inclusions of pyroxene and hornblende. The pyroxene
inclusions tend to occur in rounded grains, while the hornblende inclusions develop
their crystalline form against the plagiolcase. The inclusions in many places show a
linear arrangement along cleavage lines and twinning planes. The presence of these
inclusions and the fresh character of the plagioclase in the base of the rock indicate

1 Op. oit., p. 180.

AMI'HIBOLITES AND RELATED R<" K III I U KLL. 265

that these large crystals have developed during tin- re-crystallisation, under the
rnn-litions of the formation of the amphibolic. Tin- ;i.l>sence of any relic structure is
important negative evidence that these porphyroblasts were not phenocrysU in the
original igneous rock.

A sample of the porphyrohlastic felspar was separated and freed as far as
po>sible from the ferromagnesian inclusions by hand picking and by the use of heavy
solutions. The sample was then analysed by Mr. .1. ('. Watson in the Victorian
Survey Laboratory.

SiO 2 45-1CJ

A1 2 3 35-34

Fe a O a 1-05

MgO ... str.tr.

CaO 16-39

Na 8 O '1-70

K 2 -20

H 2 -33

H a O -00

100-93

This analysis demonstrates the highly calcic nature of the porphyroblast which
approximates to AbjAn,,.

Specimen No. 547 is a remarkable quartz-pyroxene-amphibolite. It is distin-
guished as one of the very few specimens occurring on the moraine with a weathered
crust. It is a heavy basic rock, and a rough determination of its specific gravity gave
3-25. It is rather coarse-grained, and possesses a slight schistose character. In
section it is found to be similar in structure and character to the other specimens of
pyroxene-amphibolites, except that the felspar is completely replaced by quartz. The
amount of quartz is approximately the same as the amount of felspar in the typical
amphibolites. No garnet is present in the section, but grains of ilmenite and pyrite
are unusually abundant, and partly account for the high specific gravity.

The Meso division of the amphibolites is more typically represented by a number
of amphibolites and biotite-amphibolites which are similar to some of the occurrences
tu at Cape Denison. In the typical examples the plagioclase is clear and often
untwinned, and in Nos. 946 and 959 the hornblende crystals contain numerous clear
inclusions of felspar and quartz, producing a "sieve" structure. In another case,
No. 865, there are numerous porphyroblasts of felspar, which average about 5 mm.
in width. This specimen was part of a large erratic in which the porphyroblasts were
arranged in a. linear manner and parallel with the schistosity an occurrence in strong
contrast with the scattered occurrences of metamorphic xenoliths in the amphibnlite

266 AUSTRALASIAN ANTARCTIC EXPEDITION.

in situ at Cape Denison 1 . Apart from the porphyroblastic character the rock is a
typical amphibolite, with dominating hornblende and very little brown biotite. The
felspar is clear andesine with indefinite and intermittent twinning.

A schistose variety is represented by No. 597, in which the light-coloured
constituents are sufficient to produce a greyish colour in contrast to the black colour
of the normal amphibolite. Tt is a biotite-amphibolite-schist, carrying lawsonite,
epidote, and coarse crystals of apatite. The epidote and lawsonite are interlaminated
with biotite, and the development of these minerals indicate an approach to the Epi
division of this group.

EPI DIVISION.

Specimen No. 985 is a jet black schistose chlorite-amphibolite, similar in most
respects to the preceding variety, No. 597. The biotite of No. 597 has been mostly
converted into chlorite in No. 985, and most of its felspar has been saussuritised. These
changes are considered to be Epi features, superimposed upon a typical amphibolite.
Large vein-like segregations of saussurite are present in the hand specimen of No. 985.

Another chlorite-amphibolite is specimen No. 520. Biotite is completely replaced
by chlorite, and the alteration is accompanied by the separation of a little quartz. An
interesting feature in this case is the separation of the blue glaucophane constituent
from the complex molecule in a manner similar to the schiller inclusions that are
sometimes found in hypersthene. The result is a mottled green and blue crystal of
hornblende (PI. XL. fig. 1). Apatite and ilmenite are abundant minor constituents,
and most of the felspar is cloudy and saussuritised except for small areas of clear
albite.

The same mottling of blue and green hornblende is developed in a lesser degree
in the amphibolite No. 965. This specimen possesses a pronounced " sieve " structure,
in which the hornblende is highly perforated with colourless inclusions. The plagioclase
is cloudy. Some of it possesses prominent diablastic structure, representing one stage
of its development, but most of it forms a fine granulitic aggregate of acid plagioclase.
Crystals of apatite and ilmenite are large and unusually abundant. A little chlorite
is present, derived from biotite, and pyrite is also present.

Another variety of chlorite-amphibolite is No. 247. This specimen is a schistose
rock, in which abundant lenticles or porphyroblasts of felspar are set in a dark matrix
of hornblende, mica, and felspar. The base of the rock is similar to the biotite-
amphibolites, except that biotite is replaced by chlorite. In this and the preceding
examples the chloritisation does not extend to the hornblende. Quartz is present
showing cataclasis, and, together with the abundant felspar, indicates a gradation
towards the more acid group of the plagioclase gneisses. Some of the felspar lenticles
consist of large saussuritised plagioclase, partly surrounded by clear, recrystallised
felspar. Others consist of granoblastic masses of andesine felspar, with a little quartz.

1 Op. cit, p. 48.

AMFHIBOLITES AND RELATED ROCKS STIKLNVKLL. 267

tered grains of epidote, chlorite, and muscovite occur in the felspar areas. In
this case, the features of the Epi division include the presence of epidote and the
cataclastic structures, in addition to the presence of saussurite and chlorite.

Specimen No. 828 is a similar porphyroblastic ohlorite-amphibolite, but the

r;it;icl:isis and mortar structure are much more pronounced. The margins of the
hornblende crystals have been crushed and converted into chloritic selvages. The
porphyroblasts of felspar are elongated along the schistose planes, and are surrounded
by chlorite and finely -gran ulitic quartz and felspar. This rock also contains occasional
large crystals of allanite.

Specimen No. 961 is an Epi-amphibolite, in which the hornblende crystals display
a remarkable mortar structure. It is a highly schistose rock, containing a a larger
percentage of hornblende than the typical amphibolite. The fracturing and grinding
of the fragments of hornblende have produced zones and films of green chlorite around
islands of unaltered hornblende. A blunt hornblende crystal often assumes a lenticular
shape where the chloritic brash is drawn out along the schistosity. The clear mineral
is an acid plagioclase, frequently drawn out into lenticular shapes along the schistosity.
Some crystals show a mortar structure, while others have been reduced to finely-
pulverised aggregates.

Specimen No. 233 is a chlorite-amphibolite, in which the alteration of hornblende
into chlorite has been extensive. It contains large pale green to colourless crystals of
hornblende, which are all partly converted into chlorite. The greater proportion of
tho chlorite, which is the most abundant mineral in the rock, has been formed in this
way. A subordinate quantity of chlorite appears to be pseudomorphous after biotite.
The felspar is cloudy, and occurs in granular crystals. Though mostly untwinned,
it is probably an acid plagioclase. In a vein that traverses the section, it has
r.'i r vstallised as clear albite. Ilmenite, granular sphene, and apatite are distributed
freely through the rock. Epidote is a minor constituent, except in the neighbourhood
of the vein, where it is associated with chlorite and clear albite.

Specimen No. 210 is an interesting relative of the epidote-biotite-schist, No. 153,
which occurs as an altered dyke rock at Cape Denison 1 . Its structure is more massive
and less schistose than No. 153. The biotite of No. 153 is replaced by a green pleochroic
chlorite, and epidote crystals are more numerous than in No. 153. The clear felspar
is an acid plagioclase, and no basic felspar is present. Sphene and magnetite (or
ilmenite) are important minor constituents, and are associated together in the same
manner as in the typical amphibolites. Apatite is an accessory mineral, and there
are occasional grains of calcite. The amount of felspar appears to be less in this rock
than in No. 153, tending to make this specimen a typical member of the family of
epidote-chlorite-schists.

An interesting specimen is the lawsonite-chlorite-schist, No. 936. It is one of
the few specimens from this region which shows surface weathering, being covered
with a brown iron-stained skin. A brown colour also appears on the fractured surface,

i,t,.-J -I; dp cit.,p. 3U.

268 AUSTEALASIAN ANTAKCTIC EXPEDITION.

and is due to the presence of the golden-brown mica on the schistose surface, such as
is characteristic of the lawsonite-amphibolites at Cape Denison. The ferromagnesian
constituents include pyroxene, hornblende, biotite, and chlorite. The pyroxene,
which constitutes 5 per cent, of the rock, has the granular character and pale colour
common in the plagioclase-pyroxene-gneisses. In some cases it passes over directly
into chlorite, but its alteration to green hornblende is more common. Green hornblende
forms about 14 per cent, of the rock, and the remaining four-fifths of the rock consists
chiefly of biotite with lawsonite and chlorite, saussuritised felspar, and clear acid
plagioclase. Biotite possesses the curious reddish-brown colour previously seen in the
lawsonite amphibolite (No. 635), occurring in situ at Cape Denison, and shows various
stages of decolouration in its alteration to chlorite. Colourless lobate lenticles of
lawsonite are interlaminated with both biotite and chlorite. Lawsonite is also well
developed in the chloritic and saussuritised areas. The saussuritisation of the plagioclase
has been fairly complete. The albite molecules have reappeared as clear acid plagioclase,
while the anorthite molecules have combined with water and produced lawsonite and
granular epidote. Fibrous chlorite is also abundant among these aggregates, and
there is a little quartz and white mica. Sphene, ilmenite, and pyrite are important
accessory minerals, and isolated grains of allanite and apatite are present.

Lawsonite, chlorite, and epidote indicate the relationship of this rock to the Epi
division. The presence of pyroxene indicates that the rock was previously subjected
to Kata zone metamorphism, being probably a plagioclase-pyroxene-gneiss. These
conditions are superseded by the hornblende-producing conditions of the Meso zone,
which were finally superseded again by the Epi zone conditions of recrystallisation
which now characterise the specimen.

GABBRO SCHISTS AND GNEISSES.

In addition to the typical amphibolites, the collection contains a number of
handsome hornblendic rocks of apparent coarse grain. They possess the same
mineralogical variety and composition as the amphibolites. In many cases they
possess the same average grain size, but the minerals are grouped in clusters, so that
the hand specimen assumes a mottled black-and-white character, with the appearance
of a coarse-grained rock. In these cases they are amphibolites, with the glomero-
plasmatic structure described by Loewinson-Lessing, and Holmes. 1 The mineral
grouping indicates either (1) that each cluster has been derived from an original
crystal unit, or (2) that each cluster is a minute metamorphic differentiation. The
average regularity of the areas throughout the hand specimens renders the first
explanation the more probable. In some cases this explanation is confirmed where
the white areas have retained the form of felspar crystals, and the rocks are metamorphic
representatives of gabbros. Some of the types may have been placed in order among
the amphibolites, but the " relic " gabbro structure of the hand specimens has favoured
their consideration under the heading of gabbro schists and gneisses.

1 Trav. Soo. Nat. St. Peterab., vol. xxv (1900), p. 208; A. Holmes, Q.T.G.S. Ixxiv, 1918, p. 55

AMI'!llltnl.]TK> AMI KKI.ATKD !!( K- - I I I.I.\V Kl.l..

Specimen No. 915 is an example of a garnet-biotite-amphibolite with glomero-
phisnutic structure, and is similar to No. 50 (PI. XLI, fig. 1). In section, it is found
that the dark areas of hornblende crystals consist of small lenticular clusters, together
with biotite and ilmenite. The white areas are granoblastic aggregates of basic
plagioclaae. A similar specimen is No. 990, in which the brown liiotite is replaced by
ii <-hlorite. and more of the felspar is saussuritised.

No. 339 is a garnet -bearing gabbro-gneiss (PI. XLI, fig. 2). It possesses a coarse
gneissic structure, and has a mottled appearance. The black areas consist of
aggregates of hornblende and augite, or garnet and biotite, but the four minerals are
contained in some aggregates. The white areas, which are slightly discoloured by
weathering, consist of aggregates of labradorite, partly saussuritised, and a little
quartz. The biotite aggregates contain a little lawsonite, and their rounded form
suggests their derivation from garnet 1 . The suggestion is supported by the
presence of quartz, which can be formed with the change from garnet to biotite. The
hornblende has been derived from augite, while the granular form of the metamorphic
pyroxene suggests its derivation from large platy pyroxene crystals. Ilmenite and
apatite are accessory minerals. The microscopical features are similar to those of the
hornblende-pyroxene-gneisses, with the addition of the glomeroplasmatic structure.

A specimen in which the gabbro structure is very prominent is No. 593, whose
character would be well described as a hornblende-gabbro. It is a coarse, handsome
rock, containing some large crystals of labradorite, though consisting for the most part
of coarse hornblende and felspar. Occasional crystals of green augite are associated
with the hornblende, and a few fragments of garnet are found near plates of biotite.
In thin section the labradorite has a tendency to idiomorphic outlines a contrast
to the granoblastic aggregates of many metamorphic rocks. The twin lamella; of the
labradorite are very broad and much more abundant than in previous types, but there
is a small amount of untwinned sodic felspar. The hornblende forms large crystals,
but in other respects is similar to the smaller crystals in the amphibolites. Augite is
present, and in one case forms the nucleus of a large hornblende crystal. Large
crystals of ilmenite, with sphene rims, accessory apatite, and zircon are present.
While the nature of the rock is understood by the name " hornblende-gabbro," the
rock possesses metamorphic traits in the presence of relic garnet, of secondary felspar,
saussuritised felspar, and the characteristic relation of sphene and ilmenite.

In specimen No. 367 some of the felspar areas are 2 cm. wide, and the rock has
the appearance of an augen gneiss. The section shows that the felspars have
recrystallised as granoblastic aggregates of calcic andesine, fairly free from saussurite.
ii-se crystals of epidote are present. Some of the biotite is replaced by chlorite
and is intergrown with lawsonite. The Epi zone features are, however, subordinate
in this example, and the rock may be called a plagioclase-gabbro-gneiss. The
plagioclase is thoroughly saussuritised in a somewhat similar specimen, No. 969
(PI. XLI, fig. 3).

Op. oit., p. 167.

270 AUSTRALASIAN ANTARCTIC EXPEDITION.

Specimen No. 317 is a coarse-grained rock, with a mottled black-and-white
appearance, due to glomeroplasmatic areas of hornblende and felspar of average width
of 3 to 4 mm. Each dark hornblende mass resolves itself on examination into clusters
of small crystals, associated with crystals of chlorite, ilmenite, and apatite. This
ilmenite may have been discharged from the original pyroxene in the manner
recorded in the garnet-plagioclase-pyroxene-gneiss, No. 935, from Stillwell Island 1 .
Biotite is replaced by chlorite and epidote. In the white areas are large saussuritised
felspars, in which sericite, white mica, epidote, chlorite, and zoisite are found,
in addition to clear albite. There are also areas of clear labradorite, and a few
scattered fragments of garnet. The rock may be distinguished as a saussurite-gabbro-
gneiss.

Specimen No. 728 is another variety of saussurite-gabbro-gneiss. Its felspar is
completely saussuritised, and there is a little garnet and a small development of
secondary quartz in the hornblende, probably associated with the development of
epidote and chlorite. The specimen is not schistose, and the outline of the original
felspar crystals is discernible in the hand specimen, averaging about 6-7 mm. long
and 2-3 mm. wide. The form of the felspar is even better preserved in No. 861, where
it is sufficiently obvious to suggest that the hand specimen is a hornblende-felspar-
porphyry. There is, however, complete saussuritisation of the felspar, while the
biotite and some of the hornblende is converted into epidote and chlorite. Scattered
garnets occur around the fringe of some of the hornblende clusters, and are suggestive
of the garnet roms around the pyroxene in the garnet-pyroxene-gneisses of Stillwell
Island.

The conversion of hornblende into chlorite is further advanced in specimen
No. 515, another saussurite-gabbro-gneiss. The chloritisation of the hornblende
proceeds from the edges and cleavage cracks. The section is crossed by branching
veins of lawsonite (PL XL, fig. 2), which have developed during the saussuritisation of
the felspar. These veins cross hornblende as well as felspar areas. Zoisite occurs
with the epidote in the altered areas. There are, however, considerable areas of
unaltered basic felspar.

Apart from the gabbro gneisses and the amphibolites that have been described,
there are a number of coarse-grained hornblendic rocks. With the segregation of the
lighter-coloured constituents, these rocks become representatives of the hornblende-
plagioclase-gneisses of Group III. With the segregation of the darker constituents,
they grade towards the hornblende schists of Group V. Specimen No. 721 (PI. XLI,
fig. 4) illustrates a segregation assuming the form of a vein. The vein contains large
crystals of hornblende, some of which are 3 cm. long, together with felspar and a little
quartz. A similar vein occurs on the edge of another specimen, No. 25, in which there
is a noticeable increase in grain size of the amphibolite as it approaches the edge of the
vein.

1 Op. oit., p. 174.

AMPHIBOL1TES AND KKI.ATKI) l >. K> - I I l.l.\\ KLL. 271

III. GROUP V.-MACNKSHM S1IJCATK (JNEISSKS.

As the percentage of felspar decreases the members of the Amphibolite Group
grade into the rocks of Group V, the magnesium silicate gneisses.

KATA DIVISION.

Assigned to the Kata division of Group V is Specimen No. 527, an apple-green
rock, consisting essentially of granular green pyroxene. It consists, in section, of a
granulitic mass of clear pyroxene, resembling oniphacite, with a subordinate amount
of very pale green to colourless hornblende, together with a little interstitial quartz.
Garnet is absent; and, otherwise, the rock would be classed as an eclogite in Group IV.
The development of hornblende is irregular, and sometimes is accompanied by a little
biotite. Portions of the specimen possess a distinctly dark-green shade, due to an
increased amount of hornblende. One portion of the specimen possesses a noticeable
percentage of red felspar, indicating the relationship with the amphibolite group.
This particular specimen is traversed by a vein of quartz and plagioclase, 2 cm. wide,
in which the quartz is mostly confined to the cent re of the vein, and the plagioclase
to the walls of the vein.

In Specimen No. 963 the hornblende is much more important, and the hand
specimen possesses a corresponding dark colour. This specimen was obtained from an
erratic on Cape Hunter. The main constituent of the rock is pyroxene, including
both augite and hypersthene. Some of the crystals are clear and granular, and others
are fibrous. All show more or less alteration to hornblende. There is a small
development of brown biotite accompanying the hornblende, subordinate plagioclase,
and scattered grains of ilmenite and pyrite.

These rocks are unfoliated, and therefore possess resemblances to the ultra
basic igneous rock " pyroxenite " ; but the pyroxene possesses the distinctive
character of the granular pyroxene of metamorphic origin commonly observed in the
Cape Gray series of altered basic dyke rucks.

MESO DIVISION.

As hornblende becomes the essential constituent the types are grouped in the
Meso division of Group V. Among these is Specimen No. 594, a garnet-hornblende-
schist. It is a coarse-grained, dark-green rock, which consists of green hornblende
and scattered pink garnets. In addition, there is a little interstitial felspar, as well
as scattered grains of ilmenite and apatite.

Specimen No. 513 is a heavy, dark-green, slightly schistose rock, consisting
essentially of amphibole, with scattered specks of pyrite visible in the hand specimen.
In thin section there are large talcose areas produced by the decomposition of some

272 AUSTRALASIAN ANTARCTIC EXPEDITION.

pre-existing crystals. These contain numerous inclusions of magnetite and amphibole
The magnetite appears to have been a separation product, accompanying the
formation of the talc. It is distributed as fine particles along traces of former cleavage
planes, and along a system of cracks more or less at right angles to the cleavage. In
many places these particles have coalesced to form irregularly -shaped inclusions of
magnetite. The arrangement of the discharge of magnetite suggests that the original
crystals were olivine, though the nature of some of the included amphibole indicates
the possibility of pre-existing pyroxene, partly altered to hornblende before the
curious alteration to talc. The amphibole is the most abundant mineral, and
includes green hornblende and anthophyllite. The latter has developed in long
prismatic crystals, which possess faint pleochroism from colourless to very light
brown. The cross sections show the usual hornblende cleavage and outline, and the
prismatic sections possess straight extinction. Sometimes there is a green core of
hornblende, with a colourless border zone of anthophyllite. Sometimes the hornblende
is decolourised, and there appears to be a gradation from hornblende to anthophyllite.
In addition to these minerals, the rock contains a minor amount of biotite, with
pleochroism from a pale-straw brown to a bright green, and of magnetite with
accessory grains of pyrite, apatite and quartz. The rock is an anthophyllite-schist,
on which has been superimposed some of the features of the Epi division. These
features include the development of talc, the bent and crushed appearance of some
of the biotite crystals, and the occasional fracturing and mortar structure in the
amphibole.

Specimen No. 516 is another representative of the anthophyllite-schists, and,
as in No. 513, there are only occasional grains of felspar. It is a dark-green schistose
rock, showing a brownish tinge on the weathered surface, due to the slight oxidation
of anthophyllite. In thin section, the green hornblende is the most important
mineral, and the colourless to light-brown anthophyllite is intergrown with it. The
talcose areas of the preceding specimen are absent, but some of the anthophyllite
crystals show decomposition around their edges and cleavages to a talcose material.
Brown biotite is much more abundant than in No. 513, and helps to produce the
schistose structure. Accessory grains of apatite, ilmenite, pyrite and quartz are
present.

Another representative of the anthophyllite-schists is Specimen No. 109, in
which the green hornblende is absent. This specimen is a light-brown rock, in which
the prisms of anthophyllite are arranged with their long axes in the plane of schistosity.
Some of these crystals are over 2cm. long, and they lie in all positions in this plane.
Plagioclase can be observed in the hand specimen in a proportion which makes the
inclusion of this specimen in this group doubtful. In section the anthophyllite is
slightly pleochroic from colourless to pale brown, showing the well-developed cleavage
and outline of the amphiboles. The plagioclase is labradorite, and there are odd
flakes of biotite and a few grains of magnetite (or ilmenite). The rock could possibly
be placed in the group of the amphibolites.

AMPHTBOLITRs AND RELATED R< Kfi III I \\ II I 273

A typical lioinUende-lcIs is \<i. ;>|s. a Muck, heavy, massive rock. It is fairly

se-grained, and in the hand specimen appears to consist wholly of hornblende.

In section, it is composed chiefly of fresh green hornMende, containing a few inclusions

of rounded quart/ blebs. Quart/ is the more important minor constituent, and there

in addition, accessory flakes of l>iotite. grains of felspar, calcite, apatite, and

ilmenitc.

Specimen No. 950 is a biotite-hornblende-schist . \\liich is closely related to
the biotite-amphibolites. It contains felspar and a little quart/, but in less amount
than in the typical amphibolites. The brown biotite is abundant, but not uniformly
distributed through the rock. It forms clusters, in which some of the micaceous
laminae are bent and twisted. Occasional grains of epidote are associated with the
biotite. Green hornblende is the most abundant constituent, and both the hornblende
and biotite possess pleochroic haloes, which surround minute inclusions.

A curious hornl)lende-fels is No. 941. It is a dark-green, massive rock with a
porphyroblastir appearance, due to platy crystals of hornblende, which are set in a
fine matrix. The examination reveals the large hornblendes as relic crystals which
have for the most part broken down into granulitic aggregates of a similar pale-green
hornblende (PI. XL, fig. 3). The latter forms the bulk of the thin section. A little
colourless chlorite and a little brown biotite are mixed with the granulitic hornblende.
Embedded in these aggregates are a few long, slender prisms with ragged terminations,
which are feebly pleochroic from colourless to pale green. Some show lamellar
twinning, and they possess straight extinction with yellow and grey polarisation colours
of the first order. They show cleavage, and appear to be a form of orthorhombic
amphibole. The remaining constituents are ilmenite dust, and grains of pyrite and
apatite. It would appear as if the original platy hornblende had been chloritised and
the reaction had then been reversed, reproducing the fine granulitic hornblende from
the chlorite.

Another dense hornblende-schist is No. 209, in which a subordinate amount of
liiotite, for the most part converted into chlorite, is associated with the pale bluish
green, granular hornblende. It contains large talcose areas, similar to those in No. 513,
produced by the decomposition of pre-existing olivine or pyroxene. These areas form
about half of the section, and contain magnetite in a similar manner. They are
studded with inclusions of hornblende and chlorite, but there is no trace of the
original mineral.

The family of actinolite-scnists is represented by specimen No. % J4<5, which
displays in the hand specimen a bright green mass of acicular crystals. Its appearance
forms a distinct contrast to that of the dark greenish black hornblende-schists. In
thin section it consists almost entirely of a mass of long thin a-tinolite prisms, which
are pale green to colourless. A few flakes ot biotite and a few scattered grains of
ilmenite are the only other minerals present.

274 AUSTRALASIAN ANTARCTIC EXPEDITION.

EPI DIVISION.

A larger development of chlorite occurs in specimen No. 229, which may be
called a chlorite-hornblende-schist. The formation of chlorite has been accompanied
by the development of a little epidote, and the separation of a little quartz. In places
the ends of the hornblende crystals are frayed out into needles embedded in quartz.
Some of the hornblende crystals show the effects of crushing, with an alteration toward
chlorite. Ilmenite is disseminated through the section, and there are occasional grains
of untwinned felspar. The marked development of chlorite is a feature associated
with the Epi division of this group, in which this rock has been included.

Another specimen of hornblende-schist, No. 888, contains a curious plate of
chlorite, which is composed of layers of chlorite. In each layer the micaceous laminae
of chlorite are arranged more or less at right angles to the trend of the layer ; but the
laminae in adjoining layers are set at varying angles in one another. Both chlorite and
hornblende in this specimen possess a very pale green colour, being nearly colourless.

A specimen that can be placed in this division is No. 916, a dense, dark green
pebble bearing glacial striae. It is a close relative of the epidote-chlorite-schists of
Group IV, but it is distinguished by its very subordinate amount of clear felspar. It
consists chiefly of epidote and chlorite. The iron, which has been discharged during
the conversion of biotite and hornblende into epidote and chlorite, is disseminated
throughout as numerous small particles of ilmenite. A little zoisite is associated with
the epidote. Sphene and apatite are not observed, but white mica is a minor constituent.
The rock is an epidote-chlorite-schist ; and, since chlorite is the more important
constituent, it is included in the family of chlorite schists. If epidote had been more
important, the rock would be related to the epidosites of the Epi division of Group
IX, the lime silicate gneisses.

Specimen No. 931 is a fine grained, dark-green, schistose type with thin bands
of light green epidote. The percentage of felspar and quartz is very small, and even
less than in the preceding type No. 916. The main mass of the rock consists of epidote
and hornblende, chlorite being a very minor constituent. In addition to the more
important pale green hornblende, there is a clear colourless hornblende, and both
varieties are intergrown in one crystal. The colourless hornblende shows the brighter
polarisation colours and a slightly larger extinction angle. A colourless pyroxene is
present, partially altered to hornblende, and more prominent in some layers than
others. Granular epidote is abundant through the section, and has in part segregated
into layers. The small percentage of felspar and quartz is mostly confined to
these layers. Associated with the epidote crystals are a number of zoisite crystals,
possessing grey and ultra-blue polarisation colours. Sphene is an important accessory,
and forms thin strings of crystals in the layers of epidote. One corner of the section
is crossed by a narrow vein of serpentine. The rock is an epidote-hornblende-schist,
and the development of epidote, almost to the exclusion of chlorite, produces a
resemblance towards the epidosite family.

AMPHIBOL1TES AND RELATED K<>rK- -Til. I. \VKLL 275

IV.-GROUP III.-THE PLAGIOCLASE GNEIS8E8.

With a decreased percentage of hornblende and an increased percentage of
felspar, the amphibolite types pass into less basic rocks which correspond to the
plagioclase gneisses of Group III. In many specimens quartz becomes a more important
constituent and a considerable variety of rocks, including some transitional varieties
between Groups I and IV. are placed hen-.

KATA DIVISION.
Xone of the specimens examined appear to fall into this group.

MESO DIVISION.

Among the hornblende-plagioclase-gneisses are specimens Nos. 169, 170, which
are light-coloured, coarse-grained rocks, in which the hornblende appears as small
schlieren in the light-coloured base, and forms about 16 per cent, of the rock. The
base contains quartz and alkali felspar in addition to the plagioclase. These specimens
also contain fragments of much larger schlieren composed of dark amphibolite or green
pyroxene. In the proximity of the amphibolite schlieren there is a band containing
epidote, ealcite and lawsonite along which the hornblende is partially altered. The
alteration (PI. XL, fig. 4) takes place first to a colourless hornblende and then into an
aggregate of epidote, lawsonite and calcite. The pyroxene in the pyroxenic schliere
shows alteration to hornblende, and there is in addition a separation of calcite when
fragments of pyroxene are set in a network of calcite.

Specimen No. 378 is a pink hornblende-felspar-gneiss, whose colour is due to
the pink felspar modified by the dark crystals of hornblende and mica. In section,
it has a granoblastic structure and a schistose texture. The colourless constituents
include quart/, clear microcline, plagioclase with accessory apatite. The dark
constituents are green hornblende, greenish-brown biotite, which is partly altered
lu chlorite and epidote, while split-lie is an abundant accessory mineral. The rock
may be called a hornblende-felspar-gneiss, and on account of its content of quartz
and alkali felspar it probably has a chemical composition grading towards that of
Group I, the alkali felspar gneisses.

No. 269 is a hornblende-plagioclase-gneiss, which is more basic than No. 378,
and probably grades in the opposite direction towards the group of amphibotites. No.
378 contains pink felspar in the hand specimen, but No. 259 possesses more hornblende
and white felspar. No. 259 also possesses a coarser grain size, and some of the
hornblende crystals are 4mm. wide. In section, the large hornblende crystals are
green and bluish-green in colour, ragged in outline, and frequently contain inclusions
of biotite. Some are partially broken across the biotite inclusions, and epidote haa
developed in places along the strained portions of the hornblende. Biotite is very

66-2.

276 AUSTRALASIAN ANTARCTIC EXPEDITION.

subordinate to hornblende, and epidote has developed from the crushed biotite. Some
of the epidote is finely granular, and some of it occurs in grains and crystals. Quartz
is subordinate to felspar, which consists of an acid oligoclase and untwinned felspar.
The felspar possesses myrmekitic structure, but it is free from products of decomposition.
Quartz and some of the felspar shows strong undulose extinction. Accessory grains of
apatite, allanite, and sphene are present.

Specimen No. 240 is a biotite-plagioclase-gneiss, and another example with
affinities to the amphibolite group. It is a grey schistose rock, in which the dark
constituents, biotite, and hornblende are subordinate to the white minerals. In
section, the flakes of brown biotite are arranged parallel with the schistosity, and are
in large excess over the green hornblende. There is a small amount of green chlorite
and colourless lawsonite interlaminated with the biotite. Quartz is present, and the
andesine felspar is quite clear, except for a few crystals which possess cloudy micaceous
alteration products.

An exceptional biotite-hornblende-felspar-gneiss is specimen No. 956. The
hand specimen is a dark schistose rock, showing glistening biotite, and is finer grained
than most of these rocks. In section, the hornblende is quite distinct from the
granular green variety of the typical amphibolites. It occurs in ragged, acicular
crystals, and the colour of the prisms is mostly green and pale green. Some prisms
show a pleochroism from greenish-blue to blue, indicating a change to glaucophane.
In addition to the large crystals, it also occurs as fine acicular needles. Its polarisation
colours are unusual. In many cases there is no position of absolute extinction
between crossed nicols ; but a position of 30 to 40 to the cross wires shows ultra-blue
colours. While the polarisation colours appear to be low, they are interfered with
by the colour and dispersion of the crystal. The biotite possesses an unusual colour,
which changes in cross-section from a reddish-brown to a biscuit-brown. The crystals
possess an unusually ragged outline, and contain minute dots of discharged ilmenite.
The amount of felspar appears to be a little greater than in the typical amphibolites.
The felspar crystals are mostly untwinned, some possess traces of microcline twinning
and some are lamellar twinned crystals of acid oligoclase. Calcite, ilmenite, and epidote
are minor constituents, and pyrite and apatite are accessory minerals. While the
rock is more basic than typical members of this group, it is at the same time distinct
from the amphibolites.

A different type of hornblende-quartz-plagioclase schist is No. 51. This is a
grey schistose rock, in which numerous prismatic needles of hornblende and occasional
small crystals of ilmenite are set in a grey groundmass. In section, the rock consists
of thin, elongated crystals of green hornblende, set as porphyroblasts in a finely
granulitic groundmass of quartz and felspar. The hornblende crystals are ragged
with the pronounced prismatic habit. Its colour is dark green to bluish-green.
Numerous small and isolated flakes of biotite have developed along some of the
schistose layers. The ground mass, in which the quartz appears to predominate is

AMPHIBOLITES AND RELATED ROCKS-STILLWELL. 277

very fine; and there are thin lenticles of felspar arising from it. Some of this felspar
is imr\viiim><l ; l>ut some lamellse show an extinction angle of 33 indicating a labradorite.
There are scattered crystals of ilmenite and apatite. This type of schist is distinct
from the other hornblende-plagioclase-gneisses, and the quartzitic nature of the ground
nniss suggests its derivation from a sedimentary rock.

Specimen No. 924 is another type of hornblende-plagioclase-schist, which is
also more basic than typical members of this group. The base of the rock consists
of densely-packed acicular needles of hornblende, with grains of epidote set in a
colourless base of clear felspar and quartz. The hornblende needles have a pale green
colour and an extinction of 20. The rock also possesses relic phenocrysts of quartz
and felspar, and in this way resembles a " porphyroide." Some of the felspars have
an extinction of 27 J indicating labradorite and some are studded with numerous small
crystals of epidote. These relic crystals indicate that the rock has developed from a
porphyritic igneous rock, probably a felspar porphyrite. The abundance of epidote
is a feature belonging to the Epi division of this group.

EPI DIVISION.

This division includes the group of the epidote felspar gneisses. Some of these
appear to grade into the Epi division of Group I, the Alkali-Felspar Gneisses, while
others grade toward the Epi divisions of Groups IV, V, and IX, the Amphibolites and
Edogitea, the Magnesium Silicate Gneisses and the Lime Silicate Gneisses.

Typical of these epidote gneisses is specimen No. 257, in which the colour of
the specimen is dominated by the green epidote. It is a green schistose rock with a
banded structure and a much finer grain than preceding types. It possesses a distinct
crystallisation-schistosity, and some of the bands possess a darker green colour than
the average, on account of the association of chlorite and hornblende with the epidote.
In section, the light coloured bands consist of granoblastic masses of clear felspar, with
scattered flakes of pale chlorite and epidote. The felspar consists of microcline felspar
and albite. In some bands epidote (with a little zoisite) is the most important mineral;
while in others a pale green hornblende is as abundant as the epidote. The hornblende
represents the residual mineral, which has survived the general alteration into epidote
and chlorite. Its colour is bleached prior to this alteration; and areas of residual
green mineral ;ire observed, surrounded by colourless hornblende partly converted
into chlorite. Sphene and apatite are present, and a few epidote grains possess a
reddish-brown nucleus of allanite.

In ;i similar specimen, No. 157, some of the bands of pink felspar are broader,
being as much as 5mm. wide, while others are darker and possess a larger percentage
of hornblende. In this case, and in No. .VJ-J. quartz is present. In No. 522, the
pale green hornblende changes into a colourless horn Monde prior to its alteration to
epidote and chlorite. Many of the epidote grains contain a reddish-brown nucleus
which is probably allanite.

W22 D

278 AUSTRALASIAN ANTARCTIC EXPEDITION.

A coarser-grained type of the same class of rock is No. 726, in which the pink
bands of felspar are 2^ cm. wide. No. 726 is a handsome rock, whose colour is dominated
by the pink felspar, but contains disseminated patches of bright green epidote. In
addition to the pink bands, there are bands of green chloritised hornblende with a silky
sheen. In section, there is a little cloudy untwinned felspar, which may be orthoclase,
and lamellar-twinned oligoclase-albite. The development of epidote within some
of the plagioclase indicates that it has been partly derived from felspar. Pale
decolourised hornblende is present, and the abundant chlorite and some of the epidote
has been derived from it. Quartz is absent in this case, but large grains of apatite
and zircon are present.

Of different appearance is specimen No. 598, a pink and green massive boulder
containing large porphyroblasts of pink microcline up to 3cm. in length and over 1cm.
in width, set in a rather coarsely-crystalline base of pink felspar, bright green epidote,
and dark-green chlorite. The general appearance is somewhat similar to a pink
porphyritic granite, except that there is a general absence of quartz. The resemblance
is more strictly to a felspar-porphyry, from which it has probably been derived. The
large porphyroblasts show the irregular wedge-shaped twinning of microcline, with
small inclusions of quartz, plagioclase and sphene. The character of porphyroblasts
has been confirmed by a determination of the alkalies by Mr. J. C. Watson, in the
Victorian Geological Survey laboratory, which gave :

K 2 14-04 per cent,

Na 2 -66

The felspar in the base of the rock consists of lamellar-twinned plagioclase
and subordinate untwinned felspar. Sometimes it shows cataclasis and a partial
development of myrmekitic structure, and sometimes bent lamellae. Its extinction
angle measures up to 14 and it is probably an oligoclase. The green chlorite and
epidote are bunched together in aggregates, and of these two minerals, epidote is the
more abundant, forming large idiomorphic crystals. Sphene and apatite are accessory
minerals, and the rock can be classed as an epidote-felspar-gneiss. The presence of
large crystals of microcline indicate a gradation towards the group of the alkali felspar
gneisses (Group I).

A related rock is No. 954, in which the porphyroblasts of microcline are much
smaller. The grain size is less than half that of No. 598, and the appearance is more
granulitic. The epidote and chlorite are more or less evenly distributed through the
section, producing the darker colour as opposed to the pink colour of the preceding
specimen. The aggregated amount of epidote and chlorite is less than in previous
examples, and the chlorite is more abundant than epidote. Crystals of sphene
accompany the epidote. Microcline occurs in base as well as forming the porphyro-
blasts, and contain numerous inclusions of albite, with a ragged and irregular shape,
after the manner of graphic structure. There is a considerable amount of untwinned

AMPHIBOLHK> AND RELATED R<"H- Mil I \\KLL. 279

orthoclase, and a subordinate amount of plagioclase. The abundance of microcline
and the dec Teased amount of epidote bring the rock into the group of the alkali felspar
gneisses, and it may be called a chlorite-microcline-gneisB.

Another porphyroblastic epidote-felspar-gneiss is No. 590; but it differs from
the two preceding specimens in possessing a marked schistose structure. It consists
of porphyroblasts of pink felspar, set in a base coloured green by epidote. The
porphyroblasts consist of granoblastic aggregates of untwinned and microcline felspar.
The <-loar base consists of felspar with a few grains of quartz and a subordinate
amount of plagioclase. Coarse crystals of epidote are abundant and grouped in
(luster.-, witli chlorite and a little zoisite. llmenite has been altered to leucoxene.

Specimen No. 910 is a different type, and may be described as an epidote-felspar-
breccia. It is a very dense, fine-grained, pale-greenish rock studded with minute
porphyroblasts of felspar. The thin section possesses a uniform pale-green colour,
and consists of angular crystals of felspar set in a dark semi-opaque mass of epidote.
The general appearance is that of a breccia and the felspar is water clear, partly twinned
and partly untwinned. Some of the lamellar-twinned individuals have an extinction
angle of 16. Here and there are larger crystals of epidote, which are more transparent
than the base, and consequently show brilliant polarisation colours. Under high
powers the base of the rock is seen to consist of closely packed minute grains of epidote
set in a base of clear felspar.

Specimen No. 592 is a massive, pale-greenish rock, somewhat similar in appearance
in section to the coarse-grained variety, No. 726. It contains coarse crystals of partially
altered honibjende, epidote (sometimes intergrown with allanite), chlorite and felspar.
The rock is brecciated, and the fragments are separated by veins of quartz and felspar,
darkened by finely granular epidote.

Specimen No. 591 is an epidote-felspar-gneiss, which has a pale-green colour
and a schistose structure. In section, it shows the helical structure commonly seen
in phyllites and mica schists. There are a few small felspar porphyroblasts, some of
which are cracked and broken, and show various stages in breaking down into a finely
granulitic mass. Some felspars are untwinned, and some are acid plagioclase. There
are also a few porphyroblasts of allanite. The bulk of the rock consists of the finely
granulitic base enveloping the wavy bands of dark granular epidote, similar to that in
No. 910. Chlorite is present, and interwoven with the granular epidote, and there
are occasional crystals of sphene. The specimen assumes a dark colour on the one side,
where the amount of felspar decreases and the rock passes into an epidosite.

A further stage of epidotisation exists in specimen No. 589. This specimen
rot.iins the schistose structure and possesses traces of green hornblende. The felspar
is subordinate, and there are bands of well crystallised epidote, as well as bands of
finely granular epidote. Lawsonite is also present, and has been derived from the
feNpar. In addition there is a little zoisite and abundant sphene.

280 AUSTRALASIAN ANTARCTIC EXPEDITION.

With the decrease in the percentage of felspar, the colour of the rock deepens,
and a dark-green type, like No. 599, is produced. It differs from Nos. 157, 257, and 522,
in possessing a massive structure. It has an increased percentage of chlorite, and
possesses relationships with the epidote-chlorite-schists of Group IV. Epidote is th e
most abundant mineral, and, together with the decrease in felspar, indicates a
transition towards the epidosites, the Epi division of Group IX, the lime silicate schists.
Chlorite is much more abundant than in previous specimens, and there is only a trace
of hornblende. Quartz is present, but subordinate in amount to the felspar. Both
quartz and felspar show cataclasis, and some crystals appear in polarised light as fine
granulitic aggregates. Ilmenite is present, and its association with sphene is another
point of resemblance with the amphibolite group.

With the decrease of both chlorite and felspar, a massive green rock of
epidosite is produced. No. 255 is an example of this class, consisting chiefly of epidote.
The section has a uniformly yellowish-green colour, due to the epidote. Quartz is present
but very subordinate. A little pale hornblende exists as relic crystals, and a little
chlorite is intergrown with the epidote. Sphene, with associated magnetite, is an
abundant accessory mineral, and a little calcite is present. This rock is a member of
the Epi division of Group IX, the lime silicate rocks.

V. DESCRIPTION OF PLATES.

PLATE XL.
Fig. 1. No. 520, an amphibolite in which the hornblende shows a mottled appearance,

due to the separation of blue hornblende from the green hornblende. Mag.

35 diams.

Fig. 2. Veins of lawsonite traversing a gabbro gneiss No. 515. Mag. 35 diams.
Fig. 3. Amphibolite No. 941, in which granulitic hornblende has developed from

large crystals of hornblende. The nicols are crossed, and the large crystal

which occupies most of the field is nearly in a position of extinction, while

the granulitic hornblende appears light due to its bright polarisation colours.

Mag. 35 diams.
Fig. 4. No. 169, in which a crystal of hornblende is partially altered to a mixture of

calcite, epidote, and lawsonite. Mag. 35 diams.

PLATE XLI.

Fig. 1. No. 50, hornblende-plagioclase-gabbro gneiss.
Fig. 2. No. 339, garnet-hornblende-plagioclase-gabbro gneiss.
Fig. 3. No. 969, saussurite-gabbro gneiss with augen of saussurite.
Fig. 4. No. 721, segregation vein of hornblende, felspar and subordinate quartz in
amphibolite.

[\Vith Two Plates.]

Sydney: Alfred James Kent, Government Printer 1923.

AUSTRALASIAN ANTARCTIC EXPEDITION.

SERIES A. VOL, III. PLATE XL.

Fig. 1.

Fig. 2.

Fig. 3.

Fig. 4.

;

AUSTRALASIAN ANTARCTIC EXPEDITION.

SERIES A. VOL. III. PLATE XUf

Inches.
F.g. 1.

Inches.

Fig. 2.

Inches.
Fig. 3.

Inches.
Fig. 4.

* * .

AUSTRALASIAN ANTARCTIC EXPEDITION

1911-14.

UNDER THE LEADERSHIP OP SIR DOUGLAS ttAWSON. O-B.E, B

SCIENTIFIC REPORTS.

SERIES A.
VOL. III.

GEOLOGY.

FART V.

MAGNETITE GARNET ROCKS FROM THE MORAINES,
CAPE DENISON, ADELIE LAND,

ARTHUR L. COULSON. /A.Sc. F.G.5.

WITH TWO PLATES.

PRICE: TWO SHILLINGS

Jinx* Kt. Gortrii Prfattr, Phil' 'nty-

ISSUED NOVEMBER, 1925.

MA40

Series A.

VOL. PRICE.

s. d.
I. CARTOGRAPHY AND PHYSIOGRAPHY.

This Volume will deal with the fixation of a fundamental longitude for Adelie Land by means
of " Wireless " time signals ; also with Cartography and Geographical discovery

II. OCEANOGRAPHY.

PART r. SEA-FLOOR DEPOSITS FROM SOUNDINGS.

By FREDERICK CHAPMAN, A.L.S. (Lond.),F.R.M.S.,&c., National Museum, Melbourne 060
2. DREDGING OPERATIONS. (In preparation.)
3. SOUNDINGS. (In preparation.)

4. SEA- WATER TEMPERATURE OBSERVATIONS. (In preparation.)
5. SALINITIES. (In preparation.)
6. TIDAL OBSERVATIONS. (In preparation]

III. GEOLOGY.

PART i. THE METAMORPHIC ROCKS OF ADELIE LAND.

By F. L. STILL WELL, D.Sc., Aust. Exped. Staff 220
2. THE METAMORPHIC LIMESTONES OF COMMONWEALTH BAY, ADELIE

LAND. By C. E. TILLEY, B.Sc o I 6

3. THE DOLERITES OF KING GEORGE LAND AND ADELIE LAND.

By W. R. BROWNE, D.Sc., Lecturer, Sydney University o i 6
4. AMPHIBOLITES AND RELATED ROCKS FROM THE MORAINES, CAPE

DENISON, ADELIE LAND. By F. L. STILLWELL, D.Sc 020

5. MAGNETITE GARNET ROCKS FROM THE MORAINES AT CAPE

DENISON, ADELIE LAND. By ARTHUR L. COULSON, M.Sc., F.G.S. ... 020
6. THE GRANITES OF ADELIE LAND AND KING GEORGE LAND
7. THE SEDIMENTARY ROCKS OF ADELIE LAND AND JKING GEORGE

LAND.
8. PETROLOGICAL NOTES ON THE ADELIE LAND COLLECTIONS.

IV. GEOLOGY.

PART i. THE ADELIE LAND METEORITE.

By P. G. W. BAYLEY, F.I.C., and F. L. STILLWELL, D.Sc. o i 6
2. PETROLOGY OF THE QUEEN MARY LAND AND KAISER WILHELM

LAND COLLECTIONS.
3. PETROLOGY OF ERRATICS DREDGED FROM THE OCEAN-FLOOR OF

ANTARCTIC LANDS.

4. SUMMARISED REMARKS ON THE GEOLO'GY OF ANTARCTIC LANDS
VISITED.

V. GEOLOGY.

THE GEOLOGY OF MACQUAR1E ISLAND.

VT. GLACIOLOGY.

AUSTRALASIAN ANTARCTIC EXPEDITION

1911-14.

UNDER THE LEADERSHIP OP SIR DOUGLAS AAWSON, O.B.E., B.E., D.Sc. P.R.S.

SCIENTIFIC REPORTS.

SERIES A.

VOL. III.
GEOLOGY.

PART V.

MAGNETITE GARNET ROCKS FROM THE MORAINES,
CAPE DENISON, ADELIE LAND,

ARTHUR L. COULSON, A.Sc., F.G.S.

WITH TWO PLATES.

PRICE: TWO SHILLINGS.

Alfred Jtmet Kent. GortniMai Prinlr. Pblilit,-iir SjrdiMy

ISSUED NOVEMBER. 1925.

MAGNETITE GARNET ROCKS FROM THE MORAINES,
CAPE DENISON, ADELIE LAND,

ARTHUR L. COULSON, M.Sc., F.G.S.

(WITH TWO PLATES).

CONTENTS.

PAGE.

I. Introduction 281

II. Petrography of the Magnetite Garnet Rocks

1. Mineral Composition 282

2. Description of Rock Specimens 283

3. Chemical Characters 292

4. Summary of Petrographic Characters 293

III- Position in Grubenmann's Classification of the Crystalline Schists ... 296

IV. General Discussion 299

V. A Tourmaline-bearing Magnetite Gneiss 301

VI. General Summary and Acknowledgments 303

VII. Description of Plates 305

I. -INTRODUCTION.

The rock specimens, which are described below, are part of the collection obtained
by the Australasian Antarctic Expedition from the glacial moraines at Cape Denison,
Adelie Land. They are schists and gneisses which are chiefly characterised by an
abundance of either magnetite or garnet or both. One specimen of epidote-magnetite-
schist, No. 989, shows well-developed glacial striae.

The descriptions are presented in order of relative predominance of magnetite to
garnet. In the first two examples, garnet is absent but it forms 40 per cent, of the
rocks towards the end of the series. The paper is concluded with the description of an
interesting type containing tourmaline, magnetite, and cordierite.

282

AUSTRALASIAN ANTARCTIC EXPEDITION.

The terms " gneiss " and " schist " are used in the sense denned by Holmes.*
Grubenmann's usagef of the terms " texture " and " structure " is adhered to. In
this, the structure is understood to be determined by the form and relative size of the
components of the rock. He adds that the structures appear to be functions of the
chemical composition of a rock, and of the magnitude of the temperature and pressure
and the time interval during metamorphism ; and are also dependent on the strength,
ability, and speed of crystallisation of the minerals developed. The texture is the
spacial arrangement of the constituents, and he states that the textures are not so
much dependent on the nature of the rocks as upon external circumstances.

II. PETROGEAPHY.

1. MINERAL COMPOSITION.

Rosiwal determinations of some of the specimens have been made in order to
express quantitatively their mineralogical content. These measurements were made
on sections cut at right angles to the schistosity and the results are expressed in Table I.
Generally it was found expedient to determine the quartz and felspar together. In
Table I below, iron ore includes magnetite, ilmenite, pyrite, &c., while " p " signifies the
presence of a mineral in small quantities. A Rosiwal determination of a magnetite-
garnet rock from Broken Hill, New South Wales, is added for comparative purposes.

In some rocks, mostly of fine-grained nature, the magnetite percentage alone
was determined. These form Table II, and Table III gives the specific gravities of the
remaining specimens. The figures for the various minerals as stated in percentages
present.

TABLE I.

Rock number

149

304

814

245

288

102

348

j B.H.

Specific gravity

3-29

2-99

3-07

3-27

2-97

3-33

2-98

3-81

Iron ore

Garnet ...

23-0
12-2

11-9
7-9

14-7

12-6

17-9
15-6

6-2
6-3

21-2

40-8

8-4
20-5

42-5
23-5

Quartz ... ... . .

32-4J

36-0

29-0

19-8

Felspar ...

>58-3

V 65-1

> 65-8

^55-6

45-OJ

29-0

CorcGerite

Mica

5-5

14-7

6-6

9-2

9-7

13-1

Apatite ...

1-0

0-3

1-2

0-4

2-0

14-2

Epidote

0-5

Zircon ...

Calcite

* " The Nomenclature of Petrology." A. S. Holmes, London, 1920, p. 107.

t " Structur und Textur der Metamorphischen Gesteine." U. Grubenmann. Fort, der Min., Krjst. und Pet., Band II-
1912, p. 209.

| Approximate.

MAGNETITE . \IINKT ROCKS COUL80N.
TABLE II.

283

Rock number

678

912

294

889

Specific gravity

3-12

3-31

3-14

3-11

3-13

Iron ore

23-4

32-4

33-0

18-9

26-9

TABLE III.

Rock number

788

296 .

827 (A)

933

827

181

Specific gravity

3-97

2-93

2-94

2-93

3-00

2-83

3-26

2-96

2. DESCRIPTION OF ROCK SPECIMENS.

No. 765. This is a dark, heavy, and fine grained rock, composed of a mass of
magnetite, hematite, and quartz, through which run bands of quartz and minute
hematite flakes. Small fragments of the rock are strongly magnetic. Its specific
gravity is 3-97, the highest for the group.

Microscopically, the quartz occurs in very fine grains, which show strain
polarisation colours and which contain myriads of inclusions of micaceous hematite or
" eisenglimmer." While eisenglinmier occurs abundantly within the quartz bands,
magnetite is the chief constituent of the rock. The structure is granoblastic and the
texture is schistose. The rock is a Magnetite-Schist.

No. 926. This is a dark greasy-looking rock, extremely fine grained and
possessing a well-developed crystallisation-schistositA .

Biotite is one of the chief constituents, occurring in very minute flakes and
showing no trace of any alteration to chlorite. Magnetite is fairly abundant and with
quartz makes up the bulk of the remainder of the rock ; it is usually xenoblastic in
form and generally of small dimensions. Quartz shows strain polarisation and contains
minute inclusions of magnetite, and epidote. This latter mineral is disseminated through
the rock in minute roughly circular grains. Calrite occurs in moderately large masses
and surrounds epidote, quartz, magnetite and biotite ; the characteristic cleavage is
well developed. The structure is lepidoblastic to granoblastic and the texture is
schistose. The rock is a Biotite-Magnetite-Srlii.tf.

No. 576. This is a gneissic type in which magnetite, epidote, biotite, quartz, and
pink orthoclase can be recognised in the hand specimen. The structure is heteroblastir.
Magnetite is important and occurs in xenoblastic individuals with a roughly parallel
alignment. Garnets are rare; they cannot be recognised in the hand specimen and
are confined to certain bands in the section. Such as are present are the fragmentary

284 AUSTEALASIAN ANTARCTIC EXPEDITION.

remains of large crystals which have been mostly altered to chlorite, quartz, and mag-
netite. This is a common alteration in the suite of specimens and will be dealt with
later. The colour scheme of the- biotite is X, light green, Y and Z, dark green.
Sometimes it has altered to chlorite, frequently showing the indigo blue polarisation
colours of clinochlore, and magnetite. Quartz is abundant and shows cataclasis and
strain polarisation. Orthoclase felspar is common but much kaolinized and sericitised.
Epidote occurs in numerous granular aggregates, usually almost opaque, and a very
strong light is necessary to observe its characteristic polarisation colours. Apatite is
common as small rounded grains. The rock is Micaceous-Magnetite-Gneiss.

No. 912. This is a schistose type, dark-coloured, and easily cleaved. It is
extremely fine grained and it possesses a lustrous appearance owing to the presence of
white mica. Flakes of the rock are strongly magnetic.

Microscopically, the structure is lepidoblastic. Magnetite crystals, which
constitute nearly a third of the rock, are xenoblastic in outline, especially in the larger
grains. Occasional minute flakes of micaceous hematite (eisenglimmer) are recognised
and ilmenite is intergrown with the magnetite. Garnets are moderately abundant but
they are small and much fractured and contain inclusions of magnetite. Green biotite
is common and sometimes exhibits pleochroic halos which are too indistinct to measure.
Quartz is an important constituent and contains fluid and magnetite inclusions. The
fluid inclusions generally have an arrangement parallel to the schistosity. Scaly
sericite, chlorite, epidote, and accessory apatite form the remainder of the rock, which
may be called a Magnetite-Schist.

No. 296. This is a dark greasy-looking rock, extremely fine grained and
traversed by a few lenticular bands composed mostly of quartz and epidote. It possesses
a very definite crystallisation schistosity.

The structure is both granoblastic and lepidoblastic. Magnetite is extremely
abundant, occurring in little xenoblastic individuals with linear development and also
in larger porphyroblastic grains (up to 0-3 mm.). It is disseminated through the rock
as well as segregated in thin bands. Small red flakes of micaceous hematite also occur.
Garnets are very small and idioblastic, their averages absolute diameter being about
0-4 mm. Quartz 'is the most abundant constituent and reduces the specific gravity
(2-93) below the average. Biotite is present as small flakes, while epidote is restricted
to certain bands as xenoblastic grains. Felspar is absent. The rock is Quartz-
Magnetite-Schist.

No. 989. Macroscopically the rock is dense and dark-coloured, being traversed
by a few epidotic bands. Little porphyroblasts of magnetite can be seen in the ground
mass of the rock.

The structure is granoblastic and the texture is schistose. In most respects the
rock is similar to the preceding quartz-magnetite-schist, No. 296. The garnets,
however, show two kinds of alteration, one of which has produced chlorite and the

MMJNF.TITK i;\l;NKT ROCKS mn.SnV 285

other epidote. Both changes can be observed in all stages of completion. The biotite
is irregular in shape and frequently contains inclusions of recrystallised quartz. It has
no very definite relationship to the schistosity and shows all stages of alteration to
chlorite with concomitant production of magnetite. Epidote occurs in granular
aggregates in addition to that formed directly from the garnet, and the rock may be
called an Kpidote-Magnetite-Schist.

No. 527. (A). This is somewhat similar to the micaceous magnetite gneiss,
No. 576, in the hand specimen and shows a very contorted banding. Pink orthoclase
is prominent.

The main structure is granoblastic but relic, zig-zag and lenticular structures are
observed microscopically. Magnetite is abundant and mostly in irregular xenoblasts,
while a few flakes of micaceous hematite may be observed. A few garnet relics have
survived the alteration to quartx, magnetite, and chlorite. In some cases the outline
of the original garnet is preserved (Plate II, fig. 1). Much of the chlorite has been
derived from the garnet but part has arisen from the alteration of biotite with the
accompanying production of magnetite. Muscovite is sometimes intergrown with
chlorite Fine granular quartz grains contain inclusions with an arrangement parallel
to the schistosity. Orthoclase is present as large relic crystals which are simply
twinned and untwinned. Epidote and apatite are present. The rock is a Chlorite-
Maffnetite-Gneiss.

No. 933. The hand specimen is a well-laminated greenish rock with an
abundance of biotite, chlorite, and muscovite. Biotite and chlorite form flaky aggre
gates which give the rock a spotted appearance. Quartz occurs in convolute veins
which follow the schistosity for some distance and then break across it. The laminae
of the rock are curved at one end as the result of bending and shearing rock movements
which may have accompanied the infiltration of the quartz.

The structure is lepodoblastic but relic and cataclastic structures are also
present. Magnetite is abundant. Small garnets show the usual development of quartz,
magnetite, and chlorite, when examined under the high power, but not nearly to the
same extent as in No. 527 (A). Quartz occurs both in large grains in the vein and in
small grains with an elongation parallel to the schistosity. It exhibits strain polari-
sation and the larger grains show cataclasis. Biotite contains a few pleochroic halos
around minute zircons and muscovite is very abundant. Granular epidote is largely
intergrown with chlorite. The siliceous veins contain biotite and magnetite and relic
felspars which are much kaolinised and penetrated by nr.wly formed quartz. The
twin lamellae can sometimes be picked out (Plate II, fig. 2). The rock is a Mica-
Magnetite-Schist and has the lowest specific gravity of the group (2-83).

No. 294. This is an interesting magnetite-schist with contorted banding. In
part slight shearing movements have resulted in the fracturing of the bands. Thin
quartz veins cut across the bands and are later than the fundamental metamorphism.

286 AUSTRALASIAN ANTARCTIC EXPEDITION.

The structure of the rock is granoblastic and the texture is schistose. Magnetite
forms approximately one-third of the rock. Garnets have been fairly common but are
mostly altered. By far the greater part of the chlorite is derived from the biotite which
was originally very abundant. Epidote is present and quartz is common. In the
veins, the quartz shows brush polarisation and a larger grain size. Felspars are present
but they are so completely kaolinised that it is impossible to identify them.

No. 55. This is a fine grained, dark, heavy crystalline rock of specific gravity
3-11. It possesses a strongly developed crystallisation schistosity. Glistening white
mica on the cleavage planes gives a lustrous appearance to the rock. Fragments of the
rock are strongly magnetic.

The structure is both granoblastic and lepidoblastic. Magnetite is abundant
and ilmenite is also present as shown by the presence of white leucoxene. Hematite
occurs but the bulk of the 18-9 per cent, of iron ore is magnetite. Garnets are abundant,
the chief alteration being magnetite, though chlorite is also formed. A green variety
of biotite is very common and usually associated with muscovite, which appears to be
a later product than the biotite. The latter is much altered to chlorite. Quartz shows
strain polarisation and sometimes its recrystallisation has split flakes of biotite from
their parent crystals. Felspar is present but is largely kaolinised. It sometimes
exhibits the remnants of a fine twinning. Apatite is present. The rock is a Mica-
Magnetite-Schist.

No. 889. This specimen possesses a moderate crystallisation schistosity. On
one side there is a well-developed shear face along which abundant ilmenite is developed.
In the mass of the rock, blue cordierite can be seen.

Two sections were made of this rock, one of which was at right angles to the
schistosity. The magnetite percentage (26-9) was obtained from the latter section.
Intergrown with the magnetite, is a fair amount of hematite but ilmenite is not abundant
in the sections. The rock was originally rich in garnet which has been mostly altered
to a mass of quartz, magnetite, and either biotite or muscovite. Biotite is usually
accompanied by felspar and shows alteration to magnetite. Magnetite has also been
developed along a series of sub-parallel lines in the garnet (Plate II, fig. 3), and where
these meet, the alteration is complete.

The biotite is a greenish-brown variety and contains numerous pleochroic halos.
Occasionally there is a slight development of chlorite from biotite but a much commoner
alteration has produced magnetite and muscovite. The last mineral appears to be
derived from either the garnet or the biotite. Cordierite is abundant in the section
cut at right angles to the schistosity. It is distinguished from quartz by faint pleochroic
spots and its pronounced signs of alteration. The bulk of the felspar is andesine but
some is more calcic, probably approaching labradorite. One fine example of a crystal
of andesine shows a graphic intergrowth with quartz. Apatite is abnormally biaxial
in its interference figure. The structure is heteroblastic and the rock is a Magnetite-
Garnet-Schist.

MAGNETITE GARNET ROCKS COULSON.

JST

No. 149. This rock shows a pronounced gneissic banding. It is a medium
grained rock and flakes readily owing to the amount of mica. Quartz, garnet, felspar,
mica, and magnetite can be recognised and the mineral proportions are given in Table 1.
Magnetite with intergrown ilmenite forms 23 per cent, of the rock. The ilmenitc is
recognised l>y its alteration to leucoxene. Sphene is doubtfully present. Garnet,
magnetite, and also apatite contain inclusions of one another. Some garnets are
idioblastic and others show evidence of resorption. Quartz, plagioclase, and orthoclase
are present in approximately equal proportions. The plagioclase, is mostly labradorite,
and, like the orthoclase, is slightly kaolinised. Quartz contains numerous opaque
inclusions of iron oxide.

The biotite is interesting. It is a yellowish-brown variety occurring in large
sporadic flakes which contain radioactive inclusions surrounded by beautiful pleochroic
halos (Plate II, fig. 4). All the inclusions are of considerable dimensions relative to the
halos which surround them and their longer and shorter axes have been measured.
The radii of the halos in the following table are the average of two or three measurements
of the distance from the edge of the inclusion to the outer edge of the halo. A 4 cm.
( x 45) objective lens was used in conjunction with a Beck micrometer eye-piece and the
measurements given below are in millimetres.

TABLE IV.

Radius of
Halo.

Remarks.

Dimensions
of Nucleus.

Remarks.

ui:i

Indistinct

040

Very distinct ...

035 x -020

Wedge-shaped

040

Very distinct ...

040 x -040 ..

Circular

039
039
038

Verv distinct, slight difference in colour,
In-ill}! lighter near nucleus.
Lighter area next nucleus (to about -009mm.
from tin- edw).
Very distinct

033 x -033 ...
039 x -031 ...
026 x -020

Circular.
Rectangular, straight extinction.
Wedge-shaped straight extinc-

tion (?)

042 x -037

Oval

017

010 x -010

( 'ircular

016

Indistinct

020 x -016 ...

Oval.

033 x -020

<l\al

Indistinct

020 x -013 .

Oval.

In the measurements made by Joly* a correction is made for the nucleus on the
assumption that the nuclei are zircon and that they are sensibly spherical. In this
rork, however, the nuclei are too large and such corrections would not be valid. The
halos are remarkable even in colour, but in two cases there is an appreciable difference
in colour, which, however, is too slight for accurate measurement. The inclusions
show high order polarisation colours and appear to have straight extinction, thus
suggesting zircon.

" Genera of Pleochroic Htloi."
16640 B

J. Joly. Loud. Roy. 800., Phi). Trans.. Seriei A, vol. 217. pp. 61-79. 1W17.

288 AUSTRALASIAN ANTARCTIC EXPEDITION.

Joly and Fletcher* give the radii of halos in biotite as follows :

Uranium Ra C Ra A Thorium Th C. Th X

0326 -0222 -0397 -0263

On the basis of these measurements the first six halos in the above Table IV are probably
due to Th C and the origin of the others is doubtful.

The structure of the rock is heteroblastic and it may be called a Magnetite-
Garnet-Gneiss.

No. 147. Part of this specimen is a dark magnetite gneiss and part is a light-
coloured acidic rock. It exhibits a pronounced example of lenticular structure.

The acidic part shows large crystals of quartz, fibrous sillimanite, magnetite,
biotite, plagioclase, and orthoclase. The dark banded part contains abundant magnetite,
garnet, quartz, cordierite, and felspar. Magnetite is mere abundant in the dark part
than in the light, but is especially plentiful at the junction.

In the dark gneissic portion the magnetite encloses cordierite, apatite, and
eillimanite. Ilmenite and rutile are also present. Garnet, which is absent from the
acidic part, contains inclusions of quartz, magnetite, and apatite, and also shows evidence
of resorption. The felspar has an extinction angle of 12 and a refractive index usually
greater than quartz. It thus suggests andesine. Quartz shows strain polarisation.
Biotite shows evidence of resorption and sometimes contains radioactive inclusions;
one perfect example has a halo with a radius of -037 mm. and is ascribed to Th C. f

Cordierite is abundant; at times it is altered to a pinitic mica with an absence
of any basal lamination, and at others it has given rise to a serpentinous product with
a radial arrangement and brush polarisation. Sillimanite is usually intergrown with the
biotite and appears to have developed from it, more especially in the acidic part. It
occurs in the fibrous form and also in stouter prisms, with the slowest ray parallel to the
prism axis. Orthoclase and andesine occur in the acidic part. Some crystals of epidote
are present with a darker margin which shows higher polarisation colours and stronger
pleochroism than the central portion. They usually adjoin magnetite crystals and the
darker margin contains a greater percentage of iron. The structure of the gneissic part
is heteroblastic and the rock is a Magnetite-Garnet-Schist.

No. 827. This is a gneissic type with a discontinuous banded structure, the
various bands differing greatly in mineral content. The following minerals can be
identified in the hand specimen : Quartz, magnetite, blue cordierite, felspar, sillimanite,
biotite, and pink garnet. The distribution of the magnetite is especially irregular. A
central band through the rock is composed of magnetite, altered cordierite, pink garnet,
and quartz. Biotite is plentiful on both sides of this band.

* Pieochroic Halos." Phil. Mag., 6th Series, Vol. XIX, p. 633, 1910.
t Jolly and Fletcher, op. cit., p. 633.

MAGNETITE GARNET ROCKS COULSON. 289

Magnetite has crystallised out later than the biotite and it appears to have split
off fragments of biotite by crystallisation between the cleavage flakes. The magnetite
does not appear to have been derived from the biotite as there is a general absence of
any minerals accompanying such a change. Flakes of green chlorite with a magnetite
are rare. Garnets usually show round edges due to resorption. Quartz is abundant
and shows strain polarisation. Apatite is present, and possible cyanite. Little
silliiiianite and no felspar can be observed in the section on account of the very variable
distribution of the minerals in the rock. Fragments from the hand specimen show that
the felspar is an acid plagioclase. Cordierite gives rise to a yellowish alteration product.
' )ne large zircon crystal, -20 x '15 mm., with uniaxial figure and high colours, is observed.
The structure is heteroblastic and the rock is best described as a Magnetite-Gamet-Gneiss.

No. 181. This i:; ;: dii'iiuctivc and banded variety with numerous dark porphyro-
blasts of magnetite (Plate I, fig. 1). Their longer axes are parallel to the well-developed
sdiistosity and at times measure 14 x 7 mm. One large elongated crystal of tourmaline
is noted which possesses the trigonal cross-section and characteristic absorption. A
green chloritic mica is the only other mineral occurring in large masses. A similar rock
is illustrated (Plate 1, fig. 2), in which the magnetite porphyroblasts have been elongated
into a pronounced linear structure.

Microscopically the rock is composed of bands of minute quartz and garnet,
quartz, and felspar and biotite crystals. The magnetite individuals are studded with
numerous inclusions of garnet, biotite, quartz, and altered felspar. A little rutile is
intergrown with the magnetite and occasionally there is a slight alteration to limonite.
Garnets are numerous but their size is small ; some of the grains, especially the smaller
ones, show anomalous double refraction. When a band of quartz and garnet meets a
magnetite porphyroblast, it divides into two and surrounds the porphyroblast. Green
biotite is common and associated with green chlorite. There is a distinctive yellowish-
brown biotite which is rare relative to the green variety. Muscovite is present in small
amount. Plagioclase and orthoclase are present and the former ranges down to andesine.
The felspar is frequently kaolinised. Epidote is present. Some crystals suggest
cordierite but there is an absence of brown pleochroic spots. Quartz is abundant, while
apatite is a common accessory. Xo tourmaline is present in this section. The rock is
a Porphyroblastic-Magnetite-Garnet-Gneiss.

No. 304. This is a schistose rock, to which is attached, like No. 147, part of a
coarsely crystalline vein or lens composed of quartz, felspar, garnet, magnetite, and also
micaceous hematite.

Magnetite forms 11-9 per cent, of the rock and it contains inclusions of garnet,
biotite, apatite, quartz, and cordierite. Hematite is present and a little limonite has
developed from the magnetite. No ilmenite is noted and no reaction obtained for
titanium. Garnets form 7-9 per cent, and are usually idiomorphic. Biotite is abundant,
forming 14-7 per cent., and frequently includes quartz and zircon. Quartz is abundant
and contains numerous inclusions of garnet, magnetite, and zircon and also fluid

290 AUSTRALASIAN ANTARCTIC EXPEDITION.

inclusions. It shows strain polarisation and is at times difficult to distinguish from
cordierite; the latter is usually recognised by the pleochroic brown spots. Such
extinction angles as can be measured for the felspars are high, and these, in conjunction
with a high refractory index, suggest labradorite. Though most of the felspar is
clouded, there is occasionally a clear fresh felspar with an extinction angle of 10 degrees
and a low refractive index which is probably albite. No sillimanite is present. The
structure of the rock is granoblastic and the specimen may be called a Magnetite-Garnet-
Gneiss.

No. 814. This is a banded type in which the bands are narrower at the centre of
the specimen than at either end. They are composed of quartz, garnet, biotite, felspar,
and magnetite, and the pink colour of certain bands is due to a predominance of garnet.

Magnetite with a little intergrown hematite forms 14-7 per cent, of the section.
No ilmenite occurs and no reaction was obtained for titanium. The magnetite frequently
contains inclusions of quartz, biotite, and garnet and there is sometimes a rim of
magnetite around a garnet crystal, while cracks in the garnets may be filled with
magnetite. Garnet, which was one of the first minerals to be formed, has an average
grain size of about 0-25 mm., forming 12-6 per cent, of the section. Quartz and
felspar constitute nearly two-thirds of the rock. Quartz shows strain polarisation and
is relatively clear with respect to the felspar ; it frequently contains inclusions of apatite,
zircon, and garnets. The felspar is only slightly altered and ranges from andesine to
labradorite. A little orthoclase is present. The felspar contains inclusions of quartz,
magnetite, and apatite. The biotite is a brown variety with numerous inclusions of
garnet, apatite, and sometimes magnetite. The structure is granoblastic and the rock
is a Magnetite-Garnet-Gneiss.

No. 245. This is a dark schistose type in which magnetite, garnet, biotite, and
quartz can be identified macroscopically.

Magnetite is an important constituent of the rock and a fair amount of ilmenite
is intergrown with it. Chemical tests on this and the following specimen (No. 288)
indicates the presence of titanium. Magnetite crystals enclose quartz, biotite, and garnet.
Garnets (15-6 per cent.) are almost as numerous as magnetite (17-9 per cent.) and are
more idioblastic though they are smaller in size (0-14 mm.) Some have been altered to
quartz and micaceous products and others have produced epidote, as in No. 989.
Quartz and felspar are abundant. Numerous minute gaseous inclusions in the quartz
have a roughly parallel arrangement, but there are other small reddish inclusions which
are possibly hematite or rutile. Both orthoclase and plagioclase are present ; the
latter is mostly andesine but some approaches labradorite in composition. Microcline
with its typical cross-hatching is observed and the felspar is generally free from
alteration. Green biotite forms 9-2 per cent, and shows pleochroic halos around minute
zircons. Apatite and epidote are noted. The rock is a Magnetite-Garnet-Schist.

No. 288. This rock is very similar to No. 245 in the hand specimen except
that it contains some moderately large porphyroblasts of magnetite.

MAGNETITE GARNET KOCKS COULSON. 291

The garnets are mostly of smaller dimensions (0-07 mm.) than in the preceding
No. 245 and they are sometimes enclosed in the magnetite. The larger porphyroblasts
of magnetite are confined to certain bands ; they are elongated parallel to the schistosity
and contain inclusions of quartz and appear to have a little rutile intergrown with them.
Felspar is the most abundant mineral present and it is much altered to an aggregate
of micaceous minerals and epidote. Relics of microcline with cross-hatching are
present but the greater part of the felspar is a plagioclase of composition between
andesine and labradorite. Quartz shows strain polarisation and cataclasis. Biotite is
greatly altered to chlorite and secondary magnetite. Muscovite has also been
developed from biotite and is not symmetrically disposed to the schistosity planes. A
little bleached biotite and accessory apatite occur. Granoblastic, porphyroblastic, and
cataclastic structures are the chief ones present. The rock is a Garnet-Magnetite-Schist.

No. 102. This is a massive type with dominating garnet. Most of it is fine
grained but in parts the crystals of garnet and quartz attain considerable size.

Garnets form 40-8 per cent, of the rock and occur as irregular masses traversed
by numerous cracks along which iron ore occurs (Plate II, fig. 5.) When the garnet is
altered to chlorite, a larger amount of iron ore is developed. Garnet is also intergrown
with magnetite and quartz ; and all three minerals contain inclusions of one another.
Limonite is present as a pseudomorph after magnetite and there is a fair amount of
rutile intergrown with the magnetite ; this is red by reflected light and probably absorbs
the 1-14 per cent, of TiO 2 , in the rock. Apatite is more abundant than usual and occurs
chiefly with the magnetite. Quartz is abundant and contains numerous inclusions of
minute iron ore individuals arranged in parallel lines. A few zircon crystals occur.
The structure is diablastic and the texture is massive. It is Garnet-Quartz-Magnetite
Rock.

No. 348. This is a schistose variety with bands of minute pink garnets and clear
quartz, which show up well from the otherwise greenish-black rock. Scattered through
the rock are occasional porphyroblasts of magnetite and in this respect the specimen
resembles Nos. 181 and 288.

Garnets are numerous and form one-fifth of the rock but the average size of the
grains is only 0-06 mm. The chemical analysis shows that this rock contains 8-23 per
cent, of MnO and it is probable that this is contained in the garnet. Magnetite occurs
both as xenoblastic and idioblastic individuals which generally have their longer axes
parallel to the schistosity. Felspar and quartz are present in equal proportions. The
felspars are almost entirely kaolinised and their original nature is somewhat indefinite ;
such extinction angles as are measurable indicate andesine but orthoclase is also present.
Quartz contains gaseous, garnet, and magnetite inclusions and shows strain polarisation.
Green biotite is commonly altered to chlorite. Epidote and calcite, both probably
derived from the alteration of the felspars, as well as apatite and zircon, are present.
The structure is both porphyroblastic and granoblastic and the rock is a Garnet-Magnetite-
Schist.

292

AUSTRALASIAN ANTARCTIC EXPEDITION.

3. CHEMICAL CHARACTERS.

The table below (Table V) contains the results of three chemical analyses. A
fourth analysis of the Broken Hill rock, the mineral composition of which is given in
Table I, is given for comparative purposes.

TABLE V.

II.

III.

IV.

Si0 2
A1 2 3

50-35
16-18
13-57

53-73
11-57
13-15

57-78
12-01
7-35

31-34
8-34
32-10

FeO ...

5-57

14-41

4-98

6-75

MgO .,
CaO

1-78
1-94

2-11
1-79

1-79

2-85

0-94
7-54

Na 2
K 2
H 2 +
H 2 0-
Ti0 2
C0 2

3-14

0-89
0-50
0-98
0-49
absent

0-15
absent
0-80
0-62
1-14
absent

1-70
1-47
0-69
0-52
0-73
n.d.

0-11
0-05
0-52
0-63
0-40
0-01

Zr0 2

so, 5

trace
0-48

trace
0-53
absent

0-06
0-16
absent

absent
4-67
trace

trace

Cl ...

0-09

trace

0-16

absent

Cr 2 3
NiO, CoO

absent
absent

absent
trace

MnO

4-12

0-24

8-23

7-08

BaO

absent

trace

SrO
Li 2

trace

absent

0-06

pres. (spect.)
absent

0-02

CuO .

trace

99-99

100-40

100-38

100-32

less = Cl 0-02

Specific gravity

2-96

3-33

2-98

3-809

I. No. 181, Porphyroblastic magnetite-garnet-gneiss, Adelie Land. Analyst, A. L. Coulson.

II. No. 102, Garnet-quartz-magnetite rock, Adelie Land. Analyst, A. L. Coulson.

III. No. 348, Garnet-magnetite-schist, Adelie Land. Analyst, A. L. Coulson.

IV. Quartz-garnet-magnetite rock, Broken Hill, New South Wales.* Analyst, W. G. Stone.

The analyses are noteworthy for their low silica as there is abundant free quartz
in each case. Felspar is an equally important constituent in Nos. 181 and 348 where
appreciable alkalies are present. Felspar is not recorded in No. 102 or in the Broken
Hill specimen and the total alkalies are extremely small. The remarkable high
content of magnetite and garnet is reflected in the high total iron. Lime and magnesia
are extremely low and reflect the paucity of ferromagnesian minerals. Magnesia is

" Geology of the Broken Hill District," E. C. Andrews. Mem. 8, Geol. Surv., N.S.W., p. 172.

MAGNETITK (JAKNET ROCKS- roU.suN. 293

iilinost wholly contained in tin- mini and the lime in the felspar and apatite. Practically
no lime and magnesia is available for the garnet. In No. 102, which contains 40-8 per
i cut. of garnet, it can he clearly inferred that the garnet is an iron-alumina variety.
In No. 181 there is 4-12 percent, of MnO which probably replaces part of the FeO in the
garnet. No. 348 contains an exceptional amount of 8-23 per cent, of manganese and
the composition of the garnet may approach that of spessartite. Qualitative tests
indicate the presence of considerable manganese in the chlorite magnetite schist 527A
and in the magnetite garnet schists and gneisses Nos. 304, 814, 245, and 288. Rutile is
recorded in Nos. 181 and 102 and may account for part or all of the titanium content.
Qualitative tests revealed a fair amount of titanium in the similar rocks Nos. 245 and
288, but it proved to be absent in Nos. 304 and 814. In No. 245 the titanium
percentage is ascribed to ilmenite.

Comparison with the chemical analysis of the Broken Hill rock shows great
similarity. The lower silica and higher total iron of the Broken Hill rock corresponds
with lower quartz and garnet percentages and higher iron ore than No. 102. Its high
manganese content is a marked point of resemblance with No. 348. It is distinct >
however, in its high lime and phosphorus which correspond with the measured percentage
(14-2) of apatite. Its large excess of Fe 2 O 3 over FeO appears to indicate more hematite
than in the Antarctic types.

4. SUMMARY OF PETROGRAPHICAL CHARACTERS.
The following rocks have been described :

Magnetite-schists Nos. 765, 912, 294.

Biotite-magnetite-schist No. 926.

Chlorite-magnetite-schist No. 527 A.

Mica- magnetite-schists Nos. 55, 933.

Mica-magnetite-gneiss ... ... ... No. 576.

Quartz-magnetite-schist No. 296.

Epidote-magnetite-schist No. 989.

Magnetite-garnet-schists Nos. 889, 147, 245.

Magnetite-garnet-gneiss Nos. 827, 181, 149, 304, 814.

Garnet-magnetite-schists Nos. 288, 348.

Quartz-garnet-magnetite rock No. 102.

Quartz, felspar, and mica are the chief minerals present in the rock suite in
addition to the abundant magnetite and garnet which characterise the group. The
quartz usually shows strain polarisation colours and a cataclastic structure. In a ; *

number of cases minute inclusions have a definite alignment within the individual quartz
crystals which is generally parallel to the planes of schistosity.

294 AUSTRALASIAN ANTARCTIC EXPEDITION.

Felspar is an important constituent, though it is totally absent in the magnetite-
schist No. 912, in the quartz- or epidote- or mica-magnetite-schists Nos. 296, 989, 933,
and in the quartz-garnet-magnetite rock No. 102. Orthoclase is a common constituent
but microcline only occurs in the magnetite-garnet-schist No. 245, and the garnet-
magnetite-schist No. 288. The plagioclase ranges from albite to labradorite and i^
commonly happens that an acidic and a basic plagioclase are present. Generally the
felspars are much kaolinised and sericitised.

Cordierite is not common but is observed in four magnetite-garnet schists and
gneisses No. 889, 147, 827, and 304. Such crystals as are seen in these sections do not
exhibit any measurable pleochroic halos.

Biotite is a very common constituent, the usual variety being green in colour.
Muscovite is present in some sections, notably in the chlorite-magnetite-schist No.
527 A, the mica-magnetite-schists Nos. 55, 933, and the epidote-magnetite-schist No. 889.
Chlorite is an important mineral and usually owes its origin to the alteration of biotite
or garnet. Epidote is abundant in some of the magnetite schists but it loses importance
as the ratio of garnet to magnetite increases. Apatite is a common accessory, as also is
zircon.

Magnetite usually has ilmenite or hematite or rutile intergrown with it. Limonite,
pyrite, and sphene also occur but the latter two minerals are not common. The
percentage of iron ore ranges from 33-0 in the magnetite-schist No. 294 to 6-2 in the
garnet-magnetite-schist No. 288. In most cases the crystals of magnetite are xeno-
blastic rather than idioblastic. In Nos. 181, 288, and 348 there are magnetite
porphyroblasts and in the example in Plate I, fig. 2, a pronounced linear structure has
developed from the porphyroblasts. As the porphyroblasts of magnetite have their
pnger axes parallel to the schistosity and contain numerous inclusions of garnet, quartz,
and biotite, all of which originated during, or were recrystallised by, the metamorphism,
these porphyroblasts evidently arose at the time of major metamorphism, when the
conditions of pressure and temperature probably corresponded to Grubemann's meso
.zone of metamorphism. The chloritisation of the biotite and the kaolinisation of the
felspars probably arose under later conditions of epi zone metamorphism. Still well *
has applied a conception of metamorphic differentiation to clots of biotite and epidosite
in amphibolite dykes at Cape Denison. It is possible that the same conception can be
applied to the magnetite porphyroblasts, in which case they are due to local segregation
of iron ore arising as a direct consequence of the metamorphism. In many instances in
the suite of rocks described, magnetite has been considered as arising partly from the
result of epi zone processes on the garnet and mica.

Garnet is a less constant constituent than the magnetite. It is absent in some
specimens and ranges up to 40 per cent, in No. 102, where it dominates the colour of the

* " The Metamorphic Rocka of Adelie Land." F. L. Stillwell. Aust. Ant. Exp. Scientific Reports, Series A, Vol. Ill,
Part 1 (1918), p. 58-71.

MAGNETITE GARNET ROCKS COUL80N. 295

rock. The commonest alteration of the garnet is the production of chlorite, quartz,
and magnetite which is especially well displayed in the chlorite magnetite schist
No. 527A. The reaction can be illustrated by the equation

2 MgFe 3 AlSi 3 O, 2 + 4 H 2 ; HgMgaFegAljSigO,,, + 2 Si0 2 + Fe 3 4 .

Garnet. CUorite. Quartz. Magnetite.

The two arrows going in opposite directions means that the equation representing
a reaction is reversible.

In this composition of the garnet the ratio of Mg to Fe is assumed to be as 1 to 2,
and the ratio of Fe to Al as 1 to 1 . The chlorite is also assumed to be an isomorphous
mixture of the two. silicate molecules, H 4 (MgFe) 3 Si 2 9 and H 4 (MgFe) 2 Al 2 Si0 9 in the
proportions of 1 to 1. Chemical work has indicated, however, that the garnet
contains manganese in many cases where this change is observed. On alteration,
it is probable that the manganese enters the composition of magnetite rather than the
chlorite, producing manganmagnetite, the composition of which is expressed by
(FeMn)O, Fe 2 3 .

While quartz, chlorite, and magnetite are the usual alteration products of garnet,
magnetite was the sole recognisable product, in some cases thus indicating the migration
of quartz and chlorite. In the magnetite-garnet-schist No. 889, there is a production
of quartz, magnetite, and either biotite or muscovite. This may have been attained
by a reaction between garnet and the microcline or orthoclase in the following manner :

Fe 3 Al2Si 3 O 12 + KAlSi 3 8 + 2 (OH) - H 2 KAl 3 Si 3 12 + 3 Si0 2 + Fe 3 4 .

Garnet. Orthodox. Muscovite. Quartz. Magnetite.

Stillwell * gives the following reaction which may be applicable to the case in
which biotite is produced :

2 (MgFe)O, CaO, A1 2 3 , 3Si0 2 + 2 KAlSi 3 8 ^ (KH) 2 (MgFe) 2 Al 2 (Si0 4 ) 3

Garnet. Ortkodatt. Biotite.

+ Si0 2 .

Anortkite. Quartz.

The anorthite molecule is considered to enter into, or be derived from, the composition
of the plagioclase which then becomes relatively more basic or more acid according to
the direction of the action. It is doubtful whether there is sufficient lime in the garnet
to give rise to anorthite in the present case.

Grubenmann f illustrates the production of quartz and biotite from garnet and
orthoclase by an equation in which sillimanite is produced. The latter mineral does
not appear in the present case.

Tilley J interprets the reaction as follows :
H 4 K 2 Mg 2 Fe 4 Al 6 (SiO 4 ) 9 + 3 SiO 2 ; 2 KAlSi 3 O 8 + 2 Fe 2 MgAl 2 (Si0 4 ) 3 + 2 H 2 0.

Biotite. Quartz. Ortkodate. Garnet.

Op. cit., p. 168.

f ' Die KraUllinen Schirfer." U. Grubenmann. Berlin, Vol. I (1904), p. 62.

J " The Granito-OneUM of Southern Eyre Penuuulm." C. . Tilley, Q.J.G.S., Vol. Ixxvii, Pt. 2 (1921), p. 94.

"MM <

296 AUSTRALASIAN ANTARCTIC EXPEDITION.

It was noted that epidote was sometimes developed as a result of the alteration
of the garnet. Van Hise * gives the following equation for the development of epidote
from a lime-iron-alumina garnet

2 Ca 3 Al 2 Si 3 12 . Ca 3 Fe 2 Si 3 12 + 5 C0 2 + H 2 O = 2 HCa a Al 2 FeSi 3 13 + 5 CaC0 3

+ 3 Si0 2 + k (heat liberated).

In the rocks under discussion, however, calcite is only present in No. 348 where it appears
to be derived from calcic plagioclase. In the epidote-magnetite-schist No. 989 and
the magnetite-garnet-schist No. 245 epidote is developed from the garnet but there is
no free calcite. The above equation therefore represents a garnet with more lime than
in the garnets of these rocks.

Van Hise* records that in all cases of alteration of garnet, quartz appears among
the secondary minerals. In these rocks magnetite is also as general.

III. POSITION IN GRUBENMANN'S CLASSIFICATION OF CRYSTALLINE

SCHISTS.

The described rocks can be considered in two groups (1) those in which
magnetite is the most important constituent ; and (2) those in which garnet and magnetite
are equally important.

The members of the first group can be placed in the group of iron oxide rocks
(magnetite rocks) which forms Grubenmann's Xlth group. The presence of garnet,
quartz, plagioclase, orthoclase, biotite, magnetite, &c., and the occurrence of homo-
blastic and diablastic structures, which are sometimes almost obliterated by later epi
structures, all necessitate some form of deep-seated metamorphism. In addition, a
transference to epi zone conditions is suggested in some cases by the hydrous minerals
associated with the kaolinisation and sericitisation of the felspars and with the
development of chlorite and magnetite from garnet and biotite. Thus the magnetite
schists and gneisses Nos. 576, 912, 933, 294, and 55 are considered to be epi-meso rock
types, but may fall into the family of epi-magnetite rocks of Order III. In Nos. 989,
527A, and 889 the first metamorphism may have resulted from either meso or kata
conditions. No. 889, a magnetite-garnet-schist containing cordierite, may be related
to the aluminous silicate gneisses (Group II) and the family of kata garnet gneisses.
Nos. 765, 926, and 296 appear to be pure meso zone types and belong to Order II, meso
magnetite rocks and micaceous hematite schists (Group XI).

* " Treatise on Metamorphism." C. R. Van Hise. Geol. Surv., U.S.A., monograph 47, 1904, p. 305.

MAGNETITE GARNET ROCKS COUL80N.

297

The analysed specimens belong to the second group, and Tables VI and VII
give the molecular percentages of the constituent oxides, the group values and the
projection values according to Ozann's scheme.*

TABLE VI.

II.

HI.

IV.

SiO,

58-3

61-8

64-4

A1A

11-0

7-6

7-8

FeO

21-1

24-8

18-5

CaO

2-4

2-2

3-4

McO

3-1

3-5

3-0

K.O ..

0-6

1-1

Na,0

3-5

1-8

TABLE VII.

Group and Projcc-

II.

III. IV.

tion Values.

58-3

61-8

64-4

200

4-1

0-1

2-9

0-0

4 2-4

2-2

3-4

0-4

F ...

24-2

28-3

21-5 80-0

M . ...

0-0

2-5

4-5

5-3

1-5 0-0

0-9

1-1

2-1

1-4 0-3

2-2

2-7

2-1

c 2-2

1-6

1-4

2-4

f 15-6

15-7

18-5

15-5 20

I. No. 149, Magnetite-garnet-gneiss.
II. No. 181, Porphyroblastic magnetite-garnet-gneias.

III. No. 102, Garnct-quartz-magnetite rock.

IV. No. 348, Garnet-inagnetite-schist.

V. Mean values of Grubenmann's Xltli Group.

The values for No. 149 have been obtained from the Rosiwal analysis in Table I
on the assumption that biotite has the composition K 2 FeAl 2 (Si0 4 ) 3 , that the garnet is
almandine and that quartz, orthoclase, and labradorite are present in equal proportion 8

Orubenmann, op cit., 1910, pp. 134-6.

298

AUSTRALASIAN ANTARCTIC EXPEDITION.

in the rock. The values for the other rocks are obtained directly irom the chemical
analyses in Table V. The projection values are plotted in the triangular diagram, fig. 1.

An examination of the group values of the garnet-magnetite rocks in Table VII
shows that these rocks do not fit in any of Grubenmann's groups of crystalline schists.
Column 5 gives the mean group values of Group XI, the iron oxide rocks and these
show outstanding differences, especially in regard to S and F, with the corresponding
values of the garnet-magnetite rocks. Wide as are the variation limits assigned to some
groups of the crystalline schists, it appears impossible to include these garnet-magnetite
rocks without the creation of a new group. It may, therefore, be advisable to add a
new group of garnet magnetite rocks, just as two new groups have been created (1) for
certain manganiferous rocks in India. Hezner has added a manganese silicate group
which is naturally united to the lime silicate rocks by all its chemical, mineralogies 1, and
genetic relations, and also a new manganese oxide group which best follows the Xlth
Group of iron oxide rocks.

Croup XI

en Hill

Nos. 149, 181, and 348 would be types of a new garnet magnetite rock group,
while No. 182 would be somewhat abnormal. The chief minerals of such a group are
Sarnet, magnetite, and quartz. Other minerals which may be present are biotite, felspar,
cordierite, sillimanite, and epidote. Ilmenite, rutile, and hematite may be intergrown
with magnetite. Kata zone types are represented by Nos. 149, 827, and 304. Meso
zone types include Nos. 149, 814, and 245, while Nos. 181, 288, and 348 have apparently
been subjected to epi zone metamorphism as well as meso zone metamorphism.

On the other hand, if, as will appear likely, igneous emanations are concerned in
the formation of these magnetite garnet gneisses, it becomes doubtful whether any
attempt should be made to fit them into Grubenmann's classification of the crystalline
schists.

* " Uber manganrejche Kristalline schiefer Indiens." L. Hezner, Neues Jahrbuch fur Mjnerologie, Ac., 1919, p. 28,

MAGNETITE GARNET ROCKS COULSON.

IV. GENERAL DISCUSSION.

As the specimens are morainic Ixmlders there is no field evidence bearing on the
question of their origin. The nature and character of the magnetite-schists Nos. 912
and 294, the quartz-magnetite-schist No. 296, the epidote-magnetite-schist No. 989,
the mica-magnetite-schiste Nos. 55 and 933, and possibly also the garnet-magnetite-
gneiss No. 827 indicate that they very more likely to be derivatives of sedimentary
rocks rather than igneous rocks. The probability is not so high in the case of the garnet-
magnetite rocks. The evidence of the chemical analyses of the three specimens in
Table V is not decisive. There is a considerable excess of alumina over the 1 to 1 ratio
with the lime and alkalies and this, combined with the mineral content, suggests
derivation from a sediment. On the other hand, CaO is in excess over MgO in two
analyses and NajO is in excess of K 2 O. Less weight is attached to these points and in
general it is believed that a sedimentay rather than an igneous origin is to be ascribed
to the whole class. Yet the chemical composition is clearly not that of a common
sediment.

An interesting comparative occurrence is the quartz-garnet-magnetite-schist from
Broken Hill, New South Wales. It occurs* in the neighbourhood of the Broken Hill
lode in discontinuous lenses along bands which dip conformably with the associated
garnet-sillimanite-gneisses. A specimen from a locality, one mile south of Menindie-road,
4 miles south of Broken Hill, has been studied. It is a schistose rock composed of
magnetite, garnet, quartz, and apatite in the proportions given in Table I. It may be
noted that the proportions of iron ore to garnet are approximately the reverse of that in
No. 102, its nearest relative among the Cape Denison specimens. While No. 102, with
2 per cent, of apatite, is comparatively rich in phosphorus, this is greatly exceeded by
the Broken Hill specimen '-with 14-2 per cent, of apatite.

The chemical analysis of the Broken Hill type is quoted in Table V, and the
negligible alkalies, the low magnesia, and high iron are strong pointe of resemblance.
The differences in silica percentages correspond with differences in the amount of quartz.
The percentages of lime and phosphorus are much higher than those in No. 102. corre-
sponding to the abundant apatite. The Broken Hill type contains 7-08 per cent, of
manganese and is comparable with the 8-23 per cent, in No. 348, where the garnets are
manganiferous.

There are thus striking resemblances between the Broken Hill type and the
garnet-magnetite specimens from Cape Denison. The Broken Hill type is described by
Andrewsf as a lode formed from emanations given off, apparently at a depth, from
igneous material along a crush zone, and further formed by replacement. He pointe out
(p. 174) that the garnet- magnetite type is associated with garnet-sillimanite rocks, while
the quartz-magnetite types, without garnet, are associated with mica schists. Browne J

" Geology of Broken Hill." E. C. Andrew* Mem. 8, Oeol. Surv. N.8.W.. 1923. p. 171.

t " Geology of Broken Hill District." K. 0. Andrrws. Owl. Surv. .V.S.W., No. 8, 1922, p. 181.

? Appendix I of " Geology of Broken Hill Ditri< t." W. K. Browne, p. 340.

300 AUSTRALASIAN ANTARCTIC EXPEDITION.

believes these rocks to be pegmatitic derivatives of the gabbro magma which are
intruded as a metasomatic replacement of a zone of country rock. Further, these
authors appear to regard the garnet as a primary igneous mineral.

The comparison with the Broken Hill occurrence should therefore imply that the
garnet-magnetite rocks are derived from an igneous source.

Another related group of rocks are the " Skarn " rocks of Norway and Sweden.
Skarn* is an old Swedish mining term for the sillicate gangue of certain iron ore and
sulphide deposits of Archaean age. Its use has been extended by Goldschmidtf to all
lime-iron-silicate rocks which arise through the contact metamorphism of limestone,
including andradite rocks, hedenbergite rocks and andradite-hedenbergite rocks. The
iron ore of the skarn rocks (including garnet-magnetite rocks) is arranged in layers
parallel to the schistosity or bedding. Goldschmidt is sceptical of the view that such
rocks are magmatic differentiation products a view adopted by Browne for the Broken
Hill type. Goldschmidt considers them to be pneumatolytic contact rocks determined
by three factors, (1) pneumatolytic action in depth, (2) presence of a limestone, and (3)
absorption and enrichment of the constituents of the magmatic gases.

The skarn formation from limestone takes place through increase of silica and
iron. It is supposed that iron chloride is introduced in gaseous form along joints and
fissures with the result

2 FeCl 3 + 3 CaC0 3 = Fe 2 3 + 3 CaCl 2 + 3 C0 2 .

In this way the limestone collects the iron from the issuing magmatic gases. The
silicon is probably introduced as a halogen compound which reacts with the calcite to
form quartz. Quartz then combines with calcite forming wollastonite which, when
there is sufficient iron oxide, combines to form andradite.

3 CaSi0 3 + Fe 2 3 = Ca 3 Fe 2 Si 3 12 .

Wollastonile. Andratite.

Manganese, which is a notable component of many skarn rocks, is introduced in
the same way as the iron.

MnCl 2 + CaC0 3 = MnO + CaCl 2 + C0 2 .

The MnO, however, generally enters the silicate.

Goldschmidt thus considers that the skarn rocks are the result of metasomatic
pneumatolytic processes acting on limestones, and their unusual composition is
explained by the acquisition of material from the magmatic source.

Somewhat similar rocks at Morenci, Arizona, are considered by LindgrenJ to
have been derived from limestones by the addition of large amounts of ferric oxide and
silica.

* " Nomenclature of Petrology." A. Holmes. London, 1920, p. 211.

t " Die Kontaktmetamorphose im Kristianiagebiet." V. M. Goldschmidt, Ohristiania, 1911, p. 211 et scj.
J " The Copper Deposits of the Clifton-Morenci District, Arizona." W. Lindgren, U.S. Geol. Surv., Prof. pap. 43.
1905, p. 160.

MAGNETITE GARNET ROCKS-COUL80N. 301

In the Broken Hill occurrence no associated limestone is reported, neither does
the garnet of the Adelie Land specimens correspond to andradite. Yet it is interesting
to note that the same moraines have yielded a number of specimens of metamorphic
limestones and calc-silicates rocks.* It is known, however, that similar garnet rocks
can be produced without limestone. Kemp,f in commenting on the processes of garnet
formation in contact rocks of intrusive zones, concludes that, where limestone is absent,
the emissions from the deeper parts of the eruptive supplies the lime and iron oxide for
the garnet occurrences in such rocks as quartzites, schist, and gneiss, just as the reverse
contributions of silica and iron and alumina to the limestone would lead to the same
result. In the Broken Hill type it is clear that there has also been an important addition
of phosphorus.

In the garnet-magnetite types from Cape Denison the garnet is a variety
containing magnesium, iron, manganese, and alumina. Felspar, as noted above, is
often an important constituent though it fails in the quartz-garnet-magnetite rock
No. 102. The presence of aluminous garnet in addition to felspar suggests that the
garnet- magnetite formation has occurred in a rock in which there has been sufficient
alumina and silica to satisfy all the alkalies and available lime for the formation of
felspar, and further that the garnet-magnetite formation has occurred in the presence
of excess alumina rather than excess lime. Where P 2 8 is present or introduced, the lime
would be incorporated in apatite in preference to felspar. Aluminous garnet is present
in the analogous type at Broken Hill which occurs in a sedimentary series with excess
alumina, expressed in abundant sillimanite. It is therefore possible that the Cape
Denison types have developed in aluminous sediments similar to those which are related
to the magnetite schists, Nos. 912 and 294, and the related types Nos. 296, 989 and 933.

In conclusion, it is therefore likely that the garnet-magnetite rocks of Cape
Denison have resulted from the raetamorphism of sediments, the composition of which
has been modified by igneous emanations.

V. TOURMALINE-BEARING MAGNETITE GNEISS.

In view of the probable role of igneous emanations in the formation of the
garnet-magnetite rocks the following description of No. 252, a tourmaline-bearing and
garnet-free type, has added interest.

No. 252 is a banded variety, in general fine grained, but with coarsely crystalline
bands of quartz and bluish cordierite. Tourmaline is very abundant and long needles
of sillimanite and black grains of magnetite are plainly seen in the hand specimen.
The proportions of cordierite, magnetite, tourmaline and sillmanite are very variable.
The specific gravity of the rock is 3-09.

* " The MeUmorphic Limestone* of Commonwealth By, Adelie Land." C. E. Tilley. Aurt. Ant. Ezp. Sci. RpU.
Series A., vol. iii. pt. 11, pp. L'3 1-243.

t " Ore Depoaitft at the Contacts of Intnuive Rocks and Limetone." J. F. Kemp. Journ. Qeol., vol. ii, pp. 1-13.

302

AUSTRALASIAN ANTAECTIC EXPEDITION.

A portion of the rock, which was very rich in tourmaline, was sectioned in a
direction at right angles to the schistosity. A Rosiwal analysis resulted in the following ;

Quartz, cordierite, and sillimanite ... ... 47-5

Tourmaline 30-1

Magnetite 21-4

Apatite ... 1-0

The tourmaline is a strongly pleochroic variety, changing from green to a
reddish- or brownish- violet. The flakes show a rather infrequent cleavage and contain
gas, magnetite, and minute zircon or monazite inclusions, sometimes with a definite
alignment. Where these minute zircons or monazites are present, there is occasionally
a distinct lightening of colour in the form of a halo around the inclusion. The colours
of the tourmaline sometimes vary locally in a zonal fashion. The mineral is associated
with magnetite, both occurring in bands throughout the rock.

Cordierite is abundant and exhibits a great number of well-formed pleochroic
halos around inclusions of either zircon or monazite (Plate II, fig. 6.) While the
experience of Joly and Fletcher* has indicated that halos in cordierite cannot be used
for accurate measurement, the halos in this rock are remarkably distinct and capable
of measurement. By means of a Beck micrometer eye-piece and a 4 mm. (x 45) objective,
the following results were obtained, the radii given being always the distance from the
outer edge of the halo to the edge of the inclusion. All measurements are in millimetres.

TABLE VIII.

No.

Radius of
Halo.

Remarks.

Dimensions of
Nucleus.

Remarks.

041

Distinct ...

020 x -020

Circular.

2
3

041
040

Perfect halo, very distinct
Distinct ...

031 x -022 ...
054 x -030 ...

Oval, oblique extinction.
Oval, oblique extinction.

5
6
7

040
040
040
040

Indistinct, passing through both cordierite
and sillimanite.
Indistinct, passing through both cordierite
and sillimanite.
Very distinct, lighter in colour for about
020 mm. from the inclusion.
Indistinct, fainter near nucleus for about

023 x -017 ...
054 x -018 ...
032 x -028 ...
033 x -029 ...

Prismatic, oblique extinc-
tion. Zircon.
Prismatic, oblique extinc-
tion.
Oval, oblique extinction.

Large

012 mm.
Indistinct

083 x -027

Zircon.

9
10

026
021

Fairly distinct, passing through sillimanite
and cordierite.
Indistinct

038 x -015 ...
015 x -010 .

Oval, oblique extinction.
Oval, oblique extinction.

020

Fairly distinct

017 x -Oil ...

Oval, oblique extinction.

12
13

020
016

Indistinct, faint at sides
Doubtful

016 x -Oil ...

Oval.

015

Fairly distinct ...

010 x -008

Straight extinction.

Indistinct

038 x -027

Wedge shaped.

* " Pleochroic Halo." Phil. Mag., 6th series, vol. xix, 1910, p. 635.

MAGNETITE GARNET ROCKS COULSOX.

As in the halos measured in No. 149 in biotite, corona effects are generally
absent, being noted only in two cases, and Joly's correction for nuclei, based on the
assumption that they are zircon and sensibly spherical, obviously cannot be made.
By a comparison with the list of radii given by Joly and Fletcher, it seems probable,
as the dimensions of the halos in cordierite will not differ markedly from those in
biotite, that No. 1-8 owe their origin to Th C, No. 9 to Th X and Nos. 10-12 to Ra A.

The cordierite is much intergrown with long prismatic needles of sillimanite
which make the measurement of halos sometimes very difficult. In places the halos
are so numerous that the cordierite remains brown throughout a complete revolution
of the stage. As has been noted above, the radioactive inclusions are not confined to
cordierite, some occurring in tourmaline. Monazite appears to be the more common
nucleus in No. 252, while zircon was the more frequent in No. 149. The cordierite is at
times altered to a yellowish-brown form of biotite and at others the change proceeds,
from a centre, giving a chloritic mass.

Quartz is abundant and distinguished from cordierite by the absence of halos
cleavage and decomposition. Apatite is frequent but it is restricted to certain bands
in which it is associated with magnetite. The structure of the rock is homoblastic,
and though the nature of the different bands is very variable a general name would be
Tourmaline-Magnetite-Sillimanite-Cordierite-Gneiss.

The absence of felspars and the abundance of sillimanite and cordierite clearly
indicate that the rock has been derived from a sediment. The abundance of tourmaline
indicates the action of igneous emanations and tends to confirm the similar view in
regard to magnetite and apatite.

In Grubenmann's classification it would occupy a position in the Kata division
of the group of aluminous silicate gneisses (Group II). Within this division there is a
family of the sillimanite gneisses and the family of the cordierite gneisses, but the Cape
Denison rock is clearly distinct from typical members of these families in the presence
of abundant tourmaline and magnetite.

. VI. GENERAL SUMMARY AND ACKNOWLEDGMENTS.

In all twenty-two specimens have been described from Adelie Land and treated
as crystalline schists. Two contain abundant magnetite and no garnet. Seven contain
dominating magnetite and, together with the preceding two, are associated with the iron
oxide group, No. XI, of Grubenmann's classification of the crystalline schists. In ten
apecimens garnet is approximately as important as the magnetite. The chemical
characters of these types are widely different from any defined group of crystalline
schists and a new group of garnet-magnetite rocks is suggested.

36640 I)

304 AUSTKALASIAN ANTARCTIC EXPEDITION.

The types are probably derived from sediments whose composition has been
modified by igneous emanations. Such origin is rendered the more likely by the final
type which is rich in tourmaline as well as magnetite.

Measurements of pleochroic halos and nuclei have been made in both cordierite
and biotite. The formation of these halos has been ascribed partly to Th C, partly to
Th X and Ra A.

In conclusion, the writer desires to thank Professor W. W. Watts, of the Royal
School of Mines, South Kensington, for kind assistance in the prosecution of research,
and he is also specially indebted to Dr. F. L. Stillwell, at whose suggestion the work
was undertaken and who has given the writer every possible assistance and unhesi-
tatingly helped with his great experience of Adelie Land specimens.

MAGNETITE GARNET ROCKS CQUL80N. 305

VII. DESCRIPTION OF PLATES.

PLATE I.

Fig. 1. No. 181, Porphyroblastic-Magnetite-Garnet-Gneiss. Photo, F. L.
Still well. Note that the scale photographed with the specimen in this plate is
graduated in inches.

Fig. 2. No. 650, Porphyroblastic-Magnetite-Garnet-Schist, with a well-defined
" linear " structure. Photo, F. L. Stillwell.

PLATE II.

Fig. 1. No. 527A, Chlorite-Magnetite-Schist, showing the alteration of garnet
into a mass of quartz, magnetite, and chlorite with residual garnet. Mag. 113 diams.
Photo, G. S. Sweeting.

Fig. 2. Siliceous vein crossing the Micaceous-Magnetite-Schist, No. 933. The
dark central area in the vein is felspar. Mag. 22 diams. crossed nicols. Photo, G. S.
Sweeting.

Fig. 3. No. 889, Magnetite-Garnet-Schist, showing parallel line development
with garnet. Mag. 25 diams. Photo, G. S. Sweeting.

Fig. 4. Magnetite-Garnet-Gneiss, No. 149, containing pleochroic halos around
radioactive inclusions in biotite. Mag. 48 diams. Photo, G. S. Sweeting.

Fig. 5. No. 102, a rock composed of garnet, quartz, magnetite, and apatite.
Mag. 24 diams. Photo, G. S. Sweeting.

Fig. 6. No. 252, Tourmaline-Magnetite-Gneiss, with numerous perfect pleo-
chroic halos in cordierite. Tourmaline (dark with cross cleavage), sillimanite (fibrous)
and magnetite (black) can also be seen. Mag. 34 diam. Photo, G. S. Sweeting.

[Two plate*. ]

gjtdwjr: Alfrrd Juno Kent. Covrrnmrnt Printer 12S.

AUSTRALASIAN ANTARCTIC EXPEDITION.

SERIES A. VOL. III. PLATE

Fig. 1.

Fig 2.

AUSTRALASIAN ANTARCTIC EXPEDITION.

SERIES A. VOL. III. PLATE II

Fig. 2.

Fig 3.

.Fig. 4.

Fig. 5.

Series B.

VOL. PRICE.

I. TERRESTRIAL MAGNET s. d

PART i. FIELD SU1 SD REI> POGRAPH CURVES.

By ERIC N. WEBB, D.S.O., M.C., A.C.S.E., (N.Z.) A.M.lNST.C.E.

Expedition. > r ro o

\LYSIS AND DISCUSSIONS OF MAGNETOGRAPH CURVES.

By CHARLES CHREE, M.A., Sc.D., LL.D., F.R.S.J

II. TERRESTRIAL MAGNETISM AND RELATED OBSERVATK

PART i. RECORDS OF THE AURORA POLAR

By DOUGLAS MAWSON, Kt., O.B.E. B.E., D.Sc., F.R.S. o 15 o

2. RECORD OI- TIC DISTURBANCES AND THEIR DISCUSSION.

By CHARLES CHREE, M.A. Sc.D., LL.D., F.R.S.

3. RECORD OF THE RANGE OF TK -ilON' OF WIRELESS

III. METEOROLOGY.

THE RECORD OF TH

IV. METEOROLOGY.

THE RECORD OF THI

METEOROLOGY.
PART i. RECORD OF THE Ql

2. RECORD OF THE .DE DURIN : JB-

:;CTIC CRUISES AND THREE ANTARCTIC VOYAGES.

(The reduction and tabulation < - logical Data is in progress under direction of Mr. H. G. Hunt,

Commonwealth Meteorolo^

AUSTRALASIAN ANTARCTIC EXPEDITION

1911-14

UNDER THE LEADERSHIP OF SIR DOUGLAS AVAWSON. D.Sc. P.R.S.

SCIENTIFIC REPORTS.

SERIES A.

VOL. III.

GEOLOGY.

PART VI.

PETROLOGICAL NOTES

FURTHER ROCK SPECIMENS

COLLECTED FROM IN SITU OCCURRENCES

COMMONWEALTH BAY REGION.

J. O G. GLA5TONBURY. B.A., AVSc.

WITH TWO PLATES.

PRICE: THREE SHILLINGS AND SIXr

Wholly 5ft up and printed tn Australia bv
THOVAS HENRY TENNAHT. GOVERNMENT PRINTER. . L

194O.

928JO

SERIES A.

PRICE.

VOL. s. d.

I. CARTOGRAPHY AND PHYSIOGRAPHY. Brief narrative and reference to Physiographical and
glaciological features. Geographical discoveries and Cartography. By DOUGLAS MAWSON.

II. OCEANOGRAPHY.

PART 1. SEA-FLOOR DEPOSITS FROM SOUNDINGS. By FREDERICK CHAPMAN ... 060

2. TIDAL OBSERVATIONS. By A. T. DOODSON t

3. SOUNDINGS. By J. K. DAVIS 2 6

4. HYDROLOGICAL OBSERVATIONS, MADE ON BOARD S.Y. "AURORA."

Rn'uced, Tabulated and Edited by DOUGLAS MAWSON

5. MARINE BIOLOGICAL PROGRAMME AND OTHER ZOOLOGICAL AND

BOTANICAL ACTIVITIES. By DOUGLAS MAWSON 076

III. GEOLOGY.

PART 1. THE METAMORPHIC ROCKS OF ADELIE LAND. By F. L. STILLWELL ... 2 '2
2. THE METAMORPHIC LIMESTONES OF COMMONWEALTH BAY, ADELIE

LAND. ByC. E. TILLEY 016

3. THE DOLERITES OF KING GEORGE LAND AND ADELIE LAND. By W. R.

BROWNE 016

4. AMl'HLBOLITES AND RELATED ROCKS FROM THE MORAINES, CAPE

DENISON, ADELIE LAND. By F. L. STILLWELL 020

5. MAGNETITE GARNET ROCKS FROM THE MORAINES AT CAPE DENISON,

ADELIE LAND. By ARTHUR L. COULSON 020

6. PETROLOGICAL NOTES ON FURTHER ROCK SPECIMENS. By J. 0. G.

GLASTONBURY 036

IV. GEOLOGY.

PART 1. THE ADELIE LAND METEORITE. By P. G. W. BAYLEY and F. L. STILLWELL. 1 6
2. PETROLOGY OF ROCKS FROM QUEEN MARY LAND. By S. R. NOCKOLDS.
3. GRANITES OF KING GEORGE LAND AND ADELIE LAND. By H. S. SUMMERS

and A. B. EDWARDS. Appendix by A. W. KLEEMAN 039

4. ACID EFFUSIVE AND HYPABYSSAL ROCKS FROM THE MORAINES.

By J. O. G. GLASTONBURY 026

5. BASIC IGNEOUS ROCKS AND METAMORPHIC EQUIVALENTS FROM

COMMONWEALTH BAY. By J. 0. G. GLASTONBURY 056

6. CERTAIN EPIDOTIC ROCKS FROM THE MORAINES, COMMONWEALTH BAY.

By J. 0. G. GLASTONBURY 016

7. SCHISTS AND GNEISSES FROM THE MORAINES, CAPE DENISON, ADELIE

LAND. By A. W. KLEEMAN 12

AUSTRALASIAN ANTARCTIC EXPEDITION

1911-14

UNDER THE LEADERSHIP OF SIR DOUGLAS ttAWSON. D.Sc.. P.R.S.

SCIENTIFIC REPORTS.

SERIES A.

VOL. III.

GEOLOGY.

FART VI.

PETROLOCICAL NOTES

FURTHER ROCK SPECIMENS

COLLECTED FROM IN SITU OCCURRENCES

COMMONWEALTH BAY REGION.

J. O. G. GLASTONBURY, B.A., AYSc.

WITH TWO PLATES.

PRICE: THREE SHILLINGS AND SIXPENCE.

Wholly set up and ptinted in Australia bv
THOMAS HENRY THNNANT, GOVERNMENT PRINTER, SYDNKV, NEW SOUTH WALKS, AUSTRALIA

1940.

62830 A

CONTENTS.

PAGE.

I. Notes on further Specimens from the Metamorphosed Dyke Series of Cape
Denison Nos. 142, 421, 432, 632, 633, 635A, 666, 1240

1. Introduction 309

2. Mineralogical Character of the Individual Xenoliths 310

3. Genesis of the Xenoliths 313

4. The Amphibolites of the Metamorphosed Dyke Series 313

II. Notes on Additional Rock Types from Stillwell Island

1. Introduction ... ... ... ... ... ... ... ... ... 312

2. Petrography

(a) The Garnet Gneisses Nos. 699, 980 313

(b) The Acid Hypersthenic Gneisses Unrepresented

III. Additional Petrographic Notes on Rocks from Madigan Nunatak

1. Introduction 321

2. The Plagioclase-Pyroxene Gneisses Nos. 775, 783, 788, 792, 794

(a) Modal Composition of the Rock 321

(b) A Discussion of the Mineral Characters of the Rocks 322

3. The Hypersthene-Alkali-Felspar Gneisses Nos. 778, 779, 787, 790, 791, 793,

795, 797, 798, 1226, 1227, 1254-

(a) Macroscopic Features 325

(6) Mineralogical Characters 326

4. Summary 328

Description of Plates XLIV and XLV 330

PETROLOCICAL NOTES ON FURTHER ROCK SPECIMENS,

COLLECTED FROM IN SITU OCCURRENCES.
COMMONWEALTH BAY REGION.

J. O. G. GLASTONBURY, B.A., M.Sc.

WITH TWO PLATES.

I. NOTES ON FURTHER* SPECIMENS FROM THE METAMORPHOSED DYKE

SERIES OF CAPE DENISON.

1. INTRODUCTION.

" ONE band of amphibolite, outcropping near the centre of the Cape Denison area, is
phenomenal in containing a large number of xenoliths. There are two distinct types
of material among these meta-xenoliths, and they may be distinguished as

(1) Saussuritic type ;

(2) Gneissic type.

The saussuritic type includes pale green and pale pink masses which may be
again subdivided into

(a) Those composed wholly of saussurite the individual type.

(b) Those composed of an aggregate of saussurite and hornblende the composite
type." (Stillwell : Series A, Vol. Ill, Pt. 1, p. 48.)

This statement of StillwelTs is given because the rocks to be discussed here
(Nos. 142, 421, 432, 632, 633, 635A, 666 and 1240) are all members of the dark
amphibolitic series found in situ at Cape Denison.

No. 421 is the coarsest of the set. As seen in the hand specimen it consists of
handsome blades of black hornblende set in a base of coarsely crystalline opaque felspar.
Its structure is gneissic.

Nos. 432, 666 and 1240 are examples of the amphibolite xenoliths containing
porphyritic saussurite individuals. They are of " The individual type of meta-xenolith "
variety according to Stillwell's classification.

No. 142 is another of the individual type of meta-xenolith, but owing to the few
phenocrysts of saussurite and the definite schistosity it has taken up it seemed well to give
it a further description.

*Dr. StiUwell's original report deal* with outstanding examples of the dominant type* of rock* met with in *ilu.
Rocks described in thU contribution are, in the main, closely similar to example* selected and described by Stillwell. The
present descriptions are supplied in order to render more complete the microscopical examination of the whole of the
rock collection. P.M.

310 AUSTRALASIAN ANTARCTIC EXPEDITION

Nos. 632, 633 and 635A are specimens of amphibolites from the metamorphosed
dyke series of Cape Denison. These specimens are like Nos. 634, 634A and 635
described by Stillwell in this connection. They have developed schistosity to a
considerable degree by a parallel arrangement of hornblende prisms. Both Nos. 632
and 633 are markedly fissile because of the perfection of this parallelism.

2. MINER ALOGICAL CHARACTERS OF THE INDIVIDUAL XENOLITHS.

Under the microscope Nos. 142, 421, and 431 fall into one mineralogical division
and Nos. 666 and 1240 in another. The differences are not so much in actual mineral
content as in the nature of the crystal development. In the first three of these rocks
the felspar, not only in the phenocrysts themselves, but throughout the whole
rock-mass is turbid and highly saussuritised. The amphibole present is pale green in
colour and is definitely actinolitic in character. The crystals are elongated but small.
They have not the massive, compact appearance of typical hornblende. Their colour,
as a matter of fact, tends to blue, suggesting that the glaucophane molecule has been
concentrated in them to some extent. No. 666 serves as a transition stage. It is marked
by the plagioclase intermingled with the amphibole crystals, as distinct from that in the
saussurite phenocrysts, being pellucid. This shows that the felspar of these regions has
completely recrystallised and either assumed again the matter rejected during
saussuritisation or expelled it by diffusion to be taken up in the composition of the
hornblende. The amphibole in this rock is not so actinolitic as in the first three, but it is
by no means so compact and green as in rock No. 1240.

In this rock, No. 1240, the chief difference from the others mentioned above, is the
intense green and brownish green colour of the amphibole, which now is quite definitely
hornblende. Another striking difference is the almost entire absence of saussurite which
is present in the nuclei of two large phenocrysts only.

Other mineralogical points of interest are first, the relationship of sphene and
nuclear ilmenite. The development of this is shown quite well in these rocks. In No. 142
the border of the sphene is quite narrow, sometimes less than 0-005 mm., while in the other
rocks there is a range of 0-05 mm., and perhaps even more. Secondly, a genetical
relationship between the three minerals ilmenite, sphene and zoisite (or clino-zoisite, or
epidote as the case may be) and felspar is shown by a similar serial arrangement.

Other reactions which involve sphene are best seen in Nos. 142 and 421. In these
rocks a more fibrous biotite than usual, whose pleochroism is from Z a weak brown to
X practically colourless, is found associated with grains of sphene, hornblende and
nearly colourless chlorite (Plate XLV, fig. 1). The chlorite is derived from the biotite,
its development being a possible explanation of the bleaching of the colour of the mica.
The sphene seems to be forming chiefly at the expense of the hornblende which, as a
result, is poikiloblastic in texture and studded with small inclusions of the sphene.
But a similar relation exists, to a more limited extent, between the biotite and sphene.

PETKOLOGICAL NOTES ON FURTHER ROCK SPECIMENS. 311

This is more particularly noticeable where biotite and hornblende crystals interlock,
that is, where it is more definite that the two minerals were formed by the same general
alteration of the augite of the dolerite rocks. Epidote, too, is associated with this brown
biotite, but is a rim-growth rather than an inclusion.

The minerals epidote, zoisite, clino-zoisite are usually found with white mica in the
saussurite matrix. The lawsonite, however, is not so restricted as the others ; it occurs
(in addition to the above manner) as veins which penetrate the amphibole masses and in
localised patches in the neighbourhood of the amphibole. Frequently zoisite is found
as irregular scales which thread through the saussurite. Under crossed-nicols it is seen
to form an ultra-blue background in which the other alteration minerals and relic felspar
occur.

In Nos. 666 and 1240, there is practically no sphene, and no biotite or chlorite.
This shows that when the more complex amphiboles are produced in the higher stages
of metamorphism they have absorbed into their composition much of the matter of both
sphene and biotite. The disappearance of white mica, epidote and zoisite, and the
recrystaUisation of the felspar in a more acidic form emphasise the increased complexity
of the new amphibole molecule ; and this variation in chemical content is reflected in the
different appearance of the mineral.

3. GENESIS OF THE XENOLTTHS.

This has been discussed by Stillwell on pp. 53-54 of his work. He immediately
(pp. 54-5) proceeds adequately to discuss the significance of the meta-xenoliths. In
view of this, nothing further will be given here on these matters.

4. THE AMPHIBOLITES OF THE METAMORPHOSED DYKE SERIES.
This account is concerned with the mineral composition and the microscopical
characters of the amphibolite rocks Nos. 632, 633 and 635A.

No. 632 differs from No. 633 under the microscope. Although it shows some
tendency to linearity of amphibole crystal arrangement, yet, on the whole, it is
advisable to call its texture granoblastic. In No. 633, however, there is a marked
schistosity produced by the parallel orientation of its amphibole crystals. This
schistosity is accentuated by the elongation of the crystal prisms whose average elongation
factor (length : width) is about 3 ; this is exceeded somewhat in regions where smaller
crystals are interwoven with felspar ; it is fallen short of in the more massive parts of the
rock. Again, in No. 632, there is a fairly even distribution of parts exclusively
amphibolic and parts essentially composed of both amphibole and felspar. No. 633,
on the other hand, has these mineral groups arranged in parallel alignment, producing
a completely different aspect.

312 AUSTRALASIAN ANTARCTIC EXPEDITION,

The minerals present in these two rocks are much alike. The dominant one is a
pale blue-green actinolite. This is sometimes concentrated in zones free from other
minerals, and sometimes associated with highly saussuritised felspar. The saussuritic
mass contains epidote, zoisite and clino-zoisite. No. 632 contains an occasional lath
of biotite, but this mica, as an unusually pale variety, is one of the determinative
minerals of the finer-grained regions of No. 633. Each rock contains some granules of
sphene, rare grains of ilmenite and crystals of apatite.

No. 635A is a rock in which the simpler pale-green actinolite has been recrystallised
into more complex green amphibole. This recrystallisation of amphibole has been
accompanied (or preceeded) by a recrystallisation of felspar. The hornblende crystals
are small, but no longer exclusively elongated, as a matter of fact they are usually
equi-dimensional. They do not exhibit such marked parallelism as do the crystals of
amphibole in No. 633. Sphene has practically disappeared, the biotite has a much
stronger reddish brown colour. These facts, together with the deeper green colour of the
amphibole, again emphasise the way in which the amphibole (and the mica too) gain
complexity in the higher grades of metamorphism by the absorption of the " rejection
products " of saussuritisation and of minerals like sphene. Lawsonite and the suite of
epidote minerals are practically absent.

The recrystallisation of the felspar has resulted in the formation of many grains
of quartz. The new felspar is frequently twinned, both albite and pericline multiple
twin lamellae being seen.

Some little iron ore (magnetite) and rods and needles of apatite are also present.

These three rock types can be summarised thus :
No. 632 Chlorite-albite-epidote-amphibolite.

No. 633 Chlorite-albite-epidote-amphibolite-schist.
No. 635A Amphibolite.

II. NOTES ON ADDITIONAL ROCK TYPES FROM STILLWELL ISLAND.

1. INTRODUCTION.

A discussion of some of the rock types found at Stillwell Island has already been
given in these Scientific Reports (Stillwell, Vol. Ill, Pt. 1). A brief resume of some of
the points made there is necessary for an understanding of the sequel.

" Stillwell Island is one of the largest members of the Way Archipelago.

The most conspicuous rock is a massive granitoid gneiss, often carrying abundant
dark aggregates of garnet and mica. Varieties of gneiss are also found without any
garnet at all, and the higher part of the island is formed of an acid hypersthene gneiss.

PETROLOGICAL NOTES ON FURTHER ROCK SPECIMENS. 313

In crossing the islet areas are found consisting of more strongly foliated gneisses, and the
trend of the foliation is a little west and north. Irregular bands of black gneiss, with
dyke form, exist here as at Cape Denison, and some are full of fine garnet." (pp. 145-6.)
Mention is then made of

(a) Garnet gneisses (No. 917), p. 151 ;

(b) Acid hypersthene gneisses (Nos. 979, 947), p. 155;

The present contribution is devoted to a further description of these groups with
mention of specimens not dealt with by Dr. Stillwell, but which show variations from
the main types and important differences in detail.

The specimens dealt with are Nos. 59, 699, 972, 973, 974, 976, 977, 978, 980.

It is found that Nos. 699 and 980 fall in Group (a), the Garnet Gneisses, and Nos.
591, 972, 973, 974, 976, 977 and 978 are all members of the metamorphosed basic dyke
series, Group (c).

2. PETROGRAPHY.

(a) The Garnet Gneisses.

As mentioned above the additional members to this group of gneisses which occur
both at Cape Pigeon Rocks and Stillwell Island are Nos. 699 and 980.

No. 699. Garnet-Felspar-Quartz-Gneiss.

This rock consists of alternate bands of biotite and quartz-felspar matter. Through
the rock, but concentrated to a greater extent in the vicinity of the biotite, are clusters
of deep red garnet crystals, some of which assume a high degree of crystalline form.

In some respects this rock resembles Nos. 777 and 917 described by Stillwell.
The former was collected at Garnet Point and the latter at Stillwell Island. There are
large porphyroblasts of garnet, and although they are to some extent penetrated by
biotite, they are not skeletal like those of No. 777. The mica developed is of two kinds.
One is biotite which is found as a greenish-yellow fibrous mass forming an aureole around
the garnet, and as more compact masses pleochroic in dark and light browns. The other
mica is sericite which has developed during the alteration of potassic felspar.

The areas not so immediately associated with the garnet consist of quartz,
plagioclase, orthoclase, microcline, microcline-microperthite, sericite, calcite and
accessory apatite.

The quartz is frequently undulose and contains inclusions of fluorite arranged in
long parallel lines. The fact that the fluorite shows abnormal birefringence is evidence
that the mineral has suffered internal tension. The quartz and orthoclase frequently
occur in a myrmekitic intergrowth ; other varieties of intergrowth occur, of which the

314 AUSTRALASIAN ANTARCTIC EXPEDITION.

commonest consists of blebs of quartz included in areas of orthoclase. The felspar,
when plagioclase, is very highly saussuritised. Some lamellar twin forms, both albite
and pericline, are present. The orthoclase and microcline are, in part at least,
developments due to the concentration of the potassium from earlier plagioclase which
must have been highly potassic. A further development has been the production of
considerable sericite which is almost invariably associated with secondary calcite, the
mineral which has been formed by the contemporaneous concentration of the lime
molecule of the original felspar. The existence of quartz and calcite together is evidence
of the dynamothermal metamorphism to which the rock has been subjected.

Zircon occurs as nuclei of intense pleochroic haloes in the biotite and also as
isolated fragments in the orthoclase. Its brown colour is suggestive of monazite, but
its positive uniaxial nature differentiates it from that mineral.

The structures developed are porphyroblastic, shown by the garnet, granoblastic
shown by almost the whole of the remainder of the rock, and granulitic shown by the
calcite.

No. 59. Grarnet-Plagioclase-Pyroxene-Hornblende Gneiss.

A note with this specimen made at the time of collection (by Dr. Mawson) records
that it forms a broad band about 20 feet wide with ill-defined borders where the boat
was kellicked when a landing was made upon the island.

The rock itself is gneissic. It is composed of black pyroxene and amphibole
crystals set in a mesh-work structure of brown matter, whose content is made up of pink
garnet, felspar and a very little quartz.

The mineral content of the rock is pyroxene, hornblende, biotite, ilmenite, garnet,
andesine, quartz and pyrites.

The pyroxene is non-pleochroic. It always shows some indication of alteration.
There is exhibited a serial change in different parts of the rock. The simplest is the
separation of ilmenite as fine dust particles which make the mineral very dark and patchy.
There is usually, but not always, an accompanying development of green hornblende ;
particularly is this true in the peripheral areas where a border of hornblende round the
pyroxene is quite common. This last development sometimes occurs with grains of quartz
(or very acid felspar), thus showing that these two constituents have been formed
contemporaneously by the mineralogical alteration of the pyroxene.

The rest of the mineralogical changes are of interest for they show the progressive
stages in the formation of garnet by the interaction of felspar and hornblende as a
variation there is also seen the formation of garnet from the reaction between biotite
and felspar. The usual arrangement of these minerals consists of a central core of skeletal
ilmenite interwoven with brown biotite which gives way to a surrounding zone of
hornblende the ilmenitiferous content is much lower and the hornblende, in turn,
frequently terminates in phalanges which project into reaction areas of highly acidic

PETROLOGICAL NOTES ON FURTHER ROCK SPECIMENS. 315

felspar or quartz. These reaction areas are usually found as inner zones next to the
zone of garnet. They show how the amphibole and the original basic felspar combine
to form the garnet and to give rise to a separation of quartz. That this separation of
quartz is an intermediate stage of the metamorphism and not the final one, is made
evident where garnet and the green hornblende abut without an accompanying acidic
reaction zone. The acidic material set free in the earlier reaction has been taken up
completely in the final one and hornblende and garnet are, in consequence, in intimate
juxtaposition.

Exactly similar relationships are seen to hold wherever the biotite region has been
able to react with the original felspar. There are places where phalanges of biotite
mingle with clear siliceous matter (or highly siliceous felspar), and there are other places
where the biotite and garnet abut against one another. This is good evidence of the
complementary nature of the biotite and the hornblende ; the one being the more
ilmenitic derivative of the original pyroxene and the other the component richer in iron
and more especially magnesia. They show, too, how the original pyroxene has
progressively altered, first, setting free a swarm of ilmenite grains ; secondly, producing
a granulitic pyroxenic aggregate ; thirdly, the formation of green hornblende ; fourthly,
the subsidiary production of brown biotite by the concentration of ilmenite in the
ferromagnesian, and lastly the reaction between the new ferromagnesian and the original
basic felspar with the final production of a pink garnet but with the intermediate stage of
silica liberation and later phase of absorption of this matter.

Some indication of the structure of the rock has been given by what has been
said above. But the general arrangement of the mineral masses is that these more or
less circular masses are set somewhat regularly in a felspathic base. In places two or
more ferromagnesian alteration zones have coalesced and then the same internal
structures are seen to be preserved but there is now no separation region of felspar.
The circular nature of the pyroxene and its derivative minerals sometimes gives place
to an elongated formation probably due to the effects of non-hydrostatic stresses
where the length and breadth of the masses are roughly in the ratio 2:1. Where the
pyroxene is not much altered the average grain-size is about 0-55 mm. This size is
preserved remarkably consistently throughout all the metamorphic stages, and finally
there has been produced a core of ilmenite, biotite, and hornblende about 0-5 mm. across
with a surrounding aureole of garnet from 0-05 mm. to 0-10 mm. wide. The distance
between the adjacent nuclei of these masses is about 0-8 mm., leaving an average space
of 0-10 mm. to 0-15 mm. occupied by the felspar.

No. 980. A coarse Garnet-Plagioclase-Cordierite-Biotite Gneiss.

A coarse-grained garnetiferous rock was collected from the summit of Stillwell
Island. The perfectly developed garnet crystals are set in a matrix of dense greyish
felspar. A contact region is present in the specimen examined, where a considerable
quantity of biotite comes into the rock and where the grain-size is much reduced.

316 AUSTRALASIAN ANTARCTIC EXPEDITION.

Regional metamorphism has been the chief determinative factor in the constitution
of this rock. The high temperature-pressure conditions which prevailed were followed
by thermal metamorphism where pressure no longer was one of the dominant factors
operating. This throws new light on the metamorphic transformations which have taken
place in the rocks of the island.

The large garnet crystals which dominate the slide have been formed under the
first set of conditions mentioned above; but the masses of strongly coloured brown
biotite and little accompanying flaky chlorite scattered throughout their whole mass
are more likely products of the second set of circumstances. As is true in the case of the
red garnets of the Scottish Highlands (Harker, p. 324), so here the formation of biotite
which has assumed a, decussate arrangement in some places and rosette forms in others
has from the garnet liberated ilmenite, which is seen in skeletal crystals, and some
interstitial cordierite which, though clear, is usually intensely crushed. The cordierite is
recognised by the presence of intensely pleochroic yellow haloes and its biaxial nature
(2V large, approximately 80). The alteration of the chemical nature of the minerals
of the rock has proceeded, pari passu, with the disintegration of the crystal form.
Smaller, more or less skeletal, masses of garnet are separated from one another by the
products of alteration, ilmenite, cordierite and biotite.

The pleochroism of the biotite (apparently titania-rich) is intense, the colours
being Z deep red-brown, and X very pale brown. The absorption as usual is Z>X.
Other features of the biotite are the presence of numerous pleochroic haloes with zircon
nuclei and the bent nature of the laminae. The bending shows that a shear stress has
been operative to a perceptible degree.

In the coarser parts of the rock the acidic plagioclase present occurs in large grains
which, however, have been considerably crushed, but more especially is this noticeable
in the neighbourhood of the biotite and the garnet. The plagioclase is not so clear in this
rock as in most rocks from this area, nevertheless it shows albite and pericline twins
very well developed. Bends in the twin lamellae are further evidence that shear has
occurred in the history of the rock. Quartz grains are associated with the plagioclase,
particularly in the parts where granulitisation has been most intense.

In the finer parts of the rock practically no garnet is seen at all. In its stead
is a granulitic aggregate of cordierite and clear acid felspar shot through with flakes of
brown biotite and scales of very green chlorite. The disintegration which is noticeable
in the very large garnet crystals of the coarser part of the rock has been able completely
to change the nature of the smaller garnets and this mass of cordierite, acid plagioclase,
biotite and chlorite, maintaining the position and size of the earlier garnet, is
paramorphous after the garnet. The granular nature of the aggregates is realised more
clearly under crossed nicols, for then the heterogeneously oriented granules of
cordierite form a definite mosaic.

PETROLOGICAL NOTES ON FURTHER ROCK SPECIMENS. 317

Except for an inevitable difference in size the plagioclase of the finer parts of the
rock is much the same as that found in the coarser.

(6) The Acid Hypersthene Gneisses.

This division recorded by Stillwell is unrepresented among the present series
of specimens.

Sir Douglas Mawson recorded in his diary (p. 171 of present series), " irregular
bands of black rock exist as at Cape Denison : some are not much altered, and others
are full of fine garnet."

Dr. Stillwell then continues to note that some of the rocks exhibit " remarkable
stages of incipient alteration." An alternative hypothesis seems to be suggested by
the more recent work of Harker and Wiseman in connection with the so-called epidiorites
of the Scottish Highlands : that these rocks are high grade metamorphic types, where
hornblende is giving place to two pyroxenes, garnet is developing and granulitic texture
is characteristic.

No. 972. Hornblende-Plagioclase-Pyroxene Gneiss.

This is a dense, fine-grained rock. Its apparent uniformity is resolved on close
inspection into dark masses of relatively coarser material and dark spots whose
grain-size is considerably smaller. An ill-defined band where biotite and quartz are
megascopically visible also tends to relieve the uniform nature of the rock.

The biotite and hornblende are not without signs of alteration. The grains of
these minerals are arranged in the typical decussate manner, but particularly in the case
of the hornblende, instead of being compact they are definitely poikiloblastic due to a
development of quartz. Associated with the biotite and hornblende, but more
particularly with the biotite, are diablastic and mynnekitic intergrowths of quartz and
orthoclase. The biotite prisms are still oriented in the manner referred to above, but
owing to the nature of their development unchanged pyroxene granules remain as
inclusions. The relationship between these intergrowths and the biotite (and hornblende)
is revealed by the bays and bights in the ends of the prisms and frequently by their
presence in the more crystalloblastically powerful sides of the prisms.

The granulitic areas are composed of small rounded grains of pyroxene set in a
matrix of plagioclase, probably labradorite, which has been formed by recrystallisation.

An imperfectly formed vein of quartz traversing the rock suggests that there
has been a possible acquisition from without of silica which may have been effective
in supplying this material to the ampbiboles which are usually below the meta-silicate
stage in rocks of this kind.

The existence of considerable hydrostatic pressure is evidenced by the presence
of garnet crystals, and, this pressure, no doubt, has played a considerable part in the
recrystallisation of the pyroxene.

318 AUSTRALASIAN ANTARCTIC EXPEDITION.

A thin vein of chlorite, and some flakes of calcite suggest that in the very late
history of the rock it has been subjected to a very slight metamorphism of a retrograde
nature.

The rock is a little higher grade than No. 974. The hornblende-biotite masses
are more coherent, and the granules of pyroxene are much more numerously developed.
The greater abundance of pink garnet and the presence of some sphene indicate the
greater degree of metamorphism the rock has undergone.

No. 973. Plagioclase Amphibolite.

This is another of the amphibolites which were found in the dykes which cross
the gneiss on Stillwell Island. It is a little more schistose than is usual with this suite
of rocks, but preserves the main characteristics, medium-grain size, black, dense, and
hard.

The rock differs completely from those described by Stillwell from this source,
but is much like certain plagioclase amphibolites from the moraines at Cape Denison
which have come under notice, described by the present writer.

The ferromagnesians present are almost entirely confined to green compact
hornblende whose pleochroism is strong and which has been described before in the
place referred to above. The grains are nearly equi-dimensional, they are arranged
in the decussate manner, usually they show the two amphibole cleavages quite well
but where prismatic sections are seen only one cleavage is, of course, to be perceived.

Some brown biotite, passing into chlorite is present. There is also some green
somewhat fibrous biotite, but here the alteration to chlorite is not quite so strongly
developed. The only other dark mineral component is magnetite which occurs in fairly
well-formed crystals, but irregular outlines are presented by some grains where
neighbouring particles have coalesced to make a larger individual. Occasional skeletal
crystals of this mineral occur.

The felspar which is the mineral second in abundance is andesine with Ab 66 An 35 .
It is quite clear and usually gives sharp extinction. There is no doubt of its secondary
nature. Its pellucid nature and the absence of every trace of original structure are
conclusive evidence of this. It exhibits both albite and pericline multiple twinning,
the latter sometimes appearing in very delicately defined lines which are unusually
well-developed. A little quartz is associated with the plagioclase, but its amount is
remarkably small.

Apatite occurs quite freely in small well-shaped rectangles, hexagonal basal
sections, and prism sections. It is also found in the felspar as acicular rods which
suggest that they have been able to resist the changed conditions which have so
considerably affected the rest of the rock.

PETROLOGICAL NOTES ON FURTHER ROCK SPECIMENS. 319

The presence of some few scales of calcite afford evidence that dynamic effects
have been more potent in the alteration of the rock than heat, though, of course, the
part played by the latter factor has been by no means inconsiderable.

No. 974. Hornblende-Plagioclase-Pyroxene Gneiss.

This rock, like No. 942, which was described by Stillwell, occurs in dyke-like bands
up to 10 feet wide crossing the garnet gneiss. The rock is dark, dense, and fine-grained,
although the granular nature is differential, in that there are fine-grained portions
surrounded by more coarsely crystalline rock. One face of the hand specimen shows
a remarkable development of pale brass-yellow sulphide ore, probably pyrrhotite.
Of the faces which were exposed to weathering, one has the typically indurated
appearance common to the rocks of this area, and another has been coloured yellow
by the formation of some limonite by oxidation and hydration.

The rock may have been formed in the manner Stillwell suggests for No. 942, but
the presence of an occasional crystal of pink garnet shows that the metamorphism
suffered by the rock has been a little more intense, or that there has been a slight
variation in the composition of the rock.

The rock, like No. 972, contains some hypersthene as well as monoclinic pyroxene,
calcite, well formed grains of magnetite and some crystals of apatite. An occasional
pleochroic halo, with zircon as-nucleus, is found in the biotite.

No. 977. Garnet-Plagioclase Amphibolite.

This is a moderately medium-grained, massive rock in which the most prominent
minerals are red garnet and biotite. Hornblende is quite abundant and felspathic areas
are readily observable.

The most striking feature of the rock is the presence of remarkable skeletal
crystals of ilmenite whose peripheries are composed of many curved hooks or fingers
which come from the main mass. Their relationships with the green hornblende of the
rock is made manifest by the reaction zones which exist between the two. These reaction
zones consist of a dactylitic intergrowth of felspar and pyroxene called diablastic by
Stillwell, but the term used here is preferable as it emphasises the finger-like nature of the
components which enter into the intergrowth. The ilmenite is the nuclear region where
concentration of iron (and titanium) from the hornblende is occurring : the dactylitic
region shows where the actual diffusion of matter can be regarded as occurring. The
source and the goal of the transfer are at once made evident.

Occasionally crystals of garnet are set in a rim of clear felspar which separates
it from hornblende. In the felspar are to be seen very minute, but well shaped,
rectangles of pale green amphibole which seem, in part, at least, to play a similar part

320 AUSTRALASIAN ANTARCTIC EXPEDITION,

to the pyroxene fingers in the transfer of material mentioned in the previous paragraph.
These rims of felspar, with hornblende and sometimes biotite, are the reaction zones
between the hornblende and the garnet nucleus. The possibility of felspar being
potential hornblende to some degree at least is thus emphasised, and an indication is
given that the relationship is reversible under suitable conditions.

Such masses of ilmenite, as previously referred to, are as frequently found
associated with garnet as they are with hornblende. The fingers of the dactylitic
intergrowth are still alternate bands of felspar and green pyroxene, but this time there
is evidence that the pyroxene is a diffusion path of titanium as well as lime. It is quite
usual to have the following serial relationships, hornblende connected to ilmenite nucleus
by dactylitic felspar and pyroxene in a narrow rim, and then, the ilmenite nucleus
connected to the garnet crystal by another such dactylitic rim, but this time, wider.

The garnet crystals have not assumed their perfect crystal outline, an indication
that they are still undergoing the process of formation. They are remarkably embayed,
and contain a large number of inclusions which include quartz, ilmenite, hornblende and
biotite which produce the unusual effect of poikiloblastic garnets.

The other main minerals developed are hornblende, biotite, andesine and quartz.
Accessory minerals include sphene, zircon in minute crystals, and apatite.

Rock No. 976 is merely a duplicate of this (No. 977).

No. 978. Garnet-Plagioclase-Amphibolite.

This rock, which was collected near the boat moorings, looks very much like
No. 977, except that together with an increased micaceous content there has developed
a more pronounced schistosity. This feature, however, has not been developed to such
a degree as would warrant calling the rock a schist it is still a gneiss. Another factor
which favours a reduction in the gneissic nature of the rock is the smaller amount of garnet
present, but, even so, garnet is still one of the most important minerals in the rock.
Other minerals which can be recognised by the naked eye are hornblende and felspar.

Weathering has converted the biotite of the superficial layers into a golden
coloured mica and has set free considerable quantities of oxidised iron compounds with
the result that the exposed surfaces have assumed a deep yellow-brown colour.

The microscopic examination confirms the conclusions formed from the study
of the hand specimen. Except for a slight increase in the amount of biotite present the
rock differs in no material respect from No. 977. There is the same green, compact
hornblende, garnet, ilmenite and hornblende in the same relations as before, and
leucocratic areas of quartz and plagioclase.

1'ETKOLOGICAL NOTES ON FUUTHKU K< >CK SPECl.MKNS. 321

III. ADDITIONAL PETROGRAPHIC NOTES ON ROCKS FROM

MADIGAN NUNATAK.

1. INTRODUCTION.

A brief summary of StillwelTs description of the locality and the kinds of rocks
found on the Nunatak is necessary.

He says (p. 128), " the Madigan Nunatak is situated in Lat. 67 8j' and Long.
143 20', about 30 miles distant from Cape Denison . . . , and 18J miles from Cape
Gray."

He gives photographic views of the Nunatak (Plate XXIV, figs. 1 and 2).

He further states, "it is composed of gneissic rocks whose foliation strikes
approximately north and south " ; and " two rock types are found in this area. One
is a black massive plagioclase-pyroxene-gneiss " which " seemed to form a band whose
trmd cuts at right angles across the foliation. The second type is the n. ore abundant
acid gneiss, containing blue quartz and hypersthene. In the neighbourhood of the
ant it -line it has a banded character, but in other parts the gneissic character, though
evident is less prominent."

2. THE PLAGIOCLASE-PYROXENE-GNEISS.

(a) Modal Composition of Rocks.

Stillwell has taken specimen No. 794 as the standard of this type of rock. Four
other specimens were collected by Stillwell's sledging party in the summer, 1912-13.
He has not given a description of any of these specimens, viz., Nos. 775, 783, 788, 792.
They are readily seen to be akin to the standard but differences are n< ne-t he-less readily
observed.

In the hand specimens the rocks are black, dense and fine-grained. They
resemble very fine-grained dolerites when viewed macroscopically. The minerals seen
in the hand specimens are pyroxene and felspar. As in the case of No. 794 the weathered
surfaces of these specimens are discoloured by a brown iron stain.

The textures of the rocks shown by the microscope are granoblastic. "with
subsequent modification by cataclastic effects." The average grain-size is about 0-25
nun., although No. 783 is somewhat finer, its average grain-size being approximately
0-17 mm.

:) B

322

AUSTRALASIAN ANTARCTIC EXPEDITION,

The modal mineralogical contents of the rocks (determined by use of the Leitz
Integration Table) are shown in the following table where they are compared with that
of the reference specimen, No. 794 (determined by Dr. Stillwell).

Mineral.

No. 775.

No. 783.

No. 788.

No. 792.

No. 794.

Felspar
Pyroxene
Hornblende

55-0
324
1-1

46-9
324

53-3
38-1
2-2

36-3
36-6
21-8

42-5
45-5
3-3

Biotite

13-9

Present.

0-3

Iron Ore
Total

6-1

6-8

2-9

8-4

100-0

It will be seen that the general tenor of the rocks is the same, yet, notwithstanding
this, notable variations occur. The felspar ranges from a minimum of 36-3 per cent,
in No. 792 to a maximum of 55-0 per cent, in No. 775. The pyroxene of No. 775 is
least, viz., 32-4 per cent, (the same amount as in No. 783) and that of No. 974 is most,
viz., 45-5 per cent. A remarkable variation in the amount of hornblende is shown by
the table. It is practically absent from No. 783 and yet in No. 792 it forms over one-fifth
of the rock. A correlation between the inverse of felspar and hornblende can be drawn.
Considerable variation is also shown in the biotite content, it reaches 13-9 per cent, in
No. 783 and practically disappears in Nos. 788 and 794. The iron ore content is nearly
uniform, although No. 792 differs somewhat from the rest.

A triangular graph (Fig. 16) showing percentages of felspar, F, pyroxene, P, and
metamorphic matter, M (viz., hornblende + biotite + iron ores) is given. A serial
relationship from C to E to A to B to D is seen. This shows the progression from
minimum metamorphic matter in C to the maximum in D. (The letters A, B, C, D, B
refer respectively to rock specimens Nos. 775, 783, 788, 792, 794.)

The diagram suggests that part of both the original felspar and pyroxene is taken
up in the formation of these metamorphic products (hornblende, biotite, and iron ore),
but this aspect of the matter will be treated more fully in part (c) below.

(b) A Discussion of the Mineral Characters of the Rocks.

As the Rosiwal analysis shows there are five main minerals present, felspar,
pyroxene, amphibole, biotite, and iron ore. Accessory minerals include apatite and
zircon.

The felspar present in all four rocks is plagioclase, although a little antiperthite
is found in them all.

The natures of the various plagioclases present have been determined by
refractive index methods and by the maximum extinction angles in the symmetrical
zone.

PETROLOOICAL NOTES ON FURTHER ROCK SPECIMENS. 323

Two of the rocks, viz., Nos. 783, 792, are like No. 794 of Stillwell in that they have
two plagioclases present. In No. 783 there is a very small quantity of plagioclase which
has Da = 1-538 (approx.). Accordingly its chemical composition is oligoclase \\ith
Ab 80 An 20 . Most of the plagioclase in this rock gives a maximum symmetrical
extinction (X' A 010) of 20, showing it to be andesine with Ab^ An 37 . In No. 792
some of the plagioclase has n a 1-535 and n Y 1-545. This determines its composition
as oligoclase with Abgg An 16 ; but here again, most of the plagioclase is more basic.
The extinction (X' A 010) is 28, corresponding to labradorite, Ab 48 An 52 .

The other two rocks, viz., Nos. 775, 788, carry only the more basic plagioclase,
labradorite, whose composition is approximately Ab so An 50 .

The felspar in each of the four rocks shows both albite and pericline multiple
twin lamellae. In every case the twinning is indistinct and tends to be discontinuous
along the length of the laths. No. 792 has the least indefinite twin forms found in
this suite of rocks, but even so there is a tendency towards the elimination of twinning
in the central portions of the plagioclase grains. In No. 775 the concentric circular
nature of the pericline twin lines is an indication of the deformations produced in this
mineral by stresses.

Other stress effects are evidenced by the undulose extinction and the mortar
structure associated with the peripheral granulation of the felspar.

None of the rocks shows saussuritised felspar, but every one contains felspar which
holds inclusions of several kinds. No. 775 has abundant inclusions of magnetite dust,
indiscriminately arranged flakes of biotite and green shreds of hornblende, and some
crystals of zircon. No. 783 is much the same. The felspar of No. 788 is much cleaner,
being practically free from hornblende and biotite inclusions, although it still has some
magnetite and shows a new complication, viz., the presence of granulated pyroxene
inclusions. There is a recurrence of the biotite flakes in No. 792, but here they tend
to be arranged as tongue-like forms along cracks and cleavage lines rather than
indiscriminate distributions through the grains as in No. 775. Magnetite dust is present
in this specimen and also small granules of pyroxene (cf. No. 788), which are concentrated
in the vicinity of larger grains.

These four rocks, like No. 794, contain both monoclinic and orthorhombic
pyroxenes, the latter always being in considerable excess.

The orthorhombic pyroxene in No. 775 has np = 1-710 (approx.), D.R. = 0-014.
Its elongation is positive, its optic sign negative, 2V is large. These characteristics*
are those of the hypersthene member of the enstatite-hypersthene series which carries
a molecular percentage of 30 of FeSi0 3 . In No. 783 n a is slightly less than 1-680, which
shows that the ratio MgSi0 3 : FeSi0 3 is greater, there only being 20 per cent. FeSi0 3
present. The mineral is still hypersthene.

* Wfaohell, 1927. VoL II, p. 177.
80-C

324 AUSTRALASIAN ANTARCTIC EXPEDITION

In all four rocks the hypersthene is pleockroic in light tones of green and pink ;
Z is always pale green and X light pink. The absorption is not strong, the formula is

The monoclinic pyroxene in all four cases is a pale green to colourless augite.
It is optically positive, 2V large, the extinction (Z A c) is about 45, the D.R., 0-024.

In No. 792 occasional simple twins on Oil occur among the pyroxenes.

The pyroxenes are usually clear, but the augite of No. 792 exhibits good schiller
structure, and not infrequently holds inclusions of magnetite (both as abundant dust
and larger grains), flakes of biotite, felspar (and possibly a little quartz the augite is
here poikiloblastic), and hornblende.

Granulation, usually peripheral, of the pyroxene is evident in all four rocks, but
more especially in No. 792.

Certain genetic relationships between the pyroxene and other minerals hold,
but these will be discussed below.

The hornblende of No. 792 has as limits for its refractive indices n a = 1-66,
np = 1-67. n r = 1-69. The extinction (Z A c) is 15. These data, according to
Winchell (op. cit., fig. 139, p. 224), correspond to that member of the pargasite-hornblende

series whose chemical composition is approximately :

Per cent.

NaFeSi 2 6 + Fe 2 3 ............... 16

CaMgSi 2 6 + MgSi0 3 ............... 42

CaFeSi 2 6 + FeSi0 3 ............... 42

That is to say, it is a true hornblende. It is pleochroic, with Z brown-green, Y
green-yellow, and X pale straw-yellow. The absorption is strong. The formula is
the usual one, Z > Y > X.

In the other two rocks in which hornblende occurs to any appreciable amount
(it is almost absent entirely from No. 783, see Rosiwal analysis above) it has the same
characteristics as those given for No. 972. This rock was used as the standard of reference
because of the greater abundance of hornblende in it.

The texture of the amphibole is usually granoblastic, although in places it is
found in flakes and shreds. It holds frequent inclusions of magnetite (ilmenite). Its
border is usually granulitic, in which respect it resembles the pyroxene.

The biotite of the rocks is deep reddish-brown in colour. It is pleochroic in lighter
and darker shades, and has very strong absorption, Z > Y > X.

PETROLOOICAL NOTES ON FURTHER ROCK SPECIMENS.

325

This biotite is associated with pyroxene, felspar, and ilmenite, frequently, but
not always, with hornblende, and in No. 775 at least, with scaly calcite : it is
intimately associated genetically with these associates, but see below, part (c).

It is always secondary (also see part (c)). In places, particularly in Nos. 783, 792,
there are crystals whose curved outlines with sweeping bends suggest change of direction
during growth produced by the variation of the prevailing stress-direction.

The magnetite (ilmenite) present in all these rocks has often been referred to
above. Its mode of formation and other relationships to the remaining minerals will
be treated below, see part (c).

We can assume that the original rock was a felspar-pyroxene one. If we suppose
that the metamorphic products, M, represent both original felspar and augite, and that
all the iron ores came from the augite (some, of course, may have been original), and
the amphibole and mica came equally (the simplest assumption, though, necessarily,
unsubstantiated) from the original minerals, the modal compositions of the parent rocks
are as follows :

Minerals.

No. 775.

No. 783.

No. 788.

No. 792.

No. 794.

Av. Comp.

felspar

58-3

53-9

54-4

48-4

44-4

51-9

Pyroxene

41-7

46-1

45-6

51-6

55-6

48-1

We can justifiably assume that the original plagioclase was basic, a labradorite.
It is not so easy to tell if the orthorhombic pyroxene is original, although Hatch (op. cit.,
p. 415) suggests that hypersthene-gabbro, hypersthene-diorite and hypersthene-granite
may possibly be formed by differentiation, which implies the original character of the
hvpersthnio. If it is true as has been suggested (Tyrrell, 1930, p. 139) that one of the
distinguishing features of the charnockite series rocks from norite to pyroxene-granite
is the poverty of water-formed or water-rich minerals such as biotite and hornblende,
then an accession of water must be postulated to account for the presence of these minerals
in the present rocks.

3. THE HYPERSTHENE-ALKALI-FELSPAR-GNEISS.

(a) Macroscopic Features.

Stillwell (p. 133, et seq.) has dealt with Specimens Nos. 795, 797 of this series.
There remain Specimens Nos. 778, 779, 787, 790, 791, 793, 798, 1226, 1227, 1254 to be
described.

326 AUSTKALASIAN ANTARCTIC EXPEDITION.

These members of the second type of gneiss at the Madigan Nunatak are coarse-
grained rocks in which the gneissic structure can be detected (cf., Stillwell, p. 133). This
gneissic structure is much more prominent in some specimens than in others. For
instance, very definite leucocratic and melanocratic bands are visible in Nos. 779, 787,
but the texture of Nos. 798, 1226, 1227 is granulitic, and that of Nos. 778, 790, 791, 793,
1254, is best described as intermediate. The colours of the rocks taken as a whole also
form a series from light to dark with an intermediate group where neither colour
predominates. In the leucocratic group are Nos. 778, 790 and 1254, the intermediate
group consists of Nos. 779, 791, 793, 798, 1226, 1227, and in the melanocratic group is
No. 787. The macroscopically visible minerals are quartz (which is often blue, but
sometimes brown), felspar and hypersthene. Weathering of these minerals has
produced normally a brownish-red colouration, but more complete hydration has
sometimes resulted in the production of a remarkable mustard-yellow coating which is
particularly noticeable in No. 778. There is a considerable variation in the grain-size
of these minerals. Not only is this noticeable in the case of the hypersthene (cf.,
Stillwell) but in the felspar and quartz as well.

(b) Mineralogical Characters.

These rocks consist of quartz, orthoclase and plagioclase as their most abundant
constituents, but always with important hypersthene, biotite and ilmenite. Accessories
are apatite, pyrites and zircon (or monazite).

Most of the felspar is orthoclase, which is highly perthitic. The small inclusions
of plagioclase have a higher R.I. than their host. They are usually linearly arranged in
a great number of parallel lines which preserve their parallelism extraordinarily well.
The regularity of the size and spacing of the intergrowths is remarkable. No crystal of
orthoclase has been observed which is not perthitic. These intergrowths, as a matter
of fact, frequently follow a rectangular pattern, where, again, the regularity of size,
arrangement and orientation is remarkable. In rocks where the orthoclase
overwhelmingly preponderates (e.g., No. 790) there is occasionally some multiply twinned
acid plagioclase (albiclase, Ab 90 An 10 ) which has perthitic inclusions in the sense used
by Chudoba (Chudoba, translated by Kennedy, 1933, p. 18), that is, where the host
has a lower refringence than its inclusions. In other rocks, e.g., No. 1227, where more
plagioclase is present, large areas of antiperthite are frequently found. Where the
plagioclase is more abundant it shows multiple twinning badly defined, and has the
other properties of a little more basic variety, viz., oligoclase, with Ab 80 An 20 (approx.).
These differences in the composition of the felspar units through a single individual
give a distinctive appearance to the rock even when the nicols are parallel, but more so
when they are crossed. This appearance is enhanced by the effects produced by the
extreme cataclasis which the rock has suffered.

PETROLOGICAL NOTES ON FURTHER ROCK SPECIMENS. 327

The quartz of the rock shows equally well the cataclastic effects. Often there
has been produced in the rock a crude schistosity by the parallel elongation of the
crushed portions (cf. Stillwell, p. 134). These effects are rendered more obvious under
crossed nicols when the undulose extinction of the alternate bands differentiates one
streak from the next.

Where quartz and felspar are contiguous the crushing has produced new effects.
Mynnekitic intergrowths of the two minerals appear, usually in contact with a mosaic
of equi-dimensional quartz particles on the side adjacent to this mineral and a
well-developed mortar structure on the side abutting the felspar.

D ' "*

Fig. 16.

The pyroxene present is all hypersthene, which shows distinct pleochroism in
pinks and greens, with Z green and X pink. The absorption is marked, the formula is
Z > X. The appearance of the pyroxene is determined by the degree of cataclasis it
has suffered and the extent to which it has undergone metamorphism. The second of
these factors determines, to some extent, the amount of hypersthene present in the
rock, but it is not the only influencing factor. The amount of original pyroxene is, of
course, of importance. This is seen to vary from small quantities, as in Nos. 778, 790,
1227, 1254, to considerable amounts as in Nos. 779, 791 , 1226. The degree of granulation
appears to affect the readiness with which the pyroxene changes to ilmenite and biotite.
The textural effects produced are the lenticular appearance of the larger grains (e.g.,
No. 787), elongation of the grains (e.g., No. 790), a parallel linearity of the grains (e.g.,
No. 779) and streaked rows of minute granules (e.g., No. 798).

328 AUSTRALASIAN ANTARCTIC EXPEDITION.

The oblique extinction of some of the hypersthene suggested that the mineral
might be monoclinic pyroxene. The low refractive index and the negative optic
character of the mineral differentiated it from augite, and the parallel extinction of
cleavage flakes established its orthorhombic nature. Merwin* observed the same
phenomenon, and accounted for the oblique extinction by saying that the parting or
cleavage developed was parallel to b (010). Doubtless the same explanation holds
in this similar instance.

The biotite, ilmenite and hornblende of these rocks are the products of the
metamorphism of the pyroxene (usually with some inter-action with felspar). Where
the granulation of the pyroxene has gone to the stage of the streaking-out into fine
granules the production of biotite has been most easily accomplished ; but some of the
biotite is associated with the larger masses of pyroxene. It will be remembered that
inter-action with felspar has been mentioned above as a necessary part of the formation
of the final biotite product from the pyroxene. That this is so is shown by at least
two remarkable features of these rocks. The first is the presence of biotite streaks along
nearly every crack and cleavage line of some of the felspar grains, and the second is
the presence of row upon row of parallel lines of minute dark rectangular and rhombic
grains which represent the ultimate effect of the crushing of the original pyroxene.
The alternation of regions where these rows are concentrated with bands where the
felspathic matter is free from them has developed a microscopic gneissose structure
comparable in its essential features with those of macroscopic dimensions. Frequently
there is seen a development of biotite associated with these very fine grains. The other
streaks of biotite in the felspar where the very fine pyroxene grains are absent represent
the conditions which exist when all the content of the pyroxene has been altered (e.g.,
No. 793). Another feature showing the genetic relationship that exists between the
felspar and the biotite is the usual radiating growth of small biotite crystals from a
central core of ilmenite. Each of the small biotite crystals penetrates into felspathic
material. The effect produced is singular. From the arrangement of these three
minerals there can be no doubt that the growths represents a reaction between them.

4. SUMMARY.

From this treatment of the pyroxene-plagioclase-gneisses and the hypersthene-
alkali-felspar-gneisses it will be seen that the acidity of the rock is the factor which
determines whether the metamorphic derivative of the pyroxene shall be biotite or
hornblende. In those rocks where the acid-felspar content is relatively low (e.g., No.
792) the hornblende-content is great, but where the quartz-felspar-content is great
(e.g., No. 778, analysis below) hornblende is absent and the metamorphic derivative is
biotite.

* " The Charnockite Series of Igneous Rocks ": H. S. Washington, Am. Jour. Sci., XLI (1916), p. 331.

PETROLOOICAL NOTES ON FURTHER ROCK Sl'Ki I MKNS.

329

Tlie usual occurrence of pyroxene in rocks whose nature is definitely not acidic
obscmvs this two-fold mode of its alteration. It is only in this so-called " Charnockite
Scries " that much pyroxene is found in definitely acidic rocks. It is, then, possible to find
the metamorphic derivatives of pyroxene in acid surroundings in this suite only. Even
so, the usual slight metamorphism which rocks of this suite from other regions have
suffered has permitted Washington (op. tit., p. 335), after citing the occurrence of the
cliariKK kite rocks from Norway, Ellesmere Land, New York and West Africa, to say that
" biotite is rare or accessory and quite absent in most of the types." That this is not
necessarily so in the metamorphic representatives of the suite is seen from the Rosiual
analyses of the " Intermediate Charnockites " Nos. 775, 783, where 5-4 per cent, and
13-9 per cent, of biotite is present. The amount of pyroxene in each of these rocks is
32-4 per cent. The same reaction is shown more powerfully by the following table
where the volumetric compositions of two metamorphic derivatives (Nos. 778, 787)
of " hypersthene granites " are given :

Ill

904

82-7

6-7

6-0

Biotite

1-6

11-8

Orv-

1-3

I. Rock No. 778.
H. Rock No. 787.

III. Specimen (9-658) of the Indian Survey. Occurs in central part of Magazine
Hill, St. Thomas Mount, eight miles south of Madras. (Washington, op. cil.)

It will be seen that in each of these three rocks no hornblende has developed,
but there is an appreciable quantity of biotite, especially when referred to the
hypersthene in No. 787 it is nearly twice as abundant.

It can be concluded then that amphibolization of the pyroxene is the characteristic
alteration of this mineral in basic (and related) rocks, but that the formation of biotite
is the characteristic metamorphic process in rocks which have original pyroxene and
yet are acidic. Such rocks are the so-called hypersthene-granites and diorites and it
must be among the comparatively rare metamorphic derivatives of these rocks that
we must look to see the effects which have been suggested above. The rocks here
described from Madigan Nunatak seem to be among the most interesting from this
point of view.

330 AUSTKALASIAN ANTARCTIC EXPEDITION.

DESCRIPTION OF PLATES.

PLATE XLIV.
Fig. 1. Cordierite-Garnet-Plagioclase-Biotite-Gneiss, Stillwell Island (No. 980).

The micro-photograph shows a mass of cordierite and biotite derived from
the alteration of an earlier garnet crystal. Mag. 80 diams.

2. Hypersthene-Alkali-Felspar-Gneiss, Madigan Nunatak (No. 1227). The

micro-photograph shows small biotite crystals proceeding from a
large ilmenite grain and projecting into felspar. Mag. 44 diams.

3. Hypersthene-Alkali-Felspar-Gneiss, Madigan Nunatak (No. 778). The

micro-photograph shows a synneutic aggregate of hypersthene in a
matrix of felspar. Reaction between the hypersthene and the
felspar has liberated ilmenite which has in places formed nuclei for
biotite growths which project into the felspar. Mag. 35 diams.

4. The same rock as in Fig. 3.

A portion of the field of the previous photograph has been considerably
enlarged to show more clearly the relation between the ilmenite,
biotite and felspar. Mag. 170 diams.

PLATE XLV.

Fig. 1. Amphibolite occurring in situ near the Magnetograph Hut at Cape
Denison (No. 421).

Crystals of sphene with nuclear ilmenite are shown together with biotite,
felspar and a little hornblende. Mag. 35 diams.

2. Garnet-Plagioclase-Amphibolite, in situ, Cape Denison (No. 977). The

micro-photograph shows a reaction zone where dactylitic growths
of hornblende are the medium of transference of material from
ilmenite to garnet via felspar. Massive hornblende is also seen in
the field of view. Mag. 125 diams.

3. Hornblende-Plagioclase-Pyroxene-Gneiss, in situ, Cape Denison (No.

972).

The micro-photograph shows a circular region where earlier garnet has
given place to a granulitic mass of pyroxene and felspar. Some
remnants of the original garnet crystal remain. Mag. 44 diams.

Sydney : Thomas Henry Tenuant, Government Printer 1940.

AUSTRALASIAN ANTARCTIC EXPEDITION.

SERIES A. VOL. in. PLATE XLIV.

' - ,M^ l 'T&

, -.:
\\ />\

V*.

Fig. I.

Fig. 2.

Fig. ?.

Fig. 4.

02830 U

AUSTRALASIAN ANTARCTIC EXPEDITION.

SERIES A. VOL.

i. PLATE xiv.

Fig. 2.

Fig. 3.

SERIES A fonliitutd.
VOL. PRJCB.

IV. s. d.

.T8

039

016
10. B

i.LKY 1 6

i sTAHV ! ..040

1- \ND

040

TIC
030

Tl! \D. By L. R. BLAKE and DOUGLAS MAWSOM.

SERIES B.

I. TKI;

imn

r
...J

ii.

... IB

KLAT10 .TIO

an ...

III.

Tl! of

II. .'*,

200

IV.

TH -yC. T. MAI.K;AN 1 10

... } 2
111.

UNIVERSITY OF CALIFORNIA LIBRARY
BERKELEY

Return to desk from which borrowed.
This book is DUE on the last date stamped below.

APR 10 1948

LD 21-100m-9,'47(A5702sl6)476

jaUHfai K

APR 1

M225414

THE UNIVERSITY OF CALIFORNIA UBRARY