JOURNAL OF

THE ROYAL SOCIETY

V;

OF

WESTERN AUSTRALIA

INCORPORATED

VOLUME 40 (1956-57)

4988 5 - 57-415

Journal

of the

Royal Society of Western Australia

Vol. 40

Part 1

Kelvin Medallist, 1955 : H. W. Bennetts, C.B.E., D.V.Sc.

The Kelvin Medal of the Royal Society of
Western Australia Incorporated, was inaugur-
ated in 1924, and is awarded at intervals of four
years or more for distinguished work in science
connected with Western Australia. The medal
for 1955 has been awarded to Dr. H. W. Ben-
netts, Principal of the Animal Health and
Nutrition Laboratory of the Department of
Agriculture of Western Australia.

Harold William Bennetts was born in Mel-
bourne on 18th July, 1898, and educated at
Wesley College and the University of Melbourne,
where he gained the degrees of B.V.Sc. (1919),
M.V.Sc. (1920) and D.V.Sc. (1931). After serving
in the Field Artillery in World War I. he be-
came Microbiologist in the Commonwealth De-
partment of Health and was in North Queens-
land during 1921 and 1922. In 1923 he was
appointed Lecturer and Demonstrator in Patho-
logy and Bacteriology in the Melbourne
Veterinary School and in 1925 became Veterin-
ary Pathologist in the Department of Agricul-
ture in Perth. He has remained an officer of
that department, apart from the period 1928-
1935, when he was seconded to C.SJ.R. and has
been Principal of the Animal Health and Nutri-
tion Laboratory at Nedlands since its establish-
ment in 1947. He has also been a Visiting
Lecturer in the Faculty of Agriculture of the
University of Western Australia.

Dr. Bennetts was a member of the Council
of this Society for most of the period 1929-

1944 and was President during 1934-1935. His
Presidential Address was entitled “Plants
Poisonous to Live Stock in Western Australia.”
This was a field in which he took a great deal
of interest and worked in association with the
Government Botanist (Mr. C. A. Gardner). The
fruits of their joint work are soon to appear in
a richly illustrated work “The Toxic Plants of
Western Australia.”

Dr. Bennetts first made his name in veterin-
ary research by his solution of a sheep disease
(entero-toxaemia), known in the early days as
the Beverley sheep disease and latterly as the
braxy-Iike disease, which had worried sheep-
owners and baffled veterinary investigators for
years. Subsequently he made notable contribu-
tions to the understanding and treatment of
copper deficiency and the oestrogenic effects of
subterranean clover, both major problems to the
stock-owners of Western Australia. He also
made many valuable contributions to other pro-
blems and the results of his researches between
1926 and 1955 are published in over 50 papers
in various scientific journals.

Dr. Bennetts is a Fellow of the Australian and
New Zealand Association for the Advancement
of Science and was President of Section L at
the Adelaide Meeting in 1946. He is also a
Fellow of the Australian Veterinary Association
and an Honorary Member of the Royal Society
of Medicine, London. In 1948 he was honoured
by the award of a C.B.E.

1

Bibliography

1926

A Disease Occurring in Western Australian
Sheep which resembles “Braxy.” Rep. Aust.
Ass. Advanc. Sci.^ 18; 781.

Plants Poisonous to Stock in Western Australia.
Dept. Agric. W. Aust., Bulletin 96 (with
W. M. Came and C. A. Gardner.)

The Helminths of Western Australian Stock.
J. Roy Soc. W. Aust., 13; 49.

1927

Contributions to our Knowledge of Western
Australian Poison Plants. J. Roy. Soc. W.
Aust., 14; 7.

Further Contributions to the Helminths of
Western Australian Stock. J. Roy. Soc. W.
Aust., 14; 58.

1932

Enzootic Ataxia of Lambs in Western Australia.
Aust. Vet, J., 8; 137, 183.

“Pulpy Kidney” in Lambs. Coun. Sci. Industr.
Res. Aust., Pamphlet No. 35 (with D. T.
Oxer).

Infectious Entero-Toxaemia of Sheep in West-
ern Australia. Coun. Sci. Industr.. Res
Aust., Bull. No. 57.

Infectious Entero-Toxaemia, “Pulpy-Kidney”
and other related Diseases of Sheep. Vet. J.,
88; 248.

1933

Enzootic Ataxia of Lambs in Western Australia.
Aust. Vet. J., 9; 95.

The Active Immunisation of Sheep against
Infectious Entero-toxaemia (Braxy-like
Disease) by means of B. ovitoxicus Anacul-
ture. J, Coun. Set Industr. Res. Aust. 6: 92.

Botulism of Sheep in Western Australia and its
Association with Sarcophagia. J. Coun. Sci.
Ind. Res. Aust., 6; 137.

1935

Investigation of Plants Poisonous to Stock in
Western Australia. J. Dept. Agric. W. Aust.,
12; 431.

The Prophylaxis of Enzootic Ataxia of Lambs
in Western Australia. J. Coun. Sci. Industr.
Res. Aust., 8; 61,

The Prevention of so-called “Rickets” in Lambs
and Foals. J. Dept, Agric. W. Aust., 12; 305.

The Occurrence of Bacillus pullorum septi-
caemia in Adult Hens in Western Australia.
Aust. Vet. J., 11; 223.

Plants Poisonous to Stock in Western Australia.

J. Roy. Soc. W. Aust., 21; xi.

The Hookworm Monodontus trigonocephalus
and other Sheep Parasites not recorded
previously from Western Australia. Aust,
Vet. J., 11; 113.

1937

Copper Deficiency of Sheep in Western Australia:
a Preliminary Account of the Aetiology of
Enzootic Ataxia of Lambs and an Anaemia
of Ewes. Aust. Vet. J. 12: 138 (with F. E.
Chapman) .

Copper Deficiency — the Cause of Enzootic Ataxia
(“Rickets”) of Lambs. J. Dept. Agric, W.
Aust., 14: 247.

The Control of Toxic Paralysis (Botulism) in
Sheep and Cattle. J. Dept. Agric. W. Aust.,
14; 381 (with H. T. B. Hall).

Toxic Paralysis (Botulism) of Cattle — Preven-
tion Now Possible. J. Dept. Agric. W. Aust.,
14: 127.

The Use of the Complement Fixation Test in
the Control of Bovine Pleuro-Pneumonia.
Aust. Vet. J., 13; 178.

1938

Botulism of Sheep and Cattle in Western
Australia — Its Cause and Prevention by
Immunisation. Aust. Vet. J., 14; 105 (with
H. T. B. Hall).

1939

“Falling Disease” of Cattle in the South-West of
Western Australia. Aust. Vet. J., 15: 152
(with H. T. B. Hall).

The Investigation of “Palling Disease” of Cattle
in the South-West. J. Dept. Agric. W. Aust.,
16: 156 (with H. T. B. Hall).

1940

Coast Disease in Western Australia. J. Dept.
Agric. W. Aust., 17: 41.

1941

“Falling Disease” of Cattle in the South-West
of Western Australia. Studies of Copper
Deficiency in Cattle. Aust. Vet. J. 17: 85
(with A. B. Beck, R. Harley and S. T.
Evans) .

“Falling Disease” of Cattle in the South West.
J. Dept. Agric. W. Aust., 18: 133.

1942

Enzootic Ataxia and Copper Deficiency of Sheep
in Western Australia. Coun. Sci. Industr.
Res. Aust., Bull. No. 147 (with A. B. Beck).

Studies on copper Deficiency of Cattle: The Fatal
Termination (“Falling Disease”). Aust. Vet.
J., 18: 50 (with R. Harley and S. T. Evans).

Haemoglobinmia of Calves in the South-West
of Western Australia. Aust. Vet. J., 18:
205 (with L. W. Mahaffey and A. F. Flood).

Copper Deficiency of Cattle and the Fatal
Termination “Falling Disease”. J, Dept.
Agric. W. Aust., 19: 96 (with R. Harley and
S. T. Evans).

“Stringy” Wool and Copper Deficiency in Western
Australia. J. Dept. Agric. W. Aust., 19: 7.

2

1943

IJopper Deficiency in Sheep. J. Dept. Agric. W,
Aust., 20: 40.

1944

Two Sheep Problems on Subterranean Clover
Dominant Pastures. J. Dept. Agric. W. Aust.,

Zl: 104.

1946

A Specific Breeding Problem of Sheep on
Subterranean Clover Pastures in Western
Australia. Aust. Vet. J., 22: 2 (with E. J.
Underwood and F. L. Shier).

A Breeding Problem of Sheep in the South West
Division of Western Australia. J. Dept. Agric.
W. Aust., 23: 1 (with E. J. Underwood and
F. L. Shier).

Metaplasia in the Sex Organs of Castrated Male
Sheep maintained on early Subterranean
Clover Pastures. Aust Vet. J. 22: 70.

The Pathological Approach to Problems of
Animal Disease. Rep. Aust. Ass. Adv. Sci.,
26: 231.

1947

A Further Note on Metaplasia in the Sex Organs
of Castrated Male Sheep on Subterranean
Clover Pastures. Aust. Vet. J., 23: 10.

1948

The Pathogenesis of “Falling Disease’'. Studies
on Copper Deficiency in Cattle. Aust. Vet.
J., 24: 2378 (with A. B. Beck and R. Harley).

1949

The oestrogenic Effects of Subterranean Clover
(Trifolium suhterraneum L.). Proc. Brit.
C’wealth Sci. Official Conf., Melbourne: 329
(with E. J. Underwood).

Oestrogenic Effects of Subterranean Clover (T.
subterraneaum L. var. Dwalganup) : Protec-
tive Action of Androgen in the Castrate
Male. Aust. J. Exp. Biol. Med. Sci., 27 :
105 (with June East and E. J. Underwood).

Trace Element Deficiences in Stock in Western
Australia. Proc. Brit. C'wealth Conf., Mel-
bourne: 266 (with E. J. Underwood).

The Oestrogenic Effects of Subterranean Clover.
Kept. 14th Int. Vet. Congr., 3: 165.

1950

Urinary Calculi of Sheep in Western Australia.
J. Dept. Agric. W. Aust,, 27: 129.

1951

The Oestrogenic Effects of Subterranean Clover
(Trifolium subterraneum) : Uterine Main-
tenance in the Ovariectomised Ewe on Clover
Grazing. Aust. J. Exp. Biol. Med. Sci., 29:
249 (with E. J. Underwood).

1952

Oestrogenic Hormones in Plants in Relation to
Animal Physiology. Proc. 6th. Int. Grass-
lands Congr., Philadelphia, 2: 1327 (with D.
H. Curnow).

1955

Copper and Cobalt Deficiency of Livestock in
Western Australia. J. Dept. Agric. W. Aust.,
4: 43.

Cabbage Poison. J. Dept. Agric. W. Aust., 4:
193 (with C. A. Gardner).

3

1. — Laterite and Materials of Similar Appearance in
South-Western Australia

Presidential Address, 1954

By S. E. Terrill, B.Sc., A.R.A.CX, F.G.S.

Delivered — 19th July, 1954

Introduction

The publication in 1952 of two small books,
namely, “Problems of Clay and Laterite
Genesis’* and “Laterite and Lateritic Soils”
emphasises the differences of opinion concern-
ing the nature of laterite. It is difficult, at times,
when reading these two books, to realise that
both are written about what is supposed to be
the same thing, namely. Laterite.

Laterite and lateritic gravels are to be found
advertised in the daily press “For sale” columns
as “gravel, best conglomerate,” and whatever
various scientific >vorkers may say laterite really
is. all are agreed that it is not a conglomerate,
that is, a rock made up of waterworn boulders
or pebbles set in a matrix usually of sand —
mixed with a little clay perhaps — and hardened
to a firm, solid rock.

A term in common use is “ironstone” and for
general comprehensive use for dark coloui'ed
more or less massive, vermicular or concre-
tionary, strongly coherent forms this term is
very suitable for field use whilst “ironstone
gravel” can be well applied to the loose uncon-
solidated material. This is not a strictly correct
use of the term “gravel” perhaps but is suf-
ficiently descriptive of the dark, reddish brown
to black rounded stones in a loamy sand or
sandy loam matrix. There are other forms
which are light in colour and obviously do not
contain much iron for which the term “iron-
stone” is clearly unsuitable. These forms are
generally light brown and have the appearance
of consolidated gravel or of a rock which will
readily yield gravel but it is still more or less
coherent: to these forms, for w'ant of a better
name, the term “gravelstone” is applied by the
speaker for a field use prior to a more detailed
examination in the laboratoi*y-

Such names, while not attractive perhaps, are
nevertheless descriptive and do not suffer from
any implication of origin. Nor do they serve to
cloak — as does the term “laterite” as it is
widely used — manifest differences in mineral
constitution and of structures and textures.

There is still considerable confusion at the
present time in the use of the term “laterite.”
It is necessary that there should be some sort-
ing out of the different types of rock now com-
monly included in the term by many who have
contact with these, and the development of new
names for some of them.

Mineralogical science was to see an immense
proliferation of terms during the 19th Century.
A different name would be given to a mineral
from a new' locality because it differed in some
physical aspect such as colour or form from a
previously described mineral of essentially the
same composition. That process has been re-
versed of late, following upon a better under-
standing of the essential characters of minerals
which has followed upon the X-ray studies of
the past 40 years or so. Mineralogical nomen-
clature is becoming simpler and many names
are gradually falling into disuse.

It is interesting to trace the application of
the term “laterite” from the time it was first
used by Francis Buchanan. M.D., F.R.S. up to
the present dilemma, and to try and see along
W'hat lines further progress can be made. De-
spite assertions from time to time by various
authors that this or that particular interpreta-
tion is generally accepted, there is indeed no
universal acceptance of any pai*ticular usage
even to this day, nor has there been over the
past 50 years or so.

Since there is considerable emphasis placed
on Buchanan*s usage by some, particularly by
soil scientists in Australia and elsew’here, it is
perhaps desirable to examine Buchanan’s
journal of his journeyings in Southern India.
This excursion w'as performed under the orders
of the then Governor -General of India, the
Most Noble the Marquis of Wellesley, and oc-
cupied the closing months of the 18th Century
and the beginning of the 19th Century. The
book, in three volumes, recorded his observa-
tions concerning the manner in which the
people lived: their customs, the economy, and
the nature of the countryside and the rocks
occurring therein. In this journal. Buchanan
recorded the occurrence of a peculiar rock, new
to him, to w'hich he gave the English name of
“laterite” or “brickstone” and for science, the
latinised version “lateritis.**

4

Some years ago the speaker had the good
fortune to add to his personal library a sound
copy of this journal, in its three volumes, and
it is interesting to read the original entries of
the author of the term. More particularly is this
so when one considers the nature of opinions
and statements attributed to Buchanan by
various writers on the subject of laterite termin-
ology.

As an example of the kind of thing that has
happened. Johannes Walther formed the
opinion that a red colour was the significant
criterion for laterite. Following this idea,
Hellmers believed that Buchanan considered the
red colour to be essential and that the rock
was formed as a result of volcanic action.
Neither of these concepts has been found by
the speaker among the nearly 20 entries con-
cerning laterite. not all of which are referred
to in the index.

Buchanan’s first entry concerning this
material was that of 9th December. 1800, when
he was in the vicinity of Kunamkulam south-
east of Calicut. “Cunnung colung curry
Angady.” as he called the place, in a “Nazareny
or Christian” village which he visited. He re-
corded:

“An old church is now unroofed; but

the walls, although built of indurated clay only,
continue very fresh and strong. The altar is
arched over with the same materials ”

The first entry concerning the field occur-
rence was made three days later, when in the
vicinity of Angadipuram, a few miles north of
Kunamkulam.

‘‘After crossing the river. I came to a country
like that near the Nazareny town in the Cochin
Raja’s dominions, and consisting of narrow
vallies surrounded by low bare hills. The soil,
in many places of these hills, is vei*y intractable,
and consists of a kind of indurated clay, which,
on exposure to the air, become as hard as a
brick, and serves indeed all the purposes of
stone.”

For the 20th and 21st December, 1800,
Buchanan made the following entries concerning
the iron ore which was smelted for the manu-
facture of steel.

“In all the hills of the country the ore is
found forming beds, veins, or detached masses,
in the stratum of indurated clay that is to be
afterwards described, and of which the greater
part of the hills of Malabar consists. This ore
is composed of clay, quartz in the form of
sand and of the common black iron sand. This
mixture forms small, angular nodules compacted
together and very friable. It is dug out with
a pickaxe and broken into powder with the same
instrument. It is then washed in a wooden
trough .... placed in the current of a rivulet:

. . . . The powdered ore is placed at the upper
end . . . and ... a man continually stirs it
about with his hand. The metallic sand remains
. . . . the quartz is carried to the lower end and
the clay is suspended in the water and washed
entirely away.”

Thus Buchanan recorded the primitive
character of part of the steel industry of that
time and place. Note that it is the iron ore,
which occurs in the so-called indurated clay
“forming beds, veins or detached masses” that
is smelted for the iron content, not the laterite
itself. In another part of India, Pendleton
recently saw slag heaps which he believed to
have signs of the smelting of laterite for iron.
While this laterite may be what might be
described more propei^ly as lateritic iron ore, it
is possible that it was the iron ore in the laterite
that was smelted.

After describing the furnaces and the smelting
process Buchanan went on to state, and here I
quote in full an oft-quoted entry:

“What I have called indurated clay is not the
mineral so-called by Mr. Kirwan, w'ho has not
described this of w'hich I am now writing. It
seems to be the Argilla lapidea of Wallerius I,
395, and is one of the most valuable matei'ials
for building. It is diffused in immense masses,
without any appearance of stratification and
is placed over the granite that forms the basis
of Malayala. It is full of cavities and pores
and contains a veiy large quantity of iron in
the form of red and yellow ochres. In the mass
while excluded from the air, it is so soft, that
any iron instrument readily cuts it and is dug
up in square masses wdth a pickaxe and im-
mediately cut into the shape wanted w'ith a
trowel, or large knife. It very soon after be-
comes as hard as a brick and resists the air and
water much better than any bricks that I have
seen in India. I have never observed any
animal or vegetable exuvia contained in it, but
I have heard that such have been found
immersed in its substance. As it is usually cut
into the foim of bricks for building, in several
of the native dialects it is called the brickstone
(Itica cullu). Where, however, by the w'ashing
away of the soil, part of it has been exposed to
the air, it has hardened into a rock, its colour
becomes black . . . The most proper English
name would be Laterite, from Lateritis, the
appellation that may be given to it in science.”

It is interesting to find that the reference to
“Mr. Kii’wan” is almost certainly to R. Kirwan,
who published several books on agricultural
chemistry and allied subjects and in particular,
in 1894, a book in which he described, among
things, indurated clay, citing as the typical ex-
ample the clays at Stourbridge, England, now
considered to be kaolinic in character.

The reference to Wallerius I, 395 is, with very
little doubt, a reference to page 395 of Volume
1 of the 1778 Vienna Edition of Wallerius’
“Systema Mineralogicum,” a I’eference work on
systematic mineralogy written in a form of
Latin, then the universal language of learned
men in all parts of Europe. I am indebted to
Miss Ethel Curran for assistance in the transla-
tion of the relevant passages. Wallerius gave
the Swedish, French and German names, all
meaning hardened, lithified or indurated clay.
It may perhaps be significant that the German
name is given as “steinthon” and not “stein-
lehm,” suggesting a hardened china clay or a
clay with little colouring impurity.

5

All the forms described by Wallerius — who,
incidently was Professor Royal of Chemistry,
Metallurgy and Pharmacy in the University of
Uppsala in the middle and late 18th Century —
have the appearance of clay, mostly unstratifled,
are stoney-hard and entirely lack the unctuous
feel of most clay: they cannot be softened
with water: they are softer than steel but be-
come so hard when burnt that they will strike
sparks from steel.

The group in which Wallerius placed the
argilla lapidea is one which also included soap-
stones, serpentines and potstones or altered
talcose greenstones. It would seem that Bucha-
nan considered the laterite to have some of the
properties of the group, but here again, we can-
not recognise all the properties as applying to
any one form of laterite. These rocks are all
of them light coloured, structureless for the
most part, amorphous, soft enough for use as
a substitute for chalk, can be scraped with a
knife, do not throw sparks when struck with
steel. When exposed to the air they become
harder rather than undergo disintegration, but
they do disintegrate in time, becoming more
earthy in appearance like the yellowish or greyish
clay seen In fissures of potstones or impure
talcose greenstones. When dry they absorb
water but are not softened by it. When calcined
they become so hard as to strike sparks wdth
steel and burn to light yellow or grey colours.
Fused with various salts they yield light or ash
coloured strong masses or glasses. Mixed with
clay the powders harden somewhat but with
lime and gypsum they do not fuse unless siliceous
material is added. These rocks do not effervesce
with mineral acids but some does go into solu-
tion, more with hydrochloric acid than with
nitric acid and with this more than sulphuric
acid, as is shown by the amount of precipitate
obtained with alkali carbonates. However, the
amount that goes into solution seems to be
proportional to the depth of colour of the stone.
One of the several types of argilla lapidea is
tawny to dark in colour.

I have given the properties of this rock at
length, partly in order to indicate the scanty
nature of the information there existed then
concerning some rocks. Relatively simple tests
and keen observation had to be relied upon, for
this was in the days before microscopes of any
kind were in common use, and 50 years or so
before the petrological microscope came into
being. It is well-nigh impossible to distinguish
the nature of laterite, so far as its constituent
minerals are concerned, from descriptions such
as this.

Coupled with Buchanan’s description, the re-
ferences indicate that, although it “contains a
very large quantity of iron in the form of red
and yellow ochres" these constituents, hydrous
oxides of iron, cannot be regarded as the prin-
cipal constituents of laterite. It can even be
argued that the essential constituent w'as some
mineral other than ochre, for it would seem that
Buchanan regarded the laterite as a mineral,
that is, a substance which is homogeneous.

After travelling northwards along the western ,
coast of the peninsula as far as Karwar, with |
deviations inland at several places, Buchanan ,
turned away from the ocean and ascended the ,
Ghats to the plains of southern Bombay Presi-
dency and w^estern Mysore State. He then
travelled southwards more or less parallel to
the scarp of the Ghats and then turned in a
north-easterly direction, along the valley of the
Tunga River, through Shimoga. All along this
route he reported the presence of laterite until
he reached a point a little west of Shimoga.

Near Gati, Buchanan recorded:

“ . . . a hill producing iron ore, which is wrought
to some extent. It is found in veins intermixed
with Laterite. like the ore of Angada-puram
(Angrypar) in Malahar, The ore is of the same
nature with what is usually smelted in the
peninsula; that is to say, it is a black sand
ore . .

In all, there are nearly 20 entries concerning
laterite or brickstone. most of them being merely
a record of the occurrence of laterite at the
locality referred to. In one of his last entries
concerning laterite Buchanan merely refers to
it as brickstone- In a later book concerning
Bihar he refers to it merely as brickstone also,
so that it would appear that Buchanan did not
have such a fancy for his brain-child as we have
at the present day.

The first half of the 19th Century saw the
gradual adoption of the term “laterite” by
ti’avellers, mostly geologists, more particularly
in India and the near-by countries. Rocks
having the same general characteristics as
Buchanan’s laterite were found scattered far and
wide through India and Burma, and most of
those who recorded the occurrence of laterite
seemed to have considered it some form of iron-
stone. that is. an impure hydrous iron oxide
rock. This is very understandable, for much of
it had developed over highly ferruginous crystal-
line rocks and consequently this laterite con-
tained more iron oxide minerals than the laterite
formed over granites and similar rocks, as was
that seen by Buchanan. It must be remembered
always, that this was in the days before the
development of the techniques now used for the
study of rocks. Complete chemical analyses, or
merely analyses for ten or a dozen constituents
were not available. The petrological micro-
scope had not been developed, and sure methods
of mineral identification had not yet come to
light, even for the study of comparatively simple
crystalline rocks. Even today we have no sure
method of determining the content of some rocks
with respect of the more indefinite materials
developed by weathering, such as for instance,
amorphous hydrated aluminium and iron oxides,
secondary silica and the like: approximations
are all that can be obtained at best, even
employing a whole battery of techniques such
as chemical analysis, the petrological microscope,
differential or simple thermal analysis. X-ray
diffraction, and so on. In those days reliance
had to be placed on comparatively simple tests
of a discriminatory nature only, as I have just
outlined.

6

Knowledge concerning the nature of the
laterite seen by Buchanan came slowly. South-
ern Malabar was examined by Philip Lake whose
findings were published in 1890. This geologist
could distinguish three distinct types of laterite.
Firstly, the “plateau laterite” which caps the
hill tops as a kind of “summit bed” so to speak,
which Lake considered to have been formed in
situ by the decomposition of the gneiss. A
second type was what Lake termed “terrace
laterite,” to be found on the slopes below the
“plateau laterite.” The third type “valley
laterite” occurred at still lower levels. It is the
“summit bed” that Buchanan appears to have
referred to most.

The latter half of the 19th Century saw^ the
slow development of the present day techniques
of examining and describing rocks. Chemical
analysis was applied more and more, to ascertain
the constituent elements of rocks and their
relative proportions. Following Sorby’s work of
the 1850’s the petrological microscope was de-
veloped and used to ascertain in what way those
elements were combined, that is, the mineral
expression of the chemical composition. Not
only this but the mutual relationships of the
different minerals were also studied. Rock
textures were found to have a definite meaning
in many instances, and a large body of know-
ledge gradually accumulated.

Naturally, sooner or later, someone was bound
to apply these new techniques to the study of
laterite. To Max Bauer, of the University of
Marburg, who had been examining European
bauxites late in the 19th Century, goes the
credit of first publishing work of this nature
dealing wdth tropical laterite. He micro-
scopically examined specimens of laterite and
associated crystalline rocks from Mahe, one of
the Seychelle Islands, in conjunction with
chemical analyses by Busz. The chemical
analyses showed the highly aluminous nature of
the laterite. which Bauer attributed to the
presence in quantity of the hydrated aluminium
oxide gibbsite. to be 'seen in the thin sections.
Further, he showed the close relationship there
existed between the laterite and the underlying
crystalline rocks, not only chemically as to
constituent elements, but also with regard to
micro-structure. Particularly w^as this so in
the instance of the laterite found over the
diorite. for he found that this laterite retained
in itself the structure of the diorite; the laths
of feldspar seen in the diorite were represented
by colourless rectangular areas of gibbsite and
the ferromagnesian mineral by limonite-bearing
areas. The lime and magnesia of the diorite
had disappeared from the laterite. as also the
silica. The laterite over the granite had quartz
grains of similar size and shape to those of the
granite and Bauer could recognise the granitic
stnicture retained in the laterite. Standing
alone, the evidence concerning the laterite over
the granite could be heavily discounted, for the
holocrystalline allotriomorphic equigranular
structure of granite is not particularly distinc-
tive when retained in weathered products, in
contradistinction to the structure of some basic
igneous rocks, especially the ophitic texture of
dolerites and closely related rocks, for these

structures are often readily recognisable even
when the rocks are completely altered by
weathering.

However, the evidence was such that the con-
clusion was inevitable: the laterite of Mahe was
formed in situ by the weathering of the granite
and of the diorite, and, furthermore, those
weathering processes had removed the combined
silica, the lime and the magnesia, leaving the
iron and aluminimum as hydrated oxides, and
the quartz.

About this time, geologists of the Geological
Survey of India had been paying some attention
to the problem of the formation of laterite. The
results of Max Bauer’s work was considered to
have settled the question as to the real nature
of primary lat-erite, that is. laterite which has
been produced directly by weathering processes
acting upon parent materials, before any re-
distribution of the constituents has occured
consequent upon the continued action of
weathering processes involving the movement of
ground water through the laterite mass.

The results of chemical analytical work by
the Warth brothers were published soon after
the turn of the century. Their work showed
that the term laterite had been applied by
geologists in India to a wide range of materials,
so far as chemical— and therefore mineral —
composition was concerned. On the one hand,
highly aluminous varieties existed, rich in the
hydrated aluminium oxide mineral gibbsite and
containing little limonite: on the other hand,
highly ferruginous rocks, almost free of alumina,
were also termed laterite, rocks consisting maiivly
of limonite, principally goethite; and there
wei’e all manner of varieties between these two
extremes.

These last, rich in limonite, to which
Buchanan would have applied the term “brown
hematites,” were considered best designated
“iron ore,” so, towards the close of the first
decade of this century some authors considered
the term “laterite” to be of more value as a
rock name if it were restricted to mean those
more highly aluminous rock types which lie in
composition between iron ore on the one hand
and the aluminium ore bauxite on the other.

Thus, when a chemical analysis of such a
material showed an alumina figure above 52
per cent, and less than 5 per cent, iron and a
very low silica figure it was considered to be
bauxite. On the other hand, if the ferric oxide
figure rose to something over 40 per cent, or so
it was considered to be iron ore. In between
lay laterite.

This view was by no means universally
accepted, as is shown by the letters to the
Editor of the Geological Magazine in 1910.

Even to this day we find those who advocate
strongly the restriction of the term laterite
to materials such as those ju.st described: it Is
the concept widely held by geologists and mining
engineers and is implicit in most of the papers
read at a symposium dealing with the “Problems
of Clay and Laterite Genesis” at the Annual
Meeting of the American Institute of Mining
and Metallurgical Engineers at St. Louis,
Missouri, early in 1951. The one who used the
term in a manner differing from this concept
was a soil worker, G. D. Sherman.

7

Among soil workers there is a division of
opinion. There are those who follow Pendleton’s
concept that laterite is “an illuvial horizon,
largely of iron oxides, with a slag-like cellular
or pistolitic structure, and of such a degree of
hardness that it may be quarried out and used
for building construction.” Apart from the in-
sertion of the concept of origin embodied in
the term “illuviar’ there is little if any difference
between Pendleton’s application of the term and
that of the geologists of the Indian Geological
Survey before the turn of the century and before
anything like a thorough knowledge of these
materials was obtained by the application of
the techniques of rock examination now in use.

The inclusion of a concept of origin in a
definition of a rock is considered, by the present
speaker and many othei’s, to be fundamentally
erroneous: it requirCvS one to secure satisfactory
evidence of origin where often it is not possible
to do so and to decide which of several modes
of origin fit the known data. Further, one can
never be sure that one has all the possible in-
formation which can enable one to make a valid
decision as to origin, one which cannot be ne-
gated by subsequent work. In most instances
the origin of a rock is a philosophical concept
only, and is derived from observable and measur-
able data. It cannot be sufficiently emphasized
that only observable and measurable features
should be taken into account when naming a
rock. It is possible, in some instances, for the
same minerals to be assembled in the same pro-
portions by a variety of methods. This applies
generally and not only to laterite and similar
products of weathering; it applies to granite, for
example. Not infrequently there is some struc-
ture present which suggests a particular origin
for a rock under discussion but this is not always
so, by any means. Furthermore, it is possible
for different workers to consider the same rock
to have different origins, as indeed is the case
with laterite: it would be particularly un-
fortunate if each worker used a different name
for the same rock, just because each considered
it to have a different origin.

Fox, in 1936. sought to clarify the position by
examining some of the laterite to be found in
Malabar and Canara provinces along the south
western coast of the Indian peninsula. Un-
fortunately, it would appear that a thorough
examination of all the different types of laterite
occurrences was not undertaken, especially of
that occurring on the summits of the hills, to
which Buchanan referred so often and which
Lake called “plateau laterite” some ninety years
later. Whereas the exposure at Tellicherry
examined by Fox can be interpreted simply as
a mass of very impure iron oxide, the other
occurrence, of which analyses are given of the
profile exposed in the quarries at Cheruvannur,
can be regarded as a profile in which there was
a certain amount of enrichment by hydrated
oxides of both iron and aluminium in the upper
portions. From the illustration, this occurrence
is down the slope a little distance below the
hill-top and may correspond to Lake’s “tei’race
laterite.”

Those portions of western Mysore, where Bu-
chanan reported laterite as being abundant, ap-
pear to have been entirely overlooked in dis-
cussions as to the true nature of laterite.

Amongst the laterite near Shimoga, at Kem-
niangundi, not far from the route followed by
Buchanan, there is some laterite reported by
A. M. Sen to have composition as set out in
Table I.

Table I

Laterite from Kemmangundi

A 1203

62.50

42.90

41.72

Fe20;}

5.10

25.40

30.27

Si02

0.36

10.40

2.97

H 2 O

31.90

20.70

22.64

CaO

0.56

MgO

0.03

Ti02

tr.

Parent rock: — Diorite or Hornblende diabase

For the best part of a century the term
laterite was applied to a wide range of rocks
ranging in composition from an impure iron ore
containing a large amount of limonite, to a
somewhat ferruginous bauxite. The question
arises whether the term is best applied in the
restricted sense that Pendleton so vigorously
advocates, namely an “illuvial horizon” in which
the cementing material is lai’gely iron oxide,
for which the term ferricrete, proposed by
Lamplugh in 1897 is very suitable, or whether
it should be applied in the sense that it should
lie in composition between iron ore, on the one
hand and bauxite, the ore of aluminium, on the
other.

The difficulties of ascertaining always whether
an occurrence is indeed illuvial, or what its origin
i-eally is^ are manifest. It is not sufficient to
assume that it is always of the same origin:
indeed, there is much to suggest the contrary.

To the speaker, the restricted sense, developed
by geologists early this century, appeals as being
the more useful. It is a concept which depends
solely upon characteristics of minei*al composi-
tion and of field occurrence as broad sheets,
characteristics which can be determined ap-
proximately by carefully planned chemical
analysis and by field observations of the extent
of an occurrence. It is considered that no con-
cept of origin should be involved in the name
given to any i*ock, be it granite, basalt, sand-
stone or even laterite. This is not to say that
the origin may not be indicated by a suitable
adjectival qualifier preceding the name.

The term latosol, recently introduced by C. E.
Kellogg, offers a solution to one of the problems
of the nomenclature of these materials. While
he cannot find himself at one with Kellogg in
the use of the term laterite to include sesquio-
xide-rich materials in which hydrated aluminium
oxides are low or virtually absent, the present
speaker considers the introduction of such closely
defined terms to be long overdue in pedological
science and has little doubt but that further
similarly closely defined terms will be introduced
in the course of time, to cover adequately the
different varieties of these very interesting
materials.

8

For the present purposes, the speaker will
use the term laterite to include those consoli-
dated, non-friable materials consisting essen-
tially of hydrated oxides of both iron and
aluminium, lying in composition between iron-
stone or iron ore on the one hand and the
aluminium ore bauxite on the other. Where
quartz, or any other mineral, becomes at all
prominent in the constitution of the rock, say
over ten per cent, the term is suitably modified
adjectivally.

One must apply very carefully the various
definitions depending upon ratios of alumina to
silica, either of the rock as a whole or of the
“clay fraction” only — this being selected by some
because it is considered that this fraction would
contain little or no free silica, such as quartz,
but would be truly representative of the argil-
laceous material and therefore of the material
produced by chemical alteration during weather-
ing. It must always be borne in mind that
these are merely short-cuts used to minimise
the w^ork involved in securing knowledge, suffi-
cient for the work in hand, of the chemical
constitution and therefore of the mineral con-
stitution of the rocks or soils under investiga-
tion.

South-Western Australia

Let us now turn to this south-western corner
of the Australian continent.

The features to be observed within 200 miles
of Perth conform to the general pattern of the
whole of the area and will be referred to prin-
cipally.

First it is desirable briefly to outline the
physiography and geology of the area.

Physiographically one of the most outstand-
ing features is the Darling Scarp, w'hich forms
the w^estern edge of the Great Western Plateau.
Here, along a line sub-parallel to the coast, and
about 20 miles or so inland, the land surface
drops from altitudes of 800 to 900 feet above
sea level to the coastal plains standing less than
100ft. above sea level for the most part, rising
to higher altitudes only in the belt of coastal
limestone hills, which rarely rise to an altitude
greater than 200ft.

The Great Western Plateau can be regarded
as having several elements.

One, the Darling Peneplain, is a general level
of flat-topped hills and ridges or divides be-
tween fairly broad valley systems, the tops of
the ridges standing at some 900 to 1.000ft. above
sea-level.

Rising above this peneplain are a number of
higher, comparatively isolated hills such as
Needling Hills and Mount Bakewell, remnants
of an older land surface, the roots of earlier
divides betw'een broad valley systems which
collectively form the Darling Peneplain.

Cutting down into this Old Plateau or Darling
Peneplain is a system of valleys with gentle
slopes and broad salt river flats, standing at an
altitude of some 600 ft. or so above the present
sea-level. These broad valleys with their salt
fiats are remnants of an old drainage system

which appears to have flowed in a general direc-
tion from North to South. The various branches
of the Mortlock River and the extensive salt
river flats to be seen between Kellerberrin and
Merredin afford good examples of this system.
In this region, the valley systems are very broad
and only small remnants of the old Darling
Peneplain remain. The low^er reaches of these
river systems are characterised by the much
narrower, steep-sided valleys of young rivers.

These various elements are not seen clearly
near Perth but are best seen if one goes up onto
the higher places east of the Avon River or
southwards beyond the Beaufort River.

The rocks of the western portion of the Great
Western Plateau consist mainly of granitic
gneisses and granites, granulites, metasedi-
mentary schists- — some of which are sillimanite-
and andalusite-bearing-quartzites and slates.
Cutting these are dykes of dolerite and epidio-
rite. To the east of the North Branch of the
Mortlock River and of the Avon River at York
and along the Albany Highway from 120 miles
or so from Perth onwards and extending east-
wards, the wheat belt has occurrences of moder-
ately well lithified conglomerate, argillaceous
grits and sandstones — some of which show
horizontal bedding — and very sandy clays.
These sediments lie upon the undulating eroded
surface of the crystalline rocks, and are com-
paratively young, judging by their lithification,
which is comparable with rocks of the
Plantagenet beds and are believed to be of
terrestrial origin. Gritty sediments underly
much of the high level sand -plain country of
the north-eastern and eastern wheat belt, the
sandy soils of the elevated sand-plains being
residual soils derived from these gritty sedi-
ments.

To the west of the Darling Scarp the com-
paratively low-lying coastal plains consist of a
narrow belt of sandy limestone and sand hills,
in part of aeolian origin, with a broad belt of
lenticular beds of very sandy clays and argil-
laceous sands. These yield loamy sands and
sands at the surface. In places there are patches
of fresh-water limestone. These sediments rest
upon a thick series of calcareous shales with
inter-bedded coarse grits and shales below,
these grits being the aquifers of the artesian
basin around Perth. Laterite does not occur
west of the vicinity of the Darling Scarp.

The Darling Peneplain is characterised by the
presence of a group of ferruginous and
aluminous rocks and gravels which fo) m a sheet
or cuirass upon the old rocks beneath. As ex-
posed at the surface these materials are of two
kinds. The fairly hard stone is popularly re-
ferred to as ironstone or conglomerate while the
gravel consists of an unconsolidated mixture of
ferruginous nodules with some quartz sand and
clay. The more firmly lithified materials are
commonly medium- to dark-brown, sometimes
quite light-brown. In most places they consist
of nodules set in a matrix of fine grained
material, the so-called “concretionary” laterite,
much of which is better referred to as laterite
with concretions. In places the laterite consists
almost entirely of more or less spherical pellets

9

around about 5 to 10 mm. across, firmly joined
at their points of contact: this form appears to
be confined to channels through the less porous
rock, these channels having developed along
joints inclined at any angle. They constitute
zones of free passage to lower levels for rain-
water falling on the surface above or higher up
the slope. This is a purely concretionary laterite
commonly styled “pisolitic laterite” on account
of the size and shape of its constituent nodules
resembling those of peas. Another less common
form is the so-called vesicular type which con-
sists of massive material through which irregu-
larly shaped anastomosing channels pass: this
type not uncommonly has a splotchy appear-
ance in light and dark browns, though fair uni-
formity of a dark brown is quite common.
Rarely, one finds a massive form of laterite, free
from nodules except near the surface: in places
it may be quite porous, but on the whole is
fairly massive and dense. Whatever the type of
laterite, some form of nodular structure is gen-
erally present somewhere in any exposure of
laterite and this nodular structure has come to
be regarded, by some, as the distinguishing
characteristic of laterite, so much so, that when
they see a light or dark brown or reddish brown
nodular material occurring as an extensive
layer they immediately identify it as laterite.
It is intended to return to a consideration of
this nodular structure later and to show that
the vesicular, nodular and pisolitic types are
derived from the dense, massive variety by the
continued action of waters passing through the
laterite to lower levels. Before doing so, how-
ever, it is first desirable to give some attention
to certain features of the laterite and of the
materials closely associated with it.

Firstly, concerning the composition. A num-
ber of analyses of laterite occurring in and near
the Darling Range have been published and a
few of these will suffice to indicate the general
nature of this rock (see Table II).

Table II

Laterites from the Darling Range

Wongan

Gooseberry Parkerville

Ridge

Hills

Hill

Hill

SiOa

5-96

6-41

15-83

10-30

A 1203

44-66

36-74

31-68

22-53

Fe^Os ...

19-08

39 • 80

24-95

48-56

FeO

2-52

MnO

0-06

0-05

MgO

tr.

015

0-62

CaO

tr.

0-10

0-01

NaoO

0-14

K ,6

iiU

H2O— ....

0-58

1-20

2-69

\l5-82

1120 + ....

20 • 44

13-73

19-55*

TiOa

3-10

1-98

2-07

3-24

SO 3

0-18

P2O5

tr.

tr.

0-09

CuO, ....

0-03

V2O5

0-23

Combined

100-00

100-43

100-15

100-50

SiOj ...

1-97

10-17

4-94

Analyst ....

E.S.S.

E.S.S.

S.E.T.

D.B.

K.T.P.

Date

1901

*

1912

B> ignition loss

1947

1946

These analyses show the laterite to consist
mainly of hydrated oxides of iron and of
aluminium, together with some hydrous silicate
of aluminium, some free silica and a little oxide

of titanium. The most probable minex’al ex-
pression of this chemical composition is a rock
consisting mainly of gibbsite and limonite with
some kaolin or halloysite and quartz and a little
doelterite or leucoxene or possibly traces of
residual ilmenite.

Much of the limonite is probably goethite, but
the presence of some lepidocrocite and maghe-
mite is suggested by the magnetic properties
of nodules derived from laterite. Some nodules
collected at Kalamunda. for example were quite
strongly magnetic as collected but rapidly lost
their magnetic response when heated to 800 “C
in a neutral atmosphere. This suggests the pre-
sence in the nodules of the magnetic anhydrous
iron oxide mineral maghemite, which in turn is
derived from lepidocrocite by dehydration. A
number of years ago. Professor R. T. Prider
drew attention to the polar magnetic properties
of laterite at Wattle Flat.

It will be noted that the soda and potash,
the lime and magnesia, to be found in abund-
ance in the underlying rocks, are almost com-
pletely absent from the laterite. The Darling
Range laterite manifestly conforms to the con-
cept of laterite in the restricted sense of a
material somewhere between a bauxite on the
one hand and iron ore on the other; most of
the occurrences, so far examined chemically, are
somewhat impure from the admixture of quartz
and clay.

There are certain features of the field
occurrence to which I would direct your atten-
tion.

Firstly, beneath the laterite there is com-
monly a zone of partially or completely bleached
clay. Simpson, in 1912, I’eferred to the
laterite as an efflorescence which drew ferric
oxide and alumina from the rocks below,
resulting in a layer from which most and often
practically the whole of the ferric oxide has
disappeared, leaving this stratum of white or
pale-coloured clay.

In 1915 Walther visited this State and de-
scribed the occurrence of laterite here. Figure 1
is a diagrammatic profile after that author. It
shows the massive cuirass or crust of laterite
overlying a mottled horizon which in turn over-
lies a bleached horizon which in its turn passes
into the rock beneath.

This sequence has come to be regarded as the
normal sequence of horizons of laterite-bearing
profiles when truncated. It is postulated that
above the hard laterite crust there existed in-
coherent. leached, sandy soils which have been
eroded away from those areas in which laterite
or lateritic gravels are exposed.

In evidence of this, pale yellowish grey loamy
sandy soils exist which have a brown, obviously
ferruginous horizon some eighteen inches to
three feet or so beneath. In places, local
erosion has removed the overlying loamy sands
and exposed the gravelly horizon and this
gravelly horizon has much the same appearance
as the gravel associated with the solid rock
laterite elsewhere. These gravelly horizons are
regarded by many as the same as the more
massive occurrences of laterite. It is proposed,

10

to show that they differ materially and so
cannot be regarded as the same as the
aluminium oxide-bearing laterite of the Darling
Range, for instance.

Walther’s diagram may be taken to represent
closely the features to be observed in an occur-
rence of laterite at Parkerville. At that place,
where some Quarrying has been done in the past
to secure material for roads and where a Roads
Board Hall and tennis courts are now situated,
there is exposed a section of massive laterite
developed in situ from a quartz dolerite dyke.

minerals of the parent rock are now a mixture
of dark brown limonite with a small proportion
of ironstained, presumably colourless mineral,
possibly a mixture of gibbsite and clay minerals.
A little quartz is scattered through the section,
Wlien the chemical composition of the
laterite close to the boulder is compared with
that of the quartz dolerite from which it was
formed, it may be observed that the ratio of
ferric oxide to alumina of the laterite is what
it should be if there has been no movement
of the iron with respect to the alumina. If most

earlier slopes ?

This section has much to offer in enlightment
as to the nature of laterite in the Darling Range
and the sequence of events in connection with it.

First, there is a layer of laterite, much thicker
in proportion than the layer figured in Walther’s
diagram: the topmost 18 inches or so is gravelly,
but below this the laterite is massive. Im-
mediately below the laterite there is a mottled
red and pale greenish blue clay and lower still
a pure white koalinic clay.

The laterite has the typical composition of
rocks of this group, using the term in the re-
tricted sense of a rock lying in composition
between an iron ore on the one hand and an
aluminium ore on the other. The analysis of
the rock has been published in the Society’s
Journal recently.

In the laterite there is evidence of the typical
spheroidal weathering of basic crystalline rocks
of igneous origin. Also there is immersed in
its substance a small boulder of quartz dolerite
and near its ba.se, a very large boulder which
projects down into the mottled clay horizon and
around which is a downward extension of the
laterite from the layer above. These boulders
have the appearance of being cores of
spheroidally weathered blocks of dolerite, the
original joint faces of which are still discernable.

Not only are the weathering cracks of the
parent rock preserved but in parts the micro-
structure is preserved also. In thin section the
original feldspar laths are now represented by
elongated rectangular patches of gibbsite, while
the areas occupied by the ferromagnesian

of the ferrous iron of the dolerite — all but the
amount shown to be present in the laterite — is
considered as having been converted to ferric
oxide, then the ratio of ferric oxide to alumina
remains virtually undisturbed.

The combination of these three features of
preservation, firstly, the preservation of the
structure produced by spheroidal w'eathering,
secondly the preservation of evidence of poikili-
tic structure and thirdly the retention, undis-
turbed of the iron to alumina ratio, is considered
to be sufficient evidence that this laterite has
been formed in situ, not by deposition from
solutions which have brought in the constitu-
ents from some place else, be it from some hypo-
thetical horizon directly above, or from higher
up the slopes of the hill, but that the laterite
has been formed by the removal, in solution
in circulating ground waters, of practically the
whole of the soda and potash, the lime and
magnesia and a large proportion of the com-
bined silica.

The nodular structure in the top portion of
the laterite and the gravelly .soil above it, are
considered to have developed by the continuing
action of weathering, probably under different
climatic conditions.

There is no reason to believe that the laterite
formed from the granite originated in any way
differing from the manner in which the laterite
was formed from the quartz dolerite. for the
junction of the laterite formed from the quartz
dolerite with that of the laterite alongside it.
formed from the granite, can be traced.

11

Nor is the laterite at Parkerville the only
occurrence where the structure of the parent
rock is preserved.

In an old cutting north of the present raihvay
line at Mt. Helena there is a vertical section
showing laterite derived from a basic dyke
rock preserving perfectly the structure of the
original rock, and in which the spheroidal
weathering cracks of the original basic dyke can
be detected. Further, one can run a knife down
the junction between the laterite from the basic
dyke rock and that formed from the granitic
country rock

Another occurrence of laterite which pre-
serves the structure of the parent basic dyke
rock has been found on the Beaufort Downs
property: this farm is on the road from the
Martup Hills on the Albany Highway to Wood-
anilling further east. This occurrence is neai'ly
150 miles away, to the south of those first men-
tioned.

Emphasis on basic dykes as being rocks
which yield laterite in which the structure is
well preserved is unfortunate but inescapable.
The basic dyke rock possesses a structure which,
when preserved in its weathering products, is
very readily recognisable and clearly demon-
strable to others, while granitic rocks do not
possess such a clearly defined structure. It is
true that one can consider certain features to
be seen in thin sections of some massive laterite
specimens to be relicts of the structure of the
parent granite, but the evidence is not so clear
and is open to doubt.

Once the validity is admitted of the con-
clusion that the laterite is primarily formed in
place from the basic dyke rock by. the ground
waters removing the alkalis and alkali earths
and much of the combined silica, and not by
the deposition of the laterite constituents from
solutions brought to the place from elsewhere,
certain other deductions must necessarily fol-
low. The mottled clay and the pure white,
bleached kaolin strata below are quite devoid of
any evidence of the structure of the parent
rock and further, the regeneration of such a
structure from them does not seem possible.
Consequently, they cannot be considered to be
intermediate stages in the formation of laterite
from the parent basic dyke rock.

It is concluded therefore, that the clays were
formed neither before the formation of the
laterite. nor at the same time, but subsequently
thereto, as weathering of the parent rock con-
tinued and still continues beneath the protec-
tive mantle of laterite.

There is some reason for the belief also that
underneath the laterite mantle the conditions
are even yet those which led to the formation
of a bleached clay, for, at the present day, the
granite appears to be still weathering to a
bleached, quartzose kaolin underneath the
laterite crust. Only where they are exposed
directly to the weathering agents or beneath a
thin pervious soil layer do the granites or green-
stones show evidence of weathering to red or
brown iron-bearing clays or loamy soils.

The laterite was formed at some early stage
in the development of the present land surface.

Firstly, an undulating surface, in places hilly
along divides between broad valley systems, was
produced by the erosion of granites, gneisses,
metasediments and other crystalline rocks.

More or less argillaceous grits, coarse and fine
sandstones — in places showing horizontal bed-
ding — with local conglomerate, were laid down
upon the eroded surface, in great measure filling
the valleys. The thickness of these sediments
is not known but appears to have been con-
siderable for they reach high up the sides of the
highest hills where they occur.

Following the deposition of the gritty sedi-
ments erosion developed very broad valley
systems virtually plains, with isolated hills
standing higher, the remnants of the divides
between the valleys which constituted the broad
plain country. This plain country is the Darling
Peneplain.

Following the development of the Darling
Peneplain lateritisation occurred during a de-
finite climatic phase. A change in climate fol-
lowed with the consequent change of final
product of weathering of the feldspars and
ferromagnesian minerals. Instead of the
aluminium -bearing minerals having the whole of
their combined silica removed, leaving the
hydrated aluminium oxide gibbsite, they had
only some two thirds removed, leaving hydrous
aluminium silicates, chiefly kaolin.

Local conditions have caused variations in
the nature and intensity of colouring of the
horizons beneath the laterite, because of vary-
ing proportions and state of oxidation of the
iron left along with the kaolin.

After the period of laterite development which
could well have occurred on a low-lying plain as
Woolnough has postulated, elevation has caused
a rejuvenation of erosion in the then existing
drainage pattern. This gave rise to broad
valleys in which extensive, very sandy, mostly
unst.ratified terrestrial sediments were deposited.
These sandy sediments blocked the drainage
system and have given rise to the widespread
elevated sand-plains of our wheat belt and South
West.

Following this terrestrial sedimentation, re-
juvenation of erosion has occurred in at least
two stages. At first the streams cut down to
a level about 200-300 feet below the laterite-
covered plateau. A major uplift of at least
600 feet then followed and the rejuvenation of
erosion has caused the lower reaches of the
livers to cut downw^ards. while at the same time,
the upper reaches continued to extend and are
still extending their valleys laterally, so that in
many places there is only a line of laterite-
covered, flat-topped hills to mark the divides
between broad valley systems the lower parts
of which consist of fiat, marshy or salt river
flat country standing about 600 ft. above sea
level. Woolnough’s 600 feet level. Indeed, in
many places the whole of the laterite and under-
lying clays have now been removed exposing
bare granite and gneissic hills, the so-called
‘'Rocks,” which are the very roots of the divides
between these post-peneplanation mature valleys
or of the monadnocks which stood above the
peneplain.

12

Certain features of the landscape and
associated soils call for attention at this stage.

First, attention is drawn to figure 2a which
shows diagrammatically a section of a side of
a valley in granitic country. On the left there
is the flat-topped ridge or divide between ad-
jacent valleys. It is laterite covered, the laterite
being a residual eluvial horizon which has been
further altered as to structure and composition
by the continuance of weathering action. The
soil cover is thin or absent: in places loose
boulders of laterite occur. Such soil as there
may be is full of ironstone nodules and is mainly
grey sand with a little clay and organic matter.
As one progresses down the slope, light brownish
or greyish sandy loams derived successively from
the mottled or coloured clay horizon and the
pallid or bleached clay horizon appear. The
valley may not penetrate the pallid, often

laterite and the formation of gravelly soils from
the post-laterite yellow sandstones where the
gravel and gravel-stone, at first sight, resembles
nodular or concretionaiy laterite.

Figure 3 shows the low wall of a small gravel
pit in sandstone country that occurs some 120
miles or so from Perth along the Albany High-
way. It shows a yellow argillaceous sandstone
which has developed vertical cracks down which
iron-bearing solutions have penetrated. These
solutions have impregnated the walls of the
cracks with reddish hydrous iron oxide. Near
the top of the sandstone, a little below the
sandy gravel, the columns of sandstone have
cracked horizontally with consequent penetra-
tion of iron-bearing solutions which have fol-
lowed the easier courses along the cracks. The
isolated fragments thus become impregnated by
hydrous iron oxide from all sides and the outer

thin grey sand, sandy loom,

1

brown loam with I

light brown.

brown loam

with limoniltc concretions,

1

dork purplish lumps |

yellow or

laterite boulders, bare

1

or small round

grey loam.

1

laterite masses.

1

1

limonitic concretions.

1

1

grey, reddtsh or light
brown sandy loam or
loamy sand over aiiovium.

laterite.

mottled-nodule- or concretion-
bearing light to medium brown clay.

light brown, yellow and white clays.

Fig. 2(a)

unaltered granite, granitic gneiss, epidiorite,

• dolertte, meta-sediments, consolidated argillaceous sediments, etc..

800 to 1000 ft or more
above sea level

100 to 600 ft . ^ or more.

as in fig, 2(a) I high level sandplains of grey aid yellow sands and loamy sands with
above, | poorly to well developed illuvial hori 2 on(s) of yellow to brown

I limonite-cemented argillaceous sandstone or concretions resembling
laterite overlying more or less lithified argilloceous sands and
I sandy clays, and, rarely, conglomerate at base.

I

brown loams. Igrey reddish or light

1 brown sandy loams or
I loamy sands ovei
a Iluvium.

granite, granitic gneiss, etc., weathered at contact with

Fig. 2(b).

bleached horizon, but, should it do so, red and
browm sandy loams have developed, formed by
the W'eathering of the granitic rocks exposed to
the climatic conditions and physiographic
circumstances of the veiw recent past and
present.

Where the argillaceous grits and sandstones
and very sandy clays form sandplains, the
diagram becomes slightly different, as shown in
figure 2b.

On the left is a slope such as shown in figure
2a. In the middle is a very sandy flat-topped
ridge of gritty sediments which have shared in
the mottling and bleaching which followed
lateritisation : the residual soils derived from the
grits are the very pale yellowish grey slightly
loamy sands, which constitute our high level
sand plains. On the right the valley slopes down
to the zone where residual reddish brown sandy
loams occur, similar to those of the valley
bottom of figure 2a.

Before closing, it is desirable that some
reference be made to the further weathering of

portions are constantly being removed. Such
fragments also crack across and become several
fragments, all being separated by the rootlets
and by loamy sand washed in between them by
moving water on its way down to the zone of
permanent saturation. Thus are formed “dis-
integration” type residual nodules.

In places the sandstone does not wholly dis-
integrate into separate nodules, but on the con-
trary, irregular more or less vertical “solution
channels” develop, leaving an almost nodular
mass between with channels filled with loose,
grey sand. This “nodular” mass is easily broken
to loose sandy gravel.

A similar process of disintegration occurs in
laterite, yielding loose ironstone nodules in a
very sandy matrix.

Other forms of nodular or concretionary struc-
ture have been developed in both the laterite
and in the argillaceous grits and sandstone.

In places a spotty type of mottling develops
a little below the surface, where isolated spots
become enriched with limonite to such an extent

13

that eventually they may consist largely of
limonite. Such pellets yield a “residual” type
of nodule which may be anything from a very
dense almost black and pure limonite to a lump
of dark purplish red to brownish black, friable
material with a thin, dense, comparatively hard
smooth surface, very similar to the “disintegra-
tion” type of nodule and apparently indis-
tinguishable from them. In sandy loam soils

FIG. 3.

derived from sandstones such nodules may be
very numerous forming a principal constituent
of a lower horizon in the soil profile. These
nodules have a similar oi'igin to that of similar
nodules formed in laterite, being deposited in
the matrix from solutions and often are very

similar in appearance. This has led many to
consider an horizon rich in such nodules a laterite
horizon: it is considered, however, that such
nodules should be styled “laterite” only if they
consist largely of hydrated aluminium oxide
minerals rather than the hydrated silicates as
well as hydrated iron oxide minerals. The few
such occurrences examined by the present
speaker appear to have an origin differing en-
tirely from that of the laterite itself, being the
same as the origin of nodules formed in the
laterite matrix by the continuance of w^eathering
action upon the primary laterite, which is itself
a residual skeleton derived from the parent rock
by removal of the alkalis, alkali earths and
combined silica in solution. In the one instance,
the ferruginous nodules have developed in argil-
laceous sandstone; in the other, they have de-
veloped in laterite subsequent to the formation
of the laterite: this nodular structure is not a
distinguishing characteristic of laterite, although
most laterite has developed this structure.

The development of such nodular structures
can occur during weathering in any sufficiently
porous rock or material, laterite included,
provided sparingly soluble matter be present
and that climatic and topographic conditions
are suitable. The development of nodular struc-
tures occurs in a wide variety of materials and
is merely indicative of the similarity in response
of the various rocks to the same forces of
weathering as expressed in similar climates
acting in similar topographic situations with
like drainage patterns. Different types of nodules
are formed depending upon whether a disinte-
grative or accretionary action is involved and
this, in turn, depends upon the whole of the
environment in all its complexity.

Before closing, I wish to express my thanks to
the Director of the Government Chemical Labo-
ratories. Mr. H. P. Row’ledge. for his kind per-
mission to use the facilities of the Mineral Divi-
sion outside office hours in order to carry out
certain phases of this investigation, which is
still in progress. My thanks are also due to
Miss Ethel Curran, late of Perth Modern School,
for the basic translation of the Latin of Wal-
lerius, and to Mr. and Mrs. Mira Liber for their
valued help in the translation from the German
of various articles such as those of Johannes
Walther and Max Bauer. Lastly, I thank you
for your kind attention to-night.

14

2. — Studies in the Water Relations of Plants

I. — Transpiration of Western Australian (Swan Plain) Sclerophylls

By B. J. Grieve*

Manuscript accepted — 15th November, 1955

The water economy of character plants of the
hard-leaved evergreen vegetation on the Perth
coastal plain has been studied to obtain infor-
mation on their behaviour both before and
during the long dry summer. With the excep-
tion of Eucalyptus marginata, all the sclero-
phylls so far tested (e.g., Batiksia menziesii and
B. atteuuata, Stirlingia latifolia, Hibbertia
hypericoides, Bossiaea eriocarpa, Hardenbcrgia
conipt07iiana, Kennedya prostraia. Eucalyptus
calophylla, Xantfiorrhoea preissU, Petrophila
linearis) showed in greater or less degree de-
creasing rates of transpiration with increasing
dr>' conditions. In the spring, transpiration was
high, curves being of the one-peak type in
Bossiaea, Kennedya, Batiksia aitenuata. Harden-
bergia and In summer, curves were

commonly of the two-peak type, while in late
summer values for some plants remained very
low throughout the day after an early morning-
peak. Average rates of water loss seldom ex-
ceeded 5-6 rag./g./mm. during the summer. The
relatively shallow rooting liibbertia and Bossiaea
in particular showed very low values and passed
into a state of near domancy in late summer.

The moisture content of soil at this time is low.
while the soil suction force rises above the
osmotic values of the leaves. The plants remain
in a condition of severe water stress until the
break of season rains. The tree sclerophylls.
Banksia spp. and Eucalyptus calophylla, and the
shrubs, Stirlingia, Hardenbergia and Kennedya,
with both a shallow and a deep root system,
reduced their transpiration rate in summer but
were not under conditions of marked water
stress. Stomatal movements in some plants
(e.g, Hibbertia. Bossiaea) showed reasonable
correlation with rate of water loss; in others
fe.g. Stirluigia) stomata remained open at the
University station while transpiration rate was
falling. Under the more deslccallug conditions
at the Cannington station they remained closed
during the day. A higher rate of transpiration
was found in older leaves (as against those of
the current season flush of growth) In such
plants as Banksia and Stirlingia. Slower photo
and hydro-reactions were observed In stomata
of such older leaves. Cutlcular transpiration
was found to proceed at a low level In the more
highly cutlnized sclerophylls. The osmotic
values of leaves rose with advancing summer,
while a rapid return to lower values occurred
with break of season raln.s. Experiments to
determine relative xerophytlsm have .so far
yielded Inconclusive results owing to difficulties
with water viptake by the cut-off leaves. Col-
lateral studies on non-sclerophyllous shrubs
which grow on the Perth coastal plain, indicate
that a considerable degree of physiological diver-
sity exists. PItyllanthus. for example, with its
soft thin leaves shows a high rate of water loss
in spring and earlv summer. With increasing
dryness it maintains its water balance by shed-
ding its leaves. A mesomorph, Ereehthites
hispidula maintained a high rate of transpira-
tion in soring and In .summer up to the time
it died off.

Department of Botany, University of Western Australia

Introduction

Relatively little information is available on
the water economy of Australian sclerophylls
under field conditions, the work of Wood (1923,
1924, 1934) in South Australia, standing alone
in this regard.

In an effort to extend our knowledge of this
aspect of the physiology of the sclerophyll
plants of Western Australia, transpiration and
associated studies were planned for stations
passing progressively inland from Perth on the
western coast towards the Eremaea. From these
experiments it is hoped to determine the degree
of physiological diversity existing among the
sclerophylls and to ascertain the nature of
possible ecological adaptations. In the present
paper the results obtained for the first stations
on the Swan Coastal Plain are presented.

The Research Area — Its Vegetation and Soils

The affinities of the sclerophyllous trees and
shrubs of the Swan Coastal Plain were indicated
by Diels (1906) who referred to them as “thick
shrub growths which can be compared wAh the
maquis of the Mediterranean or better still
with the stiff-leaved scrub of the Cape.” The
plants are predominantly hard-leav;d and
evergreen, herbaceous plants being poorly rep-
resented. Two stations were selecteo for the
study of character plants — one in the vicinity
of the University, and the other at Cannington
a few miles south-east of Perth — so that these
obsei'vatlons are repi’esentative of the vegeta-
tion of the metropolitan sector of the Swan
Plain.

The tree community near the University is
of mixed Jarrah (Eucalyptus viargiuata) ,MaiTi
(Eucalyptus calophylla Banksia and Casua-
rina. The associated shrub layer consists
mainly of sclerophyllous plants varying from
tall shrubs (itlO feet in height) down to shrubs
<rb 2-3 feet in height). Herbs, varying in
height from two to three feet down to a few
inches, occur in the shrub layer.

15

The soil at the University station may be
described as greyish-yellow to yellow sand
(Karrakatta Sand). It is neutral or very
slightly acid in reaction and the surface soil
is darkened with organic matter. The soil pro-
file in the main area of study is as follows: —
Sparse litter

Greyish black — 0in.-8in. Coarse sand con-
taining organic matter
Greyish yellow — 8in.-18in. Coarse sand
Light brownish-yellow to yellowish-white,
changing to yellow with depth — 18in.-
84in. Coarse sand containing a mode-
rate amount of feiTic oxide.

The climax community on the sandy ridges
at Cannington is Banksia low scrub forest with
associated shrub and herb layers. A typical
soil profile (Speck, 1952) in the Muchea sands
of the area is as follows: —

Ao Sparse litter

Ai Oin,- 3in. — Grey sand with little organic
matter

As 3in.- 8in. — Light grey sand — becoming
leached

8in.-60in. — Highly leached white sand
Bi 60in.-70in. — Definitely darkened layer of
brown sand suggesting a
slight tendency towards for-
mation of Coffee rock
70in.-80in. — Yellow brown clay streaked
with blue at depth

Several of the species selected for study were
common to each station. Those at Cannington
showed in general a higher degree of xero-
morphy.

Climate

The Perth area possesses a typical Mediter-
ranean climate, the summers being long and
dry, while the rain falls during the mild winter
period. The rains commence in' May and in-
crease in intensity during June and July. In
August there is a slight falling off and during
September and October the rainfall decreases
still further. This winter rainfall accounts for
just over 30in. of the annual average of 34.7 in.
Scanty rain (mean value rather less than lin.)
occurs during November, while December,
January, February, March and most of April
are dry months with rainfall average usually
well under lin. Evaporation figures during this

period are high. Gentilli (1948, 1950) from a
study of climatic data considered that under
the conditions of high evapo-transpiration and
little rainfall of the five summer months, the
native vegetation in the Perth area would be
under stress. Speck (1952) from observations on
soil moisture content in the Cannington area in
early summer also suggested that plants there
would be subject to conditions of water stress
for several months.

Methods

Measurement of Transpiration

Transpiration measurements were made us-
ing Huber’s (1927) “quick weighing” method in
which the loss of weight in the first 2-3 minutes
after the leaf has been severed from the plant
is considered to represent the natural transpira-
tion. This method has been widely used in
ecological and physiological w'^ork but has been
the subject of much discussion and criticism.
The main criticism has been that with the
ruptui’e of the water columns, on cutting, the
release of stress would cause the water to rush
upwards so that the leaf would temporarily be
supplied with more water, leading to heightened
transpiration. Ivanoff (1928) found such an
increase in transpiration and various other
workers including Kamp (1930) Weinmann and
Le Roux (1946) and Anderson, Hertz and
Rufelt (1954), have also recorded such an effect.
Rouschal (1938) working with northern Medi-
teiTanean sclerophylls reported that with one
exception there was a regular fall on weighing
after abscission, while Oppenheimer (1953)
working with similar plants in Palestine
reported a regular decrease with time in some
species tested and considerable irregularity in
others. Our experience here has been that
some species tested in summer showed a slow
consistent decline in weight over successive
two-minute periods, e.g., Hardenbergia, Bos-
siaea, Phyllayithus : while others sometimes
showed a suggestion of the Ivanoff type of
increase followed by a consistent fall, e.g.,.
Hibbertia. In others again (and this applied
more particularly when older leaves were
under test) the rate of water loss w^as somewhat
irregular, but remained fairly high over a
period of several minutes, e.g. Stirlingia, Bank-
sia. . Typical results for water loss over the
first few minutes are given in Table I.

TABLE I

Water Loss from Freshly Cut Leaves

Ilardenbergia

comptoniana

Boiisiaea

eriocarpa

Hibbertia

hypericoides

Stirlingia

latifoiia

Banksia

attenuata

Time

Loss

Time

Loss

Time

Loss

Time

Loss

Time

Loss

Time

Loss

(mg.)

(mg.)

(mg.)

(mg.)

(mg.)

(mg.)

12-10

15 • 5

14-50

12-10

16-10

12-40

12-12

21

15-7

i-8

14-52

2-0

12-12

i - 5

16-12

3-0

12-42

9-3

12-14

1-8

15-9

1-2

14-54

1-7

12-14

1-2

16-14

2-8

12-44

8-3

12-16

1-7

15-11

0-4

14-56

2-3

12-16

11

16-16

3-1

12-46

8-2

12-18

1-3

15-13

0-2

14-58

1-2

12-18

1-0

16-18

2-6

12-48

8-2

12-20

1-3

15-00

1-0

12-20

1-0

16-20

2-7

12-50

8-1

1 *7 • ‘’.•7

0-9

12-22

0-7

16-22

2-4

12-52

6-1

12-24

0-6

16-24

21

16-26

2»4

16

From these and other results it was con-
sidered that the most reasonable measure of
the rate of transpiration for the purposes of
this ecological study would be during the first
two minutes. The downward trend was fairly
general after this time, while the Ivanoff type
of increase, when it occured, was seldom ap-
parent before the first two minute reading.
It is held that the method is basically sound
as well as being at present the most appropriate
for field studies on transpiration.

The torsion balance used by the author was
an Oertling P type (100 mg.) with a milligram
scale and mirror so that readings could be made
to an accuracy of 0.2 mg. Leaves were cut
from the plant with a vaseline smeared razor
blade and suspended from the balance hook by
means of standard weight cotton threads.
Hinged compartments on the balance eliminated
wind effects on leaves and counterpoises at the
moments of original and final weighing, the
leaves being fully exposed for the 2 (or in some
experiments. 3) minutes specified. A stop watch
was used for timing. Parallel measurements,
with as short a time as possible between them,
were made with at least two leaves (opposite
leaflets in the case of Hardenbergia and Ken-
nedya) which were similar in age, and position
on the test plant.

Transpiration rate is expressed in terms of
fresh weight as mg./g./min., and in terms of
surface area [total area of leaf (cf. Rouschal,
1938) ] as mg./sq. dm. /min. where the leaf out-
line could conveniently be drawn and its area
obtained with a planimeter Transpiration
records were obtained on selected days dur-
ing most months of the year. More numerous
experiments v;ere done during the period of
change from the wet to the dry season.

Evaporation

Evaporation was measured using Stocker's
method (1929) as revised by Stahlfelt (1932).
The evaporation from a filter paper disk (area
27.7 sq. cm.) during a period of 1-2 minutes was
obtained, readings being made at hourly or other
intervals as required. These served to give values
which were more readily comparable with
transpiration from test plants. Piche atmo-
meters as described by Walter (1929) were used
to obtain a continuous record of evaporation.

Water Saturation Deficit

Stocker’s method (1929) was followed for the
determination of the water saturation deficit
with the modification that the petioles were
cut once only. After a saturation period of 24
hours the leaves were re-weighed and then
dried to constant weight at 100®C. The initial
and maximum water contents were obtained
by subtracting the dry weight from the initial
and maximum fresh weights and the water
saturation deficit calculated as follows: —

Maximum water content — Initial or field water content
W.S.D. = ' X 100

Maximum water content

Sub-Lethal Water Deficit

As well as determining the Water Saturation
Deficit of the leaves in the course of a day dur-
ing summer, it is necessary to know how much

water the leaf is able to give off without under-
going severe injury. This necessitates the con-
tinuation of a drying out process sufficiently
long for the first signs of death of cells to be
recognized. Oppenheimer (1932) coined the
expression Sub-lethal Deficit for this and
Rouschal (1938) applied the method in detail
to the sclerophyll vegetation near Rovigno on
the noiTh-eastern Adriatic coast. The Sub-
lethal Deficit may be defined as the maximum
water deficit which the leaf will stand without
death of more than ±5% of leaf tissue. The
concept is useful in that it allows us to deter-
mine how close to the actual danger point
natural water loss from the leaf may go and
thus gives more precise meaning to the leaf
Water Saturation Deficit.

Following Rouschal (1938), the procedure was
adopted of rapid torsion balance weighing of
8-10 separate leaves of a selected plant, then
allowing them to dry gradually in air. Periodi-
cally (intervals of 30 minutes were in general
found to be suitable) a leaf was taken and re-
weighed, notes being made on changes in colour
or appearance. The petiole was then slit longi-
tudinally and the leaf placed in water. After
2-3 hours under humid conditions it was
re-weighed to determine whether water uptake
was occurring and finally it was dried to con-
stant weight. Tills pi’ocedure was continued
for successive leaves until the point was found
where even while some water was sUll being
taken up, death cf ±5% of cells was occurring.
Determination of this sub-lethal point pre-
sented some difficulty as Rouschal’s criteria for
estimating d^ath of cells namely healthy tissue
appearing clear, and moribund tissue cloudy
when viewed in transmitted light, proved un-
suitable for most of the sclerophylls examined.
Parker’s tetrazolium chloride test (1951. 1952)
while offering advantages for precise measure-
ment of the lethal level was unsuitable for the
determination of the sub-lethal level required.
Reliance w'as finally placed upon changes in
shape and colour in the leaves and on marked
diminution of their ability to take up water after
a certain period of drying out.

Saturation water content - Dried out water content
Sub-lethal (Max. water content) (-±5% cells dead)

Deficit — X 100

M3Xi;%ium water content

The natural Water Saturation Deficit, which
is obtained from additional leaves at the same
time as the above, may then be compared with
the Sub-lethal Deficit and expressed in per cent
of this, i.e.

Natural Water Saturation Deficit

X 100

Sub-lethal Deficit

Stoviatal Aperture

Schorn’s (1929) series of infiltration liquids
(isobutyl alcohol and ethylene glycol in 11 mix-
tures varying from pure isobutyl alcohol,
through isobutyl alcohol 9: ethylene glycol 1,
isobutyl alcohol 8: ethylene glycol 2, etc. to
pure ethylene glycol) was found to be reason-
ably successful for a number of sclerophylls.

Alvim and Havis’s (1954) method of using n-
dodecane-nujol in a series of ten concentra-
tions in steps of 10% by volume from pure 7t-

17

dodecane to pure nujol was also used in
later experiments and proved valuable in
demonstrating smaller stomatal apertures than
the Schorn series could record. It also gave
more detailed information on slight changes in
stomatal apex’ture. With some plants, however,
difficulty was experienced in obtaining as clear
a reading as with the Schorn series because of
the lack of contrast between the infiltrated and
non-infiltrated areas.

Light Intensity

Light intensity was measured in foot candles
by means of an EEL photoelectric meter fitted
with suitable filters.

Temperature and Relative Humidity

Temperature and relative humidity of the
air were measured using a Sling Psychrometer.
Instead of Relative Humidity, results are ex-
pressed in percentage Saturation Deficit (100-
RH) as this latter, as Oppenheimer and Mendel
(1939) point out, is in direct relationship to
the intensity of transpiration and evaporation.

Soil Suction Force

The suction force of the soil was determined
using the method of Gradmann as modified by
Heilig (1931). The method proved suitable
when the moisture content of soils was reason-
ably low, but in the absence of a suitable
constant temperature room difficulty was
experienced in making measurements of low
suction tensions because of the condensation
of moistui*e on the walls of the vessels and on
the strips of filter paper when the vessels w'ere
taken out of the incubator to carry out w^eigh-
ings. It is believed that moisture in soils at
this stage is readily available to the plant,
forces of not more than 2 atmospheres being
involved in binding the moisture to the soil.

Soil Moisture Content

Samples were taken at 1 foot and 2 feet
depth in the vicinity of test plants. 10 g. of
soil were weighed, then dried to constant
weight. The difference between the fresh
weight and the dry weight multiplied by 10
gave the percentage of moisture in the soil
(Piper, 1944).

Periodicity, Leaf Anatomy and Root Systems

The following sclerophylls were used in this,
study of water economy: — Banksia menziesii,
B. attenuata, Stirlingia latifolia, Bossiaea erio-
carpa, Hibbertia hypericoides. Eucalyptus calo~
phylla. (these species were common to both the
University and Caruiington stations). Eucalyp-
tus marginata, Hardenbergia comptoniana,
Kennedya prostrata. In addition periodic
observations were made on sclerophylls such as
Daviesia nudiflora,. Xanthorrhoea preissii,
Acacia cyanophylla and Conostephium pendulum
and on the tomentose succulent Scaevola
canesce7is, together with the glabrous, semi-
succulent Scaevola paludosa. Phyllanthus caly-
cinus (a soft-leaved but xerophytic plant), and
the mesophyte Erechthites hispidula completed
the types of plant studied.

Periodicity

A feature of the growth of sclerophylls such
as Banksia spp. is the spring flush of growth
which continues into eaidy summer by which
time the leaves are reaching maturity and have
become thick and hard. Hibbertia, Bossiaea
and Phyllanthus commence their new growth
eai'ly in winter and continue through to early
summer, while in Stirlmgia the growth flush
begins in late spring or early summer and leaf
developm.ent continues well into the dry season.
Hardenbergia shows no marked periodicity of
growth, new' leaves continuing to appear
throughout the summer.

Ecological Anatomy — Structure

The mature leaves of Banksia species show a
thick cuticle on the upper surface and densely
matted hairs covering stomata on the lower
sui’face; Hibbertia, Bossiaea, Hardenbergia and
both species of Eucalyptus studied show a
strong development of cuticle wdth stomata
restricted to the lower surface; Stirlingia with
vertically growing leaves and overall cutiniza-
tion has stomata present on both surfaces:
older Kennedya leaves are strongly thickened
with stomata present on both surfaces but more
numerous per unit area on the low’er surface,
while in Phyllanthus the leaves are thin and
soft, the relatively few stomata per unit area
being restricted to the under surface. In Table
II the stinictural characteristics of the sclero-
phylls tested are given according to the scheme
of Evenari (1938).

TABLE II

Anatomical Features of Sclerophylls

Banksia

menziesii

Banksia

attenuata

Stirlingia

latifolia

Hibbertia

hyperi-

coides

Bossiaea

eriocarpa

Eucalyptus

calophylla

Eucalyptus

marghiata

llarden-

bergia

coinplon-

iana

Kennedya

■prostrata

Thick and ciitinized epidermis

-f-

+

4-

+

-j-

4_

Cover of thick hairs on lower

-r

_

Dense stel-

surface of leaf

late hairs

Depression of stoiuata

±

±

Small iiiter-cellular spaces

-1-

+

+

4-

4-

4-

Well developed niechanh’ul tis-

-r

+

+

+

+

-4

+

+

sues

Isobilateral structure of leaves

—

—

+

Keduction in si/.e of leaves ....

—

—

+

4-

High number of stomalcs per

unit area (s(|. mtn.)

+

—

—

+

-f

Medium

Medium

Medium

18

Root System

In Hibbertia the root system is relatively
shallow, being contained in the first two feet of
sandy soil. In Bossiaea and Kennedya the
roots are somewhat deeper penetrating, while
in Stirlingia in addition to a well branched
shallow rooting system a main root goes down
to considerable depth. In one instance such
a root w’as traced to over 8 feet without
appreciable diminution in its diameter. Har-
denbergia possesses a root stock structure
present at shallow depth but from it a strongly
developed root goes down deep into the soil.
Phyllanthus also possesses a well-defined root
stock and a dense clump of roots which, how-
ever, remain I’elatively shallow. Erechthites
possesses a shallow rooting system. The tree
types, Banksia and Eucalyptus spp. possess
both a shallow widely spreading root system
and a deeper penetrating one.

Transpiration

The results of a number of transpiration
studies are presented in Figures 1 to 8. The
purpose is to show the course of transpiration
and water balance of elements of the Swan
Plain scrub vegetation on passing from spring
to the dry summer conditions. Each point in
the curves represents the mean water loss of
two separate leaves (or in the case of Hib-
bertia, Bossiaea and Phyllanthus, of small
twigs beai'ing several leaves).

Stirlingia latifolia

This plant is characterised by fairly high
transpiration losses when water supply and
atmospheric conditions are favourable. In
early spring maximum rates as high as 14-
15 mg./g./min. have been recorded (Fig. lA),
but in the majority of the experiments the
highest values found did not exceed 10
mg./g./min. The average rate (8 a.m. to
5 p.m.) obtained from a number of experi-
ments carried out during early and late spring
periods was 4.7 mg./g./min. Rates after the
break of season rains in May and on dry days
throughout the winter when evaporation was
low, remained below those of spring. Passing
from late spring into summer, rate of water
loss gi’adually fell, maximum rates recorded
being below 4 mg./g./min. in late summer,
(Pig. IB). A daily tw'o-peak curve is
characteristic during the dry period and is to
be contrasted with the typical single peak
curve found in winter and spring experiments.
The average daily rate during dry summers
was found to lie between 2 and 2.5 mg./g./min.
in the University station and was lower
(1.2 mg./g./min.) at the Cannington station.

As will be described in more detail later, the
rate of water loss from young leaves is much
lower than that from mature leaves. As the
amount of new growth in Stirlingia in a given
season is generally small however, and as the
reduced transpiration effect tends to be mini-
mized as the spring flush leaves mature in
summer it is considered unlikely that water
loss differences between young and old leaves
exercise any marked influence upon the overall
summertime water economy of the plant.

Observations on stomatal opening in mature
leaves during spring and early summer months
showed some parallelism with transpiration
trends while in young leaves there was evi-
dence of a high degree of parallelism. Passing
into mid and late summer, mature and matur-
ing leaves of plants at the University station
showed much less con-espondence between
stomatal opening and rate of water loss. On
several occasions stomata were found to
remain widely open while transpiration was
falling. The hydro-reaction of stomata in
mature leaves was found by separate experi-
ment to be very sluggish. Stomata towards
the base of such leaves were less responsive
than those nearer the apical part. Young
Stirlingia leaves showed an interesting varia-
tion in degi*ee of stomatal aperture along their
length. In the early afternoon stomata towards
the apex of the leaf were closed, those in the
mid section were open to a medium degree and
those towards the basal part were widely open.
In the late afternoon the apical stomata
opened widely while those low'er down tended
to close. Use of the much more delicate n-
dodecane-nujol infiltration series facilitated
the study of these changes. In the more desic-
cating environment at Cannington during late
summer, stomata were frequently found to be
closed to isobuty] alcohol throughout the
major portion of the day.

The n-dodecane series was not then in use,
but in the light of comparative tests since
carried out it is likely that stomata may have
been recorded as open to 2 or 3 of this series
The low transpiration rate recorded may thus
represent more than cuticular transpiration.

The water saturation deficit (W.S.D.) was
examined in the Cannington experiments and
was found to remain low during the day wdth
a maximum of 6.9% and an average of

6.4%. In most experiments in summer similar
low W.S.D. values w^ere obtained. An excep-
tion occured in a test plant in February when
a maximum of 17.2% and an average of

11.1% for the day, was recorded. The W.S.D.
during the spring showed no marked increase
during the day indicating that water is

absorbed almost as fast as it is transpired. A

similar picture was found in early summer, but
as atmospheric and soil conditions w^orsened in
late summer, the W.S.D. tended to rise during
the morning hours. When a critical low water
content w^as reached, transpiration began to
fan. With the building up of water content
transpiration rose to give the second peak as
illustrated in Figure IB.

For Stirlingia the relatively low value of
W.S.D. extending into late summer indicated
that even though a large number of the more
shallow roots were non-functional due to
dried-out soil, the deep main root system
ensured that the plant was not subject to great
water stress.

Hardenbergia comptoniana

Hardenbergia shows a fairly high transpira-
tion rate in spring when abundant soil moist-
ure is present, but with advancing summer the
rate of water loss falls to quite low levels. The
transpiration curves in spring are of the one-

19

peak type and follow approximately the course
of evaporation and saturation deficit during
the day (Pig 2A) . Maximum transpiration
rates of about 10 mg./g./min. have been
recorded while the daily average lies between
5 and 6 mg./g./min., the peak of the curve
generally occurring between 12 noon and
2 p.m. In early summer the peak of trans-
piration, while approximately of the same
magnitude as in spring, is found to occur much
earlier in the day. Despite rising evaporation
and increasing saturation deficit, the trans-
piration rate tended to fall to a relatively low
and fairly constant level during the day. Later
on under the more severe mid-summer condi-
tions the morning peak became much shallower
(Fig. 2B) and the water loss during the rest of
the day fell to quite low levels (maximum 2.6
mg./g./min., average 1.5 mg./g./min.). The
mean daily value from a number of experi-
ments during late summer was 1.3 mg./g./min.
Water saturation deficit remained quite low,
maximum values not exceeding 8%. This
plant with advancing summer limits its trans-
piration so that water loss is fairly rapidly
made good by absorption through the deep
root system. There is therefore no difficulty
with water balance. The infiltration technique
for ascertaining stomatal apertui’e could not
be used in the case of Hardenhergia, the leaf
being of the heterobaric type.

In late summer, tests of stomatal conduc-
tance using the cobalt chloride paper method
(Milthorpe 1955) showed very low values as-
sociated with low transpiration rates as deter-
mined by the torsion balance method. It is
likely that during the summer the stomata
exercise control over water loss in Harden-
bergia.

Banksia menziesii and B. attenuata

These two species presented a contrast in
that while B. vienziesii showed quite a high
rate of water loss in late spring (maximum
daily rate 11.5; average rate 7.6 mg./g./min.;
Fig. 3A), B. attenuata lost water at a much lower
rate (maximum 5.2, average 3.5 mg./g./min..
Fig. 4A).

Passing into summer B. menziesii tended to
reduce its transpiration rate to a relatively low
level (daily average 1.7 mg./g./min. Fig., 3B)
earlier than did B. attenuata where the rate in
early summer (avei*age 11.1 mg./g./min.) rose
well above that recorded in spi'ing (Fig. 4B). It
was only in the later part of summer that B.
attenuata reduced its water loss and even then

the average daily value was high at 4.2
mg./g./min. Considerable difficulty was ex-
perienced in endeavouring to obtain a clear
picture of water saturation deficit for both
species of Bauksia. Considerable variability
was found to exist, even between matched
leaves, in their ability to take up water after '
the 2-3 minutes of exposure needed for initial i
weighings. It is therefore reported with cau-
tion that the maximum value obtained did not
exceed 8% in mature leaves and 20% in
cutinized but not fully mature leaves.

In both of the above species of Banksia the
lower surfaces of the leaves are covered with a
dense felt of hairs and the infiltration tech-
nique of determining stomatal aperture could
not be employed.

Hibbertia hypericoides and Bossiaea eriocarpa

Hibbertia. a fairly shallow rooting plant and
Bossiaea, whose root system has been found to
penetrate to 4 feet, both possess the small eri-
cold type of leaf. Owing to the small size of
the leaf, transpiration experiments were con-
ducted using twigs with several leaves at-
tached. The rates of transpiration in spring
were moderately high (see Pigs. 5A and 6A).
Transpiration rates increased on passing from
spring into early summer, while adequate soil
moisture was still present and evaporation was
not unduly high. With advancing summer the
transpiration rate of both plants markedly
declined and by late summer both showed quite
low average rates of transpiration (Pigs. 5B
and 6B). Hibbertia in late summer at the
University station gave average rates of 2.2
mg./g./min. Under the more desiccating con-
ditions of the Cannington area, leaves of Hib-
bertia plants became very revolute and turned
a yellowish colour, water loss being very low.
When twigs were cut and placed in water the
leaves took up water and recovered their
green colour within 2-3 days. This recovery
from yellow to normal green occurs in nature
after the infrequent summer rains and regu-
larly after the break of season rains. Using the
infiltration method stomatal apertui'e in Hib-
bertia could be followed reasonably well during
spring and early summer, but some difficulty
was experienced in the really dry period. Re-
curving of leaf margins and the closer packing
of the stellate hairs then made it difficult to
ascertain whether the stomata were slightly
open or completely closed. The water satura-
tion deficit in Hibbertia increased with advanc-
ing season. In spring maximum deficits of
9.6% were recorded with an average of 7.3%.

FIGURES 1-8 INCLUSIVE

The following symbols are used throughout these figures:—

T. 0 Temperature

S.D. □ Saturation Deficit

E- A Evaporation

TR. — o — Transpiration mg./g./min.

■ “ ■ - - M mg. /sq.dm. /min.

ST. - . • - - Stomatal aperture
Light intensity

20

FIG. 1. — Daily march of transpiration in Stirlingia latifolia — mature leaves. University Station.
(A) Early and late Spring, (B) Late Summer.

FIG. 2. — Daily march of transpiration in Hardenhergia comptoniana in (A) Spring and (B) Early
and Late Summer.

In Fig. 2B: — Early Summer, 4 ^.nd •

Late Summer, A and O

21

FIG. 3. — Daily march of transpiration in mature and young flush leaves of Banksia mensiesii in
(A) Late Spring and (B) Late Summer.

FIG. 4.— Daily march of transpiration in mature and young flush leaves of Banksia attenuata, in
(A) Spring and (B) Summer.

22

EVAPORATION TEMPERATURE EVAPORATION TEMPERATURE

FIG. 5. — Daily march of transpiration in Hibhertia hypericoides, in (A) Spring and (B) Summer.

FIG. 6. — Daily march of transpiration in Bossiaea eriocarpa in (A) Spring and (B) Late Summer.

23

EVAPORATION EVAPORATION TEMPERATURE

FIG. 7 . — Daily march of transpiration in Kennedya prostrata, in (A) Spring and (B) Summer.

Early Summer : — Ei and TRi
Late Summer; — E2 and TR2
Summer (after rainfall): — TR3

24

EVAPORATION

14

FIG. 8 . — Daily march of transpiration in Phyllanthus calycinus, in (A) Early and Late Spring, and
(B) Early Summer.

Early Spring: — Ei,STi and TRi
Late Spring: — E2,ST2 and TR2

FIG. 9 . — Relative transpiration (T/E) of selected plants passing from Spring into Summer, and
monthly rainfall data.

25

RAINFALL POINTS LIGHT INTENSITY

In late summer maximum deficits of 25 to 26%
were recoi'ded. In spring no significant
increase in water saturation deficit occurred
during the day indicating that water uptake
was keeping pace approximately with trans-
piration. With advancing summer the water
saturation deficit rose during the day, the
maximum value often not being reached until
2 or 3 p.m. With the transpiration peak being
reached at 8 or 9 a.m. it is clear that during
late summer the plant is for some time during
each day under considerable sti*ess.

High osmotic values obtained for extracted
leaf juice at this time confirm the condition of
stress. The plant reacts to this stress by the
recurving and incurving of the leaf margins so
that the stellate hairs are pressed close
together. With closure of stomata water loss
is reduced to a low level.

Bossiaea eriocarpa showed the lowest daily
average tran-spiration in late summer of any
plant tested in this series, namely 0.3
nig. /g. /min. For long periods at a time during
a hot day no evidence of water loss could be
obtained with the torsion balance. The rate
might then rise very suddenly, water vapor
being released as it were in a burst (Fig. 6B).
A close correspondence between stomatal
aperture and water loss in Bossiaea was estab-
lished. Where the water loss was negligible,
stomata were found to be completely closed to
isobutyl alcohol, but when the ti"anspiration
burst occurred the stomata could be shown to
be open to 1 or 2 of the infiltration series. The
stomata of Bossiaea react very rapidly in the
hydro-reaction test. Within three minutes
stomatal apertures have been observed to close
down from 5 to 3 on the isotautyl alcohol-ethy-
lene glycol scale; in another three minutes to
1, followed within a further three minutes by
complete closure. The water saturation deficit
in summer was found to be high with an
average of 49% and a maximum of 57%. This,
together with the fact that the osmotic value
of the leaves rose considerably during the sum-
mer months indicated a considerable degree of
stress even though some roots have been
observed to go down to 4 feet in the sand.

Kennedya prostrata

As may be seen from Fig. 7A, which is typi-
cal of several experiments performed, Ken-
nedya shows a one-peak curve in spring with
high maximum and average daily rates of
water loss. Stomata which are present both on
the upper and lower surface show gradual
closure during the day but appear during this
period to have little controlling effect upon
rate of water loss. Thus even when transpira-
tion was reduced to 3.8 mg./g./min. at 5 p.m.
(Fig. 7A) stomata still remained fairly widely
open. Passing into early and mid summer,
transpiration rates frequently remained quite
high and curves were of the two-peak type.
With increasing dry conditions in late summer
low transpiration rates were recorded (Fig.
7B, Ti'z) . Stomata were closed to isobutyl
alcohol series throughout the day. Water
saturation deficit values however remained low,

rising only to a maximum value of 9.6%. Rain-
fall during any part of the summer rapidly
resulted in an increase in the transpiration
rate with a tendency to return to the single
peak curve. No clear overall relationship
between stomatal aperture and water loss
could be demonstrated under spring and
early summer conditions, but in late summer
stomatal closure was an effective factor in
reducing water loss. The hydro-reaction of
stomata in Kennedya was quite rapid.

Eucalyptus marginata and E. calophylla

Marked differences were observed in the
transpiration rates of these two Eucalypts.
Eucalyptus marginata (Jarrah) frequently
showed a high rate of water loss in summer
(average 7.2 mg./g./min.), stomata often being
widely open. Under similar conditions E. calo-
phylla showed a relatively low i*ate of water
loss (average value 4.2 mg./g./min.) and the
stomata during the hotter part of the day were
closed to isobutyl alcohol. At such times the
rate of water loss was restricted to 0.3
mg./g./min. E. marginata may be regarded as
being prodigal of water. Due to its deep root-
ing system adequate water is available even in
late summer. E. calophylla even though pos-
sessing an extensive root system is more sparing
of water in summer.

Xanthorrhoea preissii, Petrophila linearis,
Daviesia nudifiora, Conostephium pendulum

These sclerophylls which were growing at
the University station were tested from time to
time. They showed fairly high rates of water
loss in spring but by mid summer the average
rates of water loss were markedly reduced.

Phyllanthus calycinus i soft-leaved xerophyte)

Phyllanthus is unusual in having quite thin
and soft though small, leaves, and yet occur-
ring as a character plant among the sclero-
phylls. In late winter and early spring the
rate of water loss is high (maximum values
of 17.7 mg./g./min. and average daily values
of 7.2 mg./g./min. were recorded in August-
September). By late spring although the type
of curve was still single peaked the time of
1 ‘eaching maximum rate had moved back to
much earlier in the day and the peak was
lower (Fig. 8A). Phyllanthus showed very
clear infiltration reactions to both the isobutyl
alcohol-ethylene glycol series and the ?i-dode-
cane-nujol series. In early spring experiments
the stomata showed some degree of closure
while transpiration rate was still rising.
Stomatal aperture then remained relatively
constant until late in the day by which time
transpiration rate had fallen to quite low
levels. In late spring and early summer stoma-
tal apertures were still found to remain at
fairly constant aperture while transpiration
was rising or falling.

In early summer the rate of water loss
remained as high as in late spring, but as
atmospheric and soil moisture conditions
worsened, defoliation of Phyllanthus plants

26

commenced and continued up the stems until
by mid-summer (late January and early Feb-
ruary) very few leaves were left. Even at the
stage where lower leaves were commencing to
yellow and fall, the stomata on all green leaves
remained fairly widely open during hot days
and closure did not occur until late in the
evening. The stomata were characterized by
very slow hydro-reactions. Photo-reactions
were faster but as cloudless conditions are
usual this reaction appears to have no eco-
logical significance.

The water saturation deficit of green leaves
still attached, while defoliation w’as occuring
lower down on the stem, showed low values up
to 2.8%. Owing to the fact that a milky latex
is present in Phyllanthus and that this may
have affected tvater uptake in the saturating
experiments, the above low values must be
viewed with caution. It seems significant, how-
ever, that leaves of Phyllanthus although so
soft and thin, do not show wilting.

Prom these results it appears that Phyllan-
thus is prodigal in the use of water and only
balances its water budget and survives the
summer by drastically reducing its transpiring
surface.

Erechthites hispidula ( mesomorph )

The water loss of Erechthites hispidula a
soft-leaved mesomorph growing at the Uni-
versity station was tested for comparison with
the sclerophylls. It showed very high trans-
piration rates through spring (average 10.7
mg./g./min.) and summer (average 13.3
mg./g./minj , up to the time when it wilted
irreversibly and died. Stomata appeared to
have little controlling influence on transpira-
tion throughout the major portion of the day.

Rate of Water Loss in Mature versus Young
Leaves

In the course of preliminary transpiration
experiments using leaves from different parts
of Stirlingia, it w'as obseiwed that the rate of
water loss was much higher from mature
leaves than from young leaves of the current
season flush of growth.

As this ran counter to the commonly expres-
sed view in transpiration literature that young
leaves transpired faster, further studies were
made of Stirlingia and the W’ork was extended
to include the two species of Banksia. These
further tests confirmed the original observa-
tion and it was found that the mature leaves
lost water at a much higher rate (both in
terms of fresh weight and area) during the
hotter parts of the day in spring and early
summer (see Pigs. 3 and 4 and Table III).
With advancing summer the spring flush of
leaves gradually assumed the normal highly
sclerophyllous form and the differences be-
tween the rates of water loss became less
apparent.

TABLE III

Rate of Water Loss From Young and Mature

Lea,ves

Spring.

Early

Summer

ing./g./inin. mg./g./min .

stirlingia latifolia ....

.... Young ....

6-8

7-2

Mature ...

12-0

12-3

BankRia menziesii ....

... Young ....

4-5

4-0

Mature ....

9-0

11-0

Banksia attenuata

. Young . .

2-0

7-0

Mature ...

4-9

15'C«

The lower rate

in the young

leaves is

due, at

least in part, to

the better

stomatal

control

associated with greater mobility of their guard
cells before the processes of lignification and
cutinization develop too far. Experiments
described earlier dealing with the degree of
stomatal aperture in young versus old Stir-
lingia leaves during the day, showed that
young leaves were much more responsive to
changes in atmospheric conditions and in
water content of leaves. The hydro-reactions
of mature leaves were also very sluggish as
compared with those of young leaves.

Cuticular Transpiration

Cuticular transpiration in Hardenbergia,
Hibhertia, Bossiaea, and Eucalyptus species
tested in early summer was found to be quite
low, varying from 0.1 to 0.3 mg./g./min. The
ratio of cuticular to overall transpiration was
highest in Eucalyptus calophylla (1 ; 60). The
lowest ratio under the conditions of these
experiments was found in Banksia attenuata
■ 1 : 36). Owing, however, to the difficulty for
Banksia of completely covering the felted hairs
on the lower surface with vaseline, (particu-
larly near the recurved leaf margins) it is
believed that the water loss recorded may not
have been completely restricted to cuticular
transpiration. Of the above listed sclerophylls
it has been shown for Hibbertia, Bossiaea and
Eucalyptus calophylla that they can com-
pletely close down their stomata under desic-
cating conditions. The first two named also
achieve more efficient protection by strong leaf
inrolling. Thus with low cuticular transpira-
tion. water loss may be reduced to a minimum.
Stirlingia and Kennedya which bear stomata
on both leaf surfaces could not be satisfac-
torily tested for cuticular transpiration. It may
be noted, however, that mature Stirlingia
leaves are heavily cutinized on both surfaces
and it seems reasonable to assume that where
stomata do close completely under late sum-
mer conditions cuticular transpiration would
be slight. Kennedya leaves are lightly
cutinized on the upper surface and dense hairs
are present on the lower surface. Under desic-
cating conditions the leaf margins tend to roll
upward and inward. Transpiration experi-
ments in late summer where stomata were
recorded as closed to isobutyl alcohol, showed
a fairly high transpiration rate early in the
morning. This declined progressively until
2 p.m. with rising evaporation. Subsequent
tests suggested that the stomata would have
been open to at least 3 on the more delicate
??--dodecane-nujol series at the start of the

27

experiment and that gradual closure would
have occurred associated with some decline in
the rate of water loss. The hairs on the lower
surface would further help to reduce trans-
piration and the upward and inward rolling of
the leaf would also give some protection to the
upper surface. Their combined operation
would tend to offset the lack of cuticular
development in Kennedya and would result in
the low transpiration rate recorded.

Course of the Transpiration Curves

Examination of the spring curves plotted in
Figs. 1 to 8 shows that with adequate moisture
in the soil at all levels and favourable climatic
conditions, transpiration in general is high.
The curves tend to run parallel to those for
eyaporation and atmospheric saturation defi-
cit. Passing into summer, with increasing
evaporation, higher atmospheric saturation
deficits and drier soil, transpiration of the
sclerophylls is seen to be gradually restricted.
In the case of Hardenhergia and Hibhertia the
late summer curves are of the one-peak type
and fall as evaporation rises during the later
morning and early afternoon hours.

Stirlingia, Kennedya, Bossiaea and Banksia
show two-peak curves. The first is in the early
morning then with rising evaporation the rate
of water loss drops. The second peak in the
afternoon may occur while evaporation is still
either rising, maintaining a high level, or fall-
ing. The increased rate of water loss at this
time may be associated with a build up of
water content in the leaf tissues but it was
not possible to obtain evidence of this.

When relative transpiration, T/E (used in a
restricted sense) was plotted against month of
year, commencing in spring, the decreasing
transpiration with passage into summer be-
came apparent (Fig. 9). It is noteworthy that
the steepest fall in relative transpiration oc-
curs by early summer. By this time the soil
at the University station at the 2 foot level is
drying out (moisture content 1.3%, suction
force 50 atmospheres). This affects a large
part of the root system of all the plants
tested. The shallow rooting Hibhertia and
Phyllanthus and to a lesser extent Bossiaea
suffer from some water lack at this stage. Stir-
lingia, Hardeyihergia, the banksias and the
eucalypts all possess as well a deep rooting
system. In all cases, how’ever, with the excep-
tion of Eucalyptus viarginata, restriction of
transpiration occurred with rising evaporation
giving low values for T/E. With more desic-
cating atmospheric and soil conditions passing
into lat.e summer, Hibhertia passed into a state
of anabiosis, while Phyllanthus retained only
a few yellowish leaves. Kennedya after the
first drop, maintained fairly high T/E values
passing into late summer, while Hardenhergia
and Stirlingia fell to rather low levels.

Osmotic Values

Detailed observations on the osmotic values
of leaves of sclerophylls will be reported else-
where. It will suffice here to note that osmotic
values rose with advancing summer while in
general a rapid return to lower values occurred

with break of season rains. Hibhertia showed
the highest values at 25-27 atmospheres. Po^s-
sibly higher values would have been obtained
in later summer but the leaves were so dry
that, with the existing sap-press, juice could
not be squeezed from them. Bossiaea also
gave high values up to 21 atmospheres. Here
again higher values would no doubt have been
obtained in late summer if sap could have been
extracted from the leaves.

Stirlingia, Banksia and Kennedya gave
values up to 21 atmospheres (in one instance
a value of 28 atmospheres was recorded for
B. menziesii) and sap could still be extracted
although with some difficulty throughout the
summer. Osmotic values up to 16 atmospheres
were recorded for Hardenhergia in late sum-
mer. The figures for Hibhertia and Bossiaea
taken together wuth the dry nature of the
leaves and the high water saturation deficit
mentioned earlier, suggest a strained wato
balance in late summer. The values for the
deeper rooting Stirlingia and Hardenhergia were
not unduly high. This fact, together with the
lower water saturation deficit observed in these
plants, suggests that they possess a reasonably
balanced water budget.

Soil Moisture and Soil Suction Force

Soil moisture values showed continuous
decline from spring into summer while soil
suction force rose. Typical results at 1 foot
depth are given in Table IV.

TABLE IV

Soil Moisture and Soil Suction Force at 1-foot

Depth

Soil Moisture (%)

University

Sept.

. 3-0

Xov.

1-8

Pec.

1-4

Jan.

1-35

Feb.

1-3

Cainiington

, 4-1

l-U

0-3

Suction Force ....

University

4 ■ 5

28-i

42-5

50-0

(Atin.)

Cannington .

. 3-8

7-6

>100

At a depth of 2 feet the moisture content
at the University station had by late January
fallen to 1.09% and the suction force risen to
75 atmospheres.

For shallow rooting plants — in particular
Hibhertia and Phyllanthus — the decrease in
soil moisture and increase in soil suction force
has considerable significance. By late summer
the suction force at the 2 foot level w^as reach-
ing values which made it impossible for the
plants to absorb water. Hibhertia whose roots
lay in this zone, after progressively reducing
its transpiration loss to low values passed into
an almost anabiotic state, while Phyllanthus
having gradually lost more and more leaves as
soil and atmospheric conditions worsened,
passed the summer in a defoliated state.

With increasing depth the moisture content
of the soil at the University station in late
summer remained high as the values in Table
V indicate.

28

TABLE V

Soil Moisture at University Station^ Late Summer

Depth of Soil Sample

1 foot
4 feet
H feet

Soil Moisture
(%)

1- 3

2- 5

4-0

Suction Force
(Atm.)

50

10

Zero

The main roots of Stirlingia and Harden-
bergia have been traced down to 8 feet in
sandy soil and judging from thetr thickness at
that depth could well continue down several
feet further. Clearly although the surface
lateral roots may be put out of action by rising
suction forces at the two foot level, the pos-
session of a deeper penetrating main root
means that these plants are unlikely to suffer
from severe water stress. The reduction in
transpiration rate in summer may be related to
slower water movement through the diffuse
porous vessels of the deep growing main root
with its subsidiary lateral system.

Discussion

All of the sclerophylls examined with one
exception, were found to reduce their rate of
water loss when passing into the dry summer
period irrespective of whether they were (a)
relatively shallow rooting types as in Bibber-
tia and Bossiaea, or (b) ones which possessed
a combination of shallow and deeper extending
roots as in Stirlingia, Hardenbergia and Ken-
nedya, or (cl the trees Banksia memiesii, B.
attenuata and Eucalyptus calophylla.

Plants of type (a) are clearly sensitive to
soil drought, while types (b) and (c) are only
partially affected. Eucalyptus marginata alone
among the sclerophylls so far examined, main-
tained a high level of water loss during
summer. The shallow rooting soft-leaved
xerophyte Phyllanthus calycinus and the meso-
phyte Erechthites hispidula also show^ed no
tendency to restrict water loss with advancing
season, but under conditions of soil drought
almost complete defoliation occurred in the
former, while in the latter case the plant
finally died.

Comparative studies of w'ater loss of sclero-
phylls during different seasons are not avail-
able for other parts of Australia, but Wood
<1923, 1924) has worked on the transpiration
of sclerophylls during summer in arid inland
South Australia. He showed that while there
was considerable individual variation, their
average rates of water loss were low (Eremo-
phila scoparia 1.15 mg./sq.dm./min.; Casuarina
lepidophloia 2.25 mg./sq.dm./min.: Acacia
aneura 1.38 mg./sq.dm./min.). The meso-
Phyte Senecio magnificus showed a rate of
water loss W'ell above that of all sclerophylls
in the area. This high rate of water loss is
paralleled by that of the mesomorph, Erech-
thites hispidula. in the S\van Plain area. The
rate of loss for this mesomorph is far higher
during spring and summer than that of any
Swan Plain sclerophyll tested. Under the field
conditions near Perth the xerophytic sclero-
phylls therefore do not conform to Maximov’s
experience at Tiflis (1929). Wood (1934) how-

ever, found high values for three Mount
Lofty sclerophylls, Eucalyptus leucoxylon. Aca-
cia pycnantha and Hakea rugosa. In field
studies of similar types of plant in Victoria
and Western Australia, the author has so far
not found such high values (Grieve, 1955).

In other areas of Mediterranean climate
many investigations on sclerophylls and as-
sociated plants have been made. The sclero-
phylls of the Swan Plain are similar in their
water loss behaviour in summer to those from
Rovigno (Rouschal, 1938), Palestine (Oppen-
heimer. 1932, 1953) and Algeria (Killian. 1931,
1932),

The osmotic values of Swan Plain sclero-
phylls so far examined agree fairly well with
Braun-Blanquet and Walter’s (1931) state-
ment that optimum figures lie between 18 and
26 atmospheres. No exceptionally high osmotic
values such as Rouschal (1938) and Oppen-
heimer (1953) record for two or three maquis
type shrubs, have so far been found. The
values obtained by Wood (1934) for sclero-
phylls in the Mount Lofty area near Adelaide
agree quite well with those obtained near
Perth. As might be expected Wood obtained
considerably higher values for sclerophylls of
arid inland South Australia.

The rise in osmotic values of Swan Plain
sclerophylls on passing from spring to summer
is similar to that recorded for sclerophylls in
ihe Mediterranean area (Rouschal (1938); Op-
penheimer (1953).

Oppenlieimer (1932. 1953) distinguishes four
types of Mediterranean maquis vegetation,
based on their water balance. Of these we
may name three into which most Swan Plain
sclerophylls and associated plants fit:

1. — Deciduous pla?its failing to show ap-

preciable stress throughout the sum-
mer — In the Mediterranean area trees
occur in this class, but the closest
Swan Plain equivalent is the soft-
leaved xerophyte, Phyllanthus caly-
cinus which avoids stress by defoliat-
ing during summer.

2. — Evergreen trees and shrubs physio-

logically active throughout the sum-
mer — Eucalyptus marginata is the
only sclerophyll so far worked on in
the Swan Plain area which fits into
this group. It maintains a fairly high
rate of water loss.

3. — Evergreen species restricting their

physiological activity considerably thus
avoiding losses of irreplaceable water,
and finally reaching a state of near
dormancy — ^The two relatively shallow
rooting genera. Hibbertia and Bossiaea
fit well into this category. Both
reduce their water loss drastically in
late summer, have high water satura-
tion deficits and high osmotic values.
Hibbertia in particular passes into a
condition of apparent anabiosis until
the break of season rains.

29

It seems that the Swan Plain sclero-
phyll types such as Stirlingia, Harderibergia>
and Eucalyptus calophylla, which possess a
shallow and a deep rooting system, which have
medium water saturation deficits and medium
osmotic vaues, but w'hich do not go into dor-
mancy during summer, must form a separate
group. They appear to come within Rouschal’s
Group 2 (1938) in that control of water loss
by stomatal closure, or by the operation of
some other internal factor, occurs even when
adequate moisture is available in the soil, at
least to the deeper penetrating part of the root
system.

The findings of von Guttenberg (1927) and
of Oppenheimer (1953) that stomata of the
sclerophyllous evergreens remained wide open
in spring, but practically closed during the day
in a dry summer, could not be duplicated for
all groups of Swan Plain sclerophylls. It is
true that Hihbertia and Bossiaea may close
their stomata completely, but in Stirlingia the
stomata frequently remained open during hot
days at the University station, while water loss
was reduced. Only under the more desiccating
conditions in late summer at Cannington were
the stomata of Stirlingia found to be closed
during most of the day. This at least suggests
that once conditions become too extreme closure
of stomata takes place and only cuticular trans-
piration occurs.

The difference in rate of water loss between
young flush and mature leaves of sclerophylls
appears to have been observed hitherto only
by Henrici (1946) in South Africa, although
Rouschal (1938) did show similar differences
between one year old and two year old leaves
of sclerophylls at Rovigno. Heni'ici noted that
for the introduced Eucalyptus stuartiana on

bright days the young leaves always transpired
less than the old. Owing to the fact that
adequate soil moisture was present at all times
on hot days during her experiments, the results
do not appear to be explainable on the grounds
of difference in stomatal behaviour between
the young flush and mature leaves as is the
case in the Swan Plain sclerophyll, Stirlingia,

The difficulties encountered in determining
water saturation deficits in some sclerophylls
have been indicated. These difficulties affected
also the attempt to apply the concept of sub-
lethal deficits (Oppenheimer 1932, Rouschal.
1938) to indicate relative drought resistance in
Swan Plain sclerophylls. Inconclusive results
for mature leaves of Banksia and Stirlingia
were obtained because very often the leaves
either failed to absorb water or absorbed it
irregularly. It was observed that even with
petioles in water rapid death of such leaves
often occun*ed. Before reaching the conclu-
sion that Banksia and Stirlingia have a low
degree of drought resistance from such results,
further experiments on infiltrating such leaves
under pressure need to be done. Hardenbergia
presented no difficulties with water uptake by
leaves and the sub-lethal deficit here indicated
that this plant would have a high degree of
drought resistance.

Acknowledgments

I am indebted to Miss A. M. Baird for
numerous discussions on the biology of sclero-
phylls in Western Australia and I also wish to
gratefully acknowledge the help given by Miss
J. Rayner in assisting in carrying out several of
the experiments and in preparing the graphs
for publication.

References

Alvim, P. de T. and Havis, J. 1954. Plant
Physiol. 29: 97.

Andersson, N., Hertz, C., and Rufelt, H. 1954.
Physiol. Plantar. 7: 753.

Braun-Blanquet. J. and Walter, H. 1931. Jb.
wiss. Bot. 74: 697.

Diels. L. 1906. “Die Pflanzenwelt von West
Australien. Die Vegetation der Erde VII.*’
(Engelmann; Leipzig.)

Evanari, M. 1938. J. Linn. Soc. (Bot.) 51: 389.
Gentilli, J. 1948. W. Aust. Naturalist 1: 120.
Gentilli. J. 1950. Westralian Farmer’s Co-Op
Gazette: 61.

Grieve, B. J. 1955. J. Roy. Soc. W. Aust.
39: 31.

Guttenberg, H. von 1927. Planta 4: 726.

Heilig, H. 1931. Zeitschr. f. Bot. 24: 225.
Huber, B. 1927. Ber. dtsch. bot. Ges. 45: 611.
Ivanoff, L. 1928. Ber. dtsch. bot. Ges. 46: 306.
Kamp. H. 1930. Jb. wiss. Bot. 72: 403.
Killian, Ch. 1931. Bull. Soc. Bot. Fr. 78: 460.
Killian, Ch. 1932. Bull. Soc. Bot. Fr. 79; 185.

Maximov, N. 1929. “The Plant in Relation to
Water” (Allen & Unwin; London).

Milthorpe, F. L. 1955. J. Exp. Bot. 6: 17.
Oppenheimer, H. 1932. Ber. dtsch. bot. Ges,
50a: 185.

Oppenlieimer, H. 1953. Palest. J. Bot., Reh.
Ser. 8: 104.

Oppenheimer, H. and Mendel, K. 1939. Agr.

Res. Sta. Rehovot, Bull. 25.

Parker. J. 1951. Bot. Gaz. 113: 210.

Parker. J. 1952. Bot. Gaz. 114: 189.

Piper. C. S. 1944. “Soil and Plant Analysis”
(University of Adelaide: Adelaide).
Rouschal. E. 1938. Jb. wiss. Bot. 87: 436.
Schorn, M. 1929 Jb. wiss. Bot. 71: 783.
Speck, N. H. 1952. M.Sc. Thesis. University
of Western Australia.

Stahlfelt, M. G. 1932. Planta 17: 22.

Stocker, O. 1929. Ber. dtsch. bot, Ges. 47: 126.
Walter, H. 1929. Ber. dtsch. bot. Ges. 47: 243.

Weinmann, H.
J. Sci. 42:

and

147.

Le Roux, M.

1946.

S. Afr.

Wood, J. G.
47: 259.

1923.

Trans. Roy.

Soc.

S. Aust.

Wood, J. G.
48: 226.

1924.

Trans. Roy.

Soc,

5, AusU

Wood, J. G.

1934.

J. Ecol. 22:

69.

30

3. — Atapozoa marshi, a compound Ascidian from Western Australia

By Beryl I. Brewin*

Manuscript accepted — 17th May, 1955

A new sub-family Atapozoinae of the family Clavelin-
idae is erected to house a new compound ascidian
Atapozoa mar.s/ii, from Western Australia. In general
features the zooids resemble those of the genus Eudis-
toma Caullery. 1909, but the presence of a stalked brood
pouch and of two median suckers in the tadpole clearly
separate the genera.

Order Aplousohranchia Lahille, 1866

Family Clavelinidae Forbes and Hanley, 1848

Sub-Family Atapozomae, a new^ sub-family

Compound ascidians. No common cloacal
apertures. Zooids with atrial siphons indepen-
dent and with a specialised brood pouch that
arises at thoracic level.

Berrill (1950) recognises three sub-families
of the Clavelinidae — Polycitorinae, Clavelininae
and Holozoinae. The species described below
belongs to none of the existing sub-families,
being separated from the first two by the pos-
session of a brood pouch and from the last by
independent opening of the atrial siphon and
the consequent absence of common cloacal
apertures.

Genus Atapozoa, n.gen.

Colonies pedunculate or sessile. Both atrial
and branchial siphons opening on the surface
of the colony, no common cloacal aperture.
Zooids hermaphrodite. A specialized brood
pouch developing at the thoracic level Tad-
pole with two median suckers.

Atapozoa marshi, n.sp. (Fig. 1)

Colonies (Fig. lA) large, fleshy, pedunculate.
Stalk tapering, up to 4.5 cm. long. 2.0 cm.
wide at base, 1.0 cm. wide at junction with
head. Head up to 4.0 cm. long. 2.8 cm. wide.
Test light greenish brown, firm, with numerous
irregularly-shaped test cells and containing
also round brown “kotballen” — masses of
foreign material. Zooids opening only on head
region over w’hich they are evenly and regu-
larly distributed. Phai*yngeal region salmon
pink, abdominial region green (due to contents).

Zooids (Fig. IBt up to 3.5 mm. long, 2.2 mm.
wide in pharyngeal region which has 18 fine
longitudinal muscle bands of 4 to 6 fibres.
Rectal-oesophageal region short. Abdominal
region about half the width of pharyngeal. A
long vascular process with a central septum
projects from abdominal region (Fig. IE).
Branchial and atrial apertures each with six
short lobes.

Pharynx with 24 tentacles of three orders of
size, regularly arranged. On the inner wall of
the phai-ynx two distinct transverse folds from
each of which a long lappet curves backwards
at the level of the fourth stigmata from the
mid-dorsal line. 3 rows of 28 to 29 stigmata,
8 to 15 times as long as wide. Parastigmatic
vessels absent. Oesophagus narrow; stomach
short, smooth-walled, very curved (Fig. IG);
intestine long, narrow; anal aperture smooth-
edged.

Zooids hermaphrodite. In specimens col-
lected 22nd October, 1952, testis in the form
of a rosette of 8-14 pear-shaped lobes situated
on right of intestinal loop just below stomach
(Fig. IB). Though the rudimentary brood

♦ University of Otago, New Zealand. Presented by Dr.
e. P. Hodgkin, Zoology Department, University of West-
ern Australia.

pouch is present in these specimens the ovary
could not be identified with certainty. Nor was
it appai*ent in specimens collected 1st Novem-
ber, 1952. or 20th December, 1952. It is sus-
pected from the number of layers on the wall
of the brood pouch that the ovary is situated
in the lower region of the atrial chamber, as
it is in Sychieioides tamaramae Kesteven. and
that the brood pouch develops around the
ovary, but the proof of this could not be
obtained (Fig. IP).

Never more than one tadpole per brood
pouch. Brood pouches become large and
remain attached to zooids (Fig. lO. Largest
tadpoles (in brood pouches of colonies collected
20.xii.51) 2.4 mm. wide in head region, 6.8 mm.
long (4.0 mm. being occupied by the tail).
Tadpoles peculiar in the possesssion of two
elongated suckers which lie one below the other
at the extreme anterior end (Pigs. 1C, ID).
Tadpoles with well developed stigmata and eye
spot show no sign of stolon or buds. This
species is quite unlike any hitherto described.
Its resemblance to Colella claviformis Herd-
man is only superficial. I am indebted to the
Australian Museum for permission to examine
the type specimen of the latter. The main dif-
ferences between it and Atapozoa marshi lie
in the structure of brood pouch (which in
Colella claviformis is merely an outbulging of
the atrial wall), the number of embryos per
brood pouch, and the arrangement of zooids in
the colony — those of Atapozoa marshi being all
at the one stage of sexual maturity, whereas
those of Colella claviformis are at different
stages of sexual maturity in the distal and
proximal regions of the head. The stalked
ascidians collected on reefs at Roebuck Bay
near Broome and depicted by Savile-Kent
(1897) resemble this species in form of colony
but differ markedly in colour from the speci-
mens in the present collection. Savile-Kent
describes them as being a “transparent grey
hue, sprinkled throughout their lower inflated
areas with minute bright blue spots. These spots
. . . are found to represent the separate

bodies of the many hundred zooids . .

Great colour range is known for many species
of ascidians and may also occur in Atapozoa
marshi.

Distribution. — Trigg Island, near Perth, Wes-
tern Australia. Collected by Mrs. L. Marsh
from the roofs of cavei'ns under reefs.

Type specivieu. — Deposited in the Australian
Museum. Sydney, No. U3843.

Acknowledgment

I am greatly Indebted to Mrs. L. Marsh for a very well
preserved material, gathered from a comparatively in-
accessible locality at different periods In the hope that
a seasonal range could be stixdied. However, it is
apparent that the questions of position of the ovary
and size of the ovum can be solved only by local scien-
tists with more or less daily access to material. The
life cycle of this ascidian will form a rewarding study.

References

Berrill, N. J. 1950. “The Tunlcata.” Roy. Soc. PuW.
Herdman, W. A, 1899. “Descriptive Catalogue of the
Tunicata in the Australian Museum, Sydney, N.S.W.’'
Liverpool .

Michaelsen, W. and Hartmeyer, R. 1930. “Die Fauna
SUdwest-Australiens." Ergeb. der Hamb. siidwest-
aust. Forsch. 1905. 5: 463-596.

Saville-Kent, 1897. “A Naturalist in Australia.” London,

31

v^ln^' side of zooid with small brood pouch

A ^ pouch in which is held an almost mature

tadpole X20. D — Anterior end of tadpole showing suckers X20. E — TS vascular
process. F— Wall of brood pouch X380. G— L.S. stomach of zooid’.

a — atrial lining,
bp — brood pouch,
em— cells of embryo
Ifc — ?inner follicle cells,
in’ — first portion of intestine.

Explanation of LetterUig

in” — second portion of intestine,
ce — oesophagus,
m — mantle wall,
s — septum,
sd — sperm duct.

32

St — stomach.

stbp — stalk of brood pouch

su — sucker.

vp — vascular process.

0-> — testis.

4. — Crustacea from the Cretaceous and Eocene of Western Australia

By M. F. Glaessner*

Manuscript accepted — 24th August, 1954

A Cretaceous Cirripede peduncle with heavily
calcified integument from the Lower Senonian
of Gingin and a new species of Decapod Crus-
tacean (Protocallianassa australica n.sp.) from
the Eocene strata at 1505 feet in the South Perth
Bore are described.

A Cirripede Peduncle from the Gingin Chalk

Some time ago a peculiar fossil from the
Senonian (Santonian) chalk of McIntyre Gully,
Gingin, was submitted for identification by Dr.
C. Teichert (Melbourne University Geology
Department Coll. No. 1993). The nature of this
fossil became obvious when Withers (1951)
described for the first time calcareous Cirri-
pede peduncles which he assigned to the genus
Euscalpellmn Hoek. Though the new fossil
differs from all four species described by
Withers it will not be given a new name as its
relations to another Cirripede whose capitular
valves occm* in the Gingin Chalk are not at
present clearly definable.

The calcareous stalk (plate, 2 a-d) is
cylindrical, about 40mm. long, gently curved,
narrowing gradually towards it upper end, with
two or three slight constrictions in the upper
half, and with the upper end slightly dilated
to an elliptical shape 15.5 X 18 mm. The lower
end. probably a fracture plane, is flat (plate,
2d). Its outline is elliptical, measuring 19.2 x
21.2 mm., with an elliptical opening (5.5 x
7.0 mm.) situated near the inner end of the
shorter diameter of the basal eclipse, which is
nearer the concave side of the longitudinal
curvature of the stalk. The upper end is fun-
nel-shaped with an iiTegular outline. The
maximum width is 17mm., and a larger open-
ing corresponds in position to the smaller one
at the lower end. The walls of the stalk are
thick and solid. An incomplete division into
plates is faintly indicated by about 25 short
radial furrows on the outer edge of the upper
surface. The outer surface of the stalk shows
wavy sub-parallel growth lines. On its lower
half there are a few widely scattered outlines
of peduncle plates. They become more fre-
quent on the upper half and cover almost com-
pletely this portion of the stalk where they
form several imbricating rows. The plates are
irregularly triangular in external view, with a

* University of Adelaide.

wide convex base and a rounded narrow upper
end. The umbo is produced below the apex
into a sharp pointed spine or hook which
curves outward and slightly downward, par-
ticularly at the concave side of the curvature
of the peduncle. The plates are completely
fused with each other and with the undivided
calcareous matter of the peduncle through
which they are scattered in the lower portion
of the fossil.

The peduncle from the Gingin Chalk
resembles Euscalpellum zelandicum Withers in
its curvature but differs in the outline of the
plates. They are elongated in the New Zealand
species, with parallel sides. According to
Withers they are regularly developed near the
base and occur sporadically near the top. The
second of these characters depends on the
orientation of the specimen, taut in any case
the first seems to exclude the possibility of
specific identity. The other species described
by Withers, E. antarcticuni Withers from the
Upper Senonian of Graham Land, E. eocenense
(Meyer) from the Eocene of the Gulf Coast of
tlie U.S.A., E. crossissiviim Withers from the
Upper Eocene of Tierra del Puego, differ in
the shape and character of their plates,
E. aiitcircticiivi being closest to E. zelandicum
and the present species.

The generic identification of the peduncles
described by Withers as Euscalpellum was
based on the occurence of capitular valves
belonging to E, eocenense together with a
“strongly plated” peduncle. This species, how-
ever, differs most markedly in the shape of the
number of peduncle plates from the present
fossil. Scalpellid capitular plates are unknown
from the other localities from which “mons-
trously developed” peduncles have been de-
scribed. Two Cirripede species occur at Gingin,
Zeugmatolepas au,stralis Withers and 5ca/-
pelluyn iNeoscalpellum) glauerti Withers. The
sub-genus Neoscalpellum is characterised by
reduced calcification of valves which makes it
unlikely that the heavily calcified peduncle
belongs to this species. On the available evi-
dence a relation between Zeugmatolepas aus-
tralis and the present specimen cannot be
excluded. The genus Zeugmatolepas possesses
“three or more whorls of subtrianguiar lower
latera with V-shaped growth lines” (Withers
1935, p. 79). The lower latera of the type
species, Z. mockleri Withers, are “sub-triangu-
lar, with angularly rounded growth lines”..

33

They resemble strikingly the uppermost ped-
uncle plates of the new specimen. In Z. aus-
tralis the lower latera are described as “tri-
angular, some acutely triangular and bowed
outwards” (Withers 1935. p. 94). This similarity
in shape between the low'er latera of Zeiigma-
tolepas and the peduncle plates of the new
specimen does not prove specific or generic
Identity, particularly when the difference in
size and also in shape between the peduncle
plates and lower latera in the Jurassic Z. con-
cinna (Morris) (Withers 1928, p. 103) is taken
into consideration . Moreover, the valves in
Zeugmatolevas are comparatively thin (Withers
1928, p. 98). Nevertheless, a taxonomic rela-
tion between these fossils which occur together
is more likely than one between the “mons-
trously developed” stalks which differ widely in
the structure of their peduncle plates. The
naming of the specimen will depend on further
discoveries of either similar fossil peduncles in
their original connection with capitular valves
or at least of loose capitular valves at one or
more of the other localities at which scalpellid
Cirripedes are at present represented only by
“monstrous” peduncles.

A Decapod Crustacean from the South Perth

Bore

In 1899 the late A. Gibb Maitland sent a
number of Crustacean remains from bores in
the Perth area to R. Etheridge jun. for iden-
tification. Twenty years later. Maitland (1919)
referred to '‘Tellina, Fusus or Triton, and Cal-
lianassa or Thalassina^’ from depths of between
1505 and 1831 feet in the South Perth Bore.

The Palaeontological Collection of the Aus-
tralian Museum in Sydney contains a Crus-
tacean on the surface of a core taken at a
depth of 1505 feet from this bore (No. F5993
‘■presented by A. G. Maitland 1899”). This is
one of the specimens examined by Etheridge
and named ''Callianassa or Thalassina:’ As it
is almost complete and identifiable and comes
from a formation from which only foraminifera
have been described, it is desirable to give a
full account of this fossil.

The fossil is preserved as a rather shadowy,
dark, flattened pellicular body, probably chi-
tinous and almost completely uncalcified with
the exception of the finger tips. The matrix is
a dark grey laminated shale, slightly sandy and
glauconitic, with interbedded lighter bands and
with microfossils including foraminifera, sponge
spicules, biyozoa and ostracodes. and organic
debris, visible under the low-power microscope
on some bedding planes. These planes show a
clear dip of 10^ The fossil is flattened, lying
on its side on a bedding plane on which few
microfossils. probably ostracodes, are indis-
tinctly visible.

Protocallianassa australica nov. sp.

Plate, 1

Description.— The abdomen and thoracopods
are clearly visible but not all the legs can be
identified and the carapace is not in its normal
position in relation to the rest of the body. A
sharp semi-elliptical ridge above the merus of
the larger cheliped may represent a cast of the

cervical groove and obscure remnants of the
carapace seem to extend upward from this line,
suggesting that the fossil is preserved in moult-
ing position (Glaessner 1929). Neither the
rostrum nor the areas on which lineae could
be seen are preserved.

The abdomen is almost completely preserved
in a strongly flexed position. The pleurae of
segments 2 to 5 are well developed, terminating
in rounded lobes. Only the pleurae of the right
side are visible and it is uncertain whether the
visible dorsal outline of the flattened abdomen
is in its median line. The outlines of the first
segment and the tail fan are not clearly pre-
served because of overlapping by the pereiopods.

The first pereiopods are heterochelous. The
right chela is larger, with an apparently gently
convex dorsal edge of the propodus which,
however, could be slightly modified by its flat-
tening. The ventral edge of which only the
distal part is clearly preserved is straight and
probably ridged. Tlie proximal edge is very
slightly Inclined downward and forward. The
immovable finger is straight and narrow and
equals in length the dactylus which is wicie
(dorso-ventrally) at its base, with a straight
ventral and a very strongly curved dorsal edge.
No teeth are visible and it is probable that
none were present. The carpus was much
shorter and probably narrower than the
propodus. Only its distal and dorsal outlines
are clearly visible. The outline of the merus
is irregularly lozenge-shaped. The ischium
was apparently rectangular and narrow. The
left chela is much smaller and its dactylus is
slender but not much shorter than that of the
right chela. The immovable finger is shorter
than the dactylus. The carpus shows clearly a
regularly curved edge extending from the
proximal joint to the ventral edge of the
propodus. The terminal joints of the remain-
ing pereiopods are not visible. The last pereio-
pods are preserved in a dorsally flexed posi-
tion as in living specimens of Callianassa.

Measurements (in mm.)
Length of abdomen (measured

along dorsal curvature) .. . 37

Larger (right) first cheliped

Length from base to tip of

dactylus 39

Length of propodus 17

Height of propodus 10

Length of dorsal edge of

propodus 12

Length of dactylus 9

Length of carpus (dorsal) . 5

Length of merus 8

Height of merus 5

Length of ischium 5

Smaller (left) first pereiopod

Length of propodus ll

Heigth of propodus 5

Length of dorsal edge of pro-
podus 6

Length of dactylus 7.5

Length of carpus (dorsal) .... 4.5

34

Comparisons . — The new species is placed in
the genus Protocallianassa Beurlen 1930, Type
species P. archiaci (H. Milne Edwards), which
is distinguished by a linea thalassinica on the
carapace together with well developed pleurae
on the third to sixth abdominal segments, uro-
pods without diaeresis, and large heterochelous
first chelipeds. It was considered by Beurlen
as intermediate between the Axiidae and Cal-
lianassidae but was placed in the latter family
as the sole representative of a subfamily
Protocallianassinae (Beurlen 1930, p. 332).
Mertin (1941) described several species from
' the Upper Cretaceous of Europe and referred
to the same genus two species from the Upper
' Cretaceous of North America. He noted that
the Lower Cretaceous ^*Callianassa^^ uncifera
Harbort closely resembles the Upper Cre-
taceous species of Protocallianassa to which
genus the only other European Lower Cre-
taceous species described as Callianassa (C^
neocomiensis Woodward and C. urgonie7isis
Lorenthey) are also likely to belong. The new
species is distinguished from all these species
by the outlines of the carpus and propodus of
its chelipeds and also by its rounded second,
third and fifth abdominal plemae. It differs
from Callianassa bakeri Glaessner (Eocene of
Victoria) of which only the chelae are known,
in their shape and ornamentation.

Mei-tin (1941, p. 209) has pointed out that
the genus Protocalliaiiassa may well extend its
range into the Cainozoic. Few complete Ter-
tiary specimens of Thalassinids are kno\\n and
many of the numerous species of Callianassa
based on chelae of widely vai*ying shapes can-
not be definitely assigned to this genus. The
present specimen is the first definite record of
a Thalassinid with well developed abdominal
pleurae from the Eocene.

Ape.— The Eocene age of the strata at 1505
feet in the South Perth Bore is proved by the
occurrence of a distinctive fauna of smaller
foraminifera in the core which contains Proto-
callianassa australica. Its microfauna in-
cludes;

Textularia sp.

Quinqueloculina sp.

Lenticulina sp.

Angulogerina cf. subangularis Parr

‘'Discorbis assulatus Cushman” (as figured
by Parr)

Eponides sp.

Alabamina westraliensis (Parr)

Anomalina cf. glabrata Cushman
Cibicides uvibonifer Parr
Cibicides spp.

Globigerina aff. bulloides d’Orb.

Globigerina niexicana Cushman

Globorotalia cJiapmani Parr

Gumbelina rugosa Parr

Ostracode fragments

Sponge spicules

Bryozoa

Fish teeth

This assemblage resembles closely the fauna
described by Parr (1938) and later studied by
Coleman (1950). It is at present the lowest
known occurence of an Eocene fauna in the
Perth Basin.

References

Beurlen, K., 1930. Vergleichende Stammes-

geschichte. Fortschr. Geol. 8: Part 26.

Coleman, P. J., 1950. Foraminiferal Investiga-
tions in the Perth Basin, Western Aus-
tralia. J. Roy. Soc. W. Aust., 36: 31-43.

Glaessner, M. F., 1929. Zur Kenntis der Hau-
tung bei fossilen Krebsen. Palacobiologica,
2: 49-56.

Maitland, A. G., 1919. A Summary of the
Geology of Western Australia. Mining
Handbook, Geol. Surney W. Aust., Mem. 1.

Mertin, H., 1941. Decapode Krebse aus dem
subhercynen und Braunschweiger Emscher
und Untersenon. Nova Acta Leop. Carol.
New Series 10, No. 68.

Parr, W. J., 1938. Upper Eocene Foraminifera
from deep borings in King’s Park, Perth,
Western Australia. J. Roy. Soc. W. Aust.,
24: 69-101.

Withers, H., 1928. Catalogue of Fossil Cir-
ripedia in the Department of Geology, vol.
1, Triassic and Jurassic. Brit. Mus. (Nat.
Hist.), London.

Withers, H., 1935. Idem, vol. 2, Cretaceous.

Withers, H., 1951, Cretaceous and Eocene ped-
uncles of the Cirripede Euscalpellum.
Bull. Brit. Mus. (Nat. Hist.), Geol., 1, No. 5.

35

■i

'yi

ji

i

i

t

I

2a

Explanation of Plate

1. — Protocallianassa australica nov. dp. Holo-

type. South Perth Bore. Core from 1505 ft.
Australian Museum. Sydney No. P5993.

2, a-d. — Cirripede peduncle. McIntyre Gully,

Gingin, Western Australia. Lower Senonian.
Melbourne University Geology Department,
No. 1993. 2a-2c X 2.5, 2d nat. size.

Journal
of the

Royal Society of Western Australia
Vol. 40 Part 2

5. — The Reaction of Plants to Growth Regulators with particular

reference to Weed Control

Presidential Address, 1955

By G. R. W. Meadly, M.Sc.*

Delivered — 18th July, 1955

The discovery, functions and composition of
plant hormones are reviewed, along with the
morphological and physiological responses of
plants to related chemicals. The use of these
chemicals in agriculture, particularly for weed
control, is discussed and the results of work on
weed and crop reaction In Western Australia
outlined.

Introduction

About a century ago Julius Sachs deduced
from his experiments on plants that special
substances are responsible for the formation and
growth of different organs. Sixty years later
Fitting (1909) made the first efforts to deter-
mine the chemical nature of plant growth stimu-
lators after finding that water extracts of dead
pollen initiated swelling of the ovaries of
orchids. He made several attempts to frac-
tionate the pollen extract but no further advance
was made until Laibach (1933) showed that
the active pollen substance was probably identi-
cal with one of the auxins.

Over the .same period discoveries were being
made in another field of plant physiology dealing
with tropisms. Darwin (1880) showed that some
influence is transmitted from the upper to the
lower part of seedlings when they are exposed
to one sided illumination. Boysen-Jensen (1913)
demonstrated that this stimulus can be trans-
mitted through a gelatin layer by cutting off the
tips of Avena coleoptiles and pasting them on
again with gelatin. Six years later Paal (1919)
showed that the stimulus did not cross a layer
of cocoa butter, mica or platinum foil and that
a tip placed on one side of the coleoptile caused
curvatures similar to those seen in phototropic
experiments. He postulated the existence of
a diffusible correlation carrier which is produced
in the tip and moves downward. Phototropic
effects were explained by an interruption in the
normal flow of the substance through inter-
ference with its action due to some change in
the protoplasm.

* Department of Agriculture, We.stern Australia.

Went (1926) was the first to collect the active
material when he found that it diffused into
gelatin from living oat coleoptile tips. This
made possible the development of a technique
for the quantitative measurement of the growth
substance before it had been isolated chemically.
An agar block containing the active material
is placed on one side of the cut coleoptile
surface after removing the tip and the amount
of curvature caused by the growth difference
of the two sides is measured. Up to a cei'tain
concentration the curvatures are proportional
to the amount of growth substance applied.
This Avena test has a high degree of sensitivity,
one twenty millionth of a milligramme of growth
substance giving a curvature of about 10 degrees.

Went’s test was of considerable assistance to
workers endeavouring to isolate and define
growth substances. In 1928 Nielsen found that
two pathogenic fungi produced in the nutrient
medium, substances which were strongly active
to the Avena test and shortly afterw^ards auxin
production was associated with other fungi and
bacteria. Attempts to isolate these substances
were deferred when the Ave^ia coleoptile was
found to react strongly to urine. By 1934 three
active substances had been isolated — auxin-a,
auxin-b and 3-indolylacetic acid. The latter
acid was subsequently isolated from yeast and
Rhizopus and for some time was considered to
be restricted to lower organisms. Evidence to
the contrary, however, continued to increase and
it was finally isolated from wheat and maize
by several workers.

Use of Growth Regulators

Following the synthesis of growth regulating
substances, applied experimental work was
undertaken in a number of directions and plants
were found to react in a variety of ways. One
of the first practical uses was associated with
vegetative propagation. With many plants thal<
are dilficuU to strike, root development may
be hastened and increased by tre^ating the cut-
tings vith growth regulating substances. These

37

chemicals are being used by nurserymen without
a great deal being known concerning their
physiological action.

The theory that a natural root inducing factor
moves from the leaves to the base of the cutting
causing root initiation has been favoured for
some time. When, for some reason this factor
does not operate it is possible for it to be replaced
by a synthetic growth regulator. Among the
better known root inducing substances are
indolylbutyric acid, naphthylacetic acid and
naphthylacetamide.

Treatment of flowers such as those of the
tomato, has under certain conditions, pro-
longed development resulting in larger fruits
which may be seedless. With beans, matura-
tion has been hastened and at the same time
the size increased. The rate of ripening of
formed fruits can also be hastened and it is
not necessary for them to remain attached to
the trees. Apples, pears, peaches and bananas
have been stimulated by very small quantities
— a few parts per million— of naphthylacetic
acid and related chemicals.

An accepted practice to improve the fruit set
of the Zante currant is to cincture or gircile
the main stem of the vine, removing a narrow
strip of the inner bark. It is thought that
this restricts the downward movement of sol-
uble food materials and results in a concenti’a-
tion above the cincture. The currant fruit is
normally seedless and apparently this stimulus
is needed to induce satisfactory fruit set and
fruit size. With cincturing of weaker vines
a marked deterioration of vigour may occur
and an alternative method of improving fruit
set has been sought. Work undertaken by L
T. Jones (1953) in the Swan Valley has shown
that satisfactory setting has followed spraying
with a solution containing 20 to 40 parts per
million of p-chlorophenoxyacetic acid. 2, 4 -
dichlorophenoxyacetic acid was somewhat less
effective and, at times, caused damage to the
bunches.

Abscission of flowers has been both retarded
and accelerated by the same growth regulator.
The reaction of greenhouse tomatoes, however,
has been sufficiently uniform for them to be
treated commercially to prevent flowers from
falling and these substances have also been used
to replace pollination. Careful control of the
treatment is essential as relatively small quan-
tities are lethal to tomatoes.

Spraying to bring about thinning wdien ex-
cessive flowering of fruit trees occurs, takes ad-
vantage of the conditions when floral abscission
is increased. The same variation occurs with
leaves and fruits but detailed investigations have
allowed growth regulators to be used in some
cases for retarding fruit drop, a practical ap-
plication that can be very useful in orchards
The stage of fruit development at the time of
application is important.

Although the properties of growtli regulat-
ing substances already mentioned are import-
ant, their herbicidal effects are of outstanding
significance. The major portion of the address
will be devoted to this aspect.

Herbicides

Chemistry

It is now generally recognised that indoly-
lacetic acid is one of the important hormones
found in plants and as it can also be synthe-
sized in the laboratory, it has been an import-
ant starting point for the preparation of sub-
stances likely to have herbicidal properties.

Studies began at the Jealotts Hill Research
Station, England, in 1936. the earliest work deal-
ing with stimulation of the rooting of cuttings
It was soon extended to include the effects of
3-indolylacetic acid and 1 -naphthylacetic
acid at different concentrations upon seed
germination, seedling development and the
growth cf established plants of many species.
The results of these experiments showed that
the higher concentrations of growth substanqgs
actually depressed growth. These observations
together with a knowledge of the variation in
rcot formation and deformation effects pro-
duced v.itti different species led to the assump-
tion that these growth substances in higher
cciicentrations would have selective phytocidal
action.

Slade. Trmpleman and Sexton (1945> found
in 1940 that applications of 251bs. naphthy-
lacetic acid per acre to oats weedy with char-
lock 'Brassica sivapis) killed the charlock. The
oats received only a slight check and recovered
fully. 84^/} of the charlock failed to germinate
and seedlings that emerged soon died. Other
experiments indicated that wheat, barley and
rye behaved similarly to oats while large plan-
tain 'Plaritago major) and yarrow or milfoil
f Achillea millefolium f resembled charlock in
their responses.

These results confirmed the selective phyto-
cidal properties of certain growth substances
and initiated the search for cheaper and more
active chemicals than naphthylacetic acid.
Within two years of this work investigators in
both England and America had recognised the
strong growth regulatory and herbicidal effects
of chlorinated phenoxyacetic acids.

Even during the early investigations a con-
siderable number of chemicals were screened.
In America the derivatives of 2. 4-dichloro-
phenoxyacetic acid (2,4-D) were favoured while
in England, where exploratory work commenced
several years earlier, the compound selected for
development was 2-methyl-4-chlorophenoxy-
acetic acid (M.C.P.A.). This was due in part at
least to the greater availability in England of
chloro-cresol as oppOvSed to chloro-phenol. Of
the many other related chemicals that have
been tried 2.4.5-trichlorophenoxyacetic acid
(2,4.5-T» has proved of greatest practical value
forming the basis of preparations for treating
woody plants.

It is interesting to note the similarity in the
structural formula of indolylacetic acid and the
synthetic hormone-like substances such as the
clUorinated phenoxyacetic acids. The main
difference in structure is that the -CHjCOOH
group of indolylacetic acid is attached to the
benzene ring through a pyrrole ring while with
the ciilorinated phenoxyacetic acids this group
is attached through an oxygen atom.

38

2.4- D acid is only slightly soluble in water
and for herbicidal purposes its derivatives are
the important chemicals. The most widely used
are the sodium and triethanolamine salts and
the esters. The sodium salt is moderately soluble
in water but “hard w’ater" having a high cal-
cium or magnesium content may cause a pre-
cipitate that will clog the filters and spraying
nozzles. The amine salts are soluble in all pro-
portions m water and are therefore well adapted
to the low volume spraying equipment, even
when high rates of chemical are required. The
ester compounds are only slightly soluble in
water but may be dissolved in some petroleum
oils. By the use of suitable emulsifiers the
minute oil droplets containing the ester can be
kept in suspension in a suitable form for spray-
ing purposes. The esters are synthesized by the
reaction between 2,4-D acid and an alcohol. As
a large range of alcohols is available it is
potentially possible to have a large range of

2.4- D esters, although naturally the cheaper and
more readily available alcohols are mainly used,
e.g., methyl, ethyl, isopropyl and butyl. The
length and structure of the alcohol portion of
the 2.4-D ester molecule affects its vapour pres-
sure and in consequence, the volatility. In
general, the longer the carbon chain in the part
of the molecule contributed by the alcohol the
lower the volatility. Where low volatility is
important, high molecular weight alcohols have
been used but they are more expensive. M.C.P.A.
has been used mainly as the sodium salt and
has similar properties to 2.4-D. With 2,4,5-T.
the ester is the main formulation, in some cases
high molecular weight alcohols being used to
reduce volatility.

As the components of proprietary lines of
herbicides have different molecular weights and
are also prepared in different concentrations it
is essential to be able to compare the active
chemical content of preparations and to have
a uniform method of expressing rates of ap-
plication. The term acid equivalent is used and
refers to that part of a 2.4-D formulation that
theoretically could be converted to the acid. The
acid equivalent may be stated as a percentage
by weight or the weight of 2.4-D acid in a given
volume.

2.4- D may be classed as a non-toxic sub-
stance. Cows and sheep were not affected by
feeding on pasture sprayed with 2.4-D and no
harmful results followed daily doses of 5.5 grams
to a cow for 106 days. The 2.4-D was not
excreted in the milk, nor was it found in the
blood serum of a calf given milk from the test
cow. There is some evidence that the growth
of aerobic micro-organisms may be inhibited
while some groups of anaerobic organisms may
be stimulated. At the usual rates of application

2.4- D does not generally reduce the total number
in the soil.

Morphological Effects

The effect on plants depends primarily on the
quantity applied and absorbed but is also in-
fluenced by a number of other factors including
the growth stage of the plant, the rate of growth
at the time of spraying and climatic conditions,
particularly temperature. As already mentioned,

at controlled rates, the growth regulating pro-
perties are apparent at much lower concentra-
tions than the toxic effects. The first reaction
is often a twisting or bending of the stems and
leaves resulting from differential growth rates
in petioles and elongating regions of the stem.
Leaves may become thickened and an increase
in turgor often results in a propping appear-
ance. Colour changes similar to autumn leaves
may occur. Stems also become thickened and
often split. The extreme response is a cessation
of growth followed by a characteristic browning
of the foliage and stems associated with the
dying of the tissue.

With sublethal doses the reaction of the plant
takes a number of forms. Besides the twisting
of the stems already mentioned, leaves produced
after the application may assume unusual
shapes. Entire leaves become deeply divided and
those normally having broad lobes may be re-
duced to narrow segments. Proliferation of
floral parts is not uncommon and fasciations
are also caused as well as adventitious roots.
Parthenocarpy may be induced resulting in
seedless fruits. In the case of cereals, irregular
spikelets and empty florets occur. These re-
sponses can be attributed to the interference of
hormonal action in the young structures by the
synthetic growth regulating substances. 2,4-D
appears to concentrate in the young embryonic
or meristematic tissues that are in a rapid state
of metabolic activity, affecting them more than
mature or relatively inactive young tissues.

Physiological Effects

Although the physiology underlying the effects
of synthetic growth substances on plants is far
from being fully understood and applied studies
have been given more attention than funda-
mental research, part of the story can be told.
The severe and diversified effects indicate that
they exert an influence on some general and
basic metabolic process. As pointed out by
Bonner and Bandurski (1952L according to this
view, the observed response, whether growth by
elongation, production of roots, suppression of
lateral buds or of flowering, increase in rate
of protoplasmic streaming or production of
changes in chemical composition, would be the
visible manifestation of the effects on this com-
mon basic process. A ready and simple ex-
planation would be an essential relationship
between enzymes and growth regulating sub-
stances. It has been shown that the action of
numerous enzymes are influenced by those sub-
stances, some being increased and others in-
hibited. but information available gives only
partial support to this theory.

A number of effects on chemical composition
and physiological processes have been recorded.
Respiration may be stimulated or retarded, the
reaction apparently depending on the concen-
tration in relation to the susceptibility of the
plant. Kelly et al. (1949) showed that the rate
of respiration in pea seedlings was increased by
concentrations within the range of 0.001 to
more than 100 parts per million of 2,4-D while
1,000 parts per million reduced the rate by 40
per cent. On the other hand a 20 per cent,
stimulation with oat seedlings required at least

39

1,000 times as much 2,4-D as for the same stimu-
lation in pea seedlings. Even when lethal doses
are applied an initial increase in respiration
may occur follow^ed by a reduction as the plant
approaches death. The influence on the respira-
tion rate is considered by some workers to be
linked with the effect of 2.4-D on water uptake.

By some means not fully understood 2,4-D up-
sets the balance between synthesis and use of
plant food, particularly carbohydrates. After
application in herbicidal quantities there is a
slight decrease in the rate of photosynthesis
associated with a steady loss in total dry weight
due mainly to a decline in the weight of starch
and starch-like substances. Usually there is an
initial increase in the sugars followed by a rapid
and steady decline. There is still some doubt
whether the depletion of sugars is suflicient to
cause death of the plant.

Both increases and decreases in the protein
content of wheat grain have been recorded with
increases predominating. When higher percent-
ages have been recorded some doubt has existed
whether there was increased synthesis of the
proteins or whether a decrease in total seed
weight accounted for the proportional increase.
There are indications of a lowering of the nitro-
gen and potassium absorption and treatment
with 2.4-D has also reduced the upward move-
ment of radioactive phosphorus to the leaves
and modified the distribution and accumulation
pattern of this radical.

It has been proposed by van Overbeek fl947)
that 2,4-D. like natm*al auxin, might affect oxi-
dative assimilation in the cell but unlike auxins
might escape inactivation by oxidases that norm-
ally regulate metabolism. The resultant increase
in the catabolic processes in the cells while the
anabolic system was blocked would cause rapid
injury to the plant. Johnson and Fults (1952)
have shown that once it has entered cells. 2,4-D
stimulates the production of scopoletin, a cou-
marin-like compound that is highly toxic.

Although the actual cause of plant mortality
following the application of growth regulating
substances has not been defined with certainty
there is no doubt that they disorganise a number
of metabolic processes.

Absorption and Translocation

It is already apparent that growth regulating
substances applied to plants produce effects at
some distance from the point of application.
This involves both absorption and translocation.
Hormones may be transported in plants by three
different mechanisms. Following absorption by
the roots they move in the xylem transpiration
stream to the leaves. In the process they may be
absorbed selectively and accumulated in adjacent
active tissue such as xylem. parenchyma and
cambium. Hormone molecules in leaves move
with carbohydrates into the phloem where they
are carried in the assimilation stream to regions
where foods are being used or stored. In living
parenchyma hormones may move from cell to
cell by a polar mechanism.

Researches undertaken by many independent
workers indicate a decided similarity between the
movement of plant hormones and 2.4-D sub-
stances. It is generally accepted that chemicals

in aqueous solution do not enter the leaf through
the stomata but diffuse through the cuticle.
Crafts (1948) suggested that, for ready entry
through the cuticle, herbicides should be in a
non-polar form. This would explain why the
non-polar estei’s of 2.4-D pass through a waxy '
cuticle more rapidly than the polar salts.

Having penetrated the cuticle the 2,4-D mole-
cules pass from a lipoidal medium into the
aqueous medium of the mesophyll cells where
somewhat polar properties are required to facili- '
tate translocation. Experiments undertaken by '
Crafts (1952) have shown that 2.4-D ions of the ‘
sodium salt enter leaves slowly; aliphatic esters
enter rapidly but do not part from the cuticle
and translocate well; and 2.4-D acid and heavy
ester molecules pass through cuticle, mesophyll
and phloem with relative ease. Having reached
the phloem without causing immediate damage
to the intervening tissue they are translocated
v;ith synthesized food materials to regions where
it is being used. The affects of the 2,4-D are
therefore likely to be felt at the most vital points
and deep penetration of perennial rooting sys-
tems is often possible.

Translocation can be quite rapid. Although
one to two hours may be required for the chemi- ,
cal to penetrate the cuticle and move across the
mesophyll. Day (1950> has shown that in the
phloem it may move at rates from 20 to 160
cm. per hour. The rate of movement is not
related to concentration of 2.4-D but is influ-
enced by the rate of food movements. High
concentrations can impede or arrest transloca-
tion by acting as contact herbicides and killing
the conducting tissue.

Differential Action

For the agriculturist the most important
characteristic of synthetic growth regulators is
their selective action on plants. Although in
general, grasses are resistant and broad leaved
dicotyledons are susceptible there are many
exceptions involving variation in tolerance of i
different species and even different varieties.
Undoubtedly a number of factors contribute to
the variation in reaction.

A waxy cuticle, particularly w^hen associated
with close parallel venation as with cereals,
causes many spray droplets to bounce and run
off. On the other hand broad leaved plants
have leaf surfaces that are more readily wet by
sprays which may spread as a thin film or
retain contact over a large pi’oportion of the
surface in the form of fine droplets. The up-
right habit of grasses, especially cereals, also i
causes more run off than leaves that are dis-
posed in a more or less horizontal position. .
Again, in the case of grasses the growing points '
are located in the crown of the plant, at or
below soil level and are therefore protected by
foliage. With broad leaved weeds the growing
points at the tips of the shoots and in the leaf
axils are usually exposed.

Selective action may also be influenced by fac-
tors affecting absorption and translocation.
Reference has been made to the fact that these
processes are affected by the relative polarity of
the chemical and the tissue. Absorption of polar
substances through the roots is uninhibited but

40

there may be some selectivity due to restriction
of entry through heavily cutinised cuticles.
Variation in translocation which is closely linked
with the plant species and stage of growth can
also result in differential reaction.

Although morphological characteristics contri-
bute. the main causes of differential action are
associated with physiological differences. The
story is only partially understood, however, and
factors such as differences in enzyme systems,
response to pH changes, chemical constitution
of the plant and cell metabolism, may be
involved. These factors are affected by the age
of the plant and also by environmental condi-
tions. such as temperature and available moist-
ure. When physiological processes are slowed
down by such factors, reaction to 2.4-D is
usually less apparent. There is some systematic
correlation between synthetic hormones and
plants, e.g., members of the Cruciferae are, in the
main, susceptible, while the Polygonaceae which
includes many broad leaved types such as docks
iRumex sppj and doublegee lEmex australis)
are relatively resistant. The correlation does not
extend very far, however, as responses have
varied with different species and even different
varieties of the one species.

Agricultural Application

Although employed primarily as selective her-
bicides, 2,4-D and related compounds can be
used as contact sprays for killing annual and
checking perennial weeds, as translocated
sprays, particularly for perennials and as tem-
porary soil sterilants. Our main interest, how-
ever, is in the selective control of annual weeds
in cereal crops. It is proposed to deal with the
various factors involved in relation to develop-
ment and research in Western Australia.

Two of the most widespread and troublesome
weeds in this State are wild turnip <Brassica
tournefortii) and wild radish iRaphauus rop-
hojiistrum f , As these are both annual species
of the Cruciferae and are closely related to
Brassica smapis with which some of the original
experiments were undertaken at Jeallots Hill,
trials with them were commenced as soon as the
chemicals became available. Much of the sub-
sequent work has been done with these two plants
along with doublegee f^Emex australis).

Type aiid Quantity of Chemical

The first commercial lines were the sodium
salts of M.C.P.A. and 2,4-D. These were soon
followed by the triethanolamine salt and various
esters of 2,4-D and latterly corresponding com-
pounds of M.C.P.A. have also been formulated.
The earlier overseas reports frequently referred
to application rates exceeding one pound of acid
equivalent per acre. With high average yields
of forty bushels ov more per acre the cost of such
treatments could be absorbed but such rates were
di£5cult to i^econcile with the relatively low
cereal yields in this State. Meadly (1951) re-
corded effective control of young wild turnip
plants in a cereal crop at Wongan Hills v/ith
both two and four ounces acid equivalent per
acre, the species being particularly susceptible.
Ti'eatment increased the yield of a heavily in-
fested crop on light land from 8.5 bushels to 13.2

bushels per acre. Corresponding yields at Dal-
wallinu were 5.7 and 11.6 bushels respectively.
Other formulations have since been proved effec-
tive and four ounces acid equivalent per acre is
used for most spraying of wild turnip although,
when conditions are favourable, somewhat lower
rates have been equally satisfactory.

The story with wild radish has been somewhat
different. With the initial trials, using the
sodium salts, four ounces acid equivalent per
acre checked the plants but a proportion re-
covered and set seed. Much more satisfactory
control was obtained at the eight ounce per acre
rate. In the years immediately following, wild
radish spraying was undertaken with the amine
and sodium salt of 2.4-D and the sodium salt of
M.C.P.A. mostly at four ounces of acid equivalent
per acre but sometimes at six. During this period
a number of indifferent results caused concern
among farmers. The reason for some, but by
no means all of these could be traced and further
detailed investigations were undertaken (Meadly
1954 >. Chemicals used were the sodium salt of
M.C.P.A. and the amine, ethyl ester and butoxy-
ethanol ester of 2,4-D. All were applied at both
four and six ounces of acid equivalent in five gal-
lons of water per acre. The four ounces of
M.C.P.A. and 2.4-D amine affected a proportion
of the wild radish plants but did not give a
satisfactory degree of control. The six ounce
rate of both gave practical control although some
plants recovered and set seed. There was no
difference betv/een the two types of ester but the
results with four ounces of acid equivalent per
acre were somewhat better than those with six
ounces of amine and M.C.P.A. The six ounce
rate of both esters gave complete control of wild
radish. Based on this work the ester is now being
used largely by farmers for this weed with in-
creased uniformity of results. It is not usually
recommended, however, when the cereal is under-
sown with a legume and this aspect of crop
tolerance will be discussed later.

The habit of doublegee and the fact that it
is an annual gave reason to believe that it
would be susceptible to moderate rates of ap-
plication of 2,4-D. This plant has belied its
appearance, however, and much research o\'er
a period of years has failed to provide a really
effective treatment. Meadly and Pearce (1954.
1955) record trials with various formulations of
2,4-D at a wide range of acid equivalent levels.
Results were variable and no treatment could
be relied upon to give a high degree of control.
With heavily infested crops six ounces acid
equivalent of 2,4-D ester per acre has suppressed
the growth of young doublegees sufficiently to
warrant its application but has not prevented
seed formation. A second treatment two weeks
after the first, even with the lower rates, gave
a total kill of doublegees in one trial at Gerald-
ton in 1954. but the same double treatment at
Beverley was not effective.

Volume of Application

When hormone-like chemicals were first ap-
plied for weed control the volumes used for
other herbicides were followed and one hundred
gallons or more of solution were applied per
acre. Such volumes, although inconvenient, are

41

acceptable for highly productive crops but
would present a problem under the conditions
applying in this State where large areas re-
quire treatment and suitable water is not
always readily available. It is simple mathe-
matics to compute that 500 acres, by no means
an excessive area for one property, would re-
quire 50,000 gallons of water if sprayed at the
rate of 100 gallons per acre. This quantity
must not only be supplied but transported to
the site of operations.

The agriculturist and engineer soon joined
forces and overcame this major difficulty by
constructing nozzles which, when fitted to
booms were capable of an even and effective
application with as little as four or five gallons
per acre. Later developments include nozzles
having spinning discs and rotating brushes and
also various types of atomisers. Most of the
cereal crop spraying in Western Australia is
now being undertaken with six to eight gallons
per acre, using booms of 30 feet or more in
length and travelling at about four miles per
hour. An average coverage is 15 acres per hour.

A more recent development is the application
of 2, 4-D by means of aircraft. With Tiger
Moths, the main type in use in this State, the
tank capacity is approximately 30 gallons and
therefore the rate used with low-volume booms
is scarcely practicable, as it would entail land-
ing and refilling after four or five acres of
spraying. A specially designed nozzle, however,
has made possible a satisfactory application of
two gallons per acre enabling 15 acres to be
treated with each flight. Much work has been
involved in calibrating aircraft, including de-
termining the effective spraying swathe and
ensuring an even dispersal of an adequate num-
ber of spray droplets.

Droplet Size

When considering droplet size two important
factors are involved. The spray must be in
such a form that it actually reaches the plant
under average conditions of wind, temperature
and humidity and upon making contact per-
mits absorption of the chemical by the plant.
Fisher et al. (1952 > reported that coarse drop-
lets estimated to be 450-500 microns gave some-
Vv^hat higher kills of mesquite (Prosopis juliftoj'a)
than either fine or medium-coarse droplets, both
of which gave comparable results. Other
workers, however, including Mullison (1953)
consider that the herbicidal effect is controlled
by the amount of 2.4-D applied rather than
the number or size of the droplets. All are
agreed that the droplet size must be such that
it does not evaporate or be dispersed by wind
between boom and weed.

The amount of potential drift depends on
wind velocity and droplet size. A still atmo-
sphere is the ideal but in practice it is neces-
sary to operate with a low velocity wind. With
ground machines the effect can be minimised
by having the boom as near the target as pos-
sible. In order to obtain complete coverage,
however, the distance is seldom less than one
foot and with some nozzles may be two feet.
With aircraft operating in Western Australia
the boom is usually six to ten feet above the crop.

Maximum wind velocities suggested are 12 miles
per hour for ground machines and seven to
eight miles per hour for aix'craft.

An interesting chart published by Edwards
and Ripper (1953> shows the relation between
droplet size and distance of drift under differ-
ent conditions. Droplets having a mass medium
diameter of 50 microns while settling 10 feet
will drift 500 feet in a 10 m.p.h. breeze, and 200
feet in a 3 m.p.h. breeze. The corresponding
figures for droplets having a m.m.d. of 400
microns were 30 feet and less than 10 feet. The
need for a large droplet size with aircraft ap-
plication is obvious but the m.m.d. is limited by
the economic application rate and the likelihood
of missing some of the weeds if the figure is
too high. With aerial spraying the tendency is
to produce coarse droplets with m.m.d. about
300-450 microns while with low volume ground
units the figure usually varies from 90-120
microns.

Growth Stage of Plant

Plants, in general, are most vulnerable to
hormone-like substances when at the seedling
stage. Later, reaction may be closely correlated
with the rate of growth. Plants checked by
adverse conditions such as low' temperatures and
lack of moisture tend to be more resistant and
difficult to kill. This would appear to be due
to the greater susceptibility of active meriste-
matic tissue.

The greatest benefit to cereal yields can also
be expected from early removal of weed com-
petition. It is necessary to ensure, however, that
a more or less complete emergence of the target
W’eed has occurred before spraying and also
that the crop is at a tolerant growth stage.

Weather Conditions

As already discussed, strong winds are very
undesirable as they accentuate spray drift, par-
ticularly wdth aircraft application. An addi-
tional hazard to nearby susceptible crops is
presented by nozzle drooling when ferrying or
turning near them. This has been reduced by
the use of individual cutouts on each boom
nozzle, at times supplemented by suction by
means of a Venturi device w'hen the main valve
is closed. In America the use of volatile esters
is not favoured because of the risk of damaging
susceptible crops, particularly cotton. In West-
ern Australia, howxver, cases of injury to crops
such as tomatoes and lupins have occurred with
the relatively low'-volatile preparations and have
been due to didft during operations rather than
subsequent volatilisation and movement. In
America, with aircraft spraying, some particles
have been carried by air currents for distances
up to four miles while in this State lupins more
than one mile aw’ay from spraying operations
have been damaged severely.

The most adverse combination for aerial
spraying is a high wind velocity and tempera-
ture associated with low^ humidity. Besides the
drift factor, significant evaporation of droplets
can take place during passage from the boom
to the herbage. On the other hand higher
temperatures favour growth and tend to make
the weeds more susceptible to the chemical.

42

During the winter, mild conditions are more
likely to favour than detract from results.
Although fine weather is recommended, rain a
few hours after treatment, particularly when
using 2.4-D ester, is unlikely to affect results.

Tolerance of Crops

Although the selective action of the synthetic
hormone-like substances is one of their out-
standing characteristics, reference has already
been made to the high susceptibility of certain
crops and there is by no means complete toler-
ance in the case of many sprayed extensively for
weed control. In this State we are concerned
mainly with cereals and a considerable amount
of research has been undertaken, to define a
safe level of treatment. Although there is some
variation in the results reported it is possible to
make conclusions of practical value.

Cereals are likely to be affected if sprayed at
the seedling stage but develop resistance when
tillering or stooling has progressed. Treatment
during this early danger period can result in a
number of malformations and growth disturb-
ances. Wheat may develop club shaped, twisted
and branched ears with an irregular arrange-
ment of spikelets. The glumes may become
fused, the number of spikelets in each group
reduced and a proportion of the florets aborted.
Symptoms also include thickened culm.s, chlor-
osis, reduction in height and delayed maturity.

SuscepMbility again increases at the “boot”
stage when the head is enclosed in the leaf
sheath. Risk of damage continues during pol-
lination but decreases at the “milk" and soft
dough stages. Late spraying does not usually
cause abnormalities but reduced grain setting
has been attributed to it. Workers are in gen-
eral agreement that wheat is least and oats
most susceptible with barley occupying an inter-
mediate position. There is also variation in
reaction to formulations. The ester of 2.4-D
is likely to be most severe, with the sodium salt
of M.C.P.A. least harmful.

During 1954. at the Merredin Research
Station three wheat varieties — Bungulla, Ben-
cubbin and Wongoondy^ — were sprayed at the
pretillering, early tillering and “boot” stages
with eight ounces acid equivalent per acre of
2,4-D amine, 2.4-D ester and sodium M.C.P.A.
Yields of grain showed no significant difference
due to variety or growth stage although there
were indications of reduced yields with the pre-
tillering and “boot” stage treatments. This tend-
ency was more noticeable following the later
spraying. With research in other countries cereal
damage is usually associated with application
rates of one pound or more of acid equivalent
per acre. Eight ounces were used in the Mer-
redin trial as higher rates are seldom applied
for weed control in cereal crops in Western
Australia. There is little doubt that external
factors including climatic and soil conditions,
by virtue of their influence on the growth pro-
cesses of cereals can influence results. Some

abnormalities have followed field spraying at
four ounces acid equivalent per acre of 2,4-D
amine but only a small proportion of the plants
have been affected.

In a country so dependent on subterranean
clover iTrifolnnn subterraneuvi) as a legume it
is important that information should be avail-
able concerning its reaction to the various
formulations, especially when first sown along
with a cereal. Although a considerable amount
of w^ork on clover tolerance has been published
little of it has dealt specifically with subter-
ranean clover and the results show a decided
difference in reaction between species. Again
most of the investigations have been under-
taken with pastures where clovers tend to be
more resistant than when growing in association
with a cereal crop.

Some exploratory work undertaken in 1952
on a pasture dominated by the Dwalganup
strain of subterranean clover indicated that this
species, except at the small seedling stage, is not
readily killed by application of 2.4-D. many
plants persisting at the 2 lb. per acre acid
equivalent rate of M.C.P.A. and Ik lb. per acre
acid equivalent rate of 2.4-D amine. It was
apparent, however, that even some of the lighter
treatments retarded the date of flowering. This
could be vital in districts having a short grow-
ing season as it would be equivalent to convert-
ing Dwalganup into a later strain from a matu-
rity viewpoint and might prevent effective seed
setting. Any reaction of this nature would be
of greatest significance in the case of spraying
to control weeds in the year of clover establish-
ment. Few. if any. dormant seeds would be
present in the soil and regeneration of the clover
could only be obtained from seed formed in
the season of sowing. As subterranean clover is
frequently established along with a cereal crop
which would normally be sprayed for weed con-
trol the reaction of the clover to chemicals under
such conditions is most important.

An experiment w^as designed to ascertain the
effects on seed setting of different formulations
applied at various rates to Dw'alganup subter-
ranean clover sown along w'ith a crop. It \vas
intended to carry out the experiment at a num-
ber of different centres but conditions during
1954 restricted the work to one property in the
Toodyay district on land being cropped for the
first time. The clover was sown at 8 pounds
per acre with a light cover crop of wheat.

The sodium salt of M.C.P.A. and the amine
salt and ethyl ester of 2,4-D were applied to the
various plots as shown in the table. The
required amount of chemical was dissolved in
water and applied at 5 gallons per acre through
a low-volume boom. At the time of spraying
the crop was approximately 9 inches tall and
stooling freely while the clover was at the 3
to 4 leaf stage. In order to estimate the effects
of the treatment 12 plant counts of one square
link were made on each plot. The position of
each count was marked and at the end of the
season the seed was removed from the individual
areas. In this way the seed set by over 2,500
plants was recorded.

43

Table Showing Seed Formation.

Treatment

Ozs. acid

No. of seeds

equivalent

set per

per acre

plant

M.C.P.A. (sodium salt)

4

6.6

8

8.5

12

7.0

16

7.4

24

3.2

!!

32

4.9

2,4-D/amme (triethanolamine)

4

5.7

,,

8

4.3

12

2.1

\\

16

1.0

24

0.8

32

0.6

2,4-D/ester (ethyl)

4

3.1

,,

8

3.0

12

1.4

,,

16

0.9

,,

24

1.1

32

1.0

Control ...

—

8.7

Significant difference

between

treatments

(P 0.05) 3.35.

The results are summarised in the table and
illustrated by the graph. With M.C.P.A. the
yield of seed per plant was significantly de-
pressed for all treatments above the 16oz. level.
In the case of 2,4-D amine the depression
occurred with all treatments above the 4oz. level,
while with 2,4-D ester all treatments caused a
significant reduction in seed yield.

The results have considerable practical value
and have made possible tentative recommenda-
tions regarding the spraying of cereal crops with
undersown clovers (Meadly and Pearce 1955).
Further similar trials during 1955 at four dif-
ferent centres should complete the information
required and allow details regarding formula-
tions and Quantities to be defined.

Other Weeds

The growth regulators have proved very useful
for the destruction of other weeds besides those
controlled selectively in cereal crops. Cape tulip
(Hovieria hreyniana and H. 77iiniata) reacts to
relatively high rates of application, corm and
cormil development being arrested. The most
effective formulation, optimum economic dosage
and most vulnerable growdh stage have been de-
fined by Meadly (1954). The ester of 2.4-D has
given best results and the most favourable time
for spraying is somewhat later for H. breyniana
than for H. ininiata. A number of other trouble-
some herbaceous weeds are also susceptible.

Among woody plants blackberry iRubxts jru~
Ucosus) has been treated successfully with the
ester of 2,4, 5-T. No advantage has been gained
by using a mixture with 2,4-D ester. Time of
application is significant, treatment at the
flowering and fruiting stage in mid-summer
being most effective. This is assumed to cor-
respond with active downward movement of food
materials in the phloem facilitating translocation
and deep penetration by the 2,4.5-T. Although
foliage spraying of mesquite (Prosopis julifiora),
including aerial applications, have not proved
effective good results have followed treatment of
the basal portion of the stump with a 1% acid
equivalent solution of 2. 4, 5-T in distillate.

Some work has also been undertaken with
York Road Poison ^Gastrolobium calycinuvi)
and regrowth of various EucalypUis species.
Results have been very variable and relatively
high rates of application are necessary. This
method of treatment is not economical for large,
dense areas but may prove advantageous under

"9

8

7

6

Number
of Seeds ^
per plant

2

1

Lo

Keu

M.C.PA.

2,4 -D/ester.
2,4 -D/awine.

-H

-j-

J

‘

1

-^1

■

4 .

**

1

.

1

1

•-J

■*

1

-r — 1

1

+.

“14

1

:

.

• T

,

t

1

'

- J

1

4-

-

;

■

'

4

8 12 16 20 24 28 32 ;

Ounces acid-ec^uivalent per

44

acre.

some conditions such as with poison plants grow-
ing in rocky situations where it is difficult to
undertake other control measures. In some cases
2,4.5-T has proved more effective and no more
costly than bashing for controlling Eucalyptus
suckers. With GastroLOhium species best results
have followed the application of 3-4 lb. acid
equivalent in 150 gallons of water per acre of
vegetation when the plants are flowering in the
Spring. With eucalypts autumn treatment has
proved more effective.

Conclusion

During the relatively short period that syn-
thetic growth regulating chemicals have been
available rapid and spectacular advances have
been made. Extensive research has been under-
taken in many countries with a natural tendency
for the concentration to be on applied rather
than fundamental aspects. The need for more
basic physiological information is now being felt
and this field is receiving more attention.

The use of these chemicals for weed control
alone has resulted in enormous savings to the
agricultural industry. Figures for Western Aus-
tralia, although not on the same scale as those
for Canada and the United States, underline this
statement. In a period of three years the area
sprayed for the control of weeds in cereals in-
creased from experimental proportions to an
estimated total of 400,000 acres in one season.
Based on a conservative estimate of increase in
yield due to treatment, the additional annual
production approaches one million bushels. Ex-
perience in this State gives support to the state-
ment of an eminent scientist that the discovery
of the herbicidal properties of growth regulating
substances is the greatest scientific contribution
to agriculture since superphosphate fertilizer.

It is likely that further researches by the
chemist and agriculturist will result in improved
formulations that will increase the effectiveness
of present work and extend the range of species
that can be controlled chemically.

References

Bonner, J. and Bandurski, R. S., 1952. Studies
of the Physiology, Pharmacology and Bio-
chemistry of the Auxins Annu. Rev. PI.
Physiol. 3: 59.

Boysen-Jensen, P., 1913. Ber. dtsch. hot. Ges.
31: 559.

Crafts, A. S., 1948. Science 108: 85.

Crafts, A. S., 1952. Symp. Rep. Amer. Soc. PL
Physiol. Meeting Sept. 8-10, 1952.

Darwin, C. and P., 1880. “The Power of
movement in plants.” (John Murray:
London.)

Day, B. E., 1950. Absorption and translocation
o f 2,4-Dichlorophenoxy-acetic acid by
bean plants. Ph. D. Thesis, Univ. of Calif.
Edwards, C. J. and Ripper, W. E.. 1953. Droplet
size, rates of application and the avoidance
of spray-drift, 1953. Proc. Brit. Weed
Cont. Conf.

Fisher, C. E. and Phillips, W. M., 1952. Ninth
Res, Rep. Nth. Cent. Weed Cont. Conf.: 70.
Fitting, H.. 1909. Z. Bot. 1: 1.

Johnson, M. and Fults. J. L. 1952. The relation
of scopoletin to the herbicidal action of 2,
4-D. Thirteenth West. Weed Cont. Conf.
Res. Prop. Rep.

Jones, L. T., 1953. Private Communication.
Kelly, Sally and G. S. Avery Jnr., 1949. The
effect of 2, 4-D acid and other physio-
logically active substances on repiration.
Avier. J. Bot. 36: 421.

Laibach, F. and Maschmann. E. 1933. J. tviss.
Bot. 78: 39.

Meadly, G. R. W.. 1951. Large scale chemical
weed control in crops. J. Dept. Agric. W.
Aust. (Second series) 28: 56.

Meadly, G. R. W., 1954. The chemical control
of Wild Radish. J. Dept. Agric. W. Aust.
'Third series) 3: 309.

Meadly, G. R. W., 1954. Weeds of Western Aus-
tralia — Cape Tulip. J. Dept. Agric. W. Aust.
(Third series) 3: 289,

Meadly, G. R. W. and Pearce, G. A., 1954. Some
chemical trials with doublegee. J. Dept.
Agric. W. Aust. (Third series) 3: 427.
Meadly, G. R. W. and Pearce, G. A., 1955. A
further chemical trial with doublegee. J.
Dept. Agric. W. Aust. (Third series) 4: 229.
Meadly, G. R. W. and Pearce, G. A., 1955. The
effect of hormone-like herbicides on
Dwalganup subteiTanean clover. J. Dept.
Agric. W. Aust. (Third series) 4: 2.

Mullison, W. R., 1953. Down to Earth 9: 11-13.
Nielsen. N., 1928. Planta 6: 376.

Paal, A., 1919. J. wiss. Bot. 56: 406.

Slade. R. E., Templeman, W. G. and Sexton,
W. A., 1945. Nature 155: 497.

Van Overbeek. J., 1947. Use of synthetic

hormones as weedkillers in tropical agri-
culture. Amer. J. Bot. 36: 421.

Went, F. W., 1926. Proc. Acad. Sci. Amst. 30:

10 .

45

i

6. — Plant Ecology of the Coastal Islands near Fremantle, W.A.

By W. M. McArthur'"

Manuscript accepted — 17th March,

The plant communities of the major coastal
islands near Fremantle. W.A,, have been
described in relation to the soils, climate and
geomorphology and a comparison drawn between
the island vegetation and that of the adjacent
mainland.

The vegetation, which is structurally a Scrub
Formation, is composed of three well defined
sub-formations, viz., tall scrub, scrub and low
scrub and several other sub-climax communities
of doubtful status.

The island vegetation is of special interest, not
only for its density of cover and high frequency
of dominants, but also for the paucity of plants
of the Myrtaceae. Papilionaceae and Epacridaceae
and the complete absence of members of the
family Proteaceae.

It has been shown that within one locality
the distribution of species Is dependent on
habitat and for comparable habitats on the
islands and the mainland the composition of
vegetation is substantially the same. The overall
distribution of species is explained by a late
Pleistocene land bridge between the islands and
the mainland.

Introduction

The vegetation of the coastal islands — Garden,
Rottnest and Carnac — is very unusual when com-
pared with that of the adjacent mainland. This
fact has long been recognized and commented
on both by local and visiting botanists, but apart
from notes on brief excursions no account has
been written. The object of this paper is to give
a description of the vegetation of the islands and
to discuss briefly its relationship to the environ-
mental factors. In addition, a compari.son is
made between the vegetation of the islands and
that of the coastal areas on the mainland be-
tween Rockingham and Scarborough. This has
led to a discussion of the present distribution
of species in the area and an interpretation of
the past history of the islands in the light of
this distribution. The geographic relation of
the islands to the mainland is shown in Pig. 1.

Garden Island has been chosen for a more de-
tailed study than the other islands because it
has remained almost unchanged by settlement.
Rottnest Isand has been inhabited from the early
history of the State and so has suffered drastic
changes in vegetation; and Carnac Island,
though its vegetation is unchanged, is considered
too small for any but shoreline conditions to
prevail.

Botany Department. University of Western Australia;
Present address: Division of Soils, C.S.I.R.O.,
University Grounds, Nedlands, W.A.

The historical records of these islands can be
traced back to 1658 ^Heeres 1899), when a Dutch
vessel, the Waekende Boeij, while looking for sur-
vivors of a ship which was wu*ecked further to
the north, anchored to the north of a small
island in latitude 3r47' S. Samuel Volkensen,
the captain of the vessel, writing to the Gov-
ernor of the East India Company in 1659,
wrote: —

In slightly under 32° south latitude there
is a large island at about three miles from
the mainland of the Southland. The island
has high mountains with a good deal of
brushwood and many tiiorn bushes so that

it is hard to go over Slightly more

to the southward is another small island.

There is no doubt that Volkersen referred to
Rottnest and Carnac Islands, although they were
not named until much later.

In 1696 three vessels under the command of
William de Vlamingh, while searching for a lost
vessel, arrived at Rottnest Island and Gerrit Col-
laert, captain of one of the vessels, wrote: —

On 31st December (1696) I again put on

shore with the skipper I found several

sorts of shrubs the greater part of which

were unkown to me There were also

a variety of trees and among them one sort
the wood of which had an aromatic
odour

and a few days later de Vlamingh wrote of: —

a very agreeable belt of trees, very thick

and about half a league in extent We

perceived that a very grateful odour came
from these trees.

This quotation suggests that the vegetation
of Rottnest Island has changed considerably and
in fact Callitris robusta thickets were probably
originally widespread. Garden and Carnac
Islands are apparently unchanged and although
Garden Island was studied by a zoologist from
the French ship Naturaliste, little useful infor-
mation on the vegetation can be obtained from
his journal.

46

Fig. 1.— Plan showing location of islands and bathymetry of the surrounding ocean.

47

In 1854 William Harvey visited the area and
writing of the vegetation of Rottnest Island
said: —

.... its land flora is remarkable for the
total absence of Proteaceae and Grass trees
and for the paucity of Myrtaceae, Epacrida-
ceae and Leguminosae ....

Harvey’s statement puts the problem in a nut-
shell, and it is the consideration of this problem,
accompanied by a description of the environ-
mental factors, which make up this paper.

A list of species and their distribution on the
islands is given in Appendix I. The authorities
for all species are also included in this list.

Geomorphology

The islands under discussion are parts of two
physiographically distinct ridges which lie
roughly parallel to the mainland coast between
32° and 32°30' south latitude. The easternmost
ridge passes through Murray Reefs, Penguin and
Seal Islands. Point Peron, Garden and Carnac
Islands and the Stragglei’s. The other, about
two miles further west, passes through Coventry
Reef, Hawley and Casuarina Shoals and Rott-
nest Island. A study of the bathymetry of the
region makes this feature even more marked
(Fig. 1). These ridges are composed of Pleisto-
cene to Recent aeolianite. They have been de-
scribed (Teichert, 1947b: Fairbridge 1950 > as
being the remnants of a peninsula which
extended about twenty miles from the present
coastline when the level of the sea was consider-
ably lower. With the rising of the sea the ridges
have become more and more dissected until now
there remain only four major land surfaces and
several small rocky islands. Erosion is still going
on, but only slowly. Point Peron has since been
reunited to the mainland by blown sand (Fair-
bridge, 1948).

The islands have a hard core of travertinized
calcareous dune rock which is generally taken
to be late Pleistocene in age (Fairbridge. 1948).
This rock is exposed extensively on Rottnest
and Carnac Islands while on Garden Island it
is largely blanketed by more recent sand dunes.

Exposures in the dunes on Garden Island show
the typical cross-bedding of aeolianite. The
upper 6-8 feet is still quite loose, but below this
there is incipient cementation and a few traver-
tinized root channels (Fig. 2). Analysis has
shown that the lower levels are only slightly
richer in calcium carbonate than the surface.

Garden Island can be divided into two physio-
graphically distinct areas. The first is the main
axis consisting of relatively high steep dunes
underlain by limestone and the second is the low
flat area making up Colpoys Point. It is sug-
gested that the latter was formed comparatively
recently by progressive additions of sand which
washes through South Passage. This is indicated
by the radial structure of Colpoys Point and the
smooth even curve of Careening Bay.

A feature of the coastline, both on the islands
and the mainland, is the presence of wave cut
platforms at different levels above the existing
,sea level. Fairbridge (1948) has shown that

there is a fair degree of correlation between the
raised platforms here and at Houtman’s Abrol-
hos Islands 300 miles to the north. This suggests
eustatic changes in sea level. The changes can
be roughly referred to world-wide changes since
the Pleistocene period (Zeuner, 1945. 1946). This
allows of accurate dating of physiographic events
in the area, the time of the 10 foot platform
having been put at 2,000 B.C. The sea level
apparently reached its maximum height at that
time and has since dropped, with periods of still-
stand at 5 feet and 2 feet. Prom this it can

Fig. 2.— Soil profile in a cutting through a dune.
Solution channels are forming in the lower layers.
Garden Island.

be assumed that individual islands and the main-
Isnd have not been linked by a direct land bridge
for at least 4.000 years, and that probably the
time when they were connected lies nearer
100.000 years ago. This would correspond with
the Wiirm exposure of Europe.

There is no permanent fresh water on the
islands and no evidence that there has been
any in the past. However, the army authorities
have put down two deep bores and several wells.
The bores are artesian and there is a continuous
flow of water to the sea.

Early visitors were unanimous in describing the
islands as being mountainous, the steep and
densely vegetated dunes giving an exaggerated
impression of height when viewed from a dis-
tance. However, the highest point on Garden
Island is slightly less than 200 feet, while Rott-
nest Island has a maximum elevation of 130
feet.

Climate

The area in the vicinity of Fremantle has a
typically Mediterranean climate. In other words,
there is a cool wet period and a warm dry period,
with short periods of transition.

Most of the climatic data were obtained from
the Rottnest Island weather station (see Appen-
dix II) and it has been assumed that these
records apply equally to conditions on Garden
and Carnac Islands and Point Peron.

48

Fig. 3.— Mean monthly wind Roses at Perth Weather Bureau. The length nf the arrows
represents the mean total number of miles from each direction on a scale of lin to 5 000
miles. The feathers indicate the speed of the wind on the Beaufort Scale. North is towards

the top of the page.

Fig. 4.— Resultant wind. Perth Weather Bureau. The length of the arrows represents the
resultant wind on a scale of lin. to 10,000 miles. North is towards the top of the page.

<0

It is only during the period May-August that
the west coast of Australia is subject to any
constant rain-bearing wind. Rottnest Island
receives 72% of its annual rainfall during these
months and 93% during the months of April
to October. Thus the distribution of rainfall
is more important than the yearly total. During
the winter there is a surplus while during the
summer there is a marked deficiency. As the
main rain comes when evaporation is at a mini-
mum and run off is negligible in the loose sand
most of it penetrates the soil. The figures quoted
for potential evapo-transpiration (Appendix ID
are calculated from a formula proposed by
Thornwaite (1948). It is suggested that condi-
tions in undisturbed scrub are such that the
actual water loss would be less than the figure
calculated.

The effectiveness of the rainfall was calculated
using Thornwaite’s (.193D formula, which is
regarded by Gentilli (1948) as being the most
suitable for Australian conditions. Rainfall
effectiveness increases very rapidly during the
latter part of April and decreases just as rapidly
in August. This leaves about half the year when
conditions become semi-arid or drier (less than
2.2). There is some lag, however, and it has
been shown (Speck, unpublished data) that some
soil water is still available until early December.
On Garden Island the position of the ground
w'ater table varies from 4 feet below the surface
in August to 14 feet in March.

It is suggested that plants obtain much of
their moisture during the summer months from
dew which forms every night, but which is never
measurable. Wallabies iSetonix bracliyurus and
Proternnodcn eugenii) have existed on the
islands for centuries without any reliable fresh
water supplies, suggesting that in summer the
dew must be considerable. The greater part of
Garden Island is covered wdth dense low' scrub
and this may have the effect of decreasing
evaporation from the soil surface.

The lower mean monthly temperature limit
for plant growth lies betw^een 45'' and 50°F.
Since the lowest mean temperature on the coastal
islands is 58 "F. in July it may be concluded that
temperature would, at the most, cause only a
slight retardation of growth. It is also signifi-
cant that the lowest recorded temperature is con-
siderably above freezing point. A striking feature
of the temperatures of the islands is the uni-
formity of the mean temperature. The daily
range varies from 10" in July to 13“ in
February.

Wind data for Western Australia are limited.
Perth Observatory being the only coastal station
having a continuous record of wind speed and
direction. However, it is considered that the
wind at Perth would be substantially the same
as on the coastal islands, except that sea breezes
would begin to blow' earlier each day and the
velocity of the w'esterly and south-w'esterly winds
w'ould be greater than at Perth.

It will be seen from the monthly wind roses
(Fig. 3> that for the nine months from July to
March the greater part of the wind is from the
south-w'estern sector. This is due partly to the
influence of the local winds, but also to the fact

that for some of this period the cyclonic winds
coincide with the local winds. During April the
predominant wind is from the east and May and
June are almost equally windy from all direc-
tions. The diagram was constructed from
records taken over ten years.

It is significant that the “resultant wind"
(Fig. 4). which has been calculated for each
month from the w'ind roses, reflects the control
of the local winds for the greater part of the
year. The cyclonic winds only become dominant
during the winter months. Vegetation on the
w'est coast of Garden Island is often aligned in
a south-west north-west direction showing the
predominating influence of the local winds
(Fig. 5). The fact that the resultant wind
makes a complete counter-clockwise rotation
during the year is explained by the annual
change in position of the belt of high pressure
systems.

Fig. 5. — Cliff side vegetation, shaped and prostrated by
wind action. Pt. Atwick. Garden Island.

The long axis of Garden Island is orientated
about 15“ w'est of north, while Rottnest has its
greatest extent east-west. The importance of
this becomes apparent when the prevailing
winds are considered. It means that for Garden
Island the maximum length of coastline is almost
at right angles to the wdnd for the greater part
of the year, while it is only the south-west
portion of Rottnest which is affected.

The contrasting orientation of the islands may
explain why Garden Island is largely covered
by loose blowm sand, while Rottnest has extensive
rock outcrops.

On the W'est coast of Garden Island the effect
of wind is very marked. Blow-outs and partial
burying of the vegetation (Pig. 6) are common,
and it is only those plants which can keep pace
with the banking sand which survive. On
exposed dunes and cliffs many shrubs have been
reduced to a prostrate form by the action of the
W'ind, each bush having a clipped appearance.
The effects of salt spray on this vegetation must
be considerable, especially where the waves break
on rocks and the spray is carried over the land
by W'ind.

50

Another wind effect may be seen in the
orientation of Callitris robusta clumps on the
north end of Garden Island (Fig. 7). From
available evidence it is suggested that this clump
has been spread through the seeds being blown
by the wind. Ring counts of trees from different
parts of the largest clump indicate that the

Fig. 6. — “Blowout” on the west coast of Garden Island.
Sand is rapidly encroaching on to stable vegetation.

direction of the spread has been towards the
north-east. During the summer months when
the seeds are being shed the winds are from the
south-west.

In this connection it is of interest to note that
cn the islands there are other species whose seeds
are distributed by the wind. Some of these are
Senecio lautus, Arctotheca nivea, Clematis
microphylla and Comesper7na integerrimum.

On the mainland the effect of wind has in
certain areas been increased by the action of
man. At Point Peron clearing of vegetation has
led to mobile sand which is now fairly extensive.
Attempts are being made to stabilise this sand by
the introduction of Marram grass (Ammopliila
arenariaf. Along the coast proper, especially on
the wide, sandy beaches, the sand is constantly
shifting and blow-outs are frequent.

It can be seen from the foregoing that the
wind, although ordinarily not strong enough to
uproot trees or break branches, has, by its very
constancy, perhaps a greater influence than
many of the other factors.

The effects of light, embracing intensity,
length of day and duration of sunshine, form an
important component of the environmental com-
plex but without controlled conditions it is
impossible to isolate its influence. It is reasonable
to assume that it is the low light intensity which
excludes most undergrowth from the dense
Acacia rostellijera and Callitris robusta com-
munities. On the other hand it is difficult to
say whether it is the lack of competition or
protection from the sun which allows the small
annual Bidiscus to flourish in the Callitris
robusta thickets.

The salient point to be made in this very brief
discussion of the climatic factor, is the seasonal
aspect of the weather. The rain comes in the
cooler part of the year when the rate of plant

51

growth is at a minimum with the result that
some plant groups are excluded or restricted.
The family Graviineae for instance, which
requires rainfall in the warm season, is repre-
sented by only one native species (Stipa
variabilis) .

Garden Island

The vegetation of this island is remarkable for
three reasons. Firstly, certain important families
of the mainland are either absent (Proteaceae)
or present in restricted numbers (Myrtaceae and
Papilionaceae). Secondly, the dominant species
present show an unusually high frequency, five-
sixths of the area available being covered by
dense scrub formed by Acacia rostellifera,
Callitris robusta and Melaleuca fiuegelii. The
third unusual feature of this vegetation is its
structure which is probably unique in Western
Australia. In many cases the scrub consists of a
single storey of dominants in contrast to the
three distinct stories found in the Eucalyptus
7narginata and E. goniphocephala communities
of the mainland (Speck 1952). The closed
canopy of the island ve.getation serves further to
distinguish it from any other community.

The vegetation on the island consists of two
ecologically and physiognomically distinct types:
a stable central area of dense scrub and an
unstable fringing zone. The fringing vegetation
is subject to constantly changing conditions and
this is reflected both in the structure and
composition.

Using the system of classification of Beadle and
Costin (1952) the following eight plant com-
munities are recognised on Garden Island: —

(i) Callitris rdbusta association (tall scrub)

(ii) Acacia ru^tellifera association (scrub)

(hi) Melaleuca huegelii association (scrub)

(iv) Myoporum insulare association (low
scrub )

(v) Acacia heteroclita association (low
scrub )

(vi) Pittosporuin phillyraeoides association
(low scrub)

(vii) Seasonal Communities

(vhi) Littoral Vegetation

The following discussions will apply to
relatively pure communities although it must be
stressed that in respect to the tall scrub and
scrub communities there are all combinations
from 100% of one to equal proportions of Acacia
rostellifera, Callitris robusta and Melaleuca
huegelii and often an admixture of Melaleuca
puhescens and Spyridhwi globulosim.

Description of the Communities

Callitris robusta Association. — T h i s com-
munity is restrictea mainly to a large clump
covering about a square mile in the northern
part of the island. There are however small
clumps in the Acacia scrub of the south end of
the island and on the eastern extremity of
Colpoys Point there is a mixed stand of Callitris
robusta and Melaleuca pubesceiis. The average
height of the trees in these clumps is 17-20 feet
although many individual trees in the open grow
to 35 feet. The average stem diameter is 2-3
inches and as most of the trees are 30-40 years
old it is apparent that the annual increments
are very small. In the clumps the trunks are
straight and unbranched and the canopy com-
pletely closed giving the community the
structural characteristics of a forest (Fig. 8).
The undergrowth is very limited, Pliyllanthus
calycinus being the only common undershrub.
In season, the tiny annual Didiscus pilosiis
flourishes in the Callitris community. The lack
of light is possibly sufficient to restrict most
other plants. There are isolated bushes of
Spyridiuni glohulosiim and rarely Eremophila
brownii and Leucopogon richei. There are

fTjg 8 . — Callitris robusta association, Garcien Island.
Note single Soorev structure.

occasional trees of Melaleuca pubescens and
Acacia rostellifera and these have taken on the
same form as the Callitris viz. tall and slender
with only a tuft on top. The creepers Cassytha
glabella and Coinesperma integerrimuvi are
present, but do not attain any large size. Within
the community seedling regeneration of Callitris
appears to be restricted, only a small percentage
of those germinated surviving beyond December.

As far as can be seen there is no obvious factor
limiting the distribution of Callitris. It grows
as individual trees and small clumps throughout
the Acacia scrub. It is scattered through the
Melaleuca huegelii community and it competes
on equal terms with Melaleuca pubescens. There
is evidence that the Callitris is encroaching on
the Acacia scrub. Ring counts of trees from
different parts of a clump show that the age
decreases from 40 years in the centre to 22 years
at the periphery. This indicates that the spread
of the clump has been a gradual process. It is
significant that the boundary between these two
communities is quite sharp. There is seldom any
merging one into the other.

Fig. 9. — Acacia rostellifera association showing the dense,
uniform canopy. Stipa variabilis and Didisciis coeruJeus
are shown at the side of the track. Garden Island.

A feature of the Callitris scrub is the thick
layer of leaf litter. There is commonly 2-3
inches of undecomposed leaves on the surface
below which is about 1-2 inches of decomposing
leaves. The soil profile below this is as follows: —

0- lin. Dark grey calcareous sand rich in

organic matter

1- 6in. Grey-brown calcareous sand

6-45in. Light yellowish grey calcareous sand

becoming very light grey with
depth

Table I shows that the nutrient status of the
soil is very high for natural conditions.

Acacia rostellifera Association. — T his is
the most widespread community on the island.
Superficially the impression is given of a pure
stand, but Callitris occurs throughout and
Melaleuca huegelii forms an integral part.
Melaleuca pubescens also appears sporadically,
while Spyridium globulosum, which often attains
a height of 8 feet, must be considered as a
co-dominant.

52

The Acacia rostellifera scrub is very uniform
in height (8-10 feet) and extremely dense and
tangled (Fig. 9). Other species of Acacia are
present, e.g.. Acacia cyaiiophylla and Acacia
hetcToclita. The former, while of rare occur-
rence. is spread throughout the association, while
the latter is mainly present in the vicinity of
the mobile dunes.

Undergrowth is usually present even though
restricted in many parts to Stipa variabilis.
Pliyllanthus calycinus and Acanthocarpus pre~
issii occur in the more open areas, Leucopogon
spp. and Diplolo.ena dampieri grow where the
scrub thins out to separate bushes. Guichenotia
ledifolia and Lasiopetaluvi aiigiLstifolium occur
towards the northern end of the island in less
dense areas, and Didiscus coeruleus (Rottnest
Daisy) grows profusely where the scrub has been
cleared.

Acacia rostellifera scrub shows a remarkable
ability to regenerate vegetatively, exposed or
damaged roots giving rise to new bushes. Many
roads made by the Army during the period 1939-
1945 are now completely over-grown and the
scrub constantly encroaches on any cleared area.
It appears to be a very vigorous community.

The soil is brown and powdery below the layer
of decomposing leaves and at 2 feet is very light
grey calcareous sand (Fig. 2). The important
point about this soli is that it supports a very
dense vigorous vegetation suggesting that the
nutrient status is high. Chemical analyses (Table
I) show a very high nitrogen level which may
be due to symbiotic nitrogen fixation. No clear-
cut evidence has been obtained to indicate nodu-
lation here. It may. however, be noted that
Wilson (1939» in U.S.A. has shown that some
Acacias do possess Rhizohia as do also some
species in Queensland (McKnight 1949). Root
nodules have been noted on several species of
Acacia found in the vicinity of Perth (Parker,
personal communication).

Melaleuca huegelii Association . — This associ-
ation is confined to several small areas to
the north of Sulphur Jetty near the eastern
coast of Garden Island where it is developed on
low lying areas. The canopy is very uniform and
its density is increased by the creepers, Cossytha
glaheda, Clematis microphylla, Coviesjiervia in-
tegerrimum, Rhagodia baccata and Hardenbergia
comptoniana, which flourish here. There is a
well defined under-storey of shrubs consisting of
Eremophila brownii, Phyllanthus calycinus,
Acanthocarpus preissii and Leucopogon richei.
Below this is a thick mat of moss and in spring
the annuals Didiscus pilosus and Parietaria
debiiis.

The soil has a fairly high concentration of
organic matter below a thin layer of decomposing
leaves, but at 2 feet is very light grey sand as
in the other communities.

Myoporum insulare Association . — This com-
munity consists of dense .scrub about 6 feet
in height and is confined to the southern end
of the island where it covers an area of a few
acres. Associated with this community are Aca-
cia cyclopis, Rhagodia baccata and Solanum
simile. Undergrowth species are Eremophila
brownii and more rarely Solanum nigrum.

The soil is a very shallow dark grey sand over-
lying limestone.

Acacia heteroclita Association. — Acacia hetero-
clita is the main component of a stand of low
scrub restricted to the limestone cliffs near
Point Atwick and northward. It stretches in-
land about a quarter of mile where it merges
very gradually into the climax island vegetation.

The main associated plants are Acacia rostel-
lifera, Spyridium globuiosum, Diplolaena dam-
pieri, Leucopogon spp., Melaleuca pubescens and
Exocarpus aphylla. Below this storey there is
a dense growth consisting mainly of Lepido-
sperma squamatum and less commonly of Stipa
variabilis, Phyilanthus calycinus and Acantho-
carpus preissii. Westringia ridiga, Dodonaea
aptera, Alyxia buxifolia, Acrotriche ovali folia
and Melaleuca pubescens also occur near the cliff
edge, these bushes often being prostrated and
elongated by wind action.

Figg. 10.— Acacia heteroclita association developed on
shallow soil. Garden Island.

The soil is seldom more than a foot in depth
and is fairly rich in organic matter (Fig. 10).
This soil is similar to that on the south end of
the island and chemical properties would be
comparable.

Pittosporum phillyraeoides Association.— ear
the south-west corner of Garden Island several
small areas of Pittosporura scrub occur. They
consist of thin straggly bushes about 4 feet 6
inches high with the foliage completely re-
stricted to a small tuft on top. The canopy,
though only 4-5 feet above the ground is very
dense and completely closed. There is very little
undergrowth and the soil, below a thick mat
of leaves, consists of light brown gritty cal-
careous sand which becomes gradually lighter
in colour with depth.

Seasonal Connminities . — T h e s e communities
are of considerable importance since they are
the feeding grounds of the Garden Island Wal-
laby iProtemnodon eugenii). On the northern
tip of the island the scrub gives way to an open
area covered by Asphodelus fistulosus, Antheri-
cum divaricaturn and Stipa variabilis. The two
former components die down to ground level
in the summer months and then regenerate

53

in the spring. There seems to be a definite
association between these plants and the intro-
duced white land snails (Bothrievibrium bulla),
which congregate thickly on the stems without
damaging the plant.

The second seasonal community, a mixed
annual meadow occurs on the south end of the
island and is developed on shallow dark grey
sand overlying limestone. Some chemical pro-
perties of this soil are shown in Table I. The
annual vegetation is rarely more than 2 inches
high and is composed of Euphorbia di'uminondii,
Geraniuvi pilosum, Erodiuvi cicutariuni and Poa
spp. In the late winter it forms a thick mat on
the area between the Acacia scrub and the cliffs
on the southern end of the island. The best
development is in the valleys protected from the
wind and salt spray. Other common plants here
are AiiagalUs arvensis, Tripteris clandestiiia.
SolamiiJt simile and 5. nigrum.

On the tops of the rocky headlands the
vegetation is subject to salt water spray and
possibly for this reason the annuals are almost
absent. The more salt resistant plants Carpo-
brotus aequilaterus, Stipa variabilis and Scripus
nodosus make up most of the ground cover and
even on these the effects of the salt are shown
by necrotic spots. The shallow soil contains
many foecal pellets which have been identified
as those of insects, probably the larvae of a
scarab beetle (Mr. J. Callaby, personal communi-
cations).

Littoral Vegetation. — This community is
developed on the beaches, cliffs and partially
fixed dunes bordering the island. The environ-
mental conditions here are extreme — conditions
which are generally recognised as being very
unfavourable to plant growth. These include
sand blast, desiccation, salt spray, intense light
and highly calcareous sand as soil. It is under-
standable then that only specialised plants,
characterised by waxy, hirsute and succulent
leaves, can survive.

The limits of this zone are very poorly defined.
It grades imperceptibly into the scrub. However,
within the zone two fairly distinct habitats
occur — the rocky cliffs and talus slopes and the
sandy beaches.

Fig. 11. — Cliff side vegetation. Garden Island.

54

Generally the succulent plants are found on
the rocky cliffs. Most common are Carpobrotus
aequilaterus (Fig. ID, Tetragonia implexicoma,
T. zeheri, Wilsonia backhousii and Nitraria
sclioberi. Where the cliff is not so steep other
plants such as Westringia rigida, Olearia
axillaris, Alyxia buxifolia and Boronia alata
occur. Where the soil is deeper and a little more
mature Lepidospernia gladiatum and ^cirpys
nodosus appear. In this habitat it seems that
the wind is the main factor limiting species.

The sandy beaches have different environ-
mental conditions. Salt spray is not so severe
but light is very intense. Plants which can
withstand sand blast such as Calocephalus
hrownii, Angianthus cunninghamii, Arctotheca
nivea and Spinifex hirsutus are the most com-
mon. Here is opportunity for primary succession
and on most beaches there is at least a rough
zonation of vegetation.

Discussion of the Soil Properties

The soils of Garden Island are comparable to
those of a considerable portion of the west coast
of Western Australia and for this reason further
discussion is warranted.

Smith (1951), referring to coastal soils in the
Margaret River area and Speck (1952) working
on similar soils in the vicinity of Perth, have
emphasised the high content of calcium car-
bonate and the consequent alkaline reaction and
high loss on ignition.

However, there has been little other recent
work and so the present investigation, while not
by any means complete, constitutes the first
detailed examination of these soils. This work
has the further advantage that the soils have
been studied under conditions in which the soil-
vegetation balance is practically undisturbed.

Table I shows the results of analyses for the
major elements, the methods employed being
those described by Piper (1947). Figures for
phosphorus have been omitted because of lack
of agreement between determinations using
different analytical methods. Indications were
that the surface horizons contained from 0.10
to 0.15% total phosphorus. This high figure
may be due in part to calcium phosphate which
occurs in some shells and foraminiferal tests.

The most important single factor affecting the
soils is the extremely high content of calcium
carbonate which commonly makes up 85% of
the solum. This means that the reaction is
always strongly alkaline and. in fact, the pH is
below 8.0 only in the upper organic horizons.
The high carbonate content serves to explain
the discrepancy between the figures for sodium
chloride and total soluble salts, the former
usually being about 60-70% of the latter. In this
case it is assumed that the total soluble salts
include a considerable proportion of bicarbonate
ions.

Chemical analyses show an unusually high
fertility under virgin conditions in Western
Australia. More important, it can be seen that
the several communities have significantly

t

different levels, the soil supporting the Acacia
rostellifera association having an extremely high
level. The high nitrogen content in this soil
suggests that Acacia rostellifera has nitrogen
fixing properties.

A remarkable feature of these soils is the Ano
and An horizons which commonly have a depth
cf 4 inches. The development and maintenance
of these horizons is probably largely due to pro-
tection from fires since they are lacking on
Rottnest Island where fires have been common.
It is suggested also that the protection from
sun and wind afforded by the dense canopy may
keep surface conditions unfavourable for decom-
position. Table I suggests that oxidation of the
Ao horizon has not proceeded very far. It is
evident that the bulk of the nutrients are in and
immediately below this horizon.

Costing (1954) and Pidgeon (1950) have
showm that the salinity of coastal soils is not as
high as would be expected and suggest that
soluble salts are removed rapidly, especially
during the wet season. Pidgeon further suggests
that it is the content of organic matter in the
soil which determines the amount of soluble salts
which can be retained. In the Garden Island
soils, which were sampled after the main winter
rains, the relation between these two components
is very marked. Analysis of the results shown in
Table I give a coefficient of 0.9425 (p <0.001) for
organic carbon wdth soluble salts and 0.9417
(p <0.001) for nitrogen with soluble salts. Thus,
taking the organic carbon and nitrogen content
as a measure of the soil organic matter, there is
a highly significant correlation. This fact needs
to be considered in relation to the perennial
beach plants, which are probably not halophytic
even though they can .survive saline conditions
for short periods. The annual plants such as
Cakile viarifima and Arctotheca nivea are only
present on the beaches during the winter when
soluble salts would be quickly leached out.

It is of interest to speculate on the rate of soil
formation on the calcareous dunes in Western
Australia. Burges and Di-over (1953) studied the
rate of podzol development in beach dunes in
New South Wales and concluded that a humus
podzol had formed from dunes containing about
3.0% calcium carbonate in about 3.000 years.
In the dunes under consideration here the
content of calcium carbonate is in the vicinity
of 80% and the time required for the soil to
reach equilibrium would be much greater. Re-
cent developments in radio-carbon dating of car-
bonates may enable the relative ages of the
coastal dunes and limestone deposits to be estab-
lished. In a highly calcareous soil a considerable
proportion of the calcium carbonate could be
leached away without a significant change in the
apparent calcium carbonate content. It is evi-
dent that the proportion of calcium carbonate
remaining to the apparent total weight of soil
would remain fairly constant. For example a
soil containing 90% calcium carbonate could lose
50% of its original weight and still show an
apparent calcium carbonate content of 80%.
This argument can be continued until the inter-
stital spaces become filled with the insoluble
silica grains.

It seems probable that under the prevailing
climatic conditions calcium carbonate would nou
remain in the solum, but would be leached out
and redeposited in the vicinity of the ground
water table leaving the upper part of. the soil
neutral or slightly acid in reaction. Such is the
case in the Tuart {Eucalyptus gomphocephala^
zone of the mainland where there is often ten
feet of yellow sand overlying limestone (Speck,
1952), In many places along the mainland coast
limestone occurs very near to the surface and
it is assumed that this represents a truncated
profile, the upper horizons of sand having been
removed by wind. This will explain the crust
of travertine which occurs so commonly.

Rottnest Island

An account of the overall geology and physio-
graphy of the islands has already been given,
but further detail is necessary for Rottnest. For
the most part it is composed of solid limestone
in contrast with the unconsolidated dunes of
Garden Island. Another feature is the low topo-
graphical level. The highest point is about 130'
above sea level and, more important from the
point of view of vegetation there are swamps and
permanent lakes which provide completely dif-
ferent environmental conditions.

The relative proportions of species on the
island has changed considerably since white man
first landed on the island. This has been brought
about by burning and clearing and to a lesser
extent by the introduction of exotic species.
However, the natural vegetation is more affected
by the accidentally introduced species than by
those intentionally introduced. Asphodelus fis-
tulosus and Ajithericum divaricatuvi which came
originally from the Mediterranean, and Euphor-
hia peplis have spread all over the island. There
are also several introduced grasses, e.g., Poa aus~
trails and Polypogon maritimus.

The present vegetation can be divided into
several well defined communities and these will
be discussed as follows: —

(i) MelaZewca piibescens association (forest)

(ii) Acaem rosfeni/era association (scrub)

(iii) Templetonia retusa association (scrub)

(iv) Acacia cuneata association (low scrub)

(v* Sti^a-Acanthocarpus community

(vi) Halophytic communities

(vii) Littoral vegetation

Description of the Communities

Melaleuca pubescens Association . — This tree
forms pure stands of considerable area on
Rottnest Island in contrast to Garden Island
where only small clumps occur. It seems in
some places that Melaleuca pubescens is the cli-
max vegetation following a succevssion from
swamps and lakes (Fig. 12) since on the higher
ground surrounding these there is often a zone
of Melaleuca pubescens. However, this tree also
grows quite commonly on soils developed over
limestone.

55

This community has a distinctive appearance
with its clean straight boles and dense rounded
crowns. In typically developed areas the trees
are 30 — 35 feet high and have a dense interlock-
ing canopy. Undergrowth is restricted to a
sparse growth of Stipa variabilis.

Fig. 12. — Melaleuca pubescens association bordering
Heischell LuKe. Rottnest Isiaud.

Acacia rostellifera Association. — This scrub
varies in height from 8 to 15 feet and has a
completely closed canopy. It is not now very
widespread on the island, but appears to be a
remnant of some former much lai'ger community.
It often occurs in the valleys but on higher
ground around the lakes it appears to be a stage
in the development of a climax.

There is evidence that Stipa-Acanthocarpus
community is spreading at the expense of this
association. Where the Acacia scrub has been
burned or cleared the grass takes over before
the Acacia can regenerate.

Templetonia retusa Association.— On Rottnest
Island this association covers quite extensive
areas mainly in the eastern sector and as on
the mainland it Is restricted to areas where
limestone is at or near the surface (Fig. 13)
The scrub is typically 6—8 feet in height and is
fairly dense. Undergrowth is quite considerable
and consists mainly of Stipa variabilis and
Senecio lauius.

Fig. 13. — Templetonia retusa association growing on
shallow soil over limestone, Rottnest Island.

Acacia cuneata Association. — This low scrub
seldom exceeds four feet in height and is very
dense. All the bushes are rounded and the
branches reach right down to the soil surface
allowing only restricted development of under-
growth.

The association is best developed on the west-
ern end of the island near the Neck where it
glows on recently vegetated sand dunes. Al-
though Acacia cuneata is the most abundant
component, Olearia axillaris and Westringia
rigida bushes together make up about 50% of the
cover, Scaevola spp. also occur, especially near
the shore. From the shore line there is fairly
well defined zonation of vegetation. Spinifex
longifolia passes into Olearia axillaris which
then merges into the A. cuneata scrub.

Stipa variabilis — A canthocarpus preissii
Ccvimunity . — This is by far the most widespread
community on the island and it is probable that
the extent can be attributed to man’s action in
burning the Acacia rostellifera scrub (Fig. 14).
The community grows apparently equally well
cn recently colonised sand dunes or on the sparse
soil overlying limestone. Generally Phyllanthus
calycinus^ is present cn the younger soils and
again, depending cn edaphic factors, both

Fig. 14. Stipa variabilis — Acanthocarpus preissii
association which has developed where Acacia scrub
has been cleared. Rottnest Island.

Tho7nasia cognata and Guichenotia ledifolia may
occur. Thomasia grows on the shallow soils over
limestone and Guichenotia ledifolia often occurs
111 the valleys. However, Guichenotia is not
restricted to this community: in one place at
l^east it appears to be a stage in the succession
from swamp vegetation to Melaleuca pubescens.
The community is the most important on the
islaind from the viewpoint of food for animals
and birds.

The Ha'ophytic Communities. — It is the
presence of this type of vegetation on Rottnest
Island which makes it so different from Garden
Island. The halophytic communities are de-
veloped on swamps and lakes W'^hich are mainly
restricted to the eastern end of the island The
low lying regions give a distinctive appearance
to the Rottnest landscape.

56

Zcnation is clearly shown in these areas.
Around each lake and swamp there are con-
centric bands of successively taller vegetation
types (Fig. 15). In fact this may be true plant
succession.

Fig — Zon^t.lyij oi vcyctariou iDorcleriug; o §vvamp on

Rottnest Island. Zones from the water's edge are
Salicornia blackiaiia. AtripJex palucloia. Scirpus nodosus,
Solanum sUnilc and Stipa variahilis.

On the lowest levels the soil is saline (6.25%
sodium chloride) calcareous mud containing algal
remains and gastropod shells. Wilsoitia humilis
and Salicornia hlacklana occur on this soil which
is waterlogged for the greater part of the year.
There is a sharp transition from this zone to a
thick mat of Salicornia below which the soil,
althcugh damp, is not always waterlogged. This
soil has 3.3% sodium chloride. The next zone
varies considerably in different localities. In
some cases the Salicornia merges into an
Arthrocnemuni bidens zone and thence into
Melaleuca pubescens or Acacia rostellifera zone.
This is generally the case in the swampy areas.
However, around the lakes the zone following
Salicornia is usually a narrow band of Atriplex
pahidosa. Rarely the Atriplex may be absent
and the Salicor^iia ends abruptly giving way to
a dense Scirpus nodosits or Scirpus nodosus-
Galmia trifida community. The soil in this zone
is brown calcareous sand in which the salt con-
tent is negligible.

In one instance the Arthrocrieviuni bidens
passes into a poorly defined zone of Stipa
variabilis in which GuichenotAa ledifolia is also
present. However, this may be lacking and either
Melaleuca pubescens or Acacia rostellifera may
grow right down to the edge of the swamp.
Around the lakes the sequence is more constant
and the Scirpus-Gahnia zone passes gradually
into a narrow band of Solanmyi simile. Where
the ground begins to rise this last band dis-
appears and. depending on whether or not the
area has been burned. Stipa — Acanthocarpus or
Melaleuca pubescens occur.

Littoral Vegetation. — These communities are
similar in all respects to those already described
on Garden Island.

Carnac Island

Carnac Island is about two miles north of
Garden Island. It forms roughly a square of
about 38 acres in area. Except for a sheltered
sandy beach on the east and two tiny beaches
on the west side the island is bounded by steep
limestone cliffs. The highest point is 60 feet
and there is very little of the island below 25
feet in height. The uniform topography means
that most of the island is exposed to the wind
and this fact is reflected in the vegetation. There
is no fresh water on the island.

It is to be expected that Carnac Island situated
as it is between Garden Island and Rottnest
Island would have similar vegetation. However,
being such a small island on which no part is
more than 200 yards from the sea, it is natural
that the dominant plants would be those which
are found only on the borders of the larger
islands.

The recognisable communities, which are
shown in Fig. 16 are as follows: —

(i) Acacia rostellifera — Olearia axillaris
association ( scrub >

( ii ) Olearia axillaris — Scaevola crassifolia
association (low scrub)

(iii) Frankeiiia 7 mucifiora — Rhagodia bac-
cata association (low scrub)

(iv) Rhagodia taccata association (low
scrub >

(v) sc7ioberi association (low scrub)

(vi) Scaevola crassifolia — Caloceplialus
browniz association (low scrub'

( vii ) Carpebrotus aequilaterzts — T etragonia
spp.—Suaeda viaritivia mat

(a) Description of the Communities

Acacia rostellifera — Olearia axillaris Asso-
ciatiozi. — This community abuts on to the eastern
beach and extends about halfway across the
island. It appears to be the climax vegetation
for this island. Acacia rostellifera is the main
component, but Olearia axillaris is spread uni-
formly throughout. Clematis microphylla does
occur here but is never well developed. Under-
growth consists of Brojnus gussoziii and Lepido-
spenzia giadiatum.

The soil consists of calcareous fine sand very
low in organic matter.

This community merges imperceptibly into the
shoreward zone of Olearia — Scaevola scrub.

Olearia axillaris — Scaevola crassifolia
Associatiozi. — This is the most variable com-
munity on the island, but the two main com-
ponents occur fairly uniformly throughout (Fig.
17'. Rhagodia baccata becomes more common
near the shoreward cliffs.

The soil is light grey calcareous sand with
limestone 8 — 12 inches below the surface.

57

Fig. 16. — The plant communities on Carnac Island.

F r a n k e 11 i a pauciflora — Rhagodia baccata
Association. — This is a very distinct and sharply
defined community restricted to South Point. It
is typically 12 — 18 inches in height and com-
pletely covers the soil. The soil consists of a
few inches of dark brown calcareous sand
overlying travertinized limestone <Fig. 18).

Rhagodia baccata Association . — This occupies
only a small area near Smith Point fFig. 19)
It grows right up to the cliff edge on the north-
west and extends about 50 yards landwards.

It is about two feet in height and is quite
dense. The soil is fairly deep and is coarse
textured throughout.

Nitraria schoberi Association. — On Carnac
Island T<liiraria schoheri forms a dense tangled
scrub all along the talus slopes of the cliffs on
the northern side. Rhagodia baccata occurs but
only rarely. The very loose sandy soil of the
talus is stabilised mainly by the cover of Nitraria.

Fig- 17. — Olearia axillaris — Scaevola crassi/oJia scrub In
the foreground merging into Acacia rosteUifera— Olearia
axillaris sciub in the background, Carxmc Island.

58

Fig. ]8. -Franheriia paiicifiora—Rliagodia haccata association, Caruac Island.

Fig. 19 . — Rhagodia baccata scrub on a talus slope, Carnac Island.

59

Scaevola crassifolia — Calocephalus brownii
Association. — On the western shore between
the two beaches is a small area where
Calocevhalns hrownii extends back from the
coast and merges with the Olearia — Scaevola
scrub to form a separate community. This may
be only a variant of the Olearia — Scaevola scrub,
but it appears different enough to warrant
separate consideration. Salicornia hlackiana
forms a mat over a small area on the point
between the beaches.

Cai’pobrotus aequilaterus — Tetragonia spp. —
Suaeda max'itima Mat. — Fringing the larger
of the two western beaches is a thick mat of
succulent vegetation. This extends back about
50 yards from the beach. Basides the three
main components E^ichylaena tomentosa occurs
frequently. The bank is quite steep but the loose
sand is held firmly by the vegetation.

One of the most striking features of the
beaches of the mainland, particularly City Beach
and Scarborough, is the presence of Arctotheca
nivea, which is most active in stabilising the
sand (Fig. 20*. This plant keeps pace with the
banking sand and in most places the bank is
several feet in height. It may well be that this
is the first stage in plant succession since this
zone is followed by successive zones of Spinifex
hirsutus and Olearia axillaris- Scaevola spp.
before reaching the low scrub which is probably
the climax dune vegetation. The scrub is made
up principally of Scaevola spp., Rhagodia bac-
cata, Myoponivi insulare, Olearia axillaris, Scir-
pus nodosus, Lepidosperma gladiatum and the
creepers Cassytha glabella and Hardenbergia
coviptoniana.

The Mainland

The mainland adjacent to the islands, with
an expanding city in its immediate hinterland,
has suffered many changes in respect to its
vegetation. However, there are remnants of the
original vegetation at various points between
Point Peron and North Beach and it is the
vegetation of these areas which will be compared
with that of the islands.

Point Peron, which was formerly a rocky island
is now joined to the mainland by blown sand
(Fail-bridge. 1948). It is controlled by the same
environmental conditions as the islands and this
is reflected in the vegetation. However, there
are some important differences. As would be
expected there has been an influx of introduced
species such as Euphorbia spp., Broinus gussonii,
Solanum sodomaeuni. Aster subidatiis, Sporo-
bolus virginicus and Tripteris clandestina.

The natural vegetation, though stiucturally
the same as that cf the islands, differs in that
certain species common to the islands are absent.
The most important of these are Melalexica
pubescens, Guichenotia ledifolia and hasio-
petaimn angttsiifoliuin. Species common to the
understoreys cf the E. goviphocephala and E.
viarginata associations do not occur on Point
Peron. It is assumed then that the absence of
these plants on the islands is not due to isolation,
but environmental conditions.

The coastal strip between Fremantle and
Rockingham was probably originally covered by
dense Acacia rostellifera scrub similar to that on
Garden Island. At present it is much denuded
by clearing and fires, but it appears to differ only
in that XaiUhorrlioea preissn occurrs commonly
throughout. The remnants of CaUitris robusia
scrub near Naval Base also shows the similarity
of this strip to the island vegetation. Relics of
a similar vegetation are to be seen along cliff
sides of the Swan River at Freshwater Bay,
Mosman Bay and Blackwail Reach.

Further to the north at Cottesloe, City Beach,
Scarborough and North Beach the vegetation is
different, the mature tuart and jarrah forest
approaching much nearer to the shoreline,
although there is an area of dense Acacia rostelli-
fera scrub between City Beach and Scarborough.

•• ’< ' »•

Fig. 20. — Beach sand stabilised by Arctotheca nivea,
City Beach.

A.lyxia buxijolia and Dedonaea aptera do not
occur in the beach vegetation of the mainland.
These are very common on the islands, but ap-
parently occur on the mainland only on Point
Peron. Within a short distance of the shoreline
in the City Beach-Scarborough area occur many
plants not usually associated with beach vegeta-
tion. These are Santalum aciiminatuvi, Hib-
bertia spp.. Melaleuca acerosa, Oxylobium
capitatum, Gonipholobiwn tornentosum, Acacia
pxdchella and Tersonia brevipes. The presence
of these is a reflection of more mature soil
conditions.

It is not until the E. gomphocephala zone is
reached that significant numbers of Proteaceous
and Leguminous plants are found. Grevillea
critlivii folia and Petrephila serruriae occur on
the seaward fringe of this zone and then Banksia
spp. and Stirlingia latifolia are associated with
the tuart. Casuarina fraseriana also occurs as
part of the understorey of the tuart. but is not
found nearer the shoreline. Agonis fiexuosa com-
nicnly grows on the calcareous dunes of the
mainland but is not found on the islands.

It is of interest to note that the coastal vegeta-
tion contains many genera and even species
which occur much further inland in the Eremean
Province, It is evident that the habitats have

60

no obvious common environmental factor, al-
though Gardner (1942) points out that the soils
which support these plants are either physically
or physiologically dry. It is likely that the solu-
tion to this problem will be found only in a
study of the physiology of the plants concerned.
The presence of calcium carbonate in the pro-
file may be part of the answer, but other factors
will also contribute.

Discussion

The vegetation of the islands has distinctive
features both structurally and floristically. The
undisturbed vegetation as already described con-
sists of dense almost impenetrable scrub in which
the upper storey is completely closed and the
understorey very sparse or absent. This is in
striking contrast to the jarrah kE. viarginata)
and tuart iE. gomphocephala > associations of
the mainland which have an open tree layer
and a rich and varied understorey of small
shrubs and harsh woody monocotyledons. These
associations are however developed on mature
soils such as do not exist on the islands.

The distribution of species both on the islands
and the mainland can, with some exceptions be
explained by habitat rather than insularity.
Thus the protected sandy beaches on the eastern
side of Garden Island, Thompson’s Bay on Rott-
nest Island and Mangles Bay near Rockingham
are chaiacterised by a zone of Spinifex longifoHa
which grows at the foot of a low fore-dune, the
scrub approaching very near to the beach. The
pposed beaches on the western sides of the
islands, Shoalwater Bay near Rockingham and
metropolitan beaches on the mainland, have a
much steeper bank, which is stabilised by Scirpiis
nodosus, Lepidosperma glaliatum and then
by Olearia axillaris and Scaevo a spp. On the
seaward side of the bank Spinifex hirsutus and
Arctotheca nivea are the only plants which can
withstand the sand blast, although there may
cccasionally be Salso.a kali and Cakile maritivia.

The rocky cliffs and headlands on the western
sides of the islands, and along the mainland
coast between the metropolitan beaches, also
have a characteristic vegetation— both in species
and habit. The salt spray tolerant plants, Car-
pobrotus. Tetragcnia, Suaeda, Enchylaena and
Frankenia paaciftora are very commonly seen
clinging to the rocks and Nitraria and Rhagodia
where there is a little more soil. Calocephalus,
Angianthus and Scaevola crassifolia are common
along the tops of exposed cliffs, but apparently
require a few inches of soil to survive. There
are some larger bushes such as Alyxia huxiiolia^
Dodonaea aptera and Westringia rigida, which,
when present in this habitat, are wind-shorn and
prostrate. Thus on Carnac Island, which is prin-
cipally bordered by rocky cliffs and of relatively
small extent, it is found that the most common
plants are Nitraria, Rhagodia, Olearia axillaris
and the succulent plants.

The stable dunes on the islands and on parts
of the mainland support a dense growth of tall
scrub or scrub. This grades shoreward into the
low scrub of the recently coloni.sed dunes or the
scattered prostrate plants of the tops of the
exposed limestone cliffs.

On Rottnest Island and on parts of the main-
land coast the limestone comes very near to the
surface and such areas are characterised by
Ternpletonia reUtsa and Thoviasia cognata.
Tevipletonia retusa does not occur on Garden
Island although there are many places whore
it would be expected.

Having established that the species distribu-
tion within any one area is dependent on habitat,
it remains to explain the overall distribution
of species, and for the explanation it is necessary
to use palaeo-geographical evidence. It has
been suggested (Fairbridge 1948) that the whole
of the continental shelf became dry land during
the period corresponding to the Wurm exposure
of Europe. This w'ould allow large areas of
aeolianite to accumulate and so for a short time
coastal conditions extended far beyond the
present shoreline allowing migration of typical
coastal species. In the ensuing period towards
the close of the Pleistocene, the level of the sea
I'ose gradually, the higher parts of the aeolianite
being cut off from the mainland. This explains
why some of the same species are found on all
the islands and the mainland, although there
are anomalies such as the absence of Melaleuca
pubescens on the coast adjacent to the islands.

The absence from the islands of such
characteristic Australian families as the Pro-
teaceae and Xanthorrhoeaceae and many of the
Myrtaceae is at first sight very striking. How-
ever, it is evident that these families are not
found on the mainland where conditions
approximate to those on the islands. The
vegetation development under similar conditions
in the.se two areas is substantially the same and
it is the environment rather than isolation which
ha.s restricted migration.

Acknowledgments

The work described in this paper has been
carried out as part of the ecological research
programme of the Botany Department of the
University of Western Australia under the
supervision of Miss A. M. Baird. The author
wishes to acknowledge his indebtedness to Dr.
B. J. Grieve, Head of the Department, who has
given encouragement and advice throughout the
work and who has critically revised the
manuscript. Thanks are due also to other
members of Botany Department staff for
generous assistance: to Mr. C. A. Gardner
(Government Botanist) and Mr. R. D. Royce of
the State National Herbarium. Perth, who
checked many of the plant names; to Dr. J.
Gentilli (Economics Department, University)
who supplied figures relating to rainfall effective-
ness and evapo-transpiration: to Dr. K. Sheard
(C.S.I.R.O. Division of Fisheries) who assisted
with transport to the islands; to Dr. D. P. Drover
of the Institute of Agriculture, University of
Western Australia, for advice and assistance on
soil analysis; to the staff of the Survey Section
at Karrakatta Army Barracks, who supplied
aerial photographs and large scale maps of
Garden Island; and finally to the lessee of
Garden Island, Mr. F. Oliver, for valuable help
and advice.

61

References

Beadle, N. C. W. and Costin, A. B. (1952). — Ecological
Classification and Nomenclature. Proc. Linn. Soc.
N.S.W. 77: 61-82.

Burges, A. and I>rover. D. P (1953).— The Rate of Podzol
Development in Sands of the Woy Woy District,
N.S.W. Aust. J. Bot. 1: 83-94.

Fairhridge. R. W. (1948). — Notes on the Geomorphology
of the Pelsart Group of Houtman's Abrolhos Islands.
J. Roy. Soc. W. Aust. 33: 1-43.

Fairhridge. R. W. (1950).— The Geology and Geo-
morphology of Pt. Peron, Western Australia. J. Roy.
Soc. W. Aust. 34: 35-72.

Gardner, C. A. (1942). — The Vegetation of Western
Australia with special references to the Climate and
Soils. J. Roy Soc. W. Aiist. 28: 11-87.

Gentilli, J. (1948). — Two Climatic systems applied to
Australia. Aust. J. Sci. 11: 13-16.

Harvey, W. H. (1854). — Some Account of the Marine
Botany of the Colony of Western Australia. Roy.
Irish Acad. Trans. 22: 522-566.

Heeres, J. E. (1899).— “The Part Borne by the Dutch in
the Discovery of Australia 1606-1765.” (Luzac and
Company: London.)

McKnight. T. (1949). — ^Efficiency of Isolates of Rhizobiura
in the Cowpea Group. Qd. J. Agric. Sci. 6: 61-76.

Oosting', H. J. (1954). — Ecological Processes and Vege-
tion of the Maritime Strand in the Southeastern
United States. Bot. ReiK 20: 226-276.

Pidgeon, lima M. (1940). — The Ecology of the Central
Coastal Area of New South Wales. III. Proc. Linn.
Soc. N.S.W. 65: 221-249.

Piper, C.S. (1947). — "Soil and Plant Analysis.” (Univer~
sity of Adelaide: Adelaide.)

Smith, R. (1951). — Soils of the Margaret and Lower
Blackw’ood Rivers, W.A. Coun. Sci. Industr. Res.
Aust. Bulletin 262.

Speck, N. H. (1952), — “Plant Ecology of the Metropolitan
Sector of the Swan Coastal Plain.” (Thesis: Uni-
versity of Western Australia.)

Teichert, C. (1947). — ^Late Quarternary Changes in the
Sea Level at Rottnest Island, Western Australia.
Proc. Roy. Soc. Viet. 59: 63-79.

Thornthwaite, C. W. (1931).— The Climates of North
America. Georgr. Rev.: 21: 631-655.

Thornthwaite, C. W. (1948).— A Rational Classification
of Climate. Georgr. Rev. 38: 55-94.

Wilson, J. K. (1939). — A Relation between Pollination
and Nodulation of the Leguminosae. J. A?ner Soc
Agron. 31: 159-170.

Zeuner. F. E. (1945). — “The Pleistocene Period, its
Climate. Chronology and Faunal Succession.” (Ray
Society: London.)

Zeuner, F. E. (1946) — “Dating the Past; An Introduction
to Geochronology.” (Methuen: London.)

Appendix I

Distribution of Plant Species on the Coastal

Islands

Rottnest Garden Carnac

Cupressaceae

Callitris rohusta (R.Br.) Mirb. .

Island Island Island

X X

Typhaceae

Typha angustifolia Linn

X

Appendix 1— continued.

Rottnest Garden Carnac;

Island

Island

Island

Gramineae

Spinifex hirsiitus Labill.

X

X

Spinifex lonqifolius R.Br.

X

X

X

Stipa variabilis Hughes

X

X

Polypagon jnaritimus Willd

X

Polypogon monspeliensis (L.)

Desf.

X

Arena fatua Linn

X

X

X

Poa annua Linn

X

X

*Poa caespitosa Forst

X

X

*Bromus gussonii Pari

X

X

X

Cyperaceae

Scirpus antarcticus Linn.

X

Scirpus nodosus Rottb

X

X

X

Lepidosperma gladiatum Labill.

X

X

X

Lepidospervia squamatum Labill.

X

X

Gahnia trifida Labill,

X

Liliaceae

Thysanotus patersoni R.Br

X

X

*Asphodelus fistulous Linn

X

X

Asparagus asparagoides Wight

X

*Anthericum divaricaium Jacq.

X

X

Dianella rei^oluta R.Br

X

X

Acanthocarpus preissii Lehm

X

X

Amaryllidaceae

Conostylis candicans Endl

X

X

Orchidaceae

Eriochilus tenuis Lindl

X

Urticaceae

Parietaria debilis Forst

X

Santalaceae

Exorcarpus aphylla R.Br

X

X

Leptonieria preissiana (Miq.) D.C.

X

X

Lcranthaceae

Loranthus miraculocus (Miq.)

var. melaJeucae (Tate) Blakely

X

Polygonaceae

*Em€x australis Steinh

X

X

Chenopodiaceae

Rhagodia baccata (Labill.) Miq.

X

X

X

Atriplex isitidia Miq

X

X

X

Atriplex paludosa R.Br

X

Salsola kali Linn

X

X

X

Salicornia blackiana Ulbrich

X

X

X

ArthTocnemum bidens Nees

X

Enchylaena tomentosa R.Br

X

X

X

Suaeda maritima (R.Br.) Miq

X

X

X

*Chenopodium murale Linn

X

X

Aizoaceae

Tetrogonia zeheri FenzL ex

Harv. et Sond

X

X

X

Tetragonia implexicorna (Miq.)

Hook. .

X

X

X

Carpobrotus aequilaterus N.E.Br.

X

X

X

Cryophytmn crystallinum (L.)

N.E.Br

X

Raunculaceae

Clematis micropliylla D.C

X

X

X

Lauraceao

Cassytha glabella R.Br

X

X

Cruciferao

Cakile maritima Scop

X

X

X

Pittosporaceae

Pittosporuvi phillyraeoides D.C.

X

X

Mimosaceae

Acacia rostellifera Benth.

X

X

X

Acacia heteroclita Meissn

X

X

Acacia cyclopis A. Cunn

X

X.

Acacia cyanophylla Lindl.

X

X

Acacia cuneata A. Cunn

X

62

Appendix 1 — continued.

Appendix 1 — continued.

Island

X
X

I'apllionaceac

Hardenbergia comptoniana R.Br.
Templetonia retusa R.Br.

Geraniaceae

Geranium pilosum Forst x

Pelargonium australe Wild x

*>ETodium cicutarimn (L.) L’Her. x

Zygophyllaceae

Nitraria schoberi Linn x

Riitaceae

Boronia alata Smith x

Dipiolaena dampieri Desf. ... x

Polygalaceae

Comesperina integerrimum Endl
Comesper7na confertum Labill

Euphorhiaceae

Phyllanthus calycinus Labill x

'Euphorbia drummondii Boiss .... x

*Ricinu$ communis Linn.

Beyeria viscosa (Labill.) Miq. .... x

Stackhousiaceac

Stackhousia pubescens A. Rich. x

Sapindaceae

Dodonaea aptera Miq x

Rhamnaceae

Spyridium globulosum (Labill.)

Benth x

Stercwliaceae

Lasiopetalum angustifolium'W.Y.

Fitz.

Guichenotia ledifolia J. Gray .... x

Thomasia cognata Steud. . . x

Frankeniaceae

Frankenia paucifiora D.C. .... x

Myrtaceae

Melaleuca pubescens Schau x

Melaleuca huegelii Endl x

Onagraceae

Epilobium glabellum Forst

UmbelUferae
Didiscus pilosiis Benth.

Didiscns coeruleus D.C. ... .... x

Epacridaceae

Leucopogon richei (Labill.) R.Br. x

Leucopogon raceviulosus R.Br. x

Acrotriche ovalifolia R.Br. .... x

Garden

Carnac

Rottnest

Garden

Carnac

Island

Island

Island

Island

Island

Gentianaceae

X

Erythraea centauriuin Pers

X

X

X

Frimulsceae

*Anagallis arvensis Linn

X

X

X

X

Samolus repens (Forst.) Pers

X

X

X

X

Apocynanceae

X

Alyxia buxifolia R.Br

X

X

Convolvulaceae

X

X

Wilsonia humilis R.Br

X

Wilso7iia backhousii Hook.

X

X

X

Labiatae

X

Westringia rigida R.Br

X

X

Solanaceae

X

Solanum nigrum Linn

X

X

X

X

Solanum simile F. Muell.

X

X

Scrophiilariaceae

X

Dischisma capitatum (Thunb.)

X

Chois

X

X

Orobanchaceae

Orobanche australiana F. Muell.

X

X

X

Myoporaceae

Myoporum insulare R.Br.

X

X

Myoporuni viscosum R.Br

X

Eremophila broionii F. Muell

X

X

Lobeliaceae

Lobelia tenuior R.Br

X

X

X

X

Goodeniaceae

.

Scaevola crassifolia Labill

X

X

X

Compositae

X

Olearia axillaris (D.C.) F. Muell.

X

X

X

X

Senecio lautus Soland.

X

X

X

X

Calocephalus brow^iii (Cass) F.

Muell.

X

X

X

*Arctotheca nivea (Less) Leeuwin

X

X

X

X

^Cryptostemma calendu'aceum

(Linn.) R.Br

X

X

Athrixia pulverulenta ) Lindl. )

X

Druce

X

diypoehaeris glabra Linn.
■■‘Sonclius asper Hill

‘■■Sonchus oleracens Linn

*Tripteris clandestina Less

*Erigeron crispens Ponnet
*Angianthus humifusus (Labill.)
Benth

Angianthus cunninghamii (D.C.)

Benth. ---

*Inula graveolens Desf

X

x

X

X

X

X

X

X

X

X

X

X

X

X

* Introduced species.

Appendix II

Climatological Data, Rottnest Island, W.A.

Jan.

Feb.

Mar.

Apl.

May.

Absolufe -Max. Temp.

101 ■()

im-o

93-0

87-0

78 • 5

Mean Max. Temp.

7H-2

(0-0

/ 3 • 5

08-0

Mean 'retii]).

71 •(•

72-2

00-4

07 ;*

02 • 9

Mean Min. Temp. ®F

04 9

().'> • .")

04 1

01-0

5 < - 9

Absolute .Min. Temp. ‘'F

.'so-o

r>r>-o

.53 • 8

51 - .5

48-2

Mean Dailv Hanm*

13-3

13-3

1 2 • 5

11-9

10 • 1

Hel. liiim.'. a, 111 . (%)

03

03

0(

09

73

Rel. Hum.. ^ p.m. (%)

04

04

0")

04

07

I'loin!. P a.m. (tenths) . .

3-7

4-0

4-4

.) 1

()-0

Floud. 3 p.tn, (tentlis)

3-3

3-7

4-0

4-9

5-9

Mean Sat. Detieit (ins. Hu)

0-2S9

0-28:j

()-22‘)

0-202

0-1 .">2

Ruinfall (ins.) -

0-23

0-38

0 • no

1-44

4-40

Wet Ilavs (No.)

2-2

1 8

r. • 0

9-1

13 0

F.fVeetiveness ( Kst.)

(1-3

Of.

0-0

1 -9

1 • J

Kviipo-trans|)imtl(m (Kst.)

4-00

4-02

3-91

2-98

2 • 25

June.

July.

Aug.

Sejit.

o.-t.

Nov.

Dee.

Year.

72-4

09-0

71-4

80-()

88-0

92-2

97-4

101 •()

04-1

02 • 4

63-0

04-8

07-2

72-0

75-4

70-1

59 • 0

57-8

58- 1

59-5

01 -0

05-8

09 1

04-0

55 •

53-2

53 • 1

54-2

55 8

59 • 5

02-8

59 0

45 -x

45-0

44-0

40-0

4S-0

50 • 0

54 0

44-0

9-0

9-2

9 • 2

10 0

11-4

12-5

12-0

11 -4

70

74

73

70

00

03

70

09

70

08

0!)

07

00

05

07

0-4

0-2

0 • 0

.5 • /

5 5

4-8

4-1

5 • 2

0 • 3

0-2

5 ■ 8

5 1

4-2

3 5

4-8

(i-r24

()-n7

0-12(

0-142

()• 170

0-223

0-271

0 • 1 05

0-49

5-99

4-42

, 2-50

1 -51

0-50

0-38

28 • 89

19-3

22 0

10-8

■ 15-0

n-1

0-9

5-0

129

11 -9

n-2

7-9

4-0

2-5

0-9

0 • 5

49-5

1 08

1 52

1 - 00

1-94

2-57

3 • 40

4-13

34-72

63

TABLE I

Chemical Properties of Garden Island Soils

Profile

Jno,*

Depth

(ins.).

pH.

Percentage

CaCO^.

Percentage

NaCi.

Percentage
Total ■
Soluble
.Salts.

Percentage

Total

Organic

Carbon.

Percentage

Total

Nitrogen.

C/N.

1

Ao

7-4

23-5

0-012

0-09

26-8

1-46

18-4

(1- 2

7-9

59-2

0 • 008

0-06

13-6

0-58

23-4

2- 9

8-2

75 • 5

0-008

0 - 04

2-9

0-19

15-2

9-19

8-4

78-3

0-002

0-04

1-9

0-15

12-7

19-32

8-4

82-9

0-006

0-04

1-9

0-08

23-8

o

Ao

7-2

41-7

0-010

0-09

21-8

0-94

23-2

0- 9

8-3

76-8

0-006

0-04

3-5

0-18

19-5

9-18

8-5

85 • 3

0-004

0-04

1-0

0-09

111

18-30

8-6

88-7

0-006

0-03

0-6

003

20-0

3

0-4

8-3

85-5

0-002

0-04

3-6

0-19

25-3

4-11

8 • 5

88-7

0 - 002

0-03

1-4

0-07

20 0

11-29

S-C

90-0

0-002

0-02

1-2

n.d.

—

4

0- 8

8-2

69-8

0-010

0-05

4-6

0-40

11-5

* 1 Soil under Acacia rostellifera association.

2 Soil under Callitria robusta association.

3 Soil under Melaleuca hucgcUi association.

4 Shallow soil on limestone. Seasonal vegetation.

64

7.— Notes on the Geology of the Carnarvon (Northwest) Basin,

Western Australia

By Curt Teichert*

Manuscript leceived — 20tli March, 1956.

Because of recent oil finds, the Carnarvon
Basin, otherwise known as Northwest Basin, so
named by A. Gibb Maitland in 1901, occupies a
position of paramount importance among the
major sedimentary basins in Australia. To ex-
ploit its possibilities detailed knowledge of its
stratigraphy and structure is a necessity. The
first comprehensive report on the geology of the
Carnarvon Basin, since Raggatt’s studies in 1936,
has recently been published by Condon (1954i
and the following comments on and corrections
to this repoi't are offered in an endeavour to
clarify parts of the stratigraphy and to point out
alternative structural interpretations.

Palaeontological Zoning

On p. 71, Condon lists a fossil assemblage
from the Bulgadoo shale in which CalceoUspoiigia
harrahiddiensis and C. acuminata are included,
together with Pseudoschistoceras simile. It
should be noted that these two species of Calceo-
lispcngia are rather widely separated strati-
graphically. C. acuminata is restricted to the
top of the Bulgadoo Shale although its occur-
rence is probably slightly lower than indicated in
a previous diagram (Teichert, 1949. pi. 26). C.
barrabiddiensis occurs with Pseudoschistoceras
simile in the Barrabiddy member ( Teichert,
1952) of the Bulgadoo Shale, about 2.100 feet
below^ the top of the formations. The two species
of Calceolispongia characterize different pala-
eontological zones, which are at least 1,500 feet
apart stratigraphically.

The most characteristic fossil of the Quinnanie
Shale (p. 75-78' is Hyperamminoides acicula
Parr which occurs in great quantities at the type
locality and in all outci'op belts of the formation
in the fault blocks to the east of Wandagee Hill.
Everywhere these foraminifera can be washed in
huge numbers out of the shales or their residual
soils. Condon does not mention the occurrence
of this fossil.

The measured section of the type Wandagee
Formation (p. 79-80) gives an inadequate picture
of the palaeontological zonation of this forma-
tion which has been carried out in great detail.
It is impossible to repeat here the details which
have been given elsewhere (Teichert, 1952). On
p. 80 Condon mentions 66 feet of medium-hard
grey quartz greywacke etc. from which he lists

*Denver, Colorado.

Calceolispongia elegantula and C. multiformis.
These two crinoid species are excellent index fos-
sils, but they do not occur together. C. elegan-
tula is abundant in and very characteristic of
the lowermost 80 feet of the Wandagee Forma-
tion (from which it is not listed by Condon):
C. muMformis is restricted to beds 165-185 feet
above the base of the formation. The position
of the bed from which Condon lists these two
species is at 102-162 feet above the base of the
formation. It has been noted before (Teichert,
1949, p. 77. pi. 15 figs. 16-21) that some basal
plates of C. multiformis resemble C. elegantula,
but care should be taken in distinguishing the
two species. This is always possible if compre-
hensive collections are available.

Stratigraphy

It is difficult from Condon’s paper to obtain
a clear picture of the relationships of the Norton
and Nalbia Greywackes and the Baker Forma-
tion. Since I am personally responsible only
for the Nalbia '‘Greywacke,” I shall restrict my
remarks to that formation. As it appears that
Condon has had difficulties in tracing the forma-
tion outside the type section, 0.9 mile west of
Quinnanie corner, the following notes on its
distribution may be helpful: at the type section,
the strike of the Nalbia Sandstone is N 20' W
and it can be followed in this direction for ap-
proximately 1 mile until it is cut off by north-
south striking fault. The same formation
outci'ops again in the next fault block to the
west where its bottom crosses the south fence
of Coolkilya Paddock 1,540 yards from the place
where the top of the foi'mation crosses the fence
further east. This is on the east flank of the
“Coolkilya syncline,” the axis of which strikes
N 10“ W with a slight northerly plunge. The
Nalbia Sandstone can be followed around the
greater part of this syncline whose center is
filled with rocks of the Coolkilya Formation. 450
yards south of the fence, in Woollies Paddock,
the Nalbia Sandstone swings around the south-
ern nose of the syncline. The northern part
of the syncline is partly covered with dune sand.
South of this sand area the total traceable out-
crop belt of the Nalbia Sandstone is at least 4.5
miles long. There are also some outcrops of it
to the north of the sand area and south of an
important cross-fault, striking N 60“ E, which
cuts off the Coolkilya syncline on the north. All
over this area the Nalbia Sandstone retains not

65

Plate 1. — Oblique aerial view of the highly faulted area south of Minilya River in the vicinity of Wandagee Hill, taken
15,000 feet, looking due north. (West Aust. Trimet Run 207. bl R) The Minilya River is seen in the upper part of pictur
a sand plain extends to the north of it. Drainage channel in middle of picture flows into Cundy Dam. drainage clmiw
in foreground into Mungadan Dam.

Indicated by back circles o;

1 Wandagee Homestead,

2 Mouribandy Dam,

3 Wandagee Wool Shed.

Indicated by Wandagee Hill Trig. Station.
Entire lines: formational boundaries.
Dashed lines: faults,
m — Mungadan Sandstone
c — Coolkilya Sanstone
n — Nalbia Sandstone

w — Wandagee Formation
q— Quinnanie Shale
d — Cundlego Sandstone
b — Bulgadoo Shale
G — Gypsum

Note that most faults in the Permian are broadly arcuat?. the convex side of the arc facing west. These are
east-dipping, normal antithetic fattlts. The fault in the west that separates Cretaceous from Permian rocks is a steeply
dipping normal fault.

The syncline to the north of Wandagee Hill has been referred to as “Coolkilya syncline” in this and previous publicatioB*-

The cross-section Fig. 1 runs E-W almost exactly through the middle of the photograph. Alluvial formations are
indicated.

Published by kind permission of the R.A.A.F.

66

only its lithological characteristics, but also its
palaeontological features with a rich Schizodus-
Oriocrassatella assemblage in the lower part and
abundant Cleiothyridina iiear the top.

North of the Minilya River the Nalbia Sand-
stone outcrops in the syncline whose axis inter-
sects the north bank of the river 650 yards
north-east of Curdamuda Well, Outcrops of the
Nalbia Sandstone, here unfossiliferous, are re-
stricted to a narrow belt on the north side of the
river.

Excellent outcrops of the Nalbia Sandstone are
also found further south in Woollies Paddock.
The belt of Nalbia Sandstone begins a little over
600 yards N 20° E of Wandagee Wool Shed where
the beds strike 85°, dipping 30'“ S. The formation
can be followed in a straight line along the strike
for 1600 yards to a position just inside Dry Pad-
dock where it is cut off by the major fault which
runs along the west side of Wandagee Hill. In
this outcrop belt both the lower pelecypod zone,
and the upper Cleiothyridina zone ai-e well
developed.

Further south the characteristic pelecypod
assemblage disappears but the Cleiothyridina
zone is well recognizable around the nose of a
little north-plunging anticline 1,600 yards south
of Wandagee Wool Shed on the road to Munga-
dan Dam, This is close to the Cretaceous bound-
ary fault and from hers the Nalbia Sandstone
can be followed first in a south-easterly direc-
tion, then swinging around the southern nose
of the north -plunging syncline which lies west of
Wandagee Hill.

East of Wandagee Hill, the Nalbia Sandstone
becomes unfossiliferous. but retains its charac-
teristic lithology. It occurs in two fault blocks
which are separated by a curved east-dipping
normal fault. The western-most of these blocks
includes Wandagee Hill itself. The Nalbia Sand-
stone emerges from underneath a Pleistocene
talus stream at a place 1.650 yards N 30' E of
Wandagee Hill Trig. Station. It swings around
Wandagee Hill in a wide, almost circular arc.
and it is cut off by a fault 3,250 yards S 20' E of
the same Trig. Station. About half way down
the outcrop belt approaches to within 40 or 50
yards of the eastern boundary of Mungadan
Paddock, but does not cross into Nalbia Paddock.
The outcrop belt in Nalbia Paddock belongs to
another fault block. On a bearing due east of
Wandagee Hill Trig. Station, this belt lies at a
distance of 2,900 yards from the latter and strikes
about N 20' W dipping west, To the north, the
strike swings further west: to the south, it swings
almost into a north-south direction. Total length
of the outcrop belt is nearly 2 miles. On both
ends, it terminates against the same curved fault
v.’hich separates this block from the one to the
west. On the whole. I should say that it would
be difficult, if not impossible, in the closely
faulted area east and north of Wandagee Hill,
to map the Nalbia Sandstone from aerial photo-
graphs. Condon seems to have confused the
Nalbia Sandstone with his “Norton Greywacke’'
and “Baker Formation.” The type localities of
these two formations are situated 30 to 40 miles
from that of the Nalbia Sandstone and there is,
as far as I know, no continuity of outcrops over
the intervening area. Correlation between the
northern Kennedy Range and the Wandagee

area can only be established by means of close
palaeontological control which has received little
attention in Condon’s report. Since Condon
refers to a small outcrop area of the Norton
Greywacke in the syncline north of the Minilya
Fiver “ 9 miles west of Wandagee Homestead”
i 650 yards north-east of Curdamuda Well, see
above), there seems to be no question that in
the Wandagee area he applied that name to the
unfossiliferous or less fossiliferous facies of the
Nalbia Sandstone.

This conclusion is also supported by his refer-
ences to the heavy mineral work by Higgins and
Carroll (1940). According to Condon, Higgins
and Carroll’s samples 1, 3 and 5 come from the
Norton Greywacke. I collected these samples
myself and to the best of my recollection, sample
3 comes from the Nalbia Sandstone, samples 1
and 5 come from somewhat higher beds, namely
from the "Fenestella nodule beds” of the Cool-
kilya Formation. It should be remembered that
Higgins and Carroll showed that there is no
striking variation in heavy mineral amounts in
the sequence below the “Wandagee Hill bed.s”

( Muiigadan Sandstone of present nomencla-
ture) and that heavy minerals, therefore, offer
little support in the correlation of these lower
beds.

While it thus seems certain that “Norton
Greywacke” is a synonym of the Nalbia Sand-
stone. perhaps Including the lowest part of the
Ccolkilya Formation, it is more difficult to be
sure which beds in the Wandagee area Condon
correlates with the Baker Formation of the Ken-
nedy Range. He says it overlies the “Norton
Greywacke” which I have shown to be essentially
identical with the Nalbia Sandstone. It would
then seem that the Baker Formation corres-
ponds to the lowest part of my Coolkilya Forma-
tion. However, this is characterized everywhere
in the Wandagee area by light-grey weathering
calcareous sandstone nodules with many speci-
mens of Feneatella- Of these, there is no men-
tion in Condon’s report. In the Coolkilya syncline
Hedcoprion, together with Pj'opmacoceras, Para-
gastrioceras and Psertdogastrioceras. was found
just above the Fenestella nodule beds, but Con-
don (p. 87) refers the beds with these fossils to
the Nalbia “Greywacke.” Part of Condon’s Nal-
bia “Greywacke” is thus older, part younger than
the “Baker Formation.” Apparently, there is
something wrong with this coi-relation and
further work is needed. If possible, this should
be carried out mainly on the ground. If no
continuity of outcrops can be established between
the Kennedy Range and the Wandagee area, it
would be better to apply a separate nomen-
clature to the two areas, unless correlations can
be made beyond reasonable doubt.

On p. 97, Condon states that “the boundary
between the Coolkilya Sandstone and the Mun-
gadan Sandstone defined by Teichert (1952), is.
not a main lithological boundary” and he, there-
fore, proposes to revise the definition of these
two formations. However, after reading his de-
scription of the upper part of the Coolkilya Sand-
stone in its type locality east of Wandagee Hill,

I have little doubt that Condon’s and my upper
boundaries of this formation are the same. The
top unit in Condon’s section is “2 feet of hard
red-brown ferruginous coarse-grained quartz

67

greywacke, with interior and exterior moulds of
spiriferids. pectenids and Oriocrassatella:’ There
actually two such red-brown horizons (which
I had called “dark purple’') only a few feet apart.
The tw’o can be followed all around the east side
01 Wandagee Hill and I chose them as markers
for the top of the Coolkilya Sandstone 'Teichert.
1952). In earlier papers (Teichert, 1949) this
horizon had been called the top of the ‘'Lino-
vroductus beds of the Wandagee Series.” There
is thus no discrepancy between Condon’s and
my interpretation of the boundary between the
Coolkilya “Greywacke” and the Mungadan Sand-
stone. Condon, however, omits to mention
that the uppermost “red-brown” horizon with
OviocransateJla also marks the last occurrence of
CalceoUspongia robusta, an important index fos-
sil of the Coolkilya Formation. Important mem-
bers of the fauna of the latter are Agathiceras,
Pseudogastrioeeras and Dictyoclostus, all of
which mark horizons in the upper part of the
formation. On p. 94. Condon lists CalceoUspongia
cf. rotundaia from the Mungadan Sandstone in a
section in the Kennedy Range. That species,
however, is restricted to the lower 165 feet of
the Wandagee Formation, and, if correctly
Identified, suggests correlation of the beds in
Question with the Wandagee rather than the
Coolkilya Formation.

Peculiarly, Condon has entirely omitted men-
tion of the calcareous eolianites (or “coastal
limestones”) of Pleistocene, most probably Wiirm,
age. These rocks are widespread along the coast
north of Warroora as far as the vicinity of Car-
dabia Homestead. They are quite similar to the
same formations well known elsewhere from
Western Australia (see various joint and indi-
vidual papers by Fairbridge and by Teichert)
and their study is important for the interpreta-
tion of the late Quaternary history of the area.
At Point Anderson, 6 miles south of Maud Land-
ing, there are clearly two generations of eolianite,
each topped by a travertinized layer which here
takes the place of the rendzina soils developed
between successive dune generations in more
southerly latitudes (Fairbridge and Teichert,
1953). The fact that the earliest visible eolianite
is now partly submerged, presents evidence for
a comparatively recent rise in sea-level. From
analogy with conditions in the south this rise
should be post- Wiirm.

Structure

Condon’s discussion of the fault mechanics of
the Carnarvon basin (pp. 135-138) is interest-
ing, but his interpretation is not supported by
the facts, at least not in those parts of the
basin known to me. I shall restrict my discus-
sion to Condon’s section C-D on Plate 2. part
of which is here reproduced as fig. lA. The
position of this section virtually coincides with
that of one published by Clarke. Prider and
Teichert (1944, p. 94) which was based on the
present writer’s at that time unpublished data,
and to which Condon does not refer. In com-
paring such sectioias, and the interpretation they
represent, it must be borne in mind that the
faults themselves are rarely exposed. Their
presence can be discovered by mapping, but
their dips can, in most cases, only be inferred
from general considerations. Condon believes

that almost all faults in the Carnarvon basin
are reversed or thrust faults and since (p. 138)
he bases important economic conclusions on this
interpretation, the matter is of more than theo-
retical interest. Condon himself records “a
marked absence within the sediments of evi-
dence of compression such as compressional
joint systems, cleavage and drag-folds,” but be-
lieves that this can be explained by assuming
that “the stress which caused the faulting was
carried mainly by the pre-Cambrian basement
and that the faults in the sediments result from
the faulting of the basement.” Nevertheless, if
there is upthrust in the basement, at least some
of the usual manifestations of thrust-faulting
should be discoverable in the sedimentary mantle
here and there.

However, to say that evidence of compression
is lacking is definitely an understatement, be-
cause wherever faulting is observed along the
section here discussed, the evidence for ten-
sional stresses is overwhelming. These are of
tlii'ee kinds: (T) abundance of calcitic veins as
fillings of tension gashes everywhere near faults;
(2) occurrence of wide tension rifts filled with
secondary gypsum; and (3) the scattering of
small blocks of sediments in random orientations
along fault zones. The latter show no signs of
crushing by compression such as one might
expect along thrust-faults, but are simply blocks
that have slumped into cracks opened up by
faulting. Gypsum zones along faults are well
developed and most easily observed in the fol-
lowing places: In Nalbia Paddock of Wandagee
Station there is a gypsum belt whose southern
end lies 2,250 yards almost due east of the Trig.
Station on Wandagee Hill. From here it runs
almost due north for about 1.500 yards, then
swinging gradually into a more north-easterly
direction. Its total length is over 1.5 miles and
Its greatest width at least 100 yards. The throw
of this fault at the southern end of the gypsum
belt is about 1.000 feet, decreasing towards the
north. In 1940 and 1941. a large, widely visible
eucalypt tree (“white gum” or “river gum”)
in the middle of this belt. If the tree is
still there, the place should be easy to find.

Another narrower and longer gypsum belt is
found in Dry and Coolkilya Paddocks on Wan-
dagee Station. It crosses the fence between the
two paddocks at 2.200 yards west of Quinnanie
Corner, where it strikes almost due north and is
about 50 feet wide. Southward it can be fol-
l()wed for about 0.5 mile until it disappears under
alluvial cover. To the north, it extends at least
1.5 miles and marks a somewhat oblique fault
which first Coolkilya Formation, then
Nalbia Sandstone, are cut off. To the west of
the fault are flats underlain by Quinnanie Shale
characterized by abundant Hyperannninoides
(see above). The throw of the fault where it
crosses the fence is approximately 1.200 feet.

These are two outstanding examples, but gyp-
sum is found along fault zones elsewhere in the
area.

Good surface evidence for east-dipping normal
faults is found along the section line (Fig. 1)
about halfway between Wandagee Hill and
Williambury where, at Trig. Station K55 and
to the east thereof, it intersects three fault

68

PLATE 2— Oblique aerial view along east coast of Salt Lake (or “Salt Marsh”), taken from
15,000 feet, looking due south from point above upper part of lake. (West Australian Trimet
Run 205. 66 R. lOj. From lower left through centre of picture trends Chirricia anticline with
12-ft. terrace along its entire western flank. Note weakly incised consequent drainage on
both flanks of anticline. The small, early Recent deltas on the terrace are not recognizable
from this altitude, but the newer Recent deltas in front of the terrace are clearly visible.

Headland in upper right of picture is Sandy Blufl’. The first river to the left (east) of it
is the Minilya River; the next river is Barrabiddy Creek.

Published by kind permission of the R.A.A.F.

69

iDlocks consisting of Callytharra Limestone and
Wooramel Sandstone. These fault blocks form
gently inclined west-sloping cuestas in which the
hard Wooramel Sandstone has protected the
softer Callytharra limestones and shales which
form steep east-facing cliffs. At least near K 55
itself, and at the eccarpment in the next block
to the east, the faults (which are not exposed ►
must be situated quite close to the base of the
escarpments. I have visited both localities
several times and examined the rocks in detail.
I am sure that there is no sign of upthrusting
along these escarpments and that, as indeed one
might expect, the cuestas owe their existence to
a westward tilt of the fault blocks, along normal
faults on both sides.

Fig. IB shows the structures which one would
expect to find along the selected section if the
hypothesis of thrust-faulting were correct.
Although I am familiar with the detailed geology
around all the faults shown in this section, I
have never seen any field evidence suggesting
structures of this kind. On the contrary, all
the evidence, circumstantial and observational,
supports the assumption of normal faults.

Fig. 1C shows how the structures can be
interpreted as the result of normal faulting.
There are some normal west dipping faults, the
only major one probably in the Precambrian to
the east. The dominating pattern is that of a
set of antithetic, east-dipping adjustment faults
(Teichert. 1948. 1952).

Antithetic faults were named and discussed by
Cloos (1928). although they were well known
before. They are due to rotational movements
of fault blocks during which each block carries
out a rotating movement around its centre of
gravity. Such movements result from tension
fracturing of sedimentary table lands, especially
in wide flexure zones, if part of the sedimentary
table is elevated. In such a situation, the stresses
normal'y lead to formation of a small number
of major faults which dip in the direction of
the regional structural dip and are therefore
called hoviothetic by Cloos, combined with a sys-
tem of numerous faults which dip against the
regional dip (hence, antithetic’ and which assist
in the release of tensional stresses. Such fault
systems have been reproduced experimentally
(Cloos, 1932) and they occur widely in nature
(Cloos, 1936. pp. 226-269). particularly along
continental margins (Cloos. 1939, p. 504).
Generally the plane of the antithetic fault lies
approximately at right angles to the plane of
the beds of the respective fault block, or it may
be slightly more gently inclined. In the faulted
country between Wandagee Hill and William-
bury. the prevailing dips of strata are between
20" and 25“ west. The faults, separating the
fault blocks, may therefore be expected to dip
65° to 70“ east, but their dips may be as low as
50°. The entire fault pattern is due to uplift of
the Precambrian basement in the east.

There is no good evidence to suggest where
exactly the coast was situated during Devonian.
Carboniferous and Permian time, except that
it must have been well to the east of the present
easternmost distribution of Palaeozoic rocks, that
is, well inside of what is now the Western Aus-
tralian Precambrian shield. In Palaeozoic times,

the present shield margin formed part of the
subsiding basement on which thick sedimentary
series accumulated. In post-Palaeozoic times,
the marginal sedimentation area was uplifted
and its basement became part of the shield.

The overall regional rise of the Precambrian
basement from the coast to the present edge of
the Western Australian shield is of the order
of 200 to 250 feet per mile. In other words, there
is an overall seaward dip of the surface of the
Piecambrian of approximately 2\ The tensional
stresses created by the uplift of the shield were
released in the manner here suggested, by a few
widely spaced west-dipping normal faults, and
a system of numerous more closely spaced east-
dipping antithetic faults.

The age of the main uplift which caused this
structure pattern is pre-Cretaceous, because the
Cretaceous transgresses over a faulted Palaeo-
zoic basement. In places there has been some
post-Cretaceous faulting, probably caused by
continued uplift of the shield. The post-Cre-
taceous fault west of Wandagee Hill is a normal
west-dipping fault which is possibly posthumous
to a pre-Cretaceous antithetic east-dipping one.
The possibility should be kept in mind that the
pre-Cretaceous tensional fault pattern which is
visible on the surface east of Wandagee Hill, con-
tinues westward below the infra-Cretaceous un-
conformity. This possibility has not been indi-
cated on the sections Fig. lA-C.

Discussing the oil possibilities of the Carnarvon
Basin, Condon suggests that the thrust faults
which he believes to exist in the basin “may
provide adequate structural closure in some
cases” (p. 154). Our interpretation affords no
basis for such optimism. Antithetic tension
faults will facilitate the escape of hydrocarbons.
The search for oil in the Carnarvon basin must,
therefore, be restricted to anticlinal structures
and to structural and stratigraphic traps below
the Cretaceous-Permian uncoiiformity. It will be
of the utmost practical importance to determine
the exact age of the dislocations of the Palaeo-
zoic rocks. The longer the interval between this
tectonic episode and the Lower Cretaceous trans-
gression. the greater the chances for Palaeozoic
oil having escaped before it could be trapped.

Age of the Anticlinal Structures

In the western part of the basin occur a num-
bei of remarkable anticlines and domal struc-
tures all of which have been named and briefly
described by Condon. He attributes their origin
to upthrusts in the basement and for some of
the major anticlines, such as Cape Range and
Giralia, he states evidence for a two-stage uplift,
both of which occurred after the deposition of
the Miocene Trealla Limestone, and before the
Pleistocene. Essentially then the formation of
these, and by implication of all other anticlines
is held to be epi-Miocene to Pliocene in age.
While I find myself in basic agreement with
these conclusions I believe that it is possible to
demonstrate that the uplift of some anticlines
took place later, or at least continued into post-
Pliocene times.

Condon describes the occurrence of marine
shell deposits in the vicinity of the Salt Lake
to heights of 20-25 feet above present lake bot-
tom. He does not. however, mention the pre-
sence of the remarkable terrace which is cut into

70

o o

X >.

k.

« i3

» e

O o

T)

c

4 . O

4> ^

tr £

o o

FIG. 1. — A, part of section C-D in Condon, 1954, PI. 2. B, same section showing fictitious struc-
tures which would be in accordance with the interpretation of thrust-faulting, but which
are not found in nature. C, same section showing interpretation of structures as an anti-
thetic fault pattern. D Devonian, C Carboniferous, Psl Permian Lyons Group, Pac Permian
Callytharra Limestone. Pab Permian Byro Group, K Cretaceous. Vertical scale approximately
2^ times horizontal scale.

71

the west flanks of the anticlines on the east side
of this generally dry lake. This terrace may be
studied especially well along the entire west side
of the Chirrida anticline (see Plate 2). where it
lies at 11-12 feet above lake bottom level and
is about 100 yards wide. It consists of loose
deposits, partly shell layers, and partly sand.
It is now incised by the many consequent strearn-
lets which run off the west flank of the anti-
cline. These little streams have cut narrow
small canyons into the limestone roof of the
anticline and at the mouth of each canyon a
small deltaic fan can be seen to rest on the ter-
race. These are the deltas which were made by
the consequent streamlets at the time when the
sea covered the terrace. They consist of lime-
stone rubble mixed with shells. After sea-level
subsided the streamlets cut through the deltas
and into the underlying terrace.

At the north end of the Salt Lake, around the
mouth of the Lyndon River, there is a river ter-
race, 12 feet above the river flat, which consists
of red. brown, and partly gypseous sand. This
terrace is particularly well developed east of the
Lyndon River where it is up to half a mile wide.
It is in the same position, hence most probably
of the same age as the terrace along the west
side of Chirrida anticline.

What is the age of the terrace and its super-
imposed deltas? The height of 11-12 feet above
lake bottom as well as the remarkable persistency
of the feature immediately suggest correlation
with the 10-ft. bench which has been so widely
recognized in the southwestern part of Western
Australia (see Teichert 1947. 1950; Fairbridge.
1950: Fairbridge and Teichert. 1953: and other
papers by the same authors) and the age of
which is regarded as early Recent. This dating
is supported by the youthful appearance of the
terrace and its deltas. The small size of the
latter would suggest that they represent a com-
paratively short span of time.

The position of the terrace poses some in-
teresting problems. Condon stated (p. 125) that
the present lake bottom is very close to ordinary
high tide level, but more recent surveys by Ray
Geophysics (Aust.) Pty. (kindly supplied by W.A.
Petroleum Pty. Ltd. through courtesy of Pro-
fessor R. T. Pi-ider) have shown that along the
Chirrida anticline the level of the lake bottom
lies approximately 2 feet below mean sea-level.
The tides along the coast near Carnaiwon are
somewhat variable, but spring range is about
4 feet (Hodgkin. 1957). The lake bottom aloiig
its eastern coast is therefore now situated at
about low water springs. The surface of
the terrace along the west side of Chirrida
anticline must have been formed somewhere
between low water and mean sea-level and it
is thus apparent that little if any movement of
this anticline has occurred since the mid-Recent
drop of sea-level, that is during the last 3,000

or 4,000 years. Conditions along other anticlines
might well be different. The study of features
attributable to Late Quaternary sea-level move-
ments, particularly the 10-ft. bench, may well
be helpful in the analysis of young movements
in this coastal belt. Where its present level
deviates appreciably from 10 ft. above Low Water
it could be deduced with great confidence that
very young movements have taken place. In
the case of Chirrida anticline the small amount
of erosion which took place prior to and during
the formation of the 12-ft. terrace seems to in-
dicate a very young age of this structural feature
v/hich, I think, must have come into existence
during the Pleistocene. There has apparently
been no measurable movement since early Recent
time.

References.

Clarke, E. de C., Prider, R. T., and Teichert, C. (1944).—
“Elements of Geology for Western
Australian students.” University
West Aust. Text Books Board,
Crawley.

Cloos, H. (1928). — Uher antithetische Bewegungen. Geol.
Rdsch. 19: 246-251.

(1932). — Zur Mechnnlk grdsserer Briiche und

Graben. Zbl.Miner. 273-286.

(1936). — “Einfuhrung in die Geologic. ” Born-

traeger, Berlin.

(1939). — Hebung — Spaltung — Vulkanismus.

Geol. Rdsch. 30, Zwischenh. 4A;
405-525.

Condon, M. A. (1954). — Progress Report on the Strati-
graphy and Structure of the
Carnarvon Basin. Western Aus-
tralia. Australia, Bur. Min. Res.
Geol., Geoph., Rep. No. 15.

Fairbridge, R. W. (1950). — The geology and geomor-
phology of Point Peron. Western
Australia. J. Roy. Soc. West. Aust.
34: 35-72.

& Teichert. C, ( 1953 ) . — Soil horizons and

marine bands in the coastal lime-
stones of Western Australia. J. Roy.
Soc. N.S.W. 86: 68-87.

Higgins, H. G., and Carroll, D. (1940). — Mineralogy of
some Permian Sediments from
Western Australia. Geol. Mag. 77:
145-160.

Hodgkin, E. P., and Di Lollo, V. (1957). — “The tides of
South-Western Australia,” J. Roy.
Soc. West Aust. 41 (in press).

Teichert, C. (1947). — Stratigraphy of Western Australia.

Bull. Amer. Ass. Petrol. Geol. 31:
1-70. (Also: J. Roy. Soc. N.S.W.
80: 81-142.)

(1948).— -Younger tectonics and erosion in

Western Australia. Geol. Mag. 85:
243-246.

(1949). — Permian crinoid CalceoLispojigia.

Geol: Soc. Am. Mem. 34: 1-132.

—(1950). — Late Quaternary changes of sea-

level at Rottnest Island, Western
Australia. Proc. Roy. Soc. Viet.
59, Pt. IT (N.S.): 63-79.

— (1952). — Carboniferous. Permian. and

Jurassic in the Northwest Basin,
Western Australia. XIX Int. Geol.
CongT., Symp. Series Gondwana.
Algiers: 115-135.

72

CONTENTS OF VOLUME 40

Part 1 ipxiblislied February 21, 1957)

Kelvin Medallist, 1955: H. W. Bennetts, C.B.E., D.V.Sc. 1

1. Laterite and Materials of Similar Appearance in South-Western Australia.

Presidential Address, 1954. By S. E. Te.-rill 4

2. Studies in the Water Relations of Plants. I. — Transpiration of Western Aus-
tralian (Swan Plain) Sclerophylls. By B. J. Grieve .. . .. 15

3. Atapozoa niarshi, a compound Ascidian from Western Australia. By Beryl I.

Brewin , . .... . .... . ... .... 31

4. Crustacea from the Cretaceous and Eocene of Western Australia. By M. F.

Glaessner . 33

Part Z f published 20th November , 1957 )

5. The Reaction of Plants to Growth Regulators with particular reference to Weed

Control. Presidential Address, 1955. By G. R. W. Meadly .. .... ... .... 37

6. Plant Ecology of the Coastal Islands near Fremantle, W.A. By W. M. McArthur 46

7. Notes on the Geology of the Carnarvon (Northwest) Basin. Western Australia.

By Curt Teichert 65

The Royal Society of Western Australia Incorporated

Council 1955-1956

President . .. D. E. White, M.Sc., D.PhiL, F.R.A.C.I.

Past President G. R. W. Meadly, M.Sc.

Vice-Presidents A. F. Wilson. M.Sc., F.G.S.

A. J. Millington, D.Sc. (Agric.).

Joint Hon. Secretaries . N. H. Brittan. B.Sc., Ph.D.

R. D. Royce, B.Sc. ( Agric. >.

Hon. Treasurer C. R. LeMesurier, A.R.A.C.I. (deceased).

R. D. Royce, B.Sc. (Agric.).

Hon. Librarian G. G. Smith, M.Sc.

Hon. Editor D. E. White, M.Sc., D.Phil., F.R.A.C.I.

G. H. Burvill, M.Ag.Sc.

L. Glauert, B.A., F.G.S.

E. P. Hodgkin, B.Sc., D.Sc., F.R.E.S.

C. F. H. Jenkins, M.A.

W. M. McArthur, B.Sc.

R. T. Prider, B.Sc., Ph.D., M.Aust.I.M.M., F.G.S.

J. Shearer, B.A., M.Sc., F.Inst.P.

A. C. Shedley, B.Sc. (For.).

Council 1956-1957

President A. F. Wilson. D.Sc.. F.G.S.

Past President D. E. White. M.Sc., D.Phil., F.R.A.C.I.

Vice-Presidents A. J. Millington. D.Sc. (Agric.).

E. P. Hodgkin, D.Sc., F.R.E.S.

Joint Hon. Secretaries B. F. Glenister, M.Sc., Ph.D.

W. M. McArthur. B.Sc.

Hon. Treasurer R. D. Royce, B.Sc. (Agric.).

Hon. Librarian G. G. Smith, M.Sc.

Hon. Editor D. E. White. M.Sc.. D.Phil.. F.R.A.C.I.

Alison M. Baird, M.Sc.

G. H. Burvill, M.Ag.Sc.

L. Glauert, B.A., F.G.S.

J. E. Glover, B.Sc.. Ph.D.

C. F. H. Jenkins, M.A.

R. J. Little.

R. T. Prider, B.Sc., Ph.D., M.Aust.I.M.M., F.G.S.

S. E. Terrill. B.Sc., A.R.A.C.I., F.G.S.