Journal
of the
Royal Society of Western Australia

Vol. 43

Part 1

1.—Complex Jointing on the Shield Margin near Darlington,
Western Australia.

By Michael J. Frost*
Manuscript received 16th June, 1959

Statistical investigation of the jointing in the
Precambrian shield to the east of the Darling
Fault shows the presence of two joint systems
indicating a period of E.-W. compression fol-
lowed by E.-W. tension. A possible third system
and some late slickensides suggest periods of

horizontal shear. Other slickensides indicate
east-side-down faulting associated with the
intrusion of the micro-gabbro dykes. It is sug-

gested that these features are associated with
the up-arching and collapse of an anticlinal
warp prior to the more westerly collapse which
produced the Darling Fault.

Introduction

The Precambrian shield of Western Australia
is abruptly terminated to the west by a series
of fracture lines of which the Darling Fault is
the most important. Immediately east of the
Darling Fault is a narrow strip of early Palaeo-
zoic or Proterozoic sediments, and to the west
geophysical work suggests the presence of be-
tween 20,000 feet and 40,000 feet of sediments
(Thyer 1951) of which the top 2,000 feet are
known from bores to consist of sandy shales,
black shales and calcareous sandstones of
Eocene and Cretaceous age. To the east the
shield forms a plateau rising to an average
height of about 1,000 feet above the western
low-lying Swan Coastal Plain. Much of the
shield is covered by laterite and is thus not
available for direct study but along the margin
of the plateau and in river valleys leading from
the scarp excellent exposures occur and some
of these have been mapped (Clarke & Williams
1926: Prider 1941; Davis 1942; Themson 1942).
From this mapping it has been possible to
cbtain some idea of the structure of the region
and the history of the marginal faulting. One
of the most recent attempts is that of Prider
(1952) who says “The Darling Scarp then is
an expression of a differentially eroded mono-
clinal structure. It was transcurrent in the late
Archaeozoic, but during post-Proterozoic times
there was the development of a downwarp to
the west (initiated by further movement on the
Darling Archaean Fault) which has been pro-
* Department of Geology, University of Canterbury,
Christchurch, New Zealand. Formerly, Department of
Geology, University of Western Australia, Nedlands,
Western Australia.

1

eressively sinking and being filled with sedi-
ments.” He also adduces evidence that there
was “a west block south and down movement
in Archaeozoic times,’ which was continued in
the late Proterozoic. Other ideas have been put
forward by Jutson (1934) and Prider (1941 and
1948). It was with the belief that a statistical
analysis of the complex jointing in this region
could add to the knowledge of the Darling Fault
that this study was undertaken.

Four quarries near Boya Siding (12 miles
east-north-east of Perth) in the valley of the
Helena River were chosen for study (Fig. 1).
These are all in an extensive batholith the
petrology, structure and extent of which have
been discussed by Wilson (1958). Flow struc-
tures are not common near the quarries but
where observed flow layers strike predominantly
N.-S. and dip steeply to the east. Locally two
varieties of granite may be distinguished, a
coarse-grained porphyritic microcline granite
and a fine even-grained granite of similar
mineralogical composition in which the former
is xenolithic. In other areas an intermediate
variety is found. It seems probable that these
do not represent separate granites but only
repeated intrusion of deeper material into an
upper partly solidified crust. Both are cut by
pegmatites. The area is also crossed by numer-
ous shear zones, many partially or completely
silicified, varying in width from several feet to
several chains. The majority of these have a
NNE.-SSW. strike and are believed to be of
earlier origin than the main jointing. Of later
origin is a swarm of NNW.-SSE. striking micro-
gabbro dykes. End-stage products from these
have coated the earlier formed joints and the
dykes themselves have caused limited alteration
of the adjoining granite but the total effect has
been slight.

Methods of Compilation and Study
Field Methods
In this study it was soon realised that natural
exposures and shallow cuttings seldom gave data
of sufficient accuracy or gave a Sufficient num-
ber of joints for any but the most important

joint systems to be evaluated. For this reason
four quarries distributed as evenly as possible
over the area were chosen for detailed investi-
gation. Their location is shown on Fig. 1.

In each quarry as many joints as possible
were measured. In order to avoid as far as
possible subjective selection the following tech-
nique was used. In any one quarry on each level
a number of points on the wall were marked
and later mapped by tape and compass. These
points were chosen on a horizontal plane, to
within several feet, in such a way as to make
the lines between any two successive points close
to and approximately parallel to the nearest
quarry wall. Offsets were taken from these lines
to the walls, and for every joint along these
lines the dip and strike were measured in the
usual way. In irregular joints the strike and
dip of the main surface rather than the mean
was taken. Joints were not measured when
there was any possibility that they had been
disturbed by blasting, soil creep, etc. All planar
features such as dyke walls, pegmatites, etc.,
were also measured. All measurements finally
used were made by the author.

At the same time as the dip and strike of
each joint was measured the following features
were noted:

Texture—the relative roughness or smooth-
ness of the joint face.

GREENMOUNT QUARRY

MOUNTAIN QUARRY

No 4 GOVERNMENT QUARRY (BOYA)
STATHAMS QUARRY

<<“ YDE®D

SCARP
0 Piste hecho s)
Miles
Perth gp

al Fremantle

Planeness—the degree of approach of the
joint surface to a plane, or the lack of
curvature.

Veneer—the nature and thickness of the
veneer on a joint.

Length—estimation of relative length.

When slickensides were present their trend
and plunge were measured and recorded. In
the majority of joints only a few of these
features could be observed.

It is of interest, having shown the methods
adopted to avoid bias, to examine the possibilities
for bias that remain. The direction and vertical
range of the line of traverse must have an im-
portant effect. Joints with strikes parallel to
the line of traverse will obviously be biassed
against in favour of those with strikes at an
angle. This is a bias that has been largely
ignored by previous authors. The bias for the
near vertical joints has been minimized to a
large extent by measuring all the joints in a
quarry, thus having traverses with a number
of different bearings. In those examples where
this was not possible, or where the quarry was
distinctly elongate, this bias must still be taken
into consideration.

Flat-lying joints with strike parallel to the

traverse present a more serious problem, as it
is obviously impossible to traverse vertically as

LOCALITY MAP

Scarp

500 Miles

Darling

Fig. 1.—Map showing location of quarries in relation to the Darling Scarp.

the angle between the shear planes has been
demonstrated by Parker (1942) who shows that
the angle decreases when the stress acting along
the axis of minimum stress becomes tensional,
and by Grigg (1936) who suggests that the
angle increases under high confining pressure.
Another point of some interest is that in no
quarry are the diagonal joints exactly bisected
by the E. joint, the two angles on either side
of the E. joint differing by up to 10°, and the
larger angles always being associated with the
same joint. This may indicate that the E. joint
was controlled by a then present ‘rift’ direction.
The average dihedral angle between the diagonal
joints is close to 60° which if the joints are
considered as shear planes is close to or little
less than that which would be expected (Hil-
genberg 1949). It is of interest to note that
if the two pairs of diagonal joints were not of
contemporaneous origin, as the evidence seems
to indicate, and if the one set in no way offsets
or deflects the other, that it might be expected
that the one which was later would be the more
regular in that it would be affected by fewer
stresses. If the data for the two diagonal sets
are examined it may be seen that the WNW.-
ENE. set is the less regular in every way. This
may be taken as some indication that it was
the earlier to form. Other points are, the sur-
prising difference between the ENE.-WNW. joint
relations in the Mountain and Boya quarries
which is not reflected in the NNE.-NNW. joint
relations, which may possibly be explained as
akove, and the facts that the dihedral angle
between the N. and E. joints is about 90° and
the angle between the two lines along which
the diagonal joints cross is about 30° varying
POM a tOlos)

TABLE 1

Table of dihedral angles between modal joint
planes in various quarries.

|
| Green- | Moun-

Joints Boya Stathams Mean
mount | tain Directions
| |

PENDS leg 882 | 20us |) 53s) (60) to, N.
ENE.-WNW. 64 | 37 95 | (116 tc

E.-WNW. | 26 Ie | 43 (56. tc

E-NNE. 60. (65?) yam TQ
NNE-NNW 128 | (134?) 105 149

E.-NNW. 68 | 68 53 res

NE. 94 86 95

B-B! 30 15 38 35

as and Bi are the hypothetical axes of intermediate
strain on the E. joint.
* Average deviation.

Support for the suggestion that the WNW.-
ENE. joints were there earlier is found in the
jointing of one of the silicified shear zones. Here,
although the rock is well jointed only the NNW.,
NNE. and EW. sets are represented. Although
alternative explanations are possible, it is be-
lieved that silicification took place between the
two periods of joint formation.

The poles of the planes forming the joint
patterns of the four quarries are plotted on the
equi-area projection of Fig. 3a. It is useful to
remember when reading this diagram that, if
we accept a standard deviation of the mean of

2°, a difference of 5° may be considered evidence
and a difference of 10° may be considered proof
of a significant variation. It will be noticed that
significant variation takes place in the position
of almost every joint. The most interesting
point is the close grouping of the NNW. joint
sets, the dominance and persistence of which
has already been noted in the field. This seems
to suggest that they were later reinforced, or
possibly even reoriented by a strong and un-
usually constant force after their original forma-
tion. It may also be significant that the joint
set with the next highest maximum is the
WNW., an adjacent set.

The last method in which the joint patterns
will be compared is that based upon the
theoretical positions of the principal axes of the
stresses which may have produced the joints.
This, regardless of its theoretical implications
which will be discussed later, is a very conveni-
ent way of comparing the orientations of the
patterns as distinct from the planes. For this
purpose it is assumed that the diagonal joints
represent shear planes intersecting on the inter-
mediate stress axis. It is also assumed that the
axis of maximum stress is normal to the axis
of intermediate stress and in the plane contain-
ing the axis of intermediate stress and bisecting
the dihedral angle between twe diagonal joints.
The axis of least stress is taken as normal to
this plane. As will be shown later, these as-
sumptions are justified and probably approach
the truth. The three principal axes are by
construction at 90° to each other. For this
reason a plane containing two of the axes and
the direction of one of the contained axes is
sufficient to fix the position of all the axes. For
the purpose of comparison the two diagonal sets
were treated separately, i.e. those with the acute
angle to the north were plotted on one diagram
(Fig. 3b), the others on another (Fig. 3c). Each
was shown by plotting, on a cyclographic pro-
jection, the plane containing the axes of maxi-
mum and minimum normal stress and on this
plane drawing a ‘V’ to point to the pole of the
axis of maximum normal! stress. This may be
considered as the cyclographic projection of a
simplified version of the figures used by Mc-
Kinstry (1949) to illustrate stress relationships.

The close similarity, in both diagrams, of the
directions for Greenmount and Mountain Quar-
ries is noticeable. In that for N.-S. maximum
normal stress Stathams is also similar while
Boya differs only in the plunge of the axis of
maximum normal stress. In the diagram for
E.-W. maximum normal stress there is a much
greater variation in the strike of this axis and
a considerable variation in the positions of the
planes for both Boya and Stathams Quarry.
This rather tends to support the argument that
the NNE. and NNW. diagonal joints are of later
formation, although it must be noted that these
data and those used before are to some extent
correlated. It may be noted that a 40° west
down rotation on a NW.-SE. axis brings all the
Boya data into a position comparable with that
from other quarries. It might also be noticed
that a similar rotation of only 35° will serve to
bring the errant data of Stathams into conform-
ity if applied to this only.

long, straight, twisted, lower lobes cordate.
Ovary globular, sessile, style straight, 3.5 mm
long.

Distinguished from other tuberous rooted
species by its open, dichotomously branched in-
florescence and by the possession of straight,
not curved anthers almost equal in length.

Type locality—Region about Mt. Lesueur, S.
of Irwin District, W.A. in sand plain vegetation
on sandy loam or clay soils.

Syntypes: S. of Mt. Lesueur N. H. Brittan 55/27
2.x.1955 (2 sheets).

The holotype (Fig. 2) is in the Herbarium,
Royal Botanic Gardens, Kew. The other sheet
(Plate II) will be lodged in the Herbarium, De-
partment of Agriculture, Perth.

(iii) Thysanotus rectantherus N H. Brittan sp.

nov.

Holotype: 15 miles S. of Kulin, N. H. Brittan
594/23-1, 6.xi.1954 (PERTH).

Herba perennis, rhizoma brevis, recta. Radices
fibrosae haud tuberosae. Folia radicalia 2-3,
20-30 cm longa, plana, glabra, ad basin mem-
branaceo marginata, 2-3 bractis membranaceis
latis acutis circumdatis.

Inflorescentia paniculata circa 14-17 cm longa,
scapum glabrum, teres, bractae angustae, acutae,
membranaceo-marginatae. Umbellae termi-
nalis 1-2 floris, bractae angustae deltoidae circa
3mm longae. Pedicellae in floris sensim dila-
tatae, articulo infra medium. Tepales exteriores
11 mm longae angustae-lanceolatae, acutae,
membranaceo-marginatae: tepales interiores el-
lipticae anguste fimbriatae. Stamina 6 aequalia,
antherae 5.5-6 mm longae, atropurpureae, vix
curvatae haud contortae, filamentis 2 mm longis,
0.75 mm supra basin antherae insertis, loculis
basalis acutis. Ovariuwm subcylindricum, sessile,
stylus 6 mm longus, rectus, apice excepto. Cap-
sulam maturam et semina haud video.

Perennial, short erect rootstock, fibrous
roots without tubers. Leaves 2-3 radical, 20-30
cm long, flat, glabrous, expanded into mem-
branous wings at the base, surrounded by 2-3
broad scarious bracts ending in acute points.
Inflorescence paniculate, ca. 14-17 em tall, scape
glabrous terete, bracts narrow-acute with mem-
branous margins. Umbels terminal 1-2 flowered,
bracts narrow-deltoid, ca. 3 mm long. Pedicels
expanding imperceptibly into the flower, articu-

lated just below the middle. Outer tepals nar-
row-lanceolate, acute, with membranous mar-
gins, 11 mm long: inner tepals elliptical with
narrow fringe. Stamens 6 equal, anthers 5.5-6
mm long, dark purple, slightly curved _hot
twisted, filaments 2 mm inserted 0.75 mm from
the base of the anthers, basal projections of the
loculi acute. Ovary subcylindrical, sessile, style
6 mm long, straight, except at the apex.
Ripe capsule and seeds not seen.

Superficially resembles some of the tuberous
rooted species, e.g., T. thyrsoideus, but is dis-
tinguished by the leaves twice as long as the
scape, by the possession of six equal and erect
anthers and the very narrow fringe to the inner
perianth segments.

Type locality—S. of Kulin, Stirling District,
W.A.

11

Syntypes—15 miles S. of Kulin, in lateritic
sand, N. H. Brittan 54/23, 6.xi.1954. (2 sheets).
The holotype (Fig. 3 and Plate III) will be
lodged in the Herbarium, Department of Agri-
culture, Perth, the other sheet is in the
Herbarium, Royal Botanic Gardens, Kew.

(iv) Thysanotus pseudojunceus N. H. Brittan
SD=nov.

Holotype: N. of Nannarup, W.A. N. H. Brittan
56/3-2, 5.xi.1956 (PERTH).

Herba perennis, rhizoma parva vel nulla,
fibrae radicales fasciculatae, gracillimae haud
tuberosae. Folia nulla, bractae radicales 2-3
circa 8-10 mm longae, acutae, membranaceo-
marginatae, purpureae. Caules teres, ad basin
purpureae 30-36 cm longae, plantae juvenis-
simae ramosae dichotomum, annosae ramosae
mcnopodiale: bractae sterilae, acutae, fuscae,
4-5 mm longae, ramis horizontaliter ad basin
vel erectis. Umbellae terminalae, pedunculatae,
2-4 floris, bractae 2-2.5 mm longae, acutae.
Pedicellae 8-9 mm longae adscendentae vel
cernuae, articulatae prope basin. Perianthum
ut in genusve, tepales exteriores 10--12 mm
longae, lanceolatae, mucronatae, anguste mem-
branaceo-marginatae, atro-purpureus supra et
infra: tepales interiores ellipticae, fimbriatae.
Stamina 6: tria antherae exteriore; 3 mm
longae, luteae vel purpureae ad basin, con-
tortae, ad latus ventras inferne squamatae: tria
antherae interiores 6 mm longae, pallido-pur-
pureae, curvatae, contortae. Ovarium cylindri-
cum sessile, stylus curvatus 7 mm longus. Cap-
sula cylindrica 5-6 x 3 mm periantho persis-
tens turgido inclusa.

Perennial, little or no rootstock, roots clus-
tered, fine, not tuberous. Leaves nil, 2-3 bracts
at base of stems ca. 8-10 mm long, acut2, mem-
branous margined, purple, Stems terete,
purple at the base green above 30-36 em long,
first branches branching dichotomously, sub-
sequent seasons monopodially branched: barren
bracts usually not less than 5 cm from root-
stock, dark, acute, 4-5 cm long, branches hori-
zontal then curving into erect position. Um-
bels pedunculate, terminal, 2-4 flowered, bracts
2-2.5 mm long, acute. Pedicels 8-9 mm long,
recurved in fruit, articulated close to base.
Perianth as in genus, cuter tepals 10-12 mm
lanceolate, mucronate, with narrow mem-
branous margins, dark purple above and _ be-
low: inner tepals elliptical, fringed. Stamens
6: 3 outer anthers 3 mm long, yellow with
purplish base, twisted, with scales on lower
cuter surface: 3 inner anthers 6 mm long, pale
purple, twisted, curved. Ovary cylindrical, ses-
sile, style curved 7 mm long. Capsule cylindri-
cal 5-6 x 3 mm, enclosed in persistent turgid
perianth.

The only Western Australian species with the
branches horizontal at the base and then
ascending. Also distinguished by the purple
colour of the outer tepals in the bud stage.

Type locality.—Coastal heaths on peaty sand
N. of Nannarup, W.A.

Syntypes.—N. of Nannarup, ca. 12 miles NE.
of Albany, N. H. Brittan 56/3, 5.xi.1956 (6
sheets). The holotype 56/3-2 (Fig. 4 and Plate
IV) will be deposited in the Herbarium, Depart-
ment of Agriculture, Perth, sheet 56/3-4 is

in the Herbarium, Royal Botanic Gardens, Kew,
another sheet will be deposited in the Herb-
arium, Royal Botanic Gardens, Melbourne. chine
remainder will be housed in the Herbarium,
Botany Department, University of Western Aus-
tralia.

Other specimens.—7 miles W. of Albany,
N. H. Brittan 14/2, 14.xii.1950; 5 miles E. ot
Alexandra Bridge, Blackwocd River, N. H. Brit-
tan 59/11, $.1.1959, both in Herbarium, Botany
Department, University of Western Australia.

(v) Thysanotus brevifolius N. H. Brittan sp. nov.

Holotype —32 miles N. of Albany, N. H. Brittan
58/32-4, 3.xi.1958 (PERTH).

Rhizoma parva, fibrae radicales haud tuber-
osae. Folia radicalia plura circa 20-25, 8 cm
longa, glabra, angusta, complanata apice
hebesce, bractae membranaceae latae ad
apicem acutae circumdatae. Scapum 18-24 cm
longum, teres, glabrum, bracta sterilis una
cirea 4 cm apicem inferne infrequentibus.
Inflorescentia multifiora, circa 20-50 floris.
Umbella una, bractae circa 5 mm longae, late
cuneatae, exteriores virides, anguste mem-
branaceo-marginatae; interiores plerumque
membranaceae. Pedicellae circa 8-10 mm
longae. Perianthum: tepales exteriores 6-7 mm
longae, 2 mm latae, late membranaceo-margi-
natae, tepales interiores ellipticae, fimbriatae.
Stamina 6: tria antherae exteriores 2.5 mm
longae, tria interiores 4 mm _longae, loculis
ad basin eminentibus, apice semicircularis.
Ovarium cylindricum, stylus rectus, ad apicem
anguste obconicus brevibus trilobatus. Capsu-
lam maturam et semina haud video.

Plant with small rootstock with fibrous roots.
Radical leaves several, ca. 20-25, up to ca. 8 cm
long, glabrous, narrow, flat with blunt apex,
surrounded by broad membranous bracts with
acute apices. Scape terete, glabrous, 18-24 cm
tall, rarely with single sterile bract some 4 cm
below the apex. Inflorescence many flowered,
ca. 20-50 flowered, umbel one, bracts broadly
cuneate ca. 5 mm long, outer ones green with
narrow membranous margins, inner ones
largely membranous. Pedicels ca. 8-10 mm
long. Perianth: outer tepals 6-7 mm long, Ca.
2mm broad with broad membranous margins;
inner tepals elliptical, fringed. Stamens 6:
outer three 2.5 mm long, inner three 4 mm
long, with short projecting lobes at the base,
rounded apex. Ovary cylindrical, style straight,
narrow obconic at the avex, shortly trilobate.
Ripe capsule and seed not seen.

Differs from T. glaucus in its single multi-
flowered umbel, flowers with six stamens and the
possession of some tuberous roots.

Type locality.—South Stirling
Stirling District.

Syntypes.—South Stirling sand plain, 32
miles N. of Albany, N. H. Brittan 58/32.
3.xi.1958 (6 sheets). The holotype 58/32-4 (Fig.
5, nos. 3 and 4) will be deposited in the Her-
barium, Department of Agriculture, Perth,
sheet 58/32-2 (Plate V) will be deposited
in the Herbarium, Royal Botanic Gardens,
Kew, sheet 58/32-3 will be deposited in the Her-
barium, Rceyal Botanic Gardens, Melbourne.
The remainder will be housed in the Botany De-
partment, University of W.A.

sand plain,

12

Other specimens.—Moist valleys, Stirling R.
F. Mueller s.d. (Fig. 5, No. 1). Source of
Blackwood River, Cronin s.n. 1889 (Fig. 5, No.
2) (MEL).

(vi) Thysanctus arenarius N H. Brittan sp. nov.

Holotype—Cape Naturaliste, N. H. Brittan
20/5, 20.xii.1950 (PERTH).

Herba perennis, rhizoma_ brevissime, fibrae
radicales paucae, crassae, haud tuberosae.
Folia radicalia tempore florationis praesentia
(viridis), 23 cm longa, plana, paucis ciliis
marginata. Caules 50-70 cm longae aut
longiores, rigidae, teres, striatae, ad basin hir-
sutae cum pilis brevibus patentibus vel glabrae
superne raro totum glabrae: ramosae monopo-
diales, ramulis strictis, adscendentibus. Um-
bellae terminales, 2- raro 3- floris. Bractae
membranaceae, acuminatae 3 mm longae. Pedi-
cellae circa 10 mm longae, prope basin articu-
latae, erectae vel cernuae. Flores ut in
genusve: tepales exteriores circa 15 mm
longae, anguste linearae, 2 mm latae, mucro-
natae; tepales interiores ellipticae, fimbriatae.
Stamina 6, declinata: tria antherae exteriores
5.5 mm longae, strictae, parum curvatae; tria
antherae interiores 9 mm longae, curvatae, con-
tortae. Ovarium subglobosum, sessile, stylus 11
mm longus, curvatus, declinatus. Capsula cylin-
drica 5 mm longa, perianthum laxo adhaerens
inferne. Caules ad nodos inferiores radicosae.

Perennial, short rootstock, roots few, thick,
not tuberous. Leaves present at time of flower-
ing, up to 23 cm long, flat, sparsely ciliate
margined. Stems 50-70 cm or longer, stiff,
terete, striate, ridges hirsute towards the base
with short patent hairs, tending to be glabrous
above, occasionally whole plant glabrous;
branching monopodiaily, branches simple as-
cending. Umbels terminal, usually 2 occa-
sionally 3 flowered. Bracts acuminate, mem-
branous, 3 mm long. Pedicels ca. 10 mm long,
articulated near the base, recurved in fruit.
Flowers as in the genus: cuter tepals ca. 15 mm
long, narrow-linear, 2 mm wide, mucronate;
inner tepals elliptical, fimbriate. Stamens 6,
declinate: 3 outer anthers 5.5 mm iong, straight
to slightly curved; 3 inner anthers 9 mm long,
curved and twisted. Ovary subglobose, sessile,
style 11 mm long, curved, declinate. Capsule
cylindrical, 5 mm long, the persistent perianth
remaining adherent above, splitting round the
capsule. Stems tending to root at nodes.

Differs from T. dichotomus in the monopo-
dial branching, hirsute ridged stems and lack
of large rhizome. It is the only W.A. species
which shows the tendency to produce vegetative
buds in the lower axils of the stems.

Type locality—Geographe Bay, N. Warren
district, on coastal sand with Agonis flexuosa.

Syntypes.—Cape Naturaliste, N. H. Brittan
20/5, 20.xii.1950 (Fig. 6, No. 1, and Plate VI)
will be deposited in the Herbarium, Department
of Agriculture, Perth: Busselton, WN. dH.
Brittan 20/1, 20.xii.1950 is in the Herbarium,
Royal Botanic Gardens, Kew: 4 miles S. of
Mandurah, N. H Brittan 55/36, 5.xii.1955 (Fig.
6, Nos. 2 and 3) will be deposited in the Her-
barium, Royal Botanic Gardens, Melbourne.

Other Specimens—Mt. Eliza, Perth, W.A.:
L. Preiss, 1564, 2.x.1839 (CP, LD, LE., MEL, S)—
all these specimens have been incorrectly at-
tributed to T. anceps Lindl. (Lehmann,
Plantae Preissianae, ii.37 1846): Shark Bay to
Murchison River, F. Mueller sn. Oct. 1877:
Greenough River, F. Mueller sn. Nov. 1877
(MEL): Claremont, W.A. Andrews, 999,
29.xii.1901 (BM): North of Muchea, N. H. Brit-
tan 52/65, 30.xi.1952: Salt River, N. H. Brit-
tan 58/30, 2.xi.1958: N. of Scott River, N. H.
Brittan 59/6, 17.1.1959 in Herbarium, Botany
Department, University of W.A.

(vii) Thysanotus spiniger N. H. Brittan sp. nov.

Holotype: Hill River, N. H. Brittan 52/39,
24.ix.1952 (vegetative) and Botany Depart-
ment experimental garden, N. H. Brittan
58/39, 8.xi.1958 (flowers) (PERTH).

Herba perennis, rhizoma crassa brevis, fibrae
radicales fasciculatae, elongatae, haud tuber-
osae. Folia plura ante florationis marce-
scentia, circa 15 cm longa, complanata, margi-
nis et dorsum brevis hirsutibus. Caules vel 40
cm longae, rigidae, teres, nudae, sulcatae, pilis
brevibus paulum recurvatis hirsutibus; copiose
ramosus, ramis brevibus dichotomose ramosus,
ramulis ultimis in aculeis brevibus terminibus.
Bractae inferne folioides, ciliato-marginatae,
bractae superne breves, acutae. Umbellae 1-2
floris, bractae breves deltoides. Perianthum:
tepales exteriores late lineari-lanceolatae,
mucronatae, anguste membranaceo-margina-
tae, circa 13 mm longae, 3 mm latae; tepales

-interiores ellipticae, fimbriatae. Stamina 6,
antherae aequales, 8 mm longae, rectae haud
ccntortae, stylus adversus declinatus. Ova-
rium subglobosum, stylus curvatus. Capsulam
maturam et semina haud video.

Perennial, short thick rhizome, roots clus-

tered, long not tuberous. Leaves flattened, sev-
eral per shoot at the start of the growing
season, up to 15 cm long, shortly hirsute on
margins and on ridges on the back, dying be-
fore flowering period. Shoots stiff, up to 40 cm
tall, naked, terete, ridged, with short, some-
what recurved hairs on the ridges; copiously
branched, side branches short and branched
again dichotomously, final branches terminat-
ing in a short prickle-like cone. Lower bracis
leaf-like, ciliate margined, upper ones short,
acute. Umbels 1-2 flowered, bracts short, del-
toid. Perianth: outer tepals broadly linear-
lanceolate mucronate, with narrow membran-
ous margins, ca. 13 mm long, 3 mm wide; inner
tepals elliptical, fimbriate. Stamens 6, equal,
anthers 8 mm long, straight not twisted, de-
clinate opposite to the style. Ovary subglobose,
style curved. Ripe capsule and seed not seen.

Differs from T. dichotomus in its more rigid
habit, the semi-pungent ends of the branches
not bearing flowers and the six, straight equal
length anthers.

Type locality—Hill River area, N.W. Avon
District.

Holotype—Hill River, N. H. Brittan 52/39,
24.ix.1952, flowers from plant cultivated in ex-
perimental garden of Botany Department, Uni-
versity of W.A., N. H. Brittan 58/39, 18.xi.1958
(Fig. 7, Nos. 2, 3, and 4 and Plate VII) will

13}

be deposited in the Herbarium, Department of
Agriculture, Perth.

Other specimen: Nr. Mogumber, N. H.
Brittan 55/13, 24.ix.1955 in Herbarium, Botany
Department, University of W.A.

(viii) Thysanotus cymesus N. H. Brittan sp. nov.

Helotype.—30 miles S. of Kulin, N. H. Brittan
58/22-1, 27.x.1958 (PERTH).

Herba perennis ?. Fibrae radicales fascicu-
latae, tuberae ad 5-7 cm ab caule productae.
Caulis vaginis foliis vetustis circumdatis. Folia
2-3, teres, glaber, 20-30 cm longa. Scapum
haud ramosum ad 20-25 cm, inflorescentia
cymeosa superne ferens. Umbellae terminales,
bractis circa 4 mm longis, deltoideis, latis mem-
branaceo-marginatis circumdatis. Umobellae
4-6 florae, pedicellis 10 mm longis, erectis vel
cernuis. Perianthum: tepales exteriores 9-10
mm longae, linearae, 1.5 mm latae, 5- venosae,
breviter mucronatae: tepales intericres 8-9 mm
longae, ellipticae, fimbriatae. Stamina 6,
antherae inaequales, rectae haud contortae:
tria interiorae 4.5 mm longae; tria exteriorae
3 mm longae; filamentae 1.5 mm longae. Ova-
rium sessile, globosum-subcylindricum. Stylus
strictus 4 mm longus. Capsulam maturam
haud video.

Perennial herb ?. Roots fibrous with few
elliptical tubers some 5-7 cm from the stock.
Stock surrounded by leaf sheaths of old leaves.
Leaves 2-3, terete, glabrous, 20-30 cm long.
Scape simple to 20-25 cm, bearing a eymose in-
florescence. Umbeis terminal, enclosed within
deltoid bracts ca. 4 mm long with wide mem-
branous margins. Umbels 4-6 flowered, pedicels
10 mm long, becoming reflexed in fruit. Outer
perianth 9-10 mm long, linear, 1.5 mm. wide,
terminating in a short mucrone, 5- veined;
inner perianth 8-9 mm long, central part ellip-
tical with fringe some 4 mm wide. Stamens 6,
anthers unequal, straight not twisted; three
outer 3 mm long, three inner 4.5 mm long,
filaments 1.5 mm long. Ovary sessile, spheri-
cal-subcylindrical. Style straight, 4 mm long.
Ripe capsule not seen.

Its affinities appear to be with T. scaber from
which it can be distinguished by the glabrous
leaves and the open cymose infiorescence.

Type locality—Sand plain between Kulin
and Lake Grace, W.A.

Syntypes.—30 miles S. of Kulin in sand plain
INGE WENN, NIG, Sip CN, YS
27.x.1958 (2 sheets). The holotype 58/22-1
(Fig 8 and Plate VIII) will be deposited in the
Herbarium, Department of Agriculture, Perth
and 58/22-2 will be deposited in the Herbarium,
Royal Botanic Gardens, Kew.

Other specimen.—Ongerup-Borden in low
heath vegetation, N. H. Brittan 58/23, 28.x.1958
in Herbarium, Botany Department, University
Oi Wie

Acknowledgments

The author wishes to express his gratitude
to Dr. R. Melville of the Herbarium, Royal
Botanic Gardens, Kew, with whom he had
helpful discussions while working at Kew in
1957, and to Mr. P. R. C. Weaver of the Uni-
versity of Western Australia for assistance with
the Latin diagnoses.

Fig. 1—Thysanotus formosus N. H. Brittan.

1.—Whole plant. 2.—Detail of sessile umbel. 3.—Flower. 4.—Detail of stamens. 5.—Umbel with capsules.
inclusive from holotype. 5.—from collection (locality unspecified) Wellstead 1900 (K).

14

nero

PLATE IV. fee
Thysanotus pseudojunceus N. H. Brittan holotype.

24 Hh

SS
_——"

Tin

\, oy,

PLATE VII
Thysanotus spiniger N. H. Brittan holotype.

27

SHIT
| 20

TIT
Ma :

=
Ee?
e-

\

ea
“ine

5 P

X)
Z
a

NS

ly

Mm

IN

meal

steers sasoe

PLATE VIII

Thysanotus cymosus N. H. Brittan holotype.
29

3.—The genus Cethegus Thorell (Mygalomorphae

.

Macrothelinae )

By Barbara York Main*
Manuscript received—l0th December, 1959

The Macrothelineae occurring in Australia
are reviewed. The species Stenygrocercus
broomi Hogg and Palaevagrus fugax Simon are
shown to be con-generic with Cethegus lugubris
Thorell. Cethegus, the prior name therefore
stands for the genus. Macrothele aculeata
Urquhart is shown to be a _ misidentified
Ctenizid. The New Caledonian species, Steny-
grocercus silvicola (Simon) is also considered
to belong in the genus Cethegus. Natural
history notes are also recorded for Cethegus
fugax (Simon).

Introduction

In the group Macrothelineae (Family Diplu-
ridae, Sub-family Macrothelinae), four genera,
each with one species, have been recorded as
occurring in Australia. These are: Cethegus
lugubris Thorell (1881), Macrothele aculeata
Urquhart (1893), Stenygrocercus broomi Hogg
(1901) and Palaevagrus fugax Simon (1908).
Examination and comparison of the type speci-
mens of C. lugubris, S. broomi and P. fugar
shows that these species are con-generic.
Thorell’s name Cethegus is the prior name and
Palaevagrus becomes a synonym. Stenygrocercus
silvicola (Simon) from New Caledonia was
originally placed in the genus Macrothele
(Simon 1889) but Simon later made it the type
species cf the new genus Stenygrocercus (Simon
1892). This species also should be included in
the genus Cethegus. Berland (1924) attributed
a male specimen to this species. The specimen
described by Urauhart (1893) as Macrothele
aculeata was a Ctenizid (to be discussed in
another paper).

The present paper redefines the genus
Cethegus, reviews the Australian species, de-
scribes the male of C. fugaxr and records natural
history observations,

Taxonomy

Simon (1892, p. 185) suggested that Cethegus
was probably a synonym of either Atraz or Had-
ronyche and Roewer (1942) placed Cethegus
with the Atraceae. Genera of the Atraceae have
relatively short, bluntly pointed posterior spin-
nerets and teeth on both margins of the cheli-
ceral furrow whereas Cethegus has long taper-
ing posterior spinnerets and teeth on the inner
margin of the cheliceral furrow only (except a
few small basal teeth cn outer margin). Roewer
(1942) lists Stanwellia Rainbow and Pulleine,
with the Macrothleae: but the presence of a
double row of teeth on the upper tarsal claws
establishes its position in the sub-family Dip-
lurinae.

*Zoology Department, University of Western Australia,
Nedlands, Western Australia.

30

Cethegus Thorell 1881
Type species: Cethegus lugubris Thorell 1881.
References and Synonymyt
Cethegus Thorell 1881. Ann. Mus. Civ. Genova.
17, pp. 240-1 (nov. gen.).

Simon 1892. Hist. Nat. des Araign. I pp:
183, 186.
Hegg 1901. Proc. Zool. Soc. Lond. 1901 (2),
{Os 2a),
Bonnet 1956. Bibliographia Araneorum
(Toulouse), Dp. 1027.
Macrothele (in part) Simon 1888. Ann. Soc.

Ent. France (6) 8, p. 245.
Palaevagrus Simon 1908. Fauna Stidwest-Aus-

tral. 1, p. 365 (nov. gen.).
Bonnet 1958. Bibliographia Araneorum
(Toulouse), p. 3301.
Stenygrocercus Simon 1892. Hist. Nat. des

Araign. 1, pp. 182, 183, 185 (nov. gen.).

Simon 1902. Hist. Nat. des Araign. (Supple-
ment) v. 968.

Hogg 1901. Proc. Zool. Soc. Lond. 1901 (2)),
pp. 265, 270.

Berland 1924. Ncva Caledonia (Sarasin and
Roux), pp. 173-4.

Bonnet 1958. Bibliographia
(Toulouse), 0. 4157.

Araneorum

Diagnosis —Flattish carapace with the eyes
raised in a small transverse compact group. Pit-
like fovea sometimes distinctly recurved. Geni-
culate chelicerae, teeth on inner margin of fur-
row only (except sometimes a few small basal
teeth on outer margin). Labium almost square,
no cuspules. Maxillae without cuspules or
spinules. Legs without scopulae in female, male
with sparse scopulae. All tarsi with ventral
spines. Posterior spinnerets long and flexible,
but without subarticulations and not exceeding
length of abdomen. Male lacks any processes
on the legs.

Cethegus lugubris Thorell 1881
Cethegus lugubris Therell 1881. Ann. Mus Civ.

Genova. 17, pp. 241-3.
Simon 1892. Hist. Nat. Araign. 1, p. 186.
Hogg 1901. Proc. Zocl. Soc. London 1901
(Ys 10, AAS.
Bonnet 1956. Bibliographia Araneorum

(Toulouse), p. 1027.

‘In the literature listed here and also under the species
names references which are to catalogue lists are
only given for Bonnet, which includes a compre-
hensive bibliography.

Types——Three specimens comprising one fe-
‘male, cne male in penultimate instar and one
.juvenile, in the Civic Museum Genova. Collected
by L. M. D’Albertis in 1875 from Somerset, Cape
York, Queensland. These specimens are in spirit,
in the one vial, and are all in deteriorated con-
dition as already reported by Thorell. All type
sspecimens a deep reddish-purple (plum) colour
with the maxillae and labium yellowish; abdo-
men without any dorsal pattern. Carapace
‘length of female 8.0 mm, width 7.9 mm. This
is thought to be the specimen described by
Thorell. The carapace lengths of the immature
‘male and juvenile specimens 6.5 mm and 5.0 mm
respectively.

Two specimens in the author’s collection, from
Reedys and Cundeelee, Western Australia were
also a deep purple colour in life and are ten-
tatively identified as lugubris. The present
author also collected four small specimens which
probably belong to lugubris, in the MacPherson
Ranges, New South Wales, near the south base
of Mt. Lindsay, nearly two miles south of the
State bcundary on the road which passes
through the border check.

No male specimens are known.

Cethegus broemi (Hogg) 1901

Stenygrocercus broomi Hogg, 1901. Proc. Zool.
Soc. Lond. 1901 (2), pp. 270-3. Text-
ibs Ske

Bonnet 1958. Bibliographia Araneorum

(Toulouse), p. 4157.

Types.—One female and one juvenile in the
British Museum of Natural History. Collected
by Dr. Brocme from Hillgrove, New South Wales.
Specimens in spirit. Specimens a uniform dark
brown. Abdomen of adult with a faint dorsal
pattern of three chevrons. Carapace length of
adult 9.0 mm, width 7.0 mm. Spinnerets of the
adult specimen asymmetrical, right posterior
spinneret shorter than left and appears to have
been regenerated.

No male specimens are known.

Cethegus fugax (Simon) 1908

Palaevagrus fugax Simon, 1908. Fauna Sud-
west-Austral. 1, p. 365.
Bonnet, 1958. Bibliographia
(Toulouse), p. 3301.

Types.—Two specimens mentioned by Simon
in his descripticn of Palaevagrus fugaxr, from
Lion Mill and Geraldton, Western Australia.
Collected by the “Hamburger stidwest-aus-
tralische Forschungreise, 1905.” The Geraldton
specimen (examined by the author) now in the
Museum at Berlin.

The specimen seen was a juvenile in spirit.
Carapace length 4.0 mm. Generally a uniform
light brown colour. Male allotype (herein
designated) collected by the author, 3 miles east
of Byford on 8th March, 1959. To be deposited
in the Western Australian Museum.

Diagnosis.—Female: Indistinguishable from
broomi and also from lugubris except possibly
on colour. Fugaz in life a light dusty brown
colour. Possibly lugubris tends to have more
numercus and heavier tarsal spines than fugax

Araneorum

31

but more specimens of lugubris require examina-
tion. Size variable. The following measure-
ments given are of a female specimen, from the
same locality as allotype, selected from the
author’s collection, to be given to the Western
Australian Museum. Carapace length, 5.3 mm.
Leg lengths respectively: I, 108 mm; II, 11.5
ervonl, INOL, iss(s Seeveqie JON, allz/{0) Tooboois letziho), ‘7/fs) iachoa,
The leg formula (obtained by dividing the length
of the leg by the length of the carapace) is thus
as follows:

3.4 2.6

Spines are present on the lateral edges of the
ventral aspect of all tarsi including palp tarsus,
ventrally on metatarsus I and II, and on all as-
pects of metatarsi III and IV. All segments of all
legs have numerous long hairs and bristles. In
many instances the bristles are scarcely distin-
guishable from spines. The largest specimens ob-
served had a carapace length cf 6.0 mm. A live
specimen is illustrated in Fig. 1, Fig. 2A gives
dorsal view of animal, Fig. 2B profile aspect, Fig.
2C sternum, Fig. 2D, spinnerets.

. 1—Cethegus fugax (Simon), female specimen.
Photograph B. Y. Main.

Male—Apart from sexually functional ex-
ternal characters, such as the male palp, di-
morphism is not conspicuous. However the legs
are proportionately longer (see formula below)
and sccpulae are well developed on the tarsi of
the third and fourth legs and divided by a
medial line of bristles. A few scopulate hairs
present on tarsi of first and second legs. Male
animals brown, slightly darker than females,
with a faint golden sheen on hairs in life.

Carapace length of allotype 6.0 mm, of other
specimens examined 5.0, 5.3, 5.6, 5.7, 5.7 and 5.9
mm. Palp as in Fig. 2E. First legs (Fig. 2F) and
second legs without any modifications or pro-

cesses. Leg lengths of allotype as follows: I,
IY sean Ihe, IPG) socks IOOL, GYR Teens Ih, 22!
mm. Leg formula:

1
2.8

8
3.3

4 2
4.0 2.9

Spines are present on the same leg segments
as in the female. No spines present on palp.
Legs generally with dense covering of long hairs
and bristles.

Localities of specimens in author’s collection.

Females and juveniles —WESTERN AUS-
TRALIA: Balladonia, 43 miles east on Eyre
Highway, 1; Bullsbrook, 7 miles from on Chit-
tering Valley Road, 1; Byford, 3 miles east, 5;
Canna, 1; Carilla, 2; Coolgardie, 27 miles west,
1; Gidgegannup, 1; Merredin, 1; Mokine, 3;
Moonera Tank, 13 miles west, 1; Morawa, 1;
Necrseman, 1; Norseman, 3 miles south, 1; Paynes
Find, 4 miles west, 1; Perth, 15 miles east on
Red Hill Road, 1; Queen Victoria Spring, 11
miles South West, 1; Wialki, 1; Widgiemooltha,
1; Zanthus, 6 miles west, 1. SOUTH AUSTRA-
LIA: Between Port Kenny and Streaky Bay, 2;
Streaky Bay, 3; Warrachie, 3; Wilpena Pound, 1.

Males—WESTERN AUSTRALIA: Beverley,
1; Byford, 3 miles east, 6 (includes allotype).

Natural History

Habitat.—Specimens cf Cethegus fugax have
been observed in varying types of habitats. In
the southern part of the sclerephyllous (jarrah)
forest region (see Gardner 1944) of Western
Australia they occur in greater density than
anywhere else. Here webs are found on lightly
lateritic soils, attached to logs and small herbs
and also along the barren banks of rcads where
they sometimes present an almost continuous
shimmering blanket of web several chains in
extent. Further east in the savanah and sclero-
phyllous woodland zones (Gardner 1944) only
isolated webs have been sighted. Again they
are usually associated with gravelly cr stony
soils or sometimes yellow sands. On the Nullar-
bor Plain and Eyre Peninsular, burrows occur
on stony limestone soil which has light cover
of litter. In south-western Western Australia
Cethegus has not been fecund in the wet karri
forest (Mesophytic Forest region of Gardner)
nor in the wetter southern part of the jarrah
region, which have a dense understory and
ground cover.

The burrows of Cethegus lugubris observed by
the authcr in the MacPherson Ranges were
amongst exposed roots of trees in an overhang-
ing bank of a valley in the jungle.

Life History.—Life history data are only avail-
able for C. fugaz and this is very slight. J. A. L.
Watson found a male specimen under a stone,
in June 1957. The author collected five recently
matured male specimens from their burrows
near Byford on 8th March, 1959. Presumably
these specimens would have run with the onset
of autumn rain in April or May. A sixth male
was collected from Byford on 31st August, 1959.
It is thought that this animal remained in its
burrow until the end of winter, due to the extra-
ordinarily dry season. No egg cocoons have

32

been observed but juvenile specimens were seen
in two adult burrows at Mokine on 23rd March,
1959. It is assumed that these hatched from
egg cocoons during the summer. From analogy
with life history patterns of other Mygalo-
morphae where the seasonal sequence has been
documented it is suggested that for Cethegus
fugax the seasonal behaviour would be as fol-
lows: Mating in late autumn through to early
winter, i.e, depending on onset cf rain with
appropriate temperature; egg laying late spring
to early summer, eggs hatch mid-summer and
juveniles vacate female burrows after first
autumn or winter rains.

Since C. lugubris occurs in a region of summer
rain with a dry winter period it is probably a
summer mating spider.

Web and burrow.—The web of Cethegus fugaxr
consists of a central porticn of vertical soil-
covered strands. These are attached at their
upper ends to low shrubs, the bases of trees,
projections of soil or rocks, to the sides of logs
cr overhang in banks, depending on the habitat.
Such strands formed from the silk bound
bundles of soil excavated from the burrow pre-

sent a dense curtain like structure. Radiating
from this section of the web are numerous
threads extending for several inches. These un-

soiled threads acting as “trap-lines” for catch-
ing prey, form an extraordinarily dense mass,
within which are several funnel like tubes open-
ing near the surface. Within the centre of the
web cr below the ground these tubes unite and
terminate in a single poorly defined burrow, lack-
ing any plastering, reinforcement or closely
woven silk lining to the walls. Rejectamenta and
cast skins are freauently found in the older parts
of the webs. Burrows cf penultimate instar
and mature males have no “‘trap-lines” and the
untidy mass of partly collapsed soil coated
threads appears like an unoccupied nest. Also
the burrow terminates in a symmetrical,
smcoth walled tube, in which possibly the male
may seal itself off during the penultimate instar.

Specimens of Cethegus lugubris, collected by
the author were taken from webs similar in
structure to those of fugdaz.

Discussion

The occurrence of Cethegus in south western
Australia is of interest since it is the only Macro-
theline genus knewn to cccur west of the Eyre
Peninsular. Hadronyche has been collected by
the author (unpublished) in the Flinders Ranges
and Eyre Peninsular in South Australia. Fur-
ther east the Macrothelinae are a prominent
component of the Mygalomorph fauna extend-
ing from Cape York to Tasmania. It is also of
interest that the genus Cethegus has a con-
tinent-wide range and occurs in varied habitats.

Acknowlec gments
Collection of specimens considered in the
preparation of this paper was done with the
help of a Research Grant from the University
cf Western Australia. A visit to the British
Museum (Natural History) was made possible
threugh an International Federation of Uni-

A, dorsal aspect,
nerets, female. E,

female.

B,

Fig. 2.—Cethegus fugax (Simon).

left lateral aspect,

right palp retrolateral aspect,

female. C., sternum, labium, right maxilla, female. D, spin-
male. F, tarsus, metatarsus, tibia of right leg, male.

33

ee

versity Women Fellowship (Alice Hamilton Fel-
lowship 1958). The assistance and working
facilities given by the staff of the Arachnida De-
partment of the British Museum of Natural
History was greatly appreciated. Dr. W. Crome
of the Berlin Museum and Professor Tortonese
of the Civic Museum of Genoa kindly loaned
type specimens. A. J. Lee and W. B. Malcolm
assisted in the field by providing transport. C. A.
Gardner, J. A. L. Watson and W. H. Butler are
thanked for specimens.

References

Berland, L. (1924).—Araignées de la Nouvelle Calédonie
et des iles Loyalty. In Sarasin and Roux,
“Nova Caledonia, Zoologie.’”’ 3: 159-255. (C.
W. Kreidel: Berlin.)

Gardner, C. A. (1944) —The vegetation of Western Aus-
tralia with special reference to the climate
and soils. J; Roy... Soc. —We “Aust. “28°
xi-1xxxvii.

34

H. R. (1901).—Australian and New Zealand spiders
of the suborder Mygalomorphae. Proc. Zool.
Soc. Lond. 1901 (2): 218-279.

(1881).—Studi sui Ragni Malesi e Papuani.
Part III. Ragni dell’ Austro-Malesia e del
Capo York, conservati nee Museo civico di
Storia Naturale de Genova. Ann. Mus. Stor.
nat. Genova. 17, vii-xxvii: 1-720.

Cc. Fr. (1942).—‘‘Katalog
(‘“Natura’’: Bremen.)

Simon, E. (1889).—Etudes arachnologiques. 21e Memoire.
XXXII. Descriptions d’espéces et de genres
nouveaux de Nouvelle Calédonie. Ann. Soc.
ent. Fr. 6(8): 237-247.

(1892).—‘“‘Histoire Naturelle des Araignées.”
1. (Libraire Encyclopédique de Roret: Paris.)

“Die Fauna
359-446.

Hogg,

fThnoren, 1.

Roewer, der Araneae.”

In
(12):

(1908).—Araneae lre partie.
Siudwest-Australiens.” 1
(Gustav Fischer: Jena.)

Urquhart, A. T. (1893).—On_ new species of Tasmanian
Araneae, Proc. Roy. Soc. Tasm. 1892: 94-130.

are however other physical variables which
influence the distribution of shore animals.
These include: the mechanical action of the
waves, the pounding and tearing action of which
inhibits some organisms and favours others; the
angle of slope of the shore; the nature of the
substrate, whether hard or soft, porous or rela-
tively impermeable to both water and filamen-
tous algae; the presence of crevices in the rock.
Perhaps these are only ‘modifying factors’ for
‘zonation, but all are a significant part of the
environment, and as such may be the immediate
-factors which favour one species rather than
sanother.

More subtle however is the danger that just
because the zonation can usefully be described
‘by a universal nomenclature, the vertical move-
ment of the water line on which this system is
‘based may be thought to produce its effect in a
cuniform manner. On the one hand movement
,of the water line is itself complex, varying both
in range and periodicity and it would be sur-
‘prising if the results it produced on the shore
»environment were uniform. I will return to this
later. On the other hand the many species
which collectively produce the zonation respond
in different ways to the physical components of
‘the varying weather of the shore: temperature,
insolation, and moisture.

i Key to Symbols
} a a

H 6 2)
i Bere mB !
As | i ef ' MWh blue-green alga
8 é Ae ay 4\® Melaraphe
De of Ss 2 §& = unifasciata
s @ B®
8 e-7) :
ee 8g, faa res
G = Sa). ee Patellanax
a ‘ 4, af a® peroni
‘4
af iw Salary Bala
“s “4 7’ at aia
a
4s Ss 4, sie a Poneroplaz
Hw oy Oy 8 i} costata
OS Te ion el @ Clavarizona
BY oo 4 ine hs] Ps hirtosa
a) m oa eC is) Laurencia and
my Bua is) an small algae
i) ae o 4 i Bt a AES Patellanaxr
H ¢ % i mM 4 (4) 4: a laticostata
=i B, Ue hy sak
ae A Haliotis
a 5 Ca | u ini a & roei

Ecklonia
radiata

- lithothamnion

‘Fig. 1—Cape Leeuwin, Western Australia. The littoral
fauna and flora of a sloping rock surface with moderate
exposure to wave action.

__) Bey to Symbot

f=)

Fay

® Cy [J
HH] Ip ee Littorina
Bs i mauritiana
“| Pill Ih plue-green algae
B
“4 Chnoospora and
é Ectocarpus
| i %¢ Tectarius

Ry, ie granularis

iff SM} Puy

1

S
:

Tetraclita
squamos4

=

we Nn

Acmea profunda
Siphonaria, sp.

Septifer
bilocularis

Echinometra
mathaei

Colobocentrotus
atratus

Stomopneustes
variolaris

Chiton rusticus

Onithochiton
maillardi

Patella
chitonoides

Laurencia
flexilis
Sargassum sp.

Hypnea
chordacea

Laurencia
obtusa

lithothamnion

Fig. 2—Poste Lafayette, Mauritius. The littoral fauna
and flora of a sloping rock surface exposed to constant
wave action.

A fourth danger in the zonational approach
is that it is static. The very uniformity recog-
nised in the zonation implies a permanance.
But the individual organisms, and the groupings
of these, which collectively make up the zones
are transient and study of their dynamic inter-
actions is more likely to be informative than
observation of apparently static distributions.

This has been brought home to me forcibly
recently. Exceptionally low tides and calm
seas in January of this year destroyed the entire
animal life of certain reef platforms on our
west coast (Hodgkin 1959). It is too early yet
to know whether these species will re-establish
themselves, but other species of animals and
plants have already replaced them in the space
left vacant after the catastrophe.

This is a particularly striking example of
sudden change in distribution, but there are
many features of the patterns of shore life
which cannot be adequately explained in terms

Larval Settlement and a Place in which to Live

The eggs and early stages of many shore
organisms are distributed by the sea water; they
are planktonic and have to settle on or attach
to the rocks before they can feed and grow.
The initial settlement may be just as important
in determining the ultimate distribution of the
animals and plants as influences which operate
on the adult organisms.

I will give two examples. ‘Most European
species of barnacles have been shown to ke ere-
garious; the cypris settles and metamorphoses
most easily in areas inhabited by adults of the
same species or a closely related one, or on sub-
strata bearing traces of a former adult popula-
tion’ (Southward 1958). Thus settlement is not
random, although new surfaces are of course
colonised. Barnacle populations tend to be self
perpetuating and dense stands result, in which
few other animals live.

The limpet Patelloida alticostata occurs alone
over considerable areas of the intertidal plat-
forms on our coast. In November 1954 a party
of Zoology students cleared every limpet from
four square yards of a reef platform at Rottnest
Island. Before clearing there were 400 to 500
limpets per square yard and no other animals
or plants on the rock. Within a few days of
their removal there was a dense growth of fila-
mentous algae which was replaced by coralline
algae over the next few weeks.

Five years after the original clearing the
limpets have still not fully re-established them-
selves. Some have managed to invade the area
from the edges and a few have settled within it,
but there is a thick growth of weed and the
limpet population is still less than half that
outside where there is no weed. The change in
environment brought about by the establish-
ment of the weed prevented settlement of young
limpets. The limpet population like the barna-
cles is normally self-perpetuating, it excludes
macroscopic algae and other animals and so
maintains its own place in which to live.

The clearing experiment was repeated on a
vastly greater scale by the catastrophe of which
I spoke earlier. The entire population of
P. alticostata was destroyed on certain platforms
at Rottnest (January 1959). During the suc-
ceeding twelve months two pulmonate limpets
have invaded the virgin fields of small algae
(there is no coralline mat). Unlike Patelloida
these limpets lay non-planktonic eggs all the
year round. It is too early yet to know what
will be the final outcome.

Other examples of what appear to be self-
perpetuating populations are seen with the large
limpet Patellanazx laticostata and the mutton
fish Haliotis roei. These, with P. alticostata,
two chitons and an anemone are the principal
fauna of the outer edge of our intertidal lime-
stone platforms. In some places all occur to-
gether, in others one is dominant to the almost
total exclusion of the others, the rock is grazed
bare of algal growth and takes a form charac-
teristic of the species. In such places each
P. laticostata has a smooth shallow channel of
four or five times its own area and often con-
tinuous with those of its neighbours, and
Haliotis occupies shallow depressions of about
its own dimensions. So long as these forma-

tions persist they certainly favour the particular
species just as the bare flat surface favours
P. alticostata. These animals are not confined
to such places, but they are much more abund-
ant in them than elsewhere.

Two rather different examples of limitation
by available places in which to live are afforded
by another limpet and a sea urchin. Patelloida
nigrosulcata lives principally on the shells of

other gastropods, chiefly the large limpet
Patellanax laticostata and the mutton fish
Haliotis roei. Usually there is only one

nigrosulcata to each shell, and each shell is
grazed bare. Although not every shell has its
nigrosulcata, most have. Unoccupied shells are
usually covered by a thick growth of algae—
again an impenetrable jungle to a young limpet
trying to find a place to settle.

The tropical sea urchin, Echinometra mathaei
is abundant in the littoral on the wave-beaten
rocky shores, such as the Mauritius shore
described earlier and on the exposed west end
of Rottnest. In such places it is confined to
characteristic burrows in the rock; each animal
has its own ‘home’. On the Mauritius shores
almost every potential home in the mid-littoral
is occupied by a sea urchin and the density of
population appears limited by the number of
available homes. On the more exposed east
coast Stomopneustes competes with Echinometra
for these homes, but on the west coast
Echinometra has the field to itself. In Decem-
ber 1957 there were vacant homes above the
upper limit of distribution of the urchins;
perhaps they are occupied at other seasons, but
at that time they were above the level of con-
tinuous splash.

Weather
Weather is of such obvious importance in
determining the distribution of intertidal

animals that it has received much attention
from workers on shore organisms. I do not
want therefore to say much here, but will give
a few examples of how weather limits the verti-
cal range of organisms.

Above the intertidal platforms of our W.A.
limestone shores the rock is generally vertical
or overhanging. Here four or five feet of the
rock face, the upper mid-littoral, is generally
dominated by two limpets—Notoacmea onychitis
and Siphonaria luzonica. These limpets estab-
lish their own ‘homes’ to which they return at
low tide, but they evidently move house periodi-
cally. Each winter some of the larger limpets
get so high on the rocks that when summer
comes, with its lower sea level, calmer seas, and
higher temperatures, the more adventurous
animals are killed by heating and desiccation.
In winter the sporelings of a variety of algae
also establish themselves on this rock face. They
never survive the lower sea ievel and calmer
seas of spring. Patelloida alticostata colonises
the lower one foot of the rock, below Notoacmea
and Siphonaria. I have never seen it dying
here, but I suspect that its upper limit may be
determined by the height to which a sufficiency
of interstitial water is retained in the rock.

There is a similar seasonal mortality among
juvenile barnacles on the granite rocks of our
south-west. In early summer many empty

juvenile shells can be found on the rock above
the main mass of Balanus nigrescens. The
obvious assumption is that they have been
unable to withstand heating and desiccation at
the higher levels, however there is heavy preda-
tion by the whelk Dicathais aegrota and this is
at least an important contributory cause of
mortality.

Food

Andrewartha and Birch say that the number
of animals in a population is seldom limited by
shortage of food. The source of food of littoral
animals is twofold: firstly the food which comes
to them in suspension in the sea water and
secondly the food which grows on the rocks.
The amount of food available to suspension
feeders certainly seems to affect their rate of
erowth, as indicated by the observations of
Barnes and Powell (1950) but how far this is
important in determining distribution in bulk
is more difficult to decide. Barnacles, mussels,
serpulids all commonly are packed tight, often
entirely covering any surface on which they
settle. The situation is very different for
erazing animals. These feed actively, mainly
on algae which grow on the rocks, and I suspect
that for them food may often be a limiting
factor.

Limpets furnish examples of natural popula-
tions which do ‘consume a large proportion of
the food available to them.’ The populations of
Patelloida alticostata of which I spoke earlier
keep the algae on which they feed grazed down
to bare rock, in fact they even remove rock
particles in the process. Between 600 and 700
small to medium sized limpets per square yard
seem to be the maximum population these
resources can support. A population of this
kind covers an area of platform at Green Island,
Rottnest. But a small part of this same plat-
form carries a population of no more than 200
large limpets per square yard, again to the
exclusion of all other life. They are still fully
utilising their food resources. Obviously food
alone cannot here be limiting the actual num-
bers of animals, though it may limit the biomass
which can be supported. The figures (Table II)
suggest that recruitment to the population is
greater in the first case than in the second and
the subsequent history supports this. Perhaps
I should have discussed this example under the
heading ‘Other animals of the same kind’.
Here the natural situation approximates the
experimental one in which the food is constantly
replenished by the experimenter.

TABLE II

Numbers and Size distribution of Patelloida alticostata
from 14 sq. yard areas of rock

Greatest P. alticostata in experimental areas:
diameter
cm Ale Ad eet = BZ Ole. C25. Dla) 2,
1.0 - 1.5 68 _— 81 _ 0 al 0 1
155 220) 29 — 51 — 6 22 0 10
2.0 - 2.5 47 — 40 — 1 aly 2 8
2.5 - 3.0 1 — 0 — 24 +23 19 27
3.0 - 3.5 o — oO — 9 G@) als} 4
Total measured 145 — 172 — 46 68 34 50
Total removed 152- lio. 139 1%s ai. /68 - ober OL

Regulation of animal numbers by other
animals of the same kind would lead me to the
dangerous ground of controversy over density
dependent and density independent factors and
this I do not propose to attempt to discuss.

Other Animals of Different Kinds

The two limpets Notoacmea onychitis and
Siphonaria luzonica mentioned earlier under
‘weather’ share the same vertical zone either as
a mixed population or with one or the other
locally dominant, and they browse on the same
sparse algal flora. Are they between them
utilising the whole of the resources available to
them? If they are, what ecological differences
are there between them which enable them to
live alongside one another instead of one
replacing the other? I cannot suggest an
answer here, but in analagous situation I can
offer a partial answer.

There are two species of Siphonaria on Rott-
nest, luzonica and baconi. Sloping intertidal
rock surfaces are the exception on W.A. lime-
stone shores, but one such shore carries a mixed
population of these two species. They are
present side by side and appear to be utilising
the same resources. However counts of the
limpets show that luzonica predominates at the
higher levels and baconi at the lower.

On the sloping rocks the two species live
together in an environment that varies continu-
ously. On the more usual stepped shores there
is a separation of the two species; luzonica lives
on the undercut face which is subject to drying
at low water, baconi is seldom common but lives
either at the foot of the undercut and onto the
platform or submerged in small intertidal rock
pools. In this case the area of overlap between
the two species is small producing a zonation
that is not evident on sloping rocks.

Where two species of animal utilise the same
resources and the resources are limited, one
animal may totally exclude the other, or they
may both survive and share the resources. In
the latter case they may intermingle freely or
a patchy distribution of the two may result.
(In laboratory experiments with insects these
situations have all been achieved by suitable
manipulation of the conditions).

I have already given an example of the first
situation. Patelloida alticostata occupies con-
siderable areas of platform from which all other
animals are excluded. Given the right condi-
tions Haliotis roei will cover the rock to the
exclusion of every other animal; there is just
such a population at Yanchep near the edge of
the reef platform.

One might include the suspension feeders in
this category too, because some of the barnacles,
rock oysters, and serpulid worms occupy belts
to the total exclusion of other macroscopic
animals, but since the adults are completely
sedentary the situation is very different. One
might be justified in regarding them as plants,
ecologically.

The second situation of two animals sharing
the resources is illustrated by the two limpets
Notoacmaea and S. luzonica mentioned above,
also by the =urchins Echinometra and
Stomopneustes on the east coast of Mauritius.
The two species of Siphonaria, while sharing a

ever they are certainly not static. A clear
example of large scale change was given by the
January catastrophe of which I spoke earlier.
The same catastrophe resulted in the colonial
anemone Actineogeton almost completely replac-
ing all other life on one platform, only to be
replaced in its turn by a mixed population more
akin to that originally occupying the area. In
another place the mussel Septifer replaced
Echinometra for a time and there has been no
sign yet of recolonisation by urchins. There
have been other great changes and it will be
instructive to see whether the original popula-
tions eventually re-establish themselves; whether
in fact they represent relatively stable terminal
communities (climaxes) or whether they are
merely adventitious associations which may
arise from time to time in this situation.

I gave earlier examples of seasonal changes in
animal and plant populations on our limestone
coast, I also mentioned the observation of
Lawson (1957) that changes in the distribution
of algae resulted from seasonal differences in
ime cf lowest low water. These are of interest
not only as evidence of constant change, but
also becaus2 they indicate that these sort of
changes may be important in producing the
more persistent patterns.

But quite apart from catastrophes and sea-
sonal fluctuations, shore populations are in
dynamic rather than static equilibrium, and even
though certain associations can be recognised as
characteristic of particular levels the compon-
ent parts are in constant flux. Shore animals
and plants do not survive long individually,
seldom more than a few years, they die or are
destroyed by various agencies and are replaced
by other organisms of the same or other kinds;
there is constant competition for the available
resources. Under these circumstances it is re-
markable that shore communities do achieve a
considerable measure of continuity. These com-
munities are often so complex that any attempt
at an analysis of the interrelations of the
organisms is likely to be unprofitable. There
are however simpler situations such as those
discussed above where a knowledge of the biology
of the few species present offers an opportunity
to understand their interrelations.

One aspect of these situations bears ampli-
fication. I have described above exampies of
self-perpetuating populations covering consid-
erable areas of rock. Patchy distributions may
also be partly explained by the inability of
interlopers to establish themselves. This is
beautifully illustrated on some Mauritius shores
where the urchin Colobocentrotus is found alone
on certain rough limestone surfaces. In these
urchin preserves of a few square yards in extent
the only visible algae are those which cling to
pinnacles of rock inaccessible to the urchins and
the only other animals are a few daring inter-
lepers belonging to more active species such as
the turban shell Turbo setosus. Grazing pres-
sure prevents macroscopic algae and young
animals of other species from obtaining a foot-
hold. T. A. and A. Stephenson (1949) illustrate
the effect of brown algae in preventing settle-
ment of barnacles.

42

Presumably in such cases the animal or plant
has some slight ecological advantage over its
neighbours. Such a situation, of two self-per-
petuating populations each with slightly differ-
ent ecological preferences must result in sharp
boundaries between them. Again this is well
illustrated from a Mauritius example. I found
there that in passing from a more to a less wave-
beaten situation Colobocentrotus was replaced
by the limpet Acmea profunda, the two were
mutually exclusive. In the more exposed part
there was an equally sharp boundary in rising
up from the band of Colobocentratus to the
Acmea above. Thus the vertical zonation
showed a sharp boundary—a common feature of
zonation.

Conclusion

I warned you I would speculate, and this I
have done. I have put forward hypotheses to
explain patterns of which I have spoken and I
believe these are justified by the observations.
To test them is quite another matter. However
many of these problems are accessible to experi-
ment in field or laboratory.

Much has been done with algae and also with
some animals to find out how weather may
determine their distribution on the shore, with
valuable results. I do not doubt that weather is
the principle component of the environment
limiting the distribution of some animals.
Clearly also shore climate ultimately determines
the broad outlines of the zonation. I have tried
however to show that observed patterns of dis-
tribution commonly result from more immediate
factors: food and grazing pressure, available
places in which to live, hazards of larval settle-
ment, predation. These will be more difficult
to study in the laboratory, particularly as they
probably seldom act singly, but they should be
accessible to analysis by relating laboratory and
field experiment to observation in the field. But
if this is to be done attention will have to be
concentrated on individual species before the
interrelations of complex associations can begin
to be understood.

The littoral environment is certainly complex
and there are considerable difficulties associated
with study of its ecology. However I do not
think it is any more complex or difficult of
analysis than many terrestrial situations which
ecologists have attacked with considerable
success.

Acknowledgments

The investigations on which these observa-
tions are based were made with the help of a
research grant from the University of Western
Australia. I am indebted to many colleagues
for the helpful discussions I have had with
them on my theme and especially to Mrs. L.
Marsh who has been associated with me in
studies of the Western Australian littoral fauna.

References

Andrewartha, H. G., and Birch, L. C. (1954)—‘“The
D'stribution and Abundance of Animals.”

(University of Chicago Press: Chicago.)

and Powell, H.T. (1950).—The development,

general morphology and subsequent elimina-

tion of barnacle populations, Balanus

crenatus and B. balanoides, after a heavy

eo Bevtlement.. J, Anim. Ecol. 19: 175-

179.

Broekhuysen, G. J. (1940) —A preliminary investigation
of the importance of desiccation, tempera-
ture, and salinity as factors controlling the

Barnes, H.

vertical distribution of certain intertidal
marine gastropods in False Bay, South
mitten. Trans, teoy.. Soc: S. Afr: 28° 255-292.

Doty, M. S. (1957).—Rocky Intertidal Surfaces. In
“Treatise on Marine Ecology and Paleoecol-
ogy Vol 1. Mem. Geol. Soc; Amer. 67: 535-
Bede

Fischer-Piette, E. (1935).—Histoire d’une mouliére. Bull.
Biol. France et Belg. 69: 152-177.

Hodgkin, E. P. (1959).—Catastrophic destruction of the
littoral fauna and flora near Fremantle,
january 1959. W. “Aust. Nat. 7: 6-14,

Kohn, A. J. (1959).—The ecology of Conus in Hawaii.
Ecol. Monogr. 29: 47-90.

Lawson, G. W. (1957).—Seasonal variation of intertidal
zonation on the coast of Ghana in relation
to tidal factors. J. Ecol. 45: 831-860.

Southward, A. J. (1958).—The zonation of plants and
animals on rocky sea shores. Biol. Rev. 33:
137-177.

Stephenson, T. A. (1943).—The causes of vertical and
horizontal distribution of organisms between
tide-marks in South Africa. Proc. Linn.
Soe. Lond. 154; 219-232.

Stephenson, T. A. and Anne (1949).—The universal
features of zonation between tide-marks on
rocky coasts. J. Ecol. 3%: 289-305.

le

1.—Intertidal limestone platform, Cape Vlaming, Rottnest Island. Patelloida alticostata dominant here before
“catastrophe” of January 1959.

2.—Limestone platform with coralline algal associa- 3.—Intert:dal basalt rock (Mauritius) with Colobo-
tion on left, P. alticostata alone on right (black centrotus on thin lithothamnion across centre;
at bottom right is shadow). blue-green al above.

PLATE I

44

% ‘(olod yooy xIg) ‘sMoainq
uTyomn Bes SUTMOYS snIyTNRW 18 s10YS pasodxea Jo gsyo01 41%

seq—p

1QL

HLV Id

“QUSII 1B 97078001910 °d
“D1DIINSOLOIU ‘d Pov1l9A00-pooM
jodurtl ou, “UOJSeWT] UO

uopies

”

‘gISUIS B SoTIIvO ynq ‘peam Jo o1eq ST

[eoIdAy Ul 0707800170] XLNUD]j]aJVd—§

‘OATAINS 8$1717/0AUWO LAWIVOION 9914} PUB NaLUWO0ZN] DIuDUOYdIG 2UO
‘IoWUINS UT 19}¥M YSTIY sAoge sUOJSOWIT UO ,seuIoY,, yedut, AIduIq—'z

‘DIDIINSOLOIU DP1011210d B SOAIVO YOts puB oIBVq dre SIETTIO

4nq ‘paIaAod-poeeM 91% $17017D]T O

M} 910ON ‘ullojye[d guO4SoUTTT Jo 4avd

Jajno UO UOJUWIeYJOUP] Surysnious pue ‘v72n780017]D “g ‘1a0L $120110H—T

45

border of the praeopercle flexible, covered with
skin, and without spines or other projections.
The heads are heavy and entirely encased in
bone, but without spiny projections anywhere.
Dorsal and anal fins long, D with 17 to 20 soft
rays, A with 15 to 18 soft rays. Lateral line
high, closely following the dorsal fin. In the
species examined both lips are fringed with
fairly long barbels though on the upper lip these
fringes are present in the middle only, not on
the sides; in Jchthyscopus fasciatus these bar-
bels seem well developed on the lower lip only;
in all representatives of other genera these bar-
bels, if present, are much smaller and more or
less rudimentary.

In other characters, there is a remarkable
diversity between the various species: a first,
spiny, dorsal fin may be present or absent, the
body may be scaled or naked, barbels on the
chin may be present or absent. These differ-
ences are of a nature and magnitude which
elsewhere in the family have been considered to
be of generic value. To me, however, the most
important fact is that, notwithstanding the
apparently considerable differences, the general
appearance of the diverse species is very much
the same. So great is their general similarity
that, in my opinion, obscuring their evident
affinity by dividing them in several genera,
would only do harm. To this comes the addit-
ional argument that even now the genus
embraces five species only, so that from the
practical viewpoint it is undesirable to chop off
a few monotypic genera. Unfortunately the
classification of fishes is more seriously oversplit
at the generic and family levels than almost
any other group of animals. It has even been
seriously contemplated to propose a uninomial
system as: ‘‘the logical outcome of the tendency
towards monotypic genera” (Hubbs 1943). Apart
from the fact that my memory is quite unable to
bear the burden of the hundreds of new generic
names annually coined for familiar species and
species which would be equally well placed in
existing genera, this oversplitting, by obscuring
true affinities, is contrary to the purpose of
binary classification. Hubbs, quoted above, is
perfectly right about the uninomial system, but

Fig.

it is the tendency towards monotypic genera
which is so absurd, not the binary system that
has proved its usefulness over a period of two
centuries. The genus, to all intents and pur-
poses, is a collective unit to indicate not diversity
but affinity. Perhaps it is a state of mind,
whether one wants to stress similarities or differ-
ences. I am perfectly aware that these argu-
ments are neither new nor original, they have
been repeated many times by workers in many
fields of zoology. However, when reading the
works of several contemporary leading ichthy-
ologists, I cannot help feeling that it is not a
superfluous luxury to raise the point again and
again.

It is most remarkable that to a genus of which
for over 150 years only one species had been
known, during the last two years four species
could be added, all well-characterized.

The genus is known from the seas of India,
China, Japan, and Australia.

Key to the species of Ichthyscopus

(The figures (counts of rays and scales) have been

taken from material personally examined; the actual
range of variation may be greater).

Body

Body naked

No barbels on chin 3
Two small barbels on chin, body more or less
uniformly pale brownish or (small specimens)
with two broad dark crossbars I. barbatus
D II.17, A 17, dorsal spines rather weak, partly
embedded in skin, upper surface dark, grey or
blackish, boldly marked with white blotches of
varying shape
D III-17, A 17, the anterior spiny part of the
dorsal fin is only at its base narrowly connected
with the soft dorsal; body spotted

i

oP ee

i, spinosus
D 20, A 18, no spiny rays in D, which is continu-
ous and not notched, body with six transverse
I. fasciatus

is

Posterior nostril roundish (Fig.
Perce ie
Posterior nostril an elongated ‘slit.

I. lebeck lebeck

5b)
lebeck sannio

ichthyscopus spinosus species nova
Differential diagnosis. Differs from all other
species of the genus by the presence of a well
developed spiny dorsal fin and by its remarkable
colour pattern with small dots all over the body.

1.—Ichthyscopus spinosus sp. nov., type, from right, ca. 2/5 & nat. size.

48

Fig. 2.—Ichthyscopus spinosus sp. nov.,

Type and unique specimen, described below,
collected near Broome, N.W. Australia, received
from the W.A. Fisheries Department on Febru-
ary 19th 1954 and presumably caught not long

| before. W.A.M. regd. no. P 3639.

Description. Figs. 1 and 2 give at a glance a
general impression of the species and are, in my
opinion, worth more than a lengthy description.
The following notes are complementary to the
figures.

D III.17, A 17, C with 10 divided rays and 2
developed undivided rays, P 16, V 1.5 spine con-
cealed in skin, scales to caudal about 50 rows.
Total length 330 mm, standard length 258 mm,
length of D at base 1164 mm, length of A at base
126 mm, length of head from tip of mandible to
hind border of opercle 107 mm, greatest width
of head 78 mm, greatest depth of body (from
D 1 to anal opening) 90 mm, width of mouth
from side to side 31 mm, height of mouth (man-
Gible) 32 mm, length interorbital sinus 24 mm,
of bony armour behind 26 mm, distance between
orbits 204 mm, length of orbits 14 mm, nostrils
close together in front of the orbits, the anter-
ior nostril surrounded by a fringe, the posterior
larger one with fringe along its anterior border
only, the posterior border being formed by the
orbit. Scales somewhat rudimentary, embedded
in skin; breast, a great part of the belly, and
the nape between the posterior border of the
head armature the lateral lines and the origin
of D naked.

Colours of preserved specimen. The ground
colour is pale creamy, slightly tinged pinkish on
the under surface and round the mouth; the
whole upper half of the body including D2, P
and C is dotted with dark brown spots, the
majority of which are roundish, but on the neck
some are rodshaped or boomerang shaped and
between the eyes they are of various curly
shapes. V unspotted, posterior half of A with
brown spots, but less dark and well-defined than
those of the back. D 1 entirely blackish brown.
An ill-defined brownish area on the cheek, an-
other above the humeral appendage, gradually

49

type, from above, ca. 2/5 * nat. size.

becoming paler in the direction of D 1; an ill-
defined dark brown band over P, dark at the
upper part of these fins, traces of a brownish
half band below the middle of the soft dorsal.

Ichthyscopus fasciatus Haysom

Ichthyscopus fasciatus Haysom, 1957, p. 139
Fig. 1—Cleveland Bay, near Townsville, North
Queensland.

Differential diagnosis. Distinguished from
other species of the genus by the presence of
six dark and very distinct transverse bands on
the body, combined with the presence of scales.

This species was described and figured well,
and is evidently quite distinct. The figure more
or less suggests the presence of a humeral spine,
but in the description mention is made of: “A
fringed humeral appendage, which hardly pro-
jects beyond the pectoral axil’.

Material. I have not seen material of this
species, which is known only from the type, a
specimen of 102 mm in standard length trawled
in Cleveland Bay and preserved in the collec-
tion of the Department of Harbours and Marine,
Brisbane.

Ichthyscopus barbatus species nova

Ichthyscopus sannio; Whitley, 1945, p. 42
(South-west Australia); Whitley, 1948, p. 27
(S.W. Western Australia).

Differential diagnosis. In general appearance
close to Ichthyscopus lebeck, but differs by its
plain coloration, shape of the operculum and
interorbital groove for the reception of the
maxillary processes, the presence of two small
barbels on the chin, the simple or nearly simple,
unbranched tentacles round the mouth, and
larger scales.

Distribution. Only known from the western
coast of South-West Australia between Duns-
borough and Rottnest Island.

Type: W.A.M. regd. no. 4338.

Description. See Figs. 3 and 4, from which a
general idea of the shape may be obtained.

Fig.

D 19 (in large specimens the anterior two or
three dorsal rays may be somewhat hardened,
but they are never spiny), A 15-16, C with 10
divided rays, P 14, V 1.5 with spine concealed in
skin, scales in 36 to 44 rows (the exact number
of rows of scales is somewhat difficult to count,
as a number of rows branch at various places).

The head and sculpture of the head show a
general similarity to J. lebeck, but differ as fol-
lows (Fig. 5a): two small barbels are always
present, one above the other, on the middle of
the chin, the lower of these being the larger;
the tentacles surrounding the mouth are un-
divided or practically so, the fringes round the
nostrils are less well developed, the naked space
between the eyes for the reception of the maxil-
lary processes is square behind, the operculum
is more strongly curved anterio-dorsally than
posterio-dorsally, the posterio-ventral border of
the maxillary makes about a right angle.

Colours of fresh specimen (P 4424). Upper
parts pale brown, under surface pinkish (white,
somewhat transparent skin with red blood shin-
ing through), exposed bony parts of head

3.—Ichthyscopus barbatus sp. nov., W.A.M. no. P 3438, from right, ca. 2/5 x nat. size.

brownish red, soft part of cheeks smoky brown,
an ill-defined brown band from across the
humeral appendage to the first five rays of the
dorsal fin, and a second band underneath the
posterior 7 rays of the dorsal fin and to base of
caudal fin: sides of caudal peduncle with a
blackish brown spot of the size of a farthing;
D light brown, blackish brown at the outer edge;
A whitish, somewhat reddish pink from blood;
C black, P dull brown with whitish base, V pink-
ish white.

Colours of preserved specimens. All the ruddy
tinge has gone, not only from the under surface
but also from the head, so that only brownish
and white remain. Small specimens are
evidently more distinctly coloured. The two
specimens of lot P 4431, notwithstanding the
fact that they had been kept in formalin for 45
years before I first examined them, still show
dark cheeks, a dark tail, and two rather well-
defined dark bands on the back. ‘The anterior
of these includes the posterior part of the
humeral appendage and the first three dorsal
rays, the posterior is under the 13th to the 17th

Fig.

50

4.—Ichthyscopus barbatus sp. nov., W.A.M. no. P 3488, from above, ca. 2/5 & nat. size.

Fig. 6—Ichthyscopus lebeck sannio Whitley, Aust. Mus. no. LB. 2152, from right, ca. 7/20 nat. siz.

dorsal ray; they are dark (blackish in the smal-
ler specimen, somewhat paler in the larger speci-
men), very distinct on the back, and fading out
down the sides, the anterior one above the base
of the pectoral fin, the posterior one lower down.

Material. Seven specimens in the W.A.

Museum.

P 672 Woodman’s Point, received from fish-
market Fremantle, 19.X1I.1919, total length
25 em, standard length 20 cm, D 19, A 16,
L1.+44.

P 1296 Busselton, received from Mr. A. J. Smith,
registered on 27.V.1933, total length 187 mm,
standard length 147 mm, D 19, A 16,
Timlest3 6)

P 3488 Dunsborough, received from Mr. Alan
Poole, registered on 9.1I.1952, total length
40 cm, standard length 32.5 cm, D 19, A 16,
L1.+41.

P 4338 Between Rottnest Island and the Strag-
glers, collected by Mr. G. Turner on
22.X%.1958, total length 41 cm, standard
length 34 cm, D 19, A 163, L.1. 37.

P 4424 Eagle Bay, collected by Mr. L. Beaman
on 2.1V.1959, total length 375 mm, standard
length 300 mm, D 19, A 15, L143.
Weight 1245 g.

P 4431 (two specimens) Rottnest Island, collec-
ted by C.S.1.R.O. on 14.X1.1954, total length
130, 157 mm, standard length 105, 124 mm,
D 19, 19, A 164, 16, L.1. 37 and 38.

Habits, etc. Only on the two most recent
specimens some information was received: speci-
men P 4338 was caught on a line with meat bait
in 25 to 30 ft of water, specimen P 4424 was
caught on prawn bait in about 9 ft of water on
sand bottom near a reef at about 15 hrs. In
the fresh specimen, and even more so when it
was first captured, as Mr. Beaman told me, the
eyes protrude, and actually stick out like horn-
lets on top of the head. In preserved speci-
mens the eyes fall back and hardly suggest the
curious position they have in life.

Discussion. Whitley (1945, 1948) identified
the specimens of this species in the Western
Australian Museum as Ichthyscopus sannio and

52

originally, judging from descriptions, this led me
to believe that barbatus would only be subspeci-
fically distinct, but actual comparison between
specimens from eastern and western Australia
showed such a number of differences as enum-
erated above, that I have no hesitation in de-
scribing it as a full species though it is obviously
not too distantly related to J. lebeck.

Ichthysccpus lebeck sannio Whitley

Ichthyscopus sannio Whitley, 1936, p. 45—
Patcnga, Broken Bay, New South Wales.

Ichythyscopus inermis; Waite, 1899, p. 28 (New
South Wales).

Ichthyscopus ‘nermis; Waite, 1899, p. 112 (off
the coast of New South Wales: many localities) ;
Borodin, 1932, p. 96 (Southport, Queensland).

Anema inerme; Waite, 1904b, p. 50 (New South
Wales): (pt.) Stead, 1906, p. 206 (Kastern Aus-
tralia: specimen mentioned from Pelican Island,
Brisbane Water, N.S.W.).

Ichthyscopus lebeck; Ogilby, 1918, p. 105
(Tewantin, Qld.); McCulloch, 1922, p. 102 (New
South Wales); (pt.) McCulloch, 1929, p. 335
(New South Wales); Munro, 1945, p. 147-148
Fig. 6 (mouth of Noosa River, Qld.); Ogilby &
Marshall, 1954, p. 84, 85 note (off the coasts of
Queensland and New South Wales).

Ichthyscopus lebeck; (pt.) de Beaufort, 1951,
p. 50 (Australia).

Differential diagnosis. The dark grey colour
with the bold white markings, the somewhat
rudimentary first dorsal fin, and the branched
tentacles at the mouth, serve to distinguish this
species from the other members of the genus.

Distribution. Seas of China, Japan, and
eastern Australia, where known from Bowen
(Queensland) down to Wollongong and Nowra
(New South Wales).

Description. See Figs. 6 and 7. D I1.17, A 17,
C with 10 divided rays, P 16, V 1.5 with spine
short and concealed in skin; scales 50-54 rows.

This form is fairly close to J. barbatus, but
differs from that species by the presence of
dorsal spines, the smaller scales, the very differ-
ent coloration, the absence of barbels on the

chin, the strongly branched almost bushlike
tentacles surrounding the mouth, the more
developed fringes round the nostrils, the poster-
iorly rounded unarmoured space between the
eyes, the dorsally evenly rounded opercles, and
the sharper posterio-ventral angle of the maxil-
lary.

Colours of preserved specimens. Generally
dark brown, mottled with white above, and
whitish below. Unencased parts of head, in-
cluding chin, and whole nuchal area brown to
greyish brown, with innumerable round whitish
dots of various sizes; moreover the whole upper
surface of the head, including the bony parts,
and the nuchal region to half way along the
dorsal fin, have a great number of tiny purely
black spots. The white dots mentioned above
gradually increase in size backwards, and change
smoothly into the pattern of the back, which is
dark brown variegated with large white blotches
of various and varying shape. Fringes round
mouth and nostrils, and at opercles and humeral
appendages, dirty pinkish, D 1 blackish, D 2 as
dorsal surface dark brown with white blotches,
C dark brownish, posteriorly more blackish, P
blackish brown with pale lower edge, V and A
colourless.

Material. Two specimens received on loan
from the Australian Museum, Sydney, from the
coast of New South Wales, total length 283, 430
mm, standard length 227, 354 mm.

Discussion. Whitley (1936) described this
form, as a species, on the basis of the following
argument: “The Stargazer recorded from Aus-
tralia as Ichthyscopus lebeck or as its synonym
inermis, has been identified as such with reser-
vations. As specimens have accumulated, it has
become more obvious that our form represents a
hitherto unnamed species which differs from
figures of the Indian type mainly in coloration,
but also in shape and proportions. A Malabar
example of the true J. lebeck, from Dr. Francis
Day’s collection, differs from all my Australian
ones in having the preocular fringes extending
backwards halfway along each side of the inter-
orbital depression, whereas Australian specimens
have the fringes restricted to the anterior part
only; they also have the opercles and vertex

less granulated than the Indian one, and there
are slight variations in fin-rays and teeth”.

The ‘slight variations of fin-rays and teeth”
without mention of the nature of these vari-
ations is, to put it mildly, not very helpful, and
as regards coloration, I find that specimens of
sSannio agree quite well with Day’s (1878) de-
scription and plate of a specimen from Canara
(Kanara of recent maps). The character of
the shape of the posterior nostril seems to hold,
however, as set forth below, so that I do not
follow Munro (1945) who synonymised sannio
with lebeck. Both Bloch & Schneider (1801)
and Cuvier & Valenciennes (1829) described and
depicted lebeck as a fish with an elongated pos-
terior nostril. Mr. Wheeler examined the
material of the British Museum (Natural His-
tory) for me, and I quote his comments (given
in litt., 27.1.1959) :

“We have three specimens from China (400,
317, 225 mm standard length), one from Bellin-
ger River estuary, N.S. Wales (225 mm) and two
from Madras (185, 71 mm). All the specimens
have fringes round the nostrils, but in the
Chinese and the Australian specimens the
second nostril is round or slightly oval, while in
the Indian material the nostril is elongate. The
fringe of this nostril extends in consequence
between the eyes and ends level with the pos-
terior edge of the eye’’.

A sketch of an Indian specimen drawn by Mr.
Wheeler served as the basis for Fig. 5c.

A specimen from Japan in the Leiden
Museum, examined by Dr. Boeseman, has a round
posterior nostril.

Unless a future closer examination reveals
more distinguishing characters, the only point
of difference between lebeck and sannio appears
to be the shape of the posterior nostril. As this
character seems to be quite constant, it deserves
nomenclatural recognition, though I consider it
much too slight for specific distinction. There-
fore I recognise sannio as a geographical race of
lebeck, with the distribution Japan, China and
eastern coast of Australia, whereas the nomin-
ate race is apparently confined to India and
Ceylon.

Fig. 7.—Ichthyscopus lebeck sannio Whitley, Aust. Mus. no. 2152, from above ca. 7/20 x nat. size.

53

Fig. 8—Ichthysccpus insperatus sp. nov. type, from right, ca. 7/10 x nat. size.

Ichthyscopus lebeck lebeck (Bloch & Schneider)

Uranoscopus lebeck Bloch & Schneider, 1801,
p. 47—Tranquebar, India (reference copied).

Differential diagnosis. Differs from sannio
and for that matter from all other species of
the genus, by the shape of its posterior nostril,
which is not roundish, but an elongated slit
Cae, Ho.

Material. None.

Discussion. Particulars on this race have been
given in the discussion of J. lebeck sannio.

Ichthyscopus insperatus species nova
Differential diagnosis. Strikingly differs from
the other species of the genus by the absence of
scales, and by its very distinct colour pattern,
with twelve vertical dark bands on body and
tail (Figs. 8 and 9).

Fig. 9.—Ichthyscopus insperatus sp. nov., type,

54

Type and unique specimen collected in Roe-
buck Bay, N.W. Australia, received from the
W.A. Fisheries Department on February 15th
1954, and presumably caught not long before.
W.A.M. regd. no. P 3638.

Description. D 17, A 17, C with 10 divided
rays, P. left 13, right 14, V 1.5, spine small and
concealed in skin. Total length about 215 mm,
standard length about 170 mm (the specimen is
curled up as the photographs show, and not-
withstanding prolonged soaking in water it
proved impossible to relax it), length of head
from tip of mandible to hind border of opercle
63 mm, greatest width of head 55 mm, greatest
depth of body (from anterior part of D to anus)
52 mm, width of mouth from side to side 24 mm,
height of mouth (mandible) 19 mm, length of
naked space between eyes 123 mm, of bony
armour behind 23 mm, distance between orbits

from above, ca. 7/10 & nat. size.

13 mm, length of orbits 83 mm, nostrils small,
close together, roundish, in front of eyes, both,
‘but particularly the anterior one, surrounded by
small fringes. Bony upper surface of head with
a somewhat reticulated appearance, as is well
shown in Fig. 9. Hind border of opercle and
‘aumeral appendage fringed, D with soft rays
only, continuous, without notch, C rounded, P
tong, pointed, 4th ray longest.

Colours of preserved specimen. Pale brown
above, darkest on the nuchal region, almost
eolourless below, with twelve dark brown cross-
oands which are dark on the back and gradually
"ade away two-thirds down the flanks towards
he belly. These bands are darkest at the
sdges, and the three anterior bands actually
consist each of a double band as their central
oarts are quite as pale as the remainder of the
surface of the body in that region. The bands
are situated as follows: the first from the an-
verior part of the humeral appendage over the
muchal region; the second from the middle of
sche humeral appendage over the nuchal region,
massing well in front of origin of D; the third
includes the membrane between the first and
jhe second dorsal ray, and goes from there
Hownwards and slightly backwards; the fourth
umcludes the third and the fourth dorsal ray,
‘and leads downwards where it fuses with the
shird band; the fifth includes the tip of the
MAfth, and sixth, and the basal portion of the
seventh dorsal ray, tapering downwards where
:s closely approaches the sixth band; the sixth
yxeads from the bases of the 9th and the 10th
florsal rays, tapering downwards (the tips of the
swo rays involved are also dark brown, but their
middle parts are colourless); the seventh goes
‘rom the 12th and 13th dorsal ray downwards;
she eighth goes from the 15th, 16th and 17th
Aorsal ray downwards; the ninth goes from the
oack of the caudal peduncle downwards; the
senth, eleventh and twelfth are on the rays of
she tail at equal distances. Moreover the tips
of the caudal rays are blackish brown, and in
its posterior part D has a subterminal blackish
Orown band on its rays; P is crossed by two
indistinct brownish bands.
Interorbital sinus white and brown vermicul-
nated, fringes of the lips pale pinkish brown
ith white spots.
Discussion. The reasons for including this
scaleless species in the genus Ichthyscopus have
Ioeen given on a previous page.

Genus Kathetostoma Gtinther

A well-defined genus characterised by the
Naked body, a strong humeral spine, and a
spineless dorsal fin; D and A usually short. D
10-18, A 11-18. The number of rays is not given
ibat the original description of K. albigutta (Bean
1892), but according to Jordan & Evermann
(1898, p. 2312) it is D 10, A 12, whereas Barbour
(1941) found D11,A11. It is difficult to under-
stand why Jordan & Evermann conclude their
essay on the species with the remark ‘1 speci-
‘nen known” as Bean in the original description
-ecords six specimens and they themselves noted
“different specimens” a few lines earlier! Bar-
Sour (1941) commented upon this, but drew the
m~wrong conclusion, as is evident from his reference

to the “unique type”, whereas actually there
were six co-types. Though I have not examined
material of the species I find it difficult to
believe that the seas round Cuba would be in-
habited by a subspecies different from that
occurring in the Gulf of Mexico, as Barbour
claimed.

The genus is known from southern Australia
(two species), New Zealand (two species),
Pacific coast of tropical America (two species),
and the Gulf of Mexico (one species).

Only the Australian species concern us here:
both occur off the Western Australian coast and
are represented in our collection; they can be
separated as follows:

1. a. D 15-16, A 14-1515, orbit posterio-medially pointed
(Fig. 5d) Feline ad ieee DE Pte Das = Kewlaeve

b. D 13-14, A 13-14, orbit roundish (Fig. Deere

Sear Cte hy 2a ANN aha PR EN PAX ac Yh K. nigrofasciatum

Kathetostoma laeve (Bloch & Schneider)

Uranoscopus laevis Bloch & Schneider, 1801,
p. 47, pl. VIII—New Holland — New South Wales
(reference copied).

Ichthyscopus laevis;
(reference copied).

Kathetostoma laeve; Giinther, 1860, p. 231
(Port Arthur, Tasmania); de Castelnau, 1872,
p. 91 (rather common on the Melbourne
market); Macleay, 1880, p. 562 (Tasmania, Mel-
bourne); Macleay, 1881, p. 197 (Tasmania,
Melbourne); Johnston, 1883, p. 115 (Tasmania,
northern coasts); Tenison-Woods, 1883, p. 192
(Melbourne); Johnston, 1891, p. 33 (Tasmania) ;
Waite, 1899, p. 113 (Port Jackson); Waite,
1904b, p. 50 (New South Wales); Stead, 1906, p.
206-207, pl. VIII (Victoria, Tasmania, New South
Wales); Waite, 1911, p. 242 (Australia); Waite &
McCulloch, 1915, p. 469-471 (New South Wales,
Victoria, Tasmania; Investigator Group, South
Australia); Waite, 1921, p. 140 Fig. 219 (no
locality mentioned — South Australia); McCul-
loch, 1922, p. 102 (no locality mentioned — New
South Wales); Waite, 1923, p. 163, Fig. (South
Australia); (pt.) McCulloch, 1929, p. 335 (Vic-
teria, New South Wales, Tasmania); Mees, 1959,
p. 9 (Esperance, W.A.).

Cathetostoma laeve;
tralian seas).

Differential diagnosis. D 15-16, A 14-153; a
very convenient character to distinguish this
species from the following is the shape of the
orbits and sculpture of the bony armour on the
dorsal surface of the head; this difference was
already noted and illustrated by Waite & Mc-
Culloch (1915) but to assist ready identification
I give sketches of both.

Distribution. Coasts of New South Wales,
Victoria, Tasmania, South Australia (off the
Investigator Group, South Australia, cf. Waite
& McCulloch, 1915, p. 471), and Western Aus-
tralia (recently recorded by me for the first
time, cf. Mees, 1959). McCulloch (1929, p. 335)
apparently overlooked the occurrence in South
Australia as published by Waite and himself
but included New Zealand in the range. Prob-
ably this record is based on Hutton (1872, p. 23),
whose K. laeve was placed in the synonymy of
K. giganteum (a species not yet described at the
time) by Phillipps (1927) and others. A con-

Swainson, 1839, p. 269

Gill, 1861, p. 114 (Aus-

venient difference between the two species seems
to be the larger number of dorsal and anal rays
in K. giganteum, and the fin-ray formula given
by Hutton for his K. laeve: D 16-17, A 17-18 is
very high for that species and fits K. giganteum.
Admittedly Haast (1873) described K. giganteum
as having D 16, A 14, but Waite (1911) on re-
examination of the type specimen found that its
actual formula was D 18, A 17 or 18. From the
fact that Hutton (1.c.) gave a range of variation
for D and A whereas he clearly stated to have
but a single specimen, it is evident that the
figures presented were copied from literature
and not taken from his specimen. Giinther’s
(1860) count for K. laeve is also high (D 17,
A 17), higher than one would expect for that
species, and if this count correctly represents
the fin-ray formula of his specimens, they are
likely to belong to K. giganteum, a species not
yet described at the time, and not to K. laeve.
Anyway, for the moment I feel justified in
accepting Phillipps’s identification and exclude
New Zealand from the range of the species. It
may be remarked that Hutton’s (1890, 1896)
identification of K. gigantewm with K. laeve,
which probably caused the confusion, was not
accepted by subsequent revisers (Waite & Mc-
Culloch 1915), and Hutton (1896, p. 315) himself
already suggested that giganteum might be a
valid species.

Material. Two specimens in the Western Aus-
tralian Museum.

P 1745 Esperance, received from Mr. P. F.
Sullivan, registered on 7.1V.1937, total length
322 mm, standard length 265 mm, D 15,
A 15.

P 3629 Esperance Bay, received from Dr. K.
Sheard, registered on 16.1.1954, total length
355 mm, standard length 290 mm, D 16s,
Awa Salon

Besides, there is in our collection an old
mounted specimen from Tasmania (no precise
locality or date).

Eight specimens received on loan from the
National Museum of Victoria range from 225 to
+ 430 mm total length, 180 to + 370 mm stand-
ard length, D 15-16, A 15-153, except one speci-
men which has evidently been damaged in its
youth and has D 11, though in front of Da few
tubercles covered with skin are apparent.

Kathetestoma nigrofasciatum Waite &
McCulloch
Kathetostoma nigrofasciatum Waite & Mc-
Culloch, 1915, p. 469, pl. XIII Fig. 1, 2—Doubtful

Island Bay, South-western Australia, 20-25
fathoms.
Kathetostoma nigrofasciatum; Waite, 1921,

p. 141, Fig. 220 (South Australia); Waite, 1923,
p. 164, Fig. (South Australia) ; McCulloch, 1929,
p. 335 (South Western Australia); Whitley,
1948, p. 27 (south coast of Western Australia).

Differential diagnosis. D 13-14, A 13-14; the
differences between this species and K. laeve are
given under that species and in the key.

Distribution. Known from the coasts of South
Australia and the south coast of Western Aus-
tralia.

Material. Three specimens in the Western

Australian Museum.

P 710 Off Bald Island, received from Chief In-

spector of Fisheries on 25.VIII.1920. Total length

238 mm, standard length 190 mm, D 14, A 13%

P 1057 “South of Rottnest” ’, W.A. Trawling
Co. Ltd., registered in October 1929. Total
length 260 mm, standard length 210 mm,
D 14, A 14.

P 2317 Albany, received from Mr. R. C. Winte-
ford, registered on 25.VIII.1941. Total
length 213 cm, standard length 17% coy
D 14, A 14.

Genus Uranoscopus Linnaeus

This genus is characterized by a scaly body,
streng humeral spine, two dorsal fins, either
entirely separated or connected at their bases,
of which the first is spiny, and by the head which
is almost entirely encased in heavy bony armour,
with spines along the lower edge of the praeop-
erculum. a

The genus occurs in all tropical and sub-

tropical seas in a fairly large number of species. —

Two species have been recorded from Aus-
tralia, but there is a possibility that these are
identical.

Uranoscopus cognatus Cantor

Uranoscopus cognatus Cantor, 1849, p. 1003—
Sea of Penang (reference copied).

Uranoscopus cognatus; Mees, 1959, p. 9 (Shark
Bay). i

Material. A single specimen of a Uranoscopus
of which the characters agree with Uranoscopus
cognatus as given by de Beaufort (1951).

P 4280 Trawled between Kok’s Island (N. end
Bernier Island) and Quobba Point (N. of
Carnarvon), Shark’s Bay, by ‘Bluefin’, W.
& W. Poole, 23-30 July 1958.

Description. The specimen mentioned above
has an overall length of 183 mm, standard length
149 mm, depth of body 40 mm, breadth of head
54 mm, D III-13, A 13, lower border of prae-
operculum with four spines, lower edge of sub-
operculum with one spine, scales in about 56
series, humeral spine well-developed, directed
obliquely upwards and backwards; somewhat
rudimentary fringes on the lips, rather better
developed on the lower lip, hind border of
opercles slightly fringed. ;

Colour of preserved specimen: brownish grey,
somewhat freckled with light greyish on the
nape. Body, C, and to a lesser extent P, with
very small black spots. C and D greyish, D 1
black with the lower half of the first spine and
its membrane white, and a small basal patch
at the end below and behind the third ray
white. .

Discussion. Two specimens ascribed to
Uranoscopus terrae-reginae Ogilby (1910) re-
ceived on loan from the Australian Museum and

1 This probably means that the company from which
the specimen was received extended its operations all
along the coast of Western Australia south of Rottnest
Island, and that the specimen may have been captured
anywhere along the south coast. Certainly there is
no proof that it was collected near Rottnest or else-
where along the west coast. The locality of provenance
of the specimen, therefore, is uncertain.

,

Johnston, R. M. (1883).—General and critical observa-
tions on the fishes of Tasmania, with a
classified catalogue of all the known species.
Pap. Roy. Soc. Tasm. for 1882: 53-144.

(1891).—Further observations upon the fishes

and fishing industries of Tasmania, together

with a revised list of indigenous species.

Pap. Roy. Soc. Tasm. for 1890; 22-46.

S. & Bollman, C. H. (1889).—Scientific re-

sults of explorations by the U.S. Fish Com-

mission steamer Albatross. No. 4. Descrip-
tions of new species of fishes collected at
the Galapagos Islands and along the coast

of the United States of Colombia, 1887-’88.

Proc. U.S. Nat. Mus. 12: 149-183.

S. & Evermann, B. W. (1898).—The fishes of

North and Middle America: a descriptive

catalogue of the species of fish-like verte-

brates found in the waters of North America,

north of the Isthmus of Panama, pt. III.

Bull. U.S. Nat. Mus. 47, pt. ITI.

S. & Hubbs, C. L. (1925).—Record of fishes

obtained by David Starr Jordan in Japan in

1922. Mem. Carnegie Mus. 10: 93-346.

_S. & Snyder, J. O. (1902)—A review of the
trachinoid fishes and their supposed allies
found in the waters of Japan. Proc. U.S.
Nat. Mus. 24: 461-497.

D. S. & Thompson, J. C, (1905).—The fish fauna
of the Tortugas Archipelago. Bull. Bur.
Fish. 24 (1904): 229-256. .

S., Evermann B. W. & Clark, H. W. (1930).—
Check list of the fishes and fishlike verte-
brates of North and Middle America north
of the northern boundary of Venezuela and
Colombia. Rep. U.S. Comm. Fish. for 1928,
pt. 2:

H. (1889).—A review of the European and
American Uranoscopidae or star-gazers.
Proc. Acad. Nat. Sci. Philad. (1889): 258-265.

C. (1758).—‘‘Systema Naturae’ Xth Ed., I
(Holmiae).

Longley, W. H. & Hildebrand, S. F. (1940).—New genera
and species of fishes from Tortugas, Florida.
Pap. Tortugas Lab. 32 (Carnegie Inst. Publ.
No. 517): 223-285.

Macleay, W. (1880).—Descriptive catalogue of the fishes

of Australia, pt. II. Proc. Linn. Soc. N.S.W.

5: 510-629.

(1881).—‘“‘Descriptive Catalogue of Australian

Fishes’ I (White, Sydney).

A. R. (1914).—Report on some fishes obtained

by the F.I.S. “Endeavour” on the coasts of

Queensland, N South Wales, Victoria, Tas-

mania, South and South-Western Australia,

pt. 2. Zool. Res. Fish. Exp. “Endeavour” 2:

77-165.

(1922).—Check list of fish

animals of New South Wales, pt. III.

Zool, 2: 86-130.

(1929) —A check-list of the fishes recorded

from Australia, pt. III. Aust. Mus. Mem. 35:

329-436.

G. F. (1959).—Additions to the fish fauna of

Western Australia—l. Bull. Fish. Dept. W.
Aust, 9, pt. 1.

- Jordan, D.

Jordan, D.

10}

Jordan,

Jordan,

Jordan,
Jordan, D.
Kirsch,

ee

Linnaeus,

McCulloch,

and _fish-like
Aust.

Mees,

58

Munro, 2. S28. 1945).—Postlarval stages of Australian
PB ri No. 1. Mem. Qld. Mus. 12: 136-153.
Myers, G. S. (1946).—Occurrence of Uranoscopoid fishes
of the Western Pacific genus Gnathagnus in
the American Atlantic fauna. Copeia, 1946,
No. 1: 42.
D. (1910).—‘‘Endeavour Series, Noid; On
some new fishes from the Queensland
coast” (Brisbane).
(1918) .—Ichthyological notes (
‘ Qld. Mus. 6: 97-105. ;
—______—__ (1954) —“The commercial fishes and fisheries

of Queensland, revised and illustrated by T.

C. Marshall” (Dept. of Harbours and Marine,

Brisbane).

O’Shaughnessy, A. W. E. (1878).—Pisces. Zool. Rec. aL

(for 1876).

W. J. (1927).—Bibliography of New Zealand
fishes. Fish. Buil., Wellington, INEZ
Stead, D. G. (1906).—‘‘Fishes of Australia; a popular

and systematic guide to the study of the

wealth within our waters’ (Sydney).
*Steindachner, F. (1877).—Ichthyologische Beitrage (V).

S.B. Akad. Wiss. Wien 74 (1876): 49-240.
=Swainson, W. (1839).—“‘The natural history and classi-

fication of fishes, amphibians and reptiles, or

monocardian animals” II (London).
*Temminck, C. J. & Schlegel, H. (1843).—Pisces, pt. 2,

in Siebold: “Fauna Japonica” (Leyde).
Tenison-Woods, J. E. (1883).—‘‘Fish and Fisheries of

New South Wales’ (Sydney).

Wade, C. B. (1945).—New fishes in the collections of
the Alan Hancock Foundation. Alan Han-

cock Pacific Exp. 9: 215-237.

Waite, E. R. (1899).—Scientific results of the trawling
expedition of H.M.C:S. “Thetis”, Fishess

Mem. Aust. Mus. 1: 25-123.

Ogilby, J.

No. 4). Mem,

Phillipps,

(1904a).—New records or recurrences of rare
fishes from Eastern Australia. No. 3. Rec.
Aust. Mus. 5: 231-244.

(1904b).—Synopsis of the fishes of New
South Wales. Mem. N.S.W. Nat. Cl. 2.
(1911).—Scientific results of the New Zea-
land government trawling expedition 1907.
Pisces—Part II. Rec. Canterbury (N.Z.)

Mus. 1: 157-272.

(1921).—Catalogue of the fishes of South
Australia. Rec. S. Aust. Mus. 2: 1-208.
(1923).—“‘The Fishes of South Australia”
(Adelaide).

Waite, E. R. & McCulloch, A. R. (1915).—The fishes of
the South Australian government trawling
cruise, 1914. Trans. Roy. Soc. S. Aust. 395
455-493.

Whitley, G. P. (1936).—More ichthyological miscellanea.
Mem. Qld. Mus. 11: 23-51.

(1945).—New sharks and fishes from Western

Australia. Aust. Zool. 11: 1-42.

(1948).—A list of the fishes of Western Aus-

tralia. Bull. Fish. Dept. W. Aust. 2.
—(1959).—More ichthyological snippets. Proc.

Roy. Zool. Soc. N.S.W. 1957/58: 11-26.

Whitley, G. P. & Phillipps, W. J. (1939) .—Descriptive
notes on some New Zealand fishes. Trans.
Proc. N.Z. Inst. 69: 228-236.

Messrs. A. Bertelsen and H. Akerstrom inform
me that the vegetation which had hitherto cov-
ered the eastern dunes was destroyed by fire in
October 1935, after which the sand began to
drift. Denudation was apparently slow for
many years, for in 1945 Serventy referred to
erosion only in the south-eastern corner of the
island. Soon after, however, another fire devas-
tated this sector. According to Mr. G. C. Barker,
the movement of sand accelerated about 1950,
reaching a peak two years later. On his return
to the island in February, 1952, he found the low
scrub behind North Point engulfed in sand blown
from the dunes to the south-east. Previously
the far northern end of the island had been
densely vegetated.

At present the blown-out areas show no sign
of regeneration. The few remaining clumps of
Spinifex longifolius, perched precariously on
pinnacles of sand up to 8 feet high, are quite
incapable of spreading. In view of the meagre
rainfall and the persistence of strong souther-
lies in summer, it seems unlikely that this cycle
of erosion will be arrested before the eastern
and northern dunes are completely levelled.

Our outline of the island (Fig. 1) shows some
minor discrepancies with that depicted on
Admiralty Chart 1723. If these are in fact indi-
cative of subsequent changes in the coastline, it
would appear that North Point has grown since
the original survey by the ‘Beagle’ and that a
high-level beach has been formed by the filling
in of the shallows along the south shore of the
north-eastern bay.

In section the central plain is like a saucer.
The soil around the higher rim is fairly deep
and is largely composed of fragments of shells.
Towards the centre it becomes shallow and
loamy. In certain areas the soil is distinctly
guano-like and contains bones of birds.

The sink-hole is a circular, steep-sided depres-
sion, with a diameter of about 10 yards and a
depth of 3-4 feet below the surrounding plain.
Its muddy floor was damp but contained no free
water at the time of our visit.

The shallow salt-lake at the northern end of
the plain is roughly circular and almost two
acres in area. Its northern and eastern banks
slope steeply for 2-3 feet down from the plain.
It contained water during our visit but was dry
in November, 1913. As its level is unaffected by
the state of the tide, the water is probably re-
plenished annually by winter rain.

Fig. 2—From eastern dunes, looking south over a big

blowout. Fishermen’s huts in left middle distance.

Vegetation

The strand vegetation is dominated by Spini-
fex longifolius R. Br., Salsola kali Linn., Atriplex
cinerea Poir and Cakile maritima Scop. Where
backed by stable dunes the Spinifex soon gives
way to Atriplex paludosa R. Br., Scaevola crassi-
folia Labill., Olearia axillaris (D.C.) F. Muell.,
Myoporum insulare R. Br. and Exocarpus spar-
tea R. Br. Where the dunes are unstable, Spini-
fex extends inland, but the bare floor of blow-
outs is colonised only by Cakile, though sparsely
and quite ineffectively. In the lee of the big
eastern blowouts Nitraria schoberi Linn. grows
luxuriantly along with dense clumps of Salsola
and Myoporum. Towards South Point where the
sand is shallow over the limestone, the following
calciphilous shrubs appear: Pimelea microce-
phala R. Br., Spyridium globulosum (Labill.)
Benth. and Acanthocarpus preissii Lehm.

Fig. 3—From northern dunes, looking southwest over

central plain to Northwest Hill. Atriplex paludosa and
Scaevola crassifolia in foreground.

The vegetation of the central plain varies with
the nature and depth of the soil. Deep shelly
soils around the rim support a low dense shrub-
bery of Rhagodia baccata Mog., Atriplex palu-
dosa and Threlkeldia diffusa R. Br. At lower
levels and with increasing clay content, the
following species become dominant: Frankenia
pauciflora D.C., Limonium salicornaceum F.

Muell. and Arthrocnemum arbuscula (R. Br.)
Mog., the last-named being replaced by
A. halocnemoides Nees wherever limestone

approaches the surface. The low-lying clayey
soils south of the salt-lake are water-logged in
winter; they carry a dense mat of Salicornia
blackiana Ulbrich, Sporobolus virginicus (Linn.)
Kunth. and Suaeda australis (R. Br.) Mogq., the
only closed community on the island.

Much of North Island is thus dominated by
low chenopodiaceous shrubs. Apart from a few
big bushes of Nitraria, the vegetation is every-
where below five, and for the greater part, below
three feet high. The aspect of the vegetation
varies seasonally. At the time of our visit winter
annuals were very much in evidence. On the
central plain in particular, the ground between
shrubs was carpeted with herbage, among
which grasses and composites were especially
conspicuous.

Reptiles

Chelonia mydas (Linnaeus).—In December
1945, Serventy observed about a hundred Green
Turtles in the north-eastern bay close to the

60

shore. He and his companions searched the
beaches but found no egg-mounds. However, A.
Bertelsen informs me that “in December and
January, hundreds of turtles come to the north
side of the island to lay their eggs”. North Is-
land would be the southernmost breeding station
of the species in Western Australia. Whether
shey still breed there is unknown, but we found
on the north-western beach the remains of a
surtle whose carapace was about three feet long.

Python spilotes (Lacépéde).—No living Carpet
Snake has ever been observed on North Island by
maturalists. Yet there is no doubt that they
occur here or did so until recently. Three of
cny informants (Geraldton fishermen) have seen
them, including G. C. Barker who has lived on
‘North Island only since 1945; however, he be-
lieves their numbers to be ‘‘very few’. We found
an intact backbone of Python with numerous
‘oetrel remains on the floor of a blowout in the
north-western sector of the island.

Heteronota bynoei Gray.—Nine specimens of
this gecko, including a very young animal, were
sollected from under slabs of limestone on the
sentral plain.

Phyllodactyius marmoratus (Gray).—Four
specimens of this gecko were obtained from the
same locality as the preceding species.

Amphibolurus barbatus (Cuvier).—Alexander
found Jew Lizards “plentiful” in 1913; Serventy
j ‘saw a few” in 1945; and subsequently G. C.
Barker has observed them. Our party found
nene during three days work all over the island.

Egernia kingii (Gray)—We saw few King
Skinks and collected only one, a juvenile from
beneath a slab of limestone.

Lygosoma (Sphenomorphus) lesueurii Duméril
land Bibron.—Several were seen in the coastal
‘Spinifexz, but we were unable to catch them.
(C'wo mummified specimens were found that
matched in all essentials with a series from West
WWallabi Island. Another and smaller skink,
gossibly Ablepharus lineo-ocellatus, occurred in
the Spinifex but likewise evaded capture.
Lygosoma (Rhodona) praepeditum Boulenger.
—Two specimens of this small worm-like skink
were found with geckoes under slabs of lime-
stone.

Birds

The following list is restricted to land-birds.
“Marine and littoral species will be discussed
2lsewhere.

Turniz varia (Latham).—In 1840 Stokes found
the central plain “covered with coarse grass,
where a great many quails were flushed.” Both
‘Alexander and Serventy described the Painted
‘(Quail as “common”. We found no trace of the
species, which presumably became extinct some
time after 1945.

Phaps elegans (Temminck).—Serventy failed
to see the Brush Bronzewing in 1945; nor have
we or any of our informants observed it. If
“Alexander was not in error when he recorded
the species for North Island, it must have be-
‘come extinct soon after his visit.

Falco cenchroides Vigors & Horsfield.—A
single kestrel was seen by Alexander, who sup-
posed that it was a visitor from the mainland.

Hirundo neoxrena Gould.—At the time of our
visit there were at least 20 Welcome Swallows
on the island, mostly in the vicinity of the
fishermen’s huts. A nest under an eave con-
tained yellow-gaped fledglings that were able to
fly. Serventy did not observe this species. As
the island entirely lacks natural nest-sites (cav-
ernous limestone and undercut cliffs) swallows
have probably become established only since the
erection of the huts (according to G. C. Barker
who built the first one in 1946, there were four
by 1950; at present there are thirteen).

Cinclorhamphus cruralis (Vigors & Horsfield).
—At least three singing males of the Brown
Songlark were observed on the central plain.

Zosterops gouldi Bonaparte.—Silvereyes were
largely confined to the few places, such as the
leeward slopes of the eastern dunes, where the
shrubbery (Nitraria and Myoporum) was rela-
tively tall and dense.

Anthus novae-seelandiae (Gmelin).—Pipits
were surprisingly more plentiful on beaches and
in dunes than on the central plain. Since Ser-
venty observed a few birds, and Alexander a
pair, the species is probably resident, rather than
a visitor from the mainland as suggested by
Alexander,

Mammals

Macropus eugenii (Desmarest).—Since pre-
vious visitors had never recorded this species—
Stokes (1846, pp. 162-4) being explicit as to its
absence—our discovery of numerous Tammar
remains was entirely unexpected. The bones,
almost wholly mandibles, were only found in the
bottom of blowouts, usually associated on the
east side of the island with rabbit skulls, and
on the north-west with petrel bones. They
probably owed their preservation to the dry cli-
mate and the calcareous nature of the sand that
had covered them.

In answer to my question, what do you know
of wallabies on North Island, A. Bertelsen re-
plied, “have seen wallabies there in 1928 to 1930,
only one or two and much larger than on East
or West Wallabies”. F. C. Burton wrote,
“Extinct. Put there by fishermen in the early
days of fishing”. My other informants had no
knowledge of wallabies on the island.

Whether these remains are those of a long
extinct natural population or of a recent intro-
duction, can only be decided after carbon-dating
of the bones. Meanwhile the first alternative
seems more likely; for the fact that the skulls
of most of the tammars had been lost indicates
that the population was not contemporary with
the rabbits, whose skulls have been preserved
in great abundance.

Neophoca cinerea (Péron & Lesueur).—Seals
have been seen on North Island by A. Bertelsen
(“several times in January”) and G. C. Barker
(Gastewaos

Oryctolagus cuniculus (Linnaeus) —Eight rab-
bits, trapped at Geraldton, were released by A.
Bertelsen in 1934. The introduction was suc-
cessful, Serventy finding them “exceedingly
numerous” eleven years later. G. C. Barker
writes, “In 1945 there were quite a few. We had
a few traps and often had a rabbit. It was
usual when we arrived in February to see a

61

dozen or so around the camp at daylight. A
walk across to the west side would show 20 or

30 rabbits. They did not appear to bur-
row but lived under bushes, and as_ their
only enemy was the sea-eagle they con-

tinued to live that way. They now appear
to be extinct. This year and last I did not notice
any, and it would appear the two cats which
have gone wild and live in the centre of the
island have cleared them out.”.

Discussion

In the not very distant past North Island more
nearly approached the Wallabi Islands in rich-
ness of fauna. Today its fauna can only be
classed as depauperate. The dismal history of
recent extinction may be divided into four
periods.

(1) Before 1840.—Petrels, probably Puffinus
pacificus, ceased for unknown reasons to breed
on the island, despite its apparent ecological
adequacy. It is also probable that a population
of the wallaby, Macropus eugenii became extinct
in this period, a period that is completely un-
documented.

(2) 1840-1913.—Unfortunately Stokes’ account
is so brief that no decision can be made whether
any animals became extinct in this period. No
naturalists and apparently few fishermen visited
the island in the 73-year period between the
visit of the ‘Beagle’ and the Percy Sladen Trust
Expedition.

(3) 1913-1945—Between the two World Wars
deep-sea fishermen from Geraldton were in-
creasingly using the island as a haven in rough
weather. As the island entirely lacks potable
water and other resources useful to fishermen,
landings were infrequent and unexploitary. To
provide themselves with an emergency source of
food, fishermen released on the island, wallabies
and rabbits, only the latter becoming estab-
lished. Early in the period the Brush Bronze-
wing seems to have become extinct. Fires in
this period began the process of denudation that
culminated in the next.

62

(4) After 1945—The post-war establishment
of the crayfishing industry has introduced new
factors. The island is now inhabited by about
40 people for several months each year. Do-
mestic cats have become feral, though there is
no evidence that they breed. With man, there
are now two efficient predators, more or less
permanently living on the island.

Faunal impoverishment, movement of sand,
and loss of vegetation have all accelerated in
this period. The following species have either
become extinct since 1945 or nearly so: Carpet
Snake, Jew Lizard, Painted Quail and European
Rabbit. While the disappearance of the rabbit
is not regretted, it does illustrate how easily
even a tenacious and abundant animal may be
exterminated on a relatively small island. The
recent establishment of a widespread species of
swallow on North Island is little consolation for
the loss of the other species.

Acknowledgments

The Director of Fisheries (Mr. A. J. Fraser)
generously placed at our disposal free ship
transport from Geraldton to the islands. Thanks
are also expressed to those fishermen who gave
information recounted in the text; to Dr. D. L.
Serventy for the use of unpublished data; and
to Mr. R. D. Royce for identifying our plant
specimens.

The expedition was financed partly by a W.A.
University Research Grant but chiefly by a gen-
erous grant from C.S.1.R.O. to Professor He
Waring for marsupial work.

I am indebted to Dr. Main, my supervisor, for
overall direction and for reading the manu-
script.

References

(1922).—The Vertebrate Fauna of
Abrolhos (Abrolhos Islands),
J. Linn, Soc. (Zool.) 34;

Wires
Houtman’s
Western Australia.
457-486.

J. (1919).—The Percy Sladen Trust Expedi-
tions to the Abrolhos Islands (Indian Ocean).
J. Linn. Soc. (Zool.) 34: 127-180.

Stokes, J. L. (1846).—‘‘Discoveries in Australia.” 2 (T. &
W. Boone: London.)

Alexander,

Dakin, W.

PLATE I
1.—Pandanus frin

ge.

2.—Paperbark fringe.

64

‘sent only in the wet season and shrubs vir-
tually absent.

At the junction of the flood plain and the
‘savannah woodland there is typically a transi-
‘tion zone or fringe dominated, according to
‘local conditions by Pandanus or Melaleuca. The
|pandanus fringe is an open association of Pan-
idanus and grassland (see Plate I, 1). The paper-
(bark fringe is likewise an open association, of
, Melaleuca and grassland, differing from the true
jpaperbark forest in that a thick herb layer
idevelops (see Plate I. 2).

In addition to savannah woodland, pandanus
\fringe and paperbark fringe, trapping was also
icarried out in a fourth habitat, thorn scrub,
consisting of isolated patches of Acacia and other
sshrubs with thick undergrowth growing on
thigher ground in the open plain.

Results
Phascogale

In 1955-1956 H. J. Frith (personal communica-
ition) found a number of P. ingrami hiding under
«sacks on a rice levee bank that extended several
{hundred yards across the flood plain from the
{jpandanus fringe. The plain was flooding rapidly
vat the time, and the animals had probably been
iforced by the rising water to leave the vegetation
;along the foot of the bank, where they had been
living.

There are two records for 1956-57, both in
{February. A pair of P. ingrami was collected
tfrom under a log lying in thorn scrub, and a

single animal was seen at the base of a pandanus
clump when it was chopped down late in the
month.

The trapping programme carried out from
September 1957 to March 1959 is shown in
Table I. It should be noted that whereas in
1957-58 no two habitats were trapped simul-
taneously, in March 1959 simultaneous trapping
was used to compare—

(1) savannah woodland and
fringe (1-18 March); and

(2) savannah woodland and _ paperbark
fringe (18-25 March.)

The captures during the two seasons are set
out in Table II.

In 1957-58 seven P. ingrami were taken in both
the pandanus and the paperbark fringes.
Although none were trapped in thorn scrub, a
single animal was observed there in January
1958. One of the specimens collected in pan-
danus fringe on March 3, 1958 had been reduced
to a skeleton by ants before it was found, so
it remained unsexed. Some of the females ate
their litters shortly after being taken into
captivity, and only two litter sizes were estab-
lished immediately after capture, one of 12 and
another of 8.

In addition to the captures made in March
1959 and presented in Table II, a female
P. ingrami without a litter was found while
undergrowth was being cleared in the pandanus
fringe in August 1958 (J. Mills, personal com-
munication.) The weights and measurements
of some animals are given in Table III.

pandanus

TABLE I

Trapping programme for small mammals
Number of trap-nights per month

at Humpty Doo, 1957-1959

1957 1958 1959 |
Habitat a | Total
Sept. Jan. Feb. March April May et | fre: |
Eg Pie wee
|
)Tropical savannah woodland 63 63 0 | 0 0 0 90 35 251
WPandanus fringe ... ....| 0 0 1S |B} 63 189 108 0 549
\Paperbark fringe Sela Oo 63 | 126 63 0 126 0 35 476
Thorn scrub_ .... ee ae Oo 0 0 0 63 0 0 0 63
TABLE II
Number of captures of Phascogale ingrami per month at Humpty Doo, 1957-1959
1957 1958 1959 |
Habitat pe te — —— - aa | Total
| | : March | March |
| Sept Jan. | Feb. | March | April | May ae iene
| | | oy pat Sas (Peres |
Tropical savannah woodland oO} 0 e eos ox 0 Ona peo
[Pandanus fringe nae eh ee av | 0 2f(L) oO | 1f(B) 2f(L) Stel U
1 | 2m 2m
[Paperbark fringe ae eal (jhe Thyab)) If(L) | 0 | 0 te. 1f(L) 3f Im
: | sma | |
Thorn serub tiie | ered PD Gace | eee 0
| |
(I ) == with pouch young or pregnant, (B) = without pouch young, f = female, m = male, U = unsexed.

NOTE.—A male was brought in by one of the farm cats on 4/1/58.

TABLE III

Quantitative data from some Phascogale ingrami
trapped at Humpty Doo, Darwin

Date
of capture Sex Details
3/3/58 F Naked pouch young 0.5 cm long
9/3/58 F Weight 4.5 g (including young):
naked pouch young 1.0 cm
long
17/5/58 F Weight 5.0 ¢g
2/3/59 M Length 14.7 cm, tail 7.2 cm
M Length 15.6 cm, tail 7.2 cm:
8/3/59 F Length 14.5 cm, tail 7.2 cm:
12 naked pouch young 0.55
cm long
21/3/59 F Length 14.7 cm, tail 66 cm:
8 naked pouch young 0.8
cm long

It can be seen that no P. ingrami has been
taken in tropical savannah woodland despite
251 trap-nights’ work there, while the rate is
one animal per 68 trap-nights in the pandanus
fringe, and one animal per 119 trap-nights in
paperbark fringe. Apparently therefore, whereas
P. ingrami is common in the pandanus fringe it
is less so in the paperbark fringe and possibly
absent in the tropical savannah woodland.
Probably it cannot live on the flood plain,
which is inundated often to a depth of several
feet, for up to six months of the year, but sight
records indicate that it occurs on the higher
patches where thorn scrub grows.

The animals appear to enter traps most readily
between January and March, the catch averag-
ing one animal per 58 trap-nights in that period
(excluding trapping in tropical savannah wood-
land) and one animal per 504 trap-nights at
other times. Presumably the animals are most
active in the wet months; certainly no litters
have been collected outside that period, whereas
every female obtained within it has been either
pregnant or carrying a litter. The only litter
born in captivity was dropped about January
26th, 1958.

Melomys

This species has been taken at only one
locality—a pandanus fringe habitat on the west
side of the Adelaide River. A male and a preg-
nant female were collected there in February
1958 during 126 trap-nights. and a single adult
male during 63 trap-nights in May of the same
year. A young male (length 29.3 cm, tail 15.0
cm) was taken during 108 trap-nights in March
1959. In addition a sight record of several
Melomys was made, 20 miles away, in pandanus
fringe on the east side of the Adelaide River, in
late May 1958. Since Melomys is an arboreal
animal these results are probably not a true
reflection of the population density, but they
do suggest that M. c. albiventer is more restricted
tc pandanus fringe than P. ingrami and that it
breeds in the wet season. The pregnant female
cropped her litter of three on February 20, 1958,
in captivity. The young mice, when raised in
Canberra, bred first at the age of seven months,
but it is possible that they would have bred
earlier had the weather been warmer. Their
three litters contained 3, 2 and 3 young.

Discussion

As has been pointed out above, both small
mammals appear to be confined to the marginal
areas lying between the flood plain and the
tropical savannah woodland, occurring also on
the isolated spots of high ground on the plain.
A consideration of certain major environmental
factors suggests an explanation of this distribu-
tion. The plain is annually flooded, and since
neither of the animals is aquatic it is reasonable
to suppose that the annual inundation is suffi-
cient to prevent a population from becoming
permanently established there.

The patches of thorn scrub on high ground
might be expected to be populated, since they
are both dry and isolated from fire by the
swamps. On the other hand, the tropical
savannah woodland is annually burnt out at the
end of the wet season by fires that consume the
undergrowth, leaving the trees alone standing.
Such fires would probably be sufficient to wipe
out any population of small mammals in the
woodland, and thus prevent their permanent
establishment. Although both fires and floods
do occasionally reach the pandanus and paper-
bark fringes, these areas have features that
might be expected to minimise their effects. AS
can be seen in Plate I, 1, clumps of young
pandanus grow around the bases of older trees
in the pandanus fringe, and form large tough
thickets. The paperbarks, on the other hand,
tend to grow on ground so damp (e.g. on drain-
age lines) that the grass never dries sufficiently
for fires to obtain a proper hold. Possibly,
however, the lower catching rate of P. ingrami
in paperbark as compared with pandanus fringe
reflects the fact that if a fire does burn through
the former very little shelter remains for the
small mammals, whereas in the pandanus fringe
the pandanus clumps provide fire-resistant
shelters large enough to shield them from the
heat of the fire.

The distribution patterns of P. ingrami and
M. c. albiventer at Humpty Doo may, therefore,
be examples of the same situation as that
reported in some Finnish birds by Sammalisto
(1957), where each of two major habitats is
unsuitable for a species, but the marginal areas
between them provide suitable conditions for it.
It will be of great interest to see whether any
measures for flood or fire control taken during
the development of the area affect the distribu-
tion of these animals.

References

Ellerman, J. R. (1941).—‘‘Families and Genera of Liv-
ing Rodents.” II (British Museum: London.)

Jones, F. W. (1925).—The mammals of South Australia.
TII Handb. Flora and Fauna, S. Aust. Govt.
Printer, Adelaide: 271-458.

Kellogg, R. (1945).—A new Australian naked-tailed rat
(Melomys). Proc. Biol. Soc. Washington. 58:
69-72.

Sammalisto, L. (1957)—The effect of the Woodland-
open peatland edge on some peatland birds in
South Finland. Ornis Fenn. 34: 8189.

Troughton, E. (1941).—‘‘Furred Animals of Australia.”
(Angus and Robertson: Sydney.)

66

wood fragments with well preserved cell struc-
tures remain. Wood in the surrounding rock
is almost all distorted and flattened.

Thin Section

Material outside the concretion.—Thin section
studies with the petrologic miscroscope show
that about 60% of the rock Surrounding the
concretion is made up of angular quartz grains,
most with a diameter of the order of 0.08 mm.
The matrix is made up partly of minute grains of
yellow-brown siderite that commonly penetrate
borders of adjacent quartz grains, giving them
a ragged appearance. The siderite grains are
generally rhombs about 0.01 mm in diameter,
and their highest refractive index (w) is close
to 1.840. Apart from siderite, the matrix con-
sists mainly of a brown to dark brown clay-sized
paste that cannot be resolved under the micro-
scope, although sericite shreds and finely divided
organic material can be recognized in places.

Flakes of muscovite are fairly common, and
are conspicuous because of their length (up to
1 mm) in the otherwise very fine-grained rock.
Many flakes bifurcate and anastomose, being
separated by aggregates of minute siderite
euhedra (see Fig. 2). It is not clear whether the
muscovite is authigenic, and has grown around
the siderite, or whether siderite has grown along
the (001) cleavage planes, forcing the flakes
apart. Optical data for the muscovite are as
follows: 8 = 1.595 + .001, y — 1.597 + 0012;
(measured on three flakes with the universal
SIDS) == BSE as oe Ovi So eelinese
properties suggest a muscovite containing vari-
able amounts of the picrophengite (Meg rich)
molecule, but little of the ferrimuscovite (Fe
rich) molecule (Winchell and Winchell 1956,
p. 368).

Carbonaceous fragments and streaks are scat-
tered throughout the rock (see Fig. 3), and are
generally intimately associated with pyrite. The

ig. 2.—Thin section (transmitted light) of calcareous
nd sideritic concretion from 4444-4447 feet. Note
plitting of the long mica flake in the centre of the
eld. Much of the mica outside the concretion has
a similar appearance. Diameter of field 1.7 mm.

Fig. 3——Thin section (transmitted light) showing tex-

ture outside the concretions in a core specimen from

4659-4668 feet. Note the ragged boundaries of many

quartz grains. The grey matrix is mainly very finely

divided siderite, and there are also black carbonaceous

streaks and black opaque iron minerals. Diameter
of field 1.7 mm.

nature of this association is best observed in
reflected light, where it can be seen that pyrite
srew in the wood cells, and has to some extent
preserved the structure of the wood. Where the
cells have not been occupied by pyrite however,
the wood has been crushed and no trace of them
remains. Pyrite is also found as minute spheri-
cal bodies (generally from 0.02 to 0.05 mm in
diameter) and anhedral grains throughout the
rest of the rock. Black opaque grains are ubi-
quitous, but angular, blood red haematite grains
are far less common. A few black opaque grains
are partly transformed to haematite, and very
rarely, there are composite grains of haematite
and pyrite. Other minerals noted include
calcareous shell fragments and finely divided
calcite, biotite, chlorite, feldspar and leucoxene.

The per cent composition of the rock outside
the concretion is visually estimated to be:

Quartz : ; : : 60
Siderite .... Va a % ey
Calcite (including shell fragments) 8
Woody material he a Ge 5
Heavy minerals (including pyrite) 5
Muscovite an ; ; 2
Clay-sized cement, fine organic detri-

tus and other minerals % 8

Concretion.—The concretion contains the
same minerals as the surrounding rock, but
calcite, both as shell fragments and as finely
divided cement, is far more abundant. The
structure of much of the organic material is far
better preserved, however, and will be briefly
described.

The structure of the wood has been well pre-
Served (see Figs. 4, 5 and 6). Some of the cells
are filled with pyrite, or partly filled by approxi-
mately spherical pyrite bodies, but the cell walls
have not been replaced. Other cells contain

Fig. 4.—Thin section (transmitted light) of wcod frag-
ments in calcareous concretion from 4668-4677 feet.
Diameter of field 1.9 mm.

minute pyrite crystals, apparently always octa-
hedra, but their small size (usually less than
0.01 mm) precludes certain determination of
the form of many of the crystals. Many wood
cells contain no pyrite, and their preservation in
the concretion contrasts with their destruction
outside it.

A few pale brown, faintly anisotropic ovoid
bodies with a length of up to 0.2 mm and a
width of about 0.1 mm are present. Their
origin is not known, for although they super-
ficially resemble coprolitic pellets, almost all of
them contain a core of black opaque mineral,
or haematite or a composite grain of the two.
In fact, it is notable that many of the haematite
grains are so enclosed.

Fig.
fragment
Note preservation

5.—Thin section (transmitted light) showing wcod

in calcareous concretion from 4441-4444 feet.

of cellular structure. Diameter of
field 0.6 mm.

70

Heavy Minerals

Heavy minerals from the disaggregated rock
were concentrated in bromoform. A _ black,
opaque, slightly magnetic, generally angular
mineral (apparently ilmenite) is abundant. The
range in magnetic susceptibility of the grains
suggests, however, that they are not pure
ilmenite, but intimate mixtures in varying pro-
portions, of ilmenite and magnetite. Strongly
magnetic black opaque grains that include some
perfect octahedra (magnetite), and red angular
grains (haematite) are less common. Some
grains are clearly composite, and are made up of
haematite and black opaque mineral. There
are also rare composite grains of haematite and
pyrite. Zircon and brown tourmaline are fairly
common as euhedral grains, but less common
as rounded grains.

wood
fragment in calcareous concretion from 4663-4677 feet.
Note the preservation of wood structure and the pyrite

Fig. 6—Thin section (reflected light) showing

filling many of the wood cells. Diameter of

field 0.7 mm.

(wh-te)

Pyrite is almost as abundant as the black
opaque grains, and it occurs in three forms, as
small approximately spherical grains, as minute
and rare octahedra, and as anhedral grains and
aggregates. The spherical grains, which com-
monly range from 0.02 to 0.05 mm in diameter,
although many are smaller, superficially resem-
ble minute berries, and are like the framboidal

pyritic bodies described by Love (1958). The
octahedra seem perfectly euhedral, but their
small size (generally less than 0.01 mm) bre-

cludes precise observations. The anhedral pyrite
attains lengths of up to 0.15 mm, and rarely
far greater: some of the larger grains may be
aggregates of the two other forms. Identifica-
tion of the pyrite was confirmed by Mr. J. G.
Kay, of the Department of Geology, University of
Western Australia, from an X-ray powder dif-
fraction pattern. It is free of marcasite.
When concentrated nitric acid is allowed to
seep between a cover slip and a glass slide on
which the spherical or subspherical pyrite grains

it is worth noting the view of Krumbein and
Garrels (1952, p. 23) that with sufficiently rapid
sedimentation, organic matter may accumulate
in an open sea neritic environment in mildly
oxidizing waters. This is not incompatible with
anaerobic or reducing conditions below the sedi-
ment-water interface. The Dingo Claystone,
probably in part a deep water facies equivalent
to the west, also contains fairly abundant, finely
divided carbonaceous matter.

Broadly speaking, preservation of organic
matter and formation of pyrite and siderite
require anaerobic or reducing conditions.
Haematite, on the other hand, forms in an
oxidizing environment. There is also a sensitive
pH control. Edwards and Baker (1951) show
evidence for neutral to mildly alkaline condi-
tions of formation for pyrite, and Krumbein and
Garrels (1952, Fig. 8) suggest that pyrite and
siderite generally form in alkaline conditions up
to a pH of 7.8, whereas calcite normally requires
slightly higher alkalinity (7.87). The factors
leading to formation and preservation of the
above minerals in a marine environment are
very likely, however, to be far more complex
than those just indicated. Evidence from the
Learmonth rocks suggests that short-lived and
variable micro-environments played a large part
in establishment of their ultimate mineralogy
and texture. Some of these micro-environments
will now be considered.

Pyrite

The intimate association of much of the
pyrite with wood fragments has been noted.
There is no evidence in these rocks that pyrite
replaces wood, as it has been stated to do else-
where by other authors. In the concretions,
where the wood has not been squashed, pyrite
completely fills some of the wood cells and in
others forms minute spherical and subspherical
grains, and even smaller crystals (apparently
always octahedra). Outside the concretions, the
cellular framework has disappeared from flatten-
ing, except where sufficient cells were filled with
pyrite to preserve evidence of the original tex-
ture. Some such wood fragments seem in re-
flected hght to have undergone very little com-
paction, for almost all the cells were so occupied,
and preliminary examination might well give
the incorrect impression that pyrite had re-
placed some of the wood. It is clear, however,
that the pyrite formed before significant com-
paction of the wood, and other evidence given
later accords with this view. Minute spherical
and subspherical grains of pyrite are also
scattered throughout the rock, outside the wood
fragments.

The octahedron is the only crystalline form
recognized here, and has been the dominant form
in authigenic pyrite observed elsewhere by the
author. It also figures prominently in descrip-
tions of authigenic pyrite in the literature,
although no special significance appears to have
been attached to the fact. The findings of
Kasizyn (1956) (as summarized in German by
Mirtsching*) may, therefore, be of interest.
Kasizyn noted that in pyrite from old metamor-
phic rocks examined by him, the forms de-

*Mirtsching, A. (1959).—Zbl. Miner. 1957, (1): 109.

72

veloped include cubes, pentagonal dodecahedra
and octahedra. Octahedral pyrite is anomalous
and contains cobalt impurities which are held
responsible for its habit. ; ;

The spherical and subspherical pyrite bodies
in the Learmonth Formation resemble those
noted in thin section by Brelie and Teich-
miiller (1953) and Balme (1956), and those
separated from the Lower Carboniferous Oil
Shale Group of Scotland by Love (1958). Love
subjected his pyrite granules to nitric acid
treatment with results very like those recorded
here. He isolated microfossils from them, and
named two new genera, Pyritosphaera and
Pyritella, although their affinities could not be
closely defined. He suggested that these
organisms generated hydrogen sulphide as a
by-product of the anaerobic decomposition of
sulphur compounds, and that the gas reacted
with iron from the surrounding medium, so
precipitating iron sulphide on the organism.
Although his genus Pyritella resembles some of
the polygonal material described in this paper,
a rather different origin is indicated for the
latter. Organic matter that almost certainly
included individual spores and spore fragments
was the most likely nucleus for precipitation
of many of the spherical pyritic grains of the
Learmonth Formation. Precipitation took place
before the spores were much compressed.

In summary then, the growth of pyrite
occurred very early in the diagenetic history of
the rock, probably on or just below the sediment-
water interface. Its formation is best accounted
for by the action of anaerobic bacteria in more
or less enclosed spaces (spores and wood Cells)
where putrefaction was proceeding. The prob-
able sequence of events has been set forth by
Balme (1956). The bacteria reduce sulphate
to sulphide at the expense of various organic
compounds, and the hydrogen sulphide so
formed reacts with ferrous hydroxide to give
troilite (FeS). Subsequently, the troilite changes
either to marcasite, in slightly acid conditions,
or pyrite, in neutral to alkaline conditions.

Siderite and Calcite

Siderite is found almost throughout the Lear-
month Formation in Rough Range No. 1, and its
usual habit (minute euhedral crystals that pene-
trate adjacent quartz grains) indicates its
authigenic origin. The siderite has changed
from greyish to brownish in the laboratory,
showing its susceptibility to oxidation, a property
noted also by Pettijohn (1957, p. 145). Its
formation in sediments, according to Krumbein
and Garrels (1952, Fig. 8) is indicative of less
strongly reducing conditions than those required
for pyrite. In one concretion of which siderite
forms most of the cement (from core 18, 4,444-
4,447 feet) the wood fragments are perhaps not
as well preserved from crushing as in the more
calcitic concretions, but they are better pre-
served than in the surrounding rock.

It seems therefore that siderite and calcite
(excluding calcite of shells) formed at about the
same time. A sufficiently dense accumulation
of either mineral resulted in a concretion: these
concretions grew before appreciable compaction
of the sediment, but after formation of the
pyrite, which they enclose. The comparatively

early growth of some types of concretion, which
preserve contained organic matter from decom-
position and crushing, has been known for some
time, and there are interesting descriptions of
the preservation of fishes (Weeks 1957) and
plants (Brelie and Teichmiiller 1953) by this
means. The rather special conditions leading
to formation of concretions are not, however,
so easily explained. Weeks suggests that in
some stagnant environments where the pH is
normally too low for precipitation of calcium
carbonate, alkalinity sufficient for deposition
may be brought about locally by ammonia evolv-
ing rapidly from decomposition of proteinaceous
(nitrogen bearing) organic matter. The
abundance of shell fragments in the concretions
of the Learmonth Formation may be significant
but the processes leading mainly to precipita-
tion of siderite instead of calcite in one con-
cretion are not understood. Presumably they
result from localized and critical combinations
of Eh and pH.

Haematite

The presence of haematite, which is scattered
sparsely through the concretions and surround-
ing rock as discrete grains, and combined with
black iron ore, seems anomalous. Organic
material, pyrite and siderite compose a mineral
suite that accords with the interpretation of an
anaerobic or reducing environment, whereas
haematite is the product of a well aerated en-
vironment. It forms from iron silicates, mag-
netite, and probably from ilmenite as it develops
rapidly in air at high temperatures from that
mineral (Karkhanavala and Momin 1959). The
presence of many grains of haematite and com-
pound grains including haematite in ovoid
pellets, perhaps of coprolitic origin, cannot be
explained, but may be due to selection by some
organism. A _ tentative explanation for the
presence of haematite is that it formed earlier
in a strongly oxidizing environment, perhaps
for example, from jaspilites on the Precambrian
Shield. It was swept into the basin of deposi-
tion with other clastics, and some of it survived
the short-lived reducing environment to which
it was then subjected.

Conclusions
Evidence from the composition and texture of
the concretionary rocks from the Learmonth
Formation in Rough Range No. 1 suggests de-

73

position in an oxygen deficient marine environ-
ment, near enough to the coast to receive a fair
amount of plant material. The basin may have
been silled during deposition of these rocks.
Pyrite formed quickly close to the sediment-
water interface, especially in micro-environments
of putrefaction, such as cells of wood fragments
and interiors of spores. Siderite then grew
rapidly in an environment assumed to have been
generally oxygen deficient and slightly alkaline.
Slight variation from these general conditions
resulted locally in precipitation of calcite. Con-
cretions, either composed mainly of siderite or
calcite, grew before appreciable compaction, pre-
serving contained organic structures. Sedimen-
tary material around the rigid concretions was
compacted by superposed sedimentary beds
which eventually attained a thickness of many
thousands of feet.

References

Balme, B. E. (1956).—Inorganic sulphur in some Aus-

tralian coals. J. Inst. Fuel 29: 21-22.

v.d., and Teichmiiller, M. (1953 ).—Beitrage

zur Geologie El Salvadors III. Mikrosko-

pische Beobachtungen an Mangrove—

Sedimenten aus El Salvador. Neues Jb.

Geol. Paladntol. 6: 244-251.

Edwards, A. B., and Baker, G. (1951).—Some occurrences
of supergene iron sulphides in relation to
their environments of deposition. J. Sed.
Petrol. 21: 34-46.

Karkhanavala, M. D., and Momin, A. C.

Jedi, Ch,

(219538) aay

alteration of ilmenite. Econ. Geol. 54: 1095-
1102.

Kasizyn, Ju. W. (1956).—On the various morphologic
types of pyrite. Symposium: Kristallo-
grafija (Russ.) 1956, 5: 159-166.

Krumbein, W. C., and Garrels, R. M. (1952).—Origin
and classification of chemical sediments
in terms of pH and oxidation—reduction
potentials. J. Geol. 60: 1-33.

Love, L. G. (1958).—Micro-organisms and the presence

of syngenetic pyrite. J. Geol. Soc.
Lond. 113: 429-440.

McWhae, J. R. H., Playford, P. E., Lindner, A. W., Glenis-
ter, B. F. and Balme, B. E. (1958).—The
stratigraphy of Western Australia. J. Geol.
Soc. Aust. 4: 1-161.

Pettijohn, F. J. (1957).—‘‘Sedimentary Rocks’ 2nd Ed.
(Harper: New York.)

Weeks, L. G. (1957).—Origin of carbonate concretions

in shales, Magdalena Valley, Columbia. Bull.

Geol. Soc. Amer. 68: 95-102.

A. N., and Winchell, H. (1956).—‘‘Elements

of Optical Mineralogy. Part II. Descriptions

of Minerals.” 4th Ed. (Wiley: New York.)

Quart.

Winchell,

recording right mandibles, but, if in one species
there are found to be four right fourth lower
premolars then it is recognized that there must
be at least four individuals present in the
sample. Another species might have the num-
ber of individuals estimated by another fea-
tune:

It will be seen that some of the species
described by Broom are not included in the
table of relative abundance (Table I). These
missing species happened not to be represented
in the sample of nine pieces of breccia. A list
of the total known fauna of the breccia is given
in addition to the table of relative abundance.

The identification of new specimens of mar-
supials of Broom’s fossil species was made by
comparing all specimens with specimens identi-
fied by him and included in his Edinburgh
Collection. No revision of the status of these
species has been made. Eudromicia lepida was
not previously obtained by him and, in this
case, the five specimens (mandibles and maxil-
lae) were compared with the excellent series
in the British Museum (Nat. Hist.)

The Muridae present a greater problem in
identification since there is a superabundance
of named forms, particularly in the Pseudomys
group of genera, some of which are certainly
synonyms. Since most of the marsupial genera
represented in the breccia are also included
among the Recent fauna of Australia, I have
assumed that most or all of the murines would
also prove to belong to Recent Australian murine
genera. Direct comparison with specimens in
the British Museum was made with all of these.

Australian Murinae may be divided into three
groups (see Tate 1951) which are: Group 1.
“modern introductions”. e.g. Rattus rattus, R.
norvegicus, Mus musculus; Group 2. “young
/endemics”, e.g. species of Rattus which have
probably evolved in Australia, including the
R. assimilis and R. lutreolus species groups;
‘Group 3. ‘old endemics’’, i.e. Murinae of genera
peculiar to Australia and the adjacent islands.

Results
Fauna

Murinae: Groups 1 and 2.—In the identifica-
tion of fossil murines, specimens of Rattus are
most easily separated from the rest by means
.of the characteristic root pattern of the first
molar (see Jones 1922) as shown in Fig. 1.
No species of Rattus were encountered in this
» deposit, nor was Mus (sens. strict.).

Murinae: Group 3.—Of the old endemics, the
following genera at present occur in continental
Australia and any might have been expected to
‘occur in the sample: Pseudomys, Gyomys,
Thetomys, Leggadina, Notomys, Mastacomys,
_Leporillus, Zyzomys, Mesembriomys, Conilurus,
Laomys. Uromys, and Melomys. Of these genera
the last two comprise the Australian mainland
‘representatives of the Uromys genus-group

which is probably papuan in origin. Tate
(1951, p. 283) believes that this group was
,derived independently from a _ Rattus-like

‘ancestor. The Australian species of the Uromys
group are mostly tropical forms and are probably
fairly recent arrivals in continental Australia.

15

eet

Upper

Fig. 1—The alveolar cavities of Australian murines
after Wood Jones (1922). (a) Rattus greyi (b) Leporillus
jones.

One would not expect to find them in a southern
(New South Wales) Pleistocene fauna and none
were obtained.

The remaining old endemic genera constitute
the Pseudomys genus group. Only three species
of this group were obtained and they were
found to belong to Mastacomys, Pseudomys,
and Gyomys. A large series of type specimens
of these genera is available in the British
Museum (Nat. Hist.) and upon comparison with
these, the specimens in the deposit were found
to agree most closely with the types of Pseu-
domys oralis Thos. and Gyomys glaucus Thos.
The Mastacomys was new and has received the
name Mastacomys wombeyensis Ride (1956b).

Marsupialia—In addition to the _ species
already described by Broom, specimens of
Eudromicia lepida were obtained.

The results of the faunal analysis are sum-
marized in the following list and in Table I.

The following species of Mammalia occur in
the deposit:

MONOTREMATA

1. Tachyglossus sp. Remains are too fragmentary
to allow of specific identification.

MARSUPIALIA

Dasyuroidea

Antechinus flavipes (Waterhouse.)
Phascogale tapoatafa (Meyer.)
Thylacinus cynocephalus (Harris.)
Perameloidea

Perameles wombeyensis Broom.
Phalangeroidea

2;
3.
4.
5
6. Cercaertus nanus (Desmarest.)

7. Eudromicia lepida (Thomas.) New record.

8. Petaurus breviceps Waterhouse.

9. Palaeopetaurus elegans Broom.

10. Burramys parvus Broom.

11. Pseudochirus antiquus Broom.

12. Potorous tridactylus (Kerr.)

13. Macropus wombeyensis Broom.

RODENTIA

Muridae

14. Pseudomys oralis Thomas. New record.

15. Gyomys glaucus Thomas. New record.

16. Mastacomys wombeyensis Ride (Ride 1956b as
new species.)

TABLE I

Relative abundance of individuals in the fauna of the
Burramys parvus breccia of the Wombeyan Caves,
N.S.W. Only those species which were found in the
test sample mentioned above are included.

Minimum

Number of number of

Species Specimens individuals
Antechinus flavipes rs 14 4
Phascogale tapoatajfa ae 2 1
Perameles wombeyensis .... Le, 3
Cercaertus nanus ... Bes 26 6
Eudromicia lepida at 8) 2
Palaeopetaurus elegans .... 1, 2
Pseudochirus antiquus _... ial 3}
Potorous tridactylus Bee 4 1
Pseudomys oralis on 10 3)
Gyomys glaucus ee. 87 Ty
Mastacomys wombeyensis 1 1
(Burramys parvus)* oe (9) (3)

Petrography of the Deposit
The petrography of the Burramys parvus
breccia has not previously been investigated and
Dr. Pamela Lamplugh Robinson of the Depart-

ment of Zoology, University College, London
kindly did this for me. The following is an
extract from her report on the material

(personal communication) :

Hand specimen.—A very porous tufa
crowded with bones, and with occasional
fragments of coarsely crystalline calcite.
No layering is apparent, and there is no
directional orientation of the bones. The
bones do not appear to be abraded by an
agent of transport such as water or wind.

Thin section.—The fragments of crystal-
line calcite, and the majority of the outer
surfaces of the bones are covered with
a thin layer of extremely fine brownish dust.
Then follows a layer or layers of calcite
containing a fine dispersion of dust. The
remaining interstices between calcite frag-
men‘s and bones may either be unfilled with
cementing calcite (in which case the cavity
commonly has a lining of a thin layer of
clear calcite) or be filled with clear cement-
ing calcite. The interior of the bones may
be hollow, or lined with a thin layer of
calcite, or completely filled in with calcite.

Discussion
Several aspects of this breccia and fossil
fauna require discussion. First, there is the

provenance of the bones and the mode of de-
position of the deposit; second, there are zoo-
geographic and _ palaeoclimatic implications;
and third, there is the age of the fauna. Finally,
certain general principles emerge.

*Three of the pieces of breccia where chosen for treat-
ment because a mandible of B. parvus was present
on the surface so that the results for this species are
not strictly comparable. If these three specimens are
ignored, the results for Burramys are six specimens
comprising at least one individual.

Provenance of the Bones and Mode of
Deposition

Great concentrations of small bones such as
occur in the Burramys breccia are a familiar
feature of many Australian cave deposits. For
example, the cave earth of Hastings Cave,
Jurien Bay, Western Australia, which is com-
parably rich in bone to the Burramys breccia,
is in places 11 ft thick. Most of the remains of
the many hundreds of thousands of individuals
which go to make up the bulk of these deposits
appear to have been transported into the caves
from outside. Further, the bones seem always
to have been moved after the bodies have de-
composed. since it is only seldom that bones
associated in life remain so in the deposit. In
Hastings Cave it appears that some of the
material is washed down into the entrance. The
entrance is a sink hole into which a collapse
has left an inclined ramp which leads from the
present day surface of the ground through the
arched entrance of the main cavern. Bones
deposited in the sloping entrance ramp or wi‘hin
the arch itself are thus washed further back
into the cave. Some layering is apparent.

In the Burramys breccia however, no stream
bedding or layering of any kind, or sorting
of the bones could be detected by Dr. Robinson,
and further, the deposit is extraordinarily free
from clay, silt and sand. These facts would
appear to discount any suggestion that the
bones were transported into the cave by water
as some have been at Hastings cave.

Dr. Robinson does not consider that the
deposit is wind-accumulated because there are
no sand grains or signs of vegetable debris which
might be expected if the remains had blown
into the cave from outside. She suggests that
animal transport is the most likely means by
which the bones accumulated and points out
that the broken state of these small bones would
appear to indicate accumulation by some preda-
tor.

The character of the matrix gives further
indication of the conditions under which the
deposit accumulated. It consists first of fine
dust which probably penetrated from outside
the cave or which may even have been derived
from the ceiling of the cavern itself by decom-
position of the limestone and the freeing of con-
tained impurities. This coating of fine dust
possibly marks arid periods during the accumu-
lation of the bulk material of the deposit which
comprises the bones and the fragments of
coarsely crystalline calcite. The bulk of the
matrix is a finely crystalline calcite cement
which was probably laid down by percolating
water. Dr. Robinson considers that deposition
was probably slow and the calcite may have been
laid down intermittently. This slow rate of
deposition was suggested to her by the great
concentration of bones in the relatively scanty
matrix and by the covering of fine dust which
must have taken some time to settle.

An additional point of importance in the
understanding of this deposit is the fact that
the majority of the remains appear to be those
of members of species of small body-size, or
of immature individuals of larger species. For
example, the larger forms represented in the

76

breccia are Pseudochirus antiquus, Potorous tri-
dactylus, Perameles wombeyensis, Thylacinus
cynocephalus, and Macropus wombeyensis. By
the examination of dental wear of isolated teeth,
and the stage of tooth eruption reached in
mandibles and maxillae, it is possible to obtain
a rough estimate of the age of individuals at
the time of death; and examination of the
material I have prepared shows that, with the
exception of a single specimen of Perameles
wombeyensis, all specimens of the first three
species for which age can be estimated are
juveniles. This great accumulation of the
remains of animals of small size would appear
to have been assembled by some definite form
of selection. In the case of T. cynocephalus
and M. wombeyensis even the juveniles are large
animals as compared with the other mammals
of the fauna and these two species appear to be
atypical of the deposit in this respect. Remains
of these are rare in the breccia and it is
possible that they represent fortuitous inclu-
sions. In recent years. I have frequently found
the mummified remains of larger mammals
among the bones of smaller ones on the surfaces
of cave floors. These caves are at present
accumulating the dead remains cf pre-
deminantly small-mammal faunas in Western
Australia and the larger bodies which occur
in them appear to be the corpses of individuals
who either seek refuge in caves in times of dis-
tress, e.g. Macropus ocydromus, Protemnodon
irma and Vulpes vulpes, or to be those of in-
dividuals which habitually frequent caves and
thus stand a reasonable chance of dying in
them, e.g. Macropus robustus and Tachyglossus
aculeata.

Broom (1896c) did not recognize that the
sample was biased and assumed that the
assemblage gave a picture of the whole fauna of
the district at that time. He noticed that most
of the forms might be classed as “feeble and
defenceless’” and he concluded from this that
they probably flourished ‘“‘owing to the absence
or scarcity of natural enemies” instead of realis-
ing that in all probability they died and were
included in the deposit because they were feeble
and defenceless.

The identification of the predator presents
a tantalizing problem. The presence of the

remains of Thylacinus in the deposit would
suggest that it and Sarcophilus might have been
responsible since they are both frequently found
together in mainland cave fillings of the late
Pleistocene. However, large carnivorous mam-
mals frequently leave some associated bones of
their prey such as bones of the feet and these
are not to be found in the deposit. The cave-
dwelling carnivorous bat Macroderma gigas
presents a further possiblity. Mr. A. M.
Douglas and I have examined the accumulated
debris of living colonies of these bats in parts
of the Pilbara district of Western Australia in
recent years. These deposits are characteristic
in that they frequently contain the remains of
Macroderma itself and moreover contain a large
proportion of avian remains but neither of these
characters is possessed by the Burramys breccia.
Finally, owls make use of caves as roosting
places and owl pellets of Tyto alba and Ninox
connivens which I have examined lead me to

ae

believe that these birds of prey are mainly
responsible for the accumulation of the bones
in the Burramys breccia. Main (1959) has
come to a similar conclusion with respect to
the extensive deposits of small mammal bones
in the caves of the Western Australian
aeolianite, and Dr. J. T. Robinson of the Trans-
vaal Museum has told me that almost identical
deposits are at present being formed by owls
in the caves of the Transvaal.

Zoogeography and Palaeoclimate.

The fauna of the Burramys breccia can be
divided into three groups:

M1) Those which are extinct today e.g. Bur-
ramys parvus, Palaeopetaurus elegans,
Mastacomys wombeyensis, Perameles
wombeyensis, Pseudochirus antiquus,
and Macropus wombeyensis.

Those which certainly occurred in the
area in historic times e.g. Tachyglossus,
Antechinus flavipes, Phascogale tapoa-
tafa, Cercaertus nanus, Petaurus
breviceps, Potorous tridactylus, Pseu-
domys oralis. Possibly Perameles
wombeyensis, Pseudochirus antiquus
and Macrcpus wombeyensis may prove
to be chrono-subspecies of Recent
species like Potorous tridactylus anti-
quus and as such could be listed here.
Gyomys glaucus should also probably be
included in this group although its
present distribution is Southern Queens-
land (Tate 1951).

Those which have only existed in his-
toric times in Tasmania, e.g. Thylacinus
cynocephalus and Eudromicia lepida.

The conclusions which can be derived from
the evidence of these groups are as follows:

(1) Some of the extinct forms have not as
yet been found in any other known
deposit of Quaternary or Tertiary age
and nothing is known of their distri-
bution. Nothing can be derived from
the presence of these.

Of those species which certainly
occurred in the area in historic times,
Tachyglossus and Phascogale tapoatafa
are widespread in Australia and
although local forms of them probably
have specific requirements in relation
to climate it has not been possible to
relate the morphology of the fossils
to that of present geographical races
and climatic information has not re-
sulted. Petaurus breviceps is similarly
distributed through both summer and
winter rainfall areas in Australia, but
here there is some evidence that it is
confined to areas of reasonably high
rainfall, e.g. south-eastern and eastern
Australia, Cape York, the Northern
Territory, the Kimberley District, and
New Guinea. Cercaertus nanus,
Potorous tridactylus, Antechinus fla-
vipes, Pseudomys oralis and Gyomys
glaucus are as far as I can determine
confined to areas of abundant rainfall
or at least reliable winter rainfall (see

(2)

(3)

(2)

Keast 1959, Fig. 4). This would seem to
indicate that the climate of the area at
the time of deposition was not much
more arid than it is at present.

Two of the mammals of the fauna
(Thylacinus cynocephalus and Eudro-
micia lepida) are today only found in
Tasmania. A case may be made that
the mainland extinction of the Thyla-
eine followed the introduction of the
Dingo into continental Australia in the
sub-modern period, but it is possible that
here we do not have cause and effect.
Climatic change may be involved. In
addition, although we possess scant
ecological knowledge, it seems unlikely
that the mainland population of
Eudromicia lepida has become extinct
through competition. Its closest rela-
tive, and apparent competitor in the
mainland Pleistocene, is Cercaertus
nanus which survives it on the mainland
at present but both species still co-exist
in modern Tasmania. The extinction
of E. lepida and Thylacinus on the
mainland may well be the result of a
general environmental change which
has resulted from slowly increasing
aridity. There is evidence that this
has gone on since the last pluvial period
and that the present day climate in
parts of Australia is as arid, or even
more arid, than any period in the
Pleistocene (Tindale 1955).

The occurrence of these two Tas-
manian forms in the fauna increases
the probability that the climate was
somewhat wetter and colder at the time
of deposition than it is at present.

The occurrence in this area of a
Pleistocene fauna which required a
colder and moister environment than
the area possesses today is not surpris-
ing. Only a hundred miles or so to the
south, the Kosciusko region was
glaciated three times during the Pleis-
tocene and glaciers extended down to
4,800 feet. They covered some 150
square miles (David 1950). At approxi-
mately 2,000 feet, and only 150 miles
from the centre of the glaciations, the
Wombeyan Caves must on several occa-
sions have had a periglacial climate
which would have been colder and
wetter than at present.

(3)

The Age of the Fauna

The advances of geophysics have been such
that modern palaeontologists can, in many cases,
know the absolute ages of their materials. In
the case of the Burramys breccia this has not

yet been possible. Insufficient bone, and no
plant remains, are available for C™ dating.
However, the palaeoclimatic evidence, which

indicates a slightly colder and wetter climate
than the area enjoys today, suggests, when
taken in conjunction with the nature of the
fauna, an age somewhat later than the last
pluvial period of the Pleistocene.

78

Dr. K. P. Oakley of the British Museum (Nat.
Hist.) has tried to obtain data by physical
methods through which an age relative to other
known Australian cave deposits might he
achieved, but no significant information re-
sulted. This was largely because sufficient
reliable material upon which comparison might
be based was not available.

If an attempt is made to correlate faunas of
known cave deposits little success is achieved,
simply because relevant data are not available.
The reasons for this are twofold. First, most
authors working in these deposits, with the
exception of a few like Finlayson (1933), have
not recorded on a total faunal basis, merely
confining themselves to descriptions of some of
the specimens. Secondly, most of the Pleisto-
cene fossils which have been described are of
relatively large animals which are not com-
parable with these of the Wombeyan Caves.
For example, the most obvious recorded faunal
assemblage with which to compare the Burramys
fauna is that of the Wellington Caves, but the
recorded fauna from them is one of large mam-
mals. However, the remains of small mammals
do occur in this Pleistocene fauna but they are
difficult to prepare by manual methods, and
when prepared in this way are often even more
difficult to identify. Lydekker (1885, p. 227)
catalogued specimens of Mastacomys fuscus,
Conilurus albipes, and Pseudomys lineolatus
from the Wellington Caves, in the British
Museum Collection; these identifications have
been confirmed by me and I have obtained
further specimens of P. lineolatus and M. fuscus
from a piece of breccia from these caves which
was collected by D. M. S. Watson and is now
in his collection.

In the geographically close Wellington and
Wombeyan Caves it may be chronologically
significant that both Pseudomys and Mastacomys
are represen’ed by different species, but, in the
case of Pseudomys, since both species (see Tate
1951) are extant today it may merely reflect a
slight ecological difference between the two
areas. Another alternative is that the two
species are biologically one since my specific
identifications merely record that each speci-
men is morphologically closer to the type speci-
mien of its assigned name than to any other.

The differences between the species — of
Mastaccmys can be similarly dismissed. At
first sight the fact that M. wombeyensis of the
Wombeyan Caves’ fauna is extinct and not
otherwise known, could possibly imply an even
greater age for the fauna than is indicated by
the general faunal picture and the evidence of
climate. Mastacomys fuscus and its subspecies
are widely distributed in Pleistocene and Recent
south-eastern Australia, and it even extends
into cave deposits as far north-west as the
Flinders Range of South Australia where it is
associated with such typical giant Pleistocene
forms as Thylacoleo (see Ride 1956b). In the
Wellington Caves it appears to be associated
with Diprotodon, Nototherium, Thylacoleo,
Sthenurus and Procoptodon etc., but there is
insufficient stratigraphic evidence to be certain.
The Burramys fauna could thus be older than
these faunas. Unfortunately the validity of
the species Mastacomys wombeyensis still re-

quires confirmation. It is still only known from
a single specimen in this fauna which contains
no M. fuscus. It differs from M. fuscus in two
characters, one of which (the great width of the
cheek plate in almost unworn teeth) is so dis-
tinctly different from that of all other speci-
mens of M. fuscus (including M. f. mordicus)
known to me, that I consider it unlikely that it
is not a separate species. The other character,
an extra cusp on the third molar, is possibly
less reliable. Extra cusps on the molars of the
murids of the Pseuwdomys group are not un-
common, for example, the presence of a sub-
sidiary cusp on the inner front edge of M’ is
one of the distinguishing characters of the
genus Thetomys, but 30% of all specimens of
Pseudomys and Notomys in the collections of the
British Museum (Nat. Hist.) also have this as
an “abnormality”. However, that a character
is unreliable in one genus need not necessarily
render it so in another. The presence of two
apparently unrelated abnormal conditions in a
unique specimen is unlikely, but more material is
needed to establish the validity of the species.

Even if we accept the validity of M. wom-
beyensis, it is in no way morphologically
ancestral to M. fuscus (see Ride 1956b, pp. 436,
7) and there are no phylogenetic reasons as to
why it should occur earlier in time. There can
be no reason why the occurrence of two species
of Mastacomys in eastern New South Wales
curing the Pleistecene should not be syn-
chronous.

There can be little doubt that the great abun-
dance of murine fossils in the cave breccias
will be of great help in future faunal compari-
sons. Before they can be used, however, we must
have a realistic taxonomy of them. Keys to
their identification which depend on characters
cther than numbers of mammae etc., and which
provide for the statistical appreciation of in-
dividual variation, must also be made before
the working palaeontologist can use _ these
species because working collections which are
comprehensive enough are not generally avail-
able for comparison.

A further distinction between the faunas of
the Wombeyan and Wellington caves is one
pointed out by Broom (1896c). Trichosurus
vulpecula is absent from the Burramys breccia
while it is present in the Wellington Caves fauna
(Brit. Mus. (N.H.) No. M10789). The absence
of T. vulpecula from the Burramys breccia led
Broom to suggest that the deposit accumulated
before the species came into the district. Absence
from the palaeontological record is always an
unsound basis for argument and in this case
it is particularly so because the fauna repre-
sented is clearly only a selected part of the
whole. Even if Trichosurus were resident in the
area at the time of deposition, adult specimens
of it would fall well outside the size range of the
included species and it would at most only be
represented by occasional juvenile individuals.
Trichosurus is widespread throughout the
greater part of Recent Australia. It is one of
the most successful and adaptable phalan-
gercids. If Broom is right in his contention,
then the Wombeyan Caves’ fauna is an earlier
one than that of the characteristic “giant”

719

fauna of the Wellington Caves. A similar fauna

of giant marsupials including Nototherium and
Sthenurus (Mammoth Cave) in Western Aus-
tralia has recently been dated as >37,000 years
nee personal communication C"™
ate).

In conclusion then, we may say that com-
parison of the Wombeyan Caves fauna with
Recent faunas, together with palaeoclimatic con-
iderations suggests that the fauna is Upper
Pleistocene and probably dates from the period
since the last pluvial. More slender arguments
can be brought forward that the fauna is older
than that of the Wellington Caves.

General Considerations.

During the preceding paragraphs one fact
has clearly emerged and that is that we do not
yet know enough about our faunas to provide
a basis for any real comparative discussion.
The work which will form the necessary pre-
liminary to the quantitative comparison of
faunal assemblages, and their palaeoecological
interpretation, is yet to be done. We do not
even know much of the ecological interrelation-
ships within modern Australian mammalian
communities, and this knowledge must neces-
sarily form one of the premisses of any logical
argument in the interpretation of the environ-
mental conditions under which fossil faunas
have lived.

In our present state of knowledge, not even
fully valid qualitative comparisons can be made
because of the uncertain value of many of cur
species. A number of the species of our fossil and
mcdern mammals are suspect because authors
have not made adequate comparisons with other
known, and obviously similar, forms. Further,
although type specimens are the basis of zco-
logical names, zoological species comprise popu-
lations with ranges of variation which, if un-
known, can still be more or less predicted sta-
tistically in so far as measurable characters are
concerned. It is against these ranges of in-
dividual variation that specimens which are
suspected of belonging to new species must be
compared and not merely with the types. The
types may actually represent peripheral ex-
amples in the range of the species. New names
made withcut biologically intelligent comparison
co not advance our knowledge. They merely
contribute to the existing confusion.

Acknowledgments

I wish to acknowledge with gratitude the
help given me by Drs. K. P. Oakley and P.
Lamplugh Robinson. Mr. A. M. Douglas
examined deposits left by colonies of Macro-
derma with me, and Mr. B. E. Balme and Dr.
J. Glover ascertained the thickness of the Hast-
ing’s Cave deposits for me. Drs. A. J. Cain and
A. T. Hopwood have spent much time with me
in most useful discussion. I am also grateful
to Dr. E. Lundelius for allowing me to quote his
unpublished C™ date for the Mammoth Cave
fauna. Finally, the work would not have been
possible except for the way in which the Profes-
sor of Anatomy, the University of Edinburgh;
Professor D. M. S. Watson, F. R. S.; and the
British Museum (Natural History); made their

Perth and Nedlands may be considered as
coastal as they are both about four miles from
the ocean, The sulphur values are lower than
the 8-10 lb/acre/year previously reported by
Hingston (1958) and are more comparable to
the values of Hutton and Leslie (1958) in Vic-
toria.

At both localities in 1959 the sulphur values
are considerably higher than those of the two
previous years, despite a much lower rainfall.
No explanation can be given, and the fact that
the analytical method used was different is not
considered to be a contributing factor. It has
been mentioned above that the titration method
was adopted in 1959 because of the difficulty of
obtaining a satisfactory colorimetric determina-
tion due to impurities in a batch of reagents.
When this factor was rectified a trial was done
to estimate the errors involved. The standard
error between the titration and colorimetric
method was 0.084 p.p.m. sulphur. The local
data of Hingston covered two year’s observations
and there was a difference of 2 lb/acre/year.
The data of Hutton and Leslie apparently are
only for a single year. A greater amount of
atmospheric dust in the dry year of 1959 may
account for the higher sulphur figures.

Eriksson (1958) quotes a figure of 10 Kg/
hectare (approx. 9 Ib/acre) as being the
characteristic sulphur value for unpolluted air.
The results given here do not support this con-
tention in contrast to the previous findings of
Hingston in Western Australia. The data are
more comparable to those of Hutton and Leslie
for two Victorian coastal stations.

82

Acknowledgments

The author is indebted to Mr. G. W. Mackey,
Deputy Director, Commonwealth Bureau of
Meteorology, Perth, for permitting the rain-
water collection unit to be established at the
Perth Weather Bureau; Mrs. D. Brocx and Miss
W. Marriner for their assistance with the
analyses.

The project was financed by a Research Grant
from the University of Western Australia.

References

Bond, R. D. (1955).—Determination of low concentra-
tions of sulphate with barium chloride and
ethylene diamine tetra-acetic acid (E.D.T.A.).
Chem. & Ind. (Rev.) No. 30: 941-942.

Eriksson, E. (1952)—Composition of atmospheric pre-
cipitation 2. Sulfur, chloride, iodine com-
pounds. Bibliography. Tellus 4: 280-303.

Eriksson, E. (1958).—The chemical climate and saline
soils in the arid zones. Climatology, Reviews
of Research, U.N.E.S.C.O. 1958: 147-180.
Presented U.N.E.S.C.O., Symp. on Arid Zone
Climatology, Canberra, 1956. (Arid zone
research 10.)

Hingston, F. J. (1958).—The major ions in Western Aus-

tralian rainwaters. C.S.I.R.O. Aust. Div.
Soils, Divl. Rep. No. 1/58.
Hutton, J. T., and Leslie, T. I. (1958).—Accession of

non-nitrogenous ions dissolved in rain-
water to soils in Victoria. Aust. J. Agric.
Res. 9: 492-507.

Johnson, C. M., and Nishita, H. (1952).—Microestima-
tion of sulfur in plant materials, soils, and
irrigation waters. Analyt. Chem. 24: 736-742.

Stephens, C. G., and Donald, C. M. (1958).—Australian
soils and their responses to fertilisers.
Advance. Agron. 10: 167-256.

11._Stratigraphy of the Boogardie Group

By D. J. Forman*
Manuscript received—22nd September, 1959

The Boogardie Group at Mount Magnet,
Western Australia forms part of a succession
of Archean rocks known as the “Older Green-
stones.” The group overlies the Lennonville
Beds and other older rocks and is overlain
by an unknown thickness of sedimentary and
igneous rock. It has developed minerals typical
of the greenschist facies of metamorphism.

Nine formations make up the group. These
are in ascending order: Poverty Flat Formation,
Jupiter Jaspilite, Mount Magnet Greenstone,
Three Boys Formation, Perseverance Jaspilite,
Mars Greenstone, Hill 50 Jaspilite, Havelock
Greenstone and Saturn Formation. Within these
formations fine-grained sediments, volcanic
rock and jaspilite are conformably interbedded.
They were probably deposited within an eugeo-
syncline which itself provided a source for
clastic and chemical material. Instability dur-
ing deposition has caused: buried valley struc-
ture; buried hill structure; rapid thickening
of units; slump structures and _ brecciation.
Ripple marking and cross lamination in silt-
stone and calcareous nodules in jaspilite were
used to determine sequence.

During intrusion of granite the group was
folded, metamorphosed and _ fractured. bj aVe%
intrusion of porphyry and quartz blows followed
and was accompanied by mineralization. Later,
during taphrogenesis, dolerite was intruded.

Erosion has since peneplaned the area. There
is evidence of a more rugged topography before
lateritization in the Tertiary.

Introduction

Boogardie is situated approximately 3 miles
N. 65° W. from Mount Magnet and approxi-
mately 300 miles NNW. of Perth, Western Aus-
tralia. Gold-bearing ore is being produced at
the present time (1958) from two mines at
Boogardie. These are the Hill 50 and Hill
50 Eclipse gold mines. To the end of 1956
the Hill 50 gold mine had crushed 982,105.90
tons for a return of 448,601.75 fine ounces of
gold.

The Archean rocks of the district are steeply
dipping and consist of lavas, sedimentary
greenstones and jaspilites, probably deposited
in an eugeosyncline. A stratigraphic subdivision
of these rocks has been made and their pro-
posed nomenclature is set out herein.

General Stratigraphy

The interbedded lavas and sediments of the
Mount Magnet district are tightly folded and
intruded by granite. They fall within the Yil-
garn Geosyncline of Prider (1952) and are con-
sequently correlated with the Older Greenstone
Phase of Prider (1948). :

The distribution of rocks near Boogardie is
shown in Fig. 1. Two major stratigraphic units
have been delineated—the Lennonville Beds and
the Boogardie Group.

*Bureau of Mineral Resources, Geology and Geophysics,
Canberra, A.C.T. Formerly Department of Geology,
University of Western Australia, Nedlands, Western
Australia.

83

The Lennonville Beds

The name Lennonville Beds is proposed herein
for the sequence of greenstones, jaspilites, cherts
and (?) conglomeratic quartzites, of Pre-
cambrian age, hounded above by the Boogardie
Group and below by an unknown thickness of
sediments.

The Lennonville Beds to the north of the Hill
50 gold mine are composed in stratigraphic
order cf the following lithologies:

4. 2,500-7,500 feet of amphibolite, chlori-
tized quartz dolerite and chlorite schist,
with thin (6 ins.-1 foot) jaspilite beds.
45 feet + of quartzite, conglomerate,
and conglomeratic quartzite.

il) Wee se Oil Classe,

90 feet + of interbedded greenstone and
jaspilite with thin cherty members,
grading in the upper 30 feet to jaspilite
with rare cherty members.

The jaspilites and cherts are coarser in grain
size than those nearer Boogardie. Stylolitic
seams parallel the bedding in the cherts, and
adjacent to the seams _ recrystallization is
marked. Lateral gradation of interbedded chert
and jaspilite is best explained by a sedimentary
facies change for the following reasons:—

There is no apparent change in thickness
when jaspilite changes to chert; there is rapid
change in iron content and close association of
jaspilite with chert; stylolites in banded iron
formations generally mark an increase in iron
content, not a complete absence of this element.

The stratigraphic position of the quartzite, (?)
cenglomeratic quartzite and (?) conglomerate is
in doubt, but it is thought that they are a con-
formable sedimentary sequence within the
Lennonville Beds. They vary lithologically from
conglomeratic quartzite to even-grained quart-
zite. Under the microscope these rocks show
little evidence of the fragmental texture seen
in hand-specimen. Large fragments in the
conglomerate or breccia are similar to the
matrix, being distinguished from it only by a
lack of ferruginous cement. Bedding and the
recrystallized nature of the rock suggest it is
conformable with the jaspilites. However, some
specimens resemble the superficial deposits
known elsewhere in Western Australia as “billy.”

It may be considered that the conglomeratic
quartzite and even-grained quartzites are con-
formable with the jaspilites and that a super-
ficial deposit was formed on these rocks in
comparatively recent times. This explanation
is accepted until the problem of the two rock
types of roughly similar composition but dif-
ferent metamorphic grade can be solved.

Intrusive Igneous Rocks
The Porphyritic Dyke Rocks

The “porphyries and porphyrites” of Boogar-
die are hypabyssal rocks ranging from leuco-
dacite porphyry to porphyritic micro-diorite
(andesite) and micro-diorite (andesite). They
intrude the Boogardie Group near the Hill 50
gold mine and trend in two general directions:
30° east of north, parallel to the Boogardie
“Breaks,” and 50° west of north, subparallel to
the strike of the sediments and the ‘Main Fault”
of the Hill 50 gold mine.

The porphyrites are dominantly post-folding,
and pos‘t-faulting intrusives, although they are
displaced by minor flat-dipping reverse faults in
the mine, and are also sheared and jointed.
Flow alignment of chlorite pseudomorphs (pos-
sibly after hornblende) in one porphyrite dyke
indicates that they were intruded in what is
now a vertical direction.

Intrusion of vorphyrite may have occurred
over a considerable period, as Finucane (1953, p.
232) shows the “Main Fault” (which itself is
said to contain a porphyrite filling in places)
cutting the porphyrite filling the Boogardie
“Breaks.” Finucane (1953, p. 232) claims that
a little gold has occasionally been found in the
porphyry, so that they are tentatively regarded
as pre-ore.

Granite

The granite varies between a fine- to medium-,
even-grained gneissic adamellite and a fine-
to medium-, even-grained massive adamellite.
The igneous contact with amphibolites on the
west limb of the syncline is well-exposed two
miles west of the Hill 50 gold mine. The linea-
tion developed in the amphibolites is identical
with that in some of the amphibolitic xenoliths
within the granite itself. There are numerous
cross-cutting aplogranite dykes in the green-
stone adjacent to the contact. In these a folia-
tion is developed which is not parallel to the
dyke trend itself but is subparallel to the schis-
tosity of the enclosing amphibolites. The folia-
tion developed in the granite, trending approxi-
mately 20°, is parallel to the contact and to the
direction of elongation of xenoliths of amphi-
bolite and metajaspilite.

The similarity of structural features noted
above, both in the granites and the adjacent
amphibolitic schists indicates that the intrusion
of the granite, the forma*ion of the amphi-
boli‘es, and the development of the lineation
and the foliation were broadly synchronous.
Therefore, the writer believes the regional fold-
ing was caused by, and took place during, the
period of granite intrusion.

The absence of hornblende xenocrysts in the
granite and the lack of feldspathized amphi-
bolitic schists indicate that the granite was
magmatically emplaced.

Quartz Blows

Quartz blows concordantly and discordantly
intrude the schists and gneissic granite of the
Mount Magnet district.

The quartz “reefs” are usually poorly
mineralized. Gold, pyrite, stibnite, cervantite,
stibiconite, scheelite, psilomelane, pyrolusite and

86

fuchsite were noted by Jutson many years ago
and tourmaline occurs in many of them. They
are apparently lenticular, rarely greater than
10-15 feet wide, and some may be traced for
several hundreds of feet along their strike.
In the Boogardie district they are rarely longer
than 20-30 feet.

Dolerite

A quartz dolerite dyke intrudes the Boogardie
Group at the Hill 50 gold mine. The dyke has
an east-west trend and maintains a constant
width of 18 in. to 2 feet.

The absence of alteration in this rock, which
intrudes chloritic schists, chloritized quartz
delerite and most probably partially altered
porphyrite dykes, indicates that it is later in
age than any of them and the lack of alteration
and the inclusion of a xenolith of pyrrhotized
jaspilite, demonstrates that the dolerite is post-
ore in age. Structurally, it follows a trend
foreign to the normal mine structures.

Metamorphism

The ‘“greenstenes” in the vicinity of the Hill
50 gold mine have developed minerals typical
of the greenschist facies of metamorphism. Two
and a half miles west-north-west of the mine
the “greenstones” are intruded by granite and
their mineral content is more typical of the
albite-epidote or lower amphibolite facies of
metamorphism.

Amphibole has been proven to develop in rocks
originally of igneous character and also in
rocks originally of sedimentary character with-
cut evidence of directional stresses. The re-
sultant amphibolites are probably metasomatic
in origin and occur in areas of low grade
metamorphism as well as in areas of moderate
grade metamorphism.

Superficial Deposits

Superficial deposits of conglomerate, traver-
tine, laterite and alluvium are the products of
weathering and erosion of the older rocks of
the region. ;

Stratigraphy of the Boogardie Group

Within this section will be found the defini-
tion, facies variations, geographic distribution
and petrography of the formations of the Boo-
gardie Group. The petrography is limited to
the component greenstones. The jaspilites show
little variation and their petrographic descrip-
tion will be found in the following section under
Petrology.

The Boogardie Group (Fig. 2) consists of nine
formations with the following bottom-to-top
sequence :—

Poverty Flat Formation

Poverty Flat is the name of an area apprcxi-
mately two miles south-east of the Hill 50 gold
mine, and characterized by a number of thin
jaspilite bars along which fault intersections
have yielded occasional bonanzas.

Definition—The Poverty Flat Formation is
defined as that sequence of minor jaspilite bars
and chert-carbonate-chlorite schists, which lies
between and is bounded by the Lennonville Beds
and the Jupiter Jaspilite.

highly weathered and lateritized. Jaspilites of
this formation generally outcrop well, but be-
cause of large areas of poorly outcropping
greenstones and the faulted nature of the type
section, the indicated total thickness is approxi-
miate only.

The type section in stratigraphically descend-
ing order is given immediately below.

Three Boys Formation
Mount Magnet Greenstone:

Thickness
Unit Feet
9—Greenstone 360
8—Jaspilite 4
7—Greenstone 139
6—Jaspilite 2
5—Greenstone 44
4—Jasp ‘lite 4
3—Greenstone es ats ae ms 70
2—Jaspilite cae aah aS he ; 1
1—Massive, fine-grained purplish-red,
highly weathered greenstone _.... 26

Estimated total
Magnet Greenstone

Jupiter Jaspilite

thickness, Mount

Geographic distribution—The Mount Magnet
Greenstone persists throughout the area mapped
in detail. Unit 9 of the type section thins from
a thickness of 500’ + in the south of the area
to almost 140 feet at the junction of G.M.L.’s
1389M and 1435M. Thus the distinctive width
of greenstone which marks the top of this
formation in the south becomes less distinctive
further north and the precise upper limit is
more difficult to recognize.

On G.M.L. 1480M penecontemporaneous tec-
tonics seem to be indicated in both this forma-
tion and the underlying Jupiter Jaspilite, by
the presence of a buried hill-like structure, by
the greenstone lenses appearing in the Jupiter
Jaspilite and by intraformational folding sug-
gestive of slumping. Should this evidence indi-
cate tectonic instability, the rapid thickening
and thinning of beds is readily explained as a
broader consequence.

Facies variations.—Unit 9 varies in the manner
indicated above. Jaspilite members of this
formation are variable in both thickness and
position. For example one jaspilite member
changes in thickness from 4 to 10’ between
G.M.L. 1435M and G.M.L. 1480M. For this reason,
certain correlation of individual jaspilite mem-
bers from one fault block to the next is diffi-
cult.

Three Boys Formation

Three Boys is the name of G.M.L. 1322M, a
lease about one half-mile south-east of the Hill
50 gold mine.

Definition The name Three Boys Formation
is here proposed for the sequence of jaspilites
and greenstones lying between and bounded by
the Mount Magnet Greenstone and the Per-
severance Jaspilite.

The type section is exposed west of the Hill
50 main shaft, between co-ordinates 11044E
12888N and 11350E 13000N on GML. 1438M
(Mars) and G.M.L. 1356M.

88

Because of the highly weathered nature of
the fine-grained greenstone at the surface, no
specimens were taken from the type locality.
The lithologies described below are identified
from the workings of the Hill 50 gold mine.

The following is a description of the type
section in stratigraphically descending order.

Perseverance Jaspilite

Three Boys Formation:
Thickness
Unit Feet
14—Fine - grained plagioclase - magnetite -
carbonate - quartz - chlorite - sericite
schist and moderately massive fine-
grained dark greenish grey, feldspar-

bearing sericite-iron ore-carbonate-
quartz-chlorite rock a a ae 40
13—Jaspilite ind aa are Bi ate 2

12—Q uar t z-magnetite-carbonate-chlorite
schist ae ie ae er oe 60
11—Jaspilite 2
10—Greenstone 30
9—Jaspilite 6
8—Greenstone 45
7—Jaspilite 3
6—Greenstone 18
5—Jaspilite 5)
4—Greenstone 9
3—Jaspilite a
2—Greenstone ee

1—Jasp!lite oe eas : Be
Total estimated thickness: 310

Mount Magnet Greenstone

Geographic distributicn—The Three Boys
Formation persists throughout the area mapped
in detail, and beyond to the south-east. To
the north-west of the area the base of the Three
Boys Formation is difficult to distinguish from
the top of the Mount Magnet Greenstone.

Facies changes——The thickness and position
of the various units of this formation change
in such a manner that their correlation from
fault block to fault block is uncertain. Within
the area mapped in detail the formation as a
whole converges from 500’ thickness in the
north-west to approximately 250’ thickness in
the south-east. Over the lateral distance of
half-a-mile of this convergence individual
jaspilite members thicken as greenstone mem-
bers become thinner.

Perseverance Jaspilite

Perseverance is the name of G.M.L. 1505M
in which the main workings have been in the
formation here defined as the Perseverance
Jaspilite.

Definition—The name Perseverance Jaspilite
is proposed herein for that jaspilite lying be-
tween and bounded by the Three Boys Forma-
tion and the Mars Greenstone.

The type section is on G.M.L. 1438M (Mars)
at co-ordinate 11116E 12744N. At this locality
it is a banded brown, grey and black rock con-
taining sparse cavities near its base, marking
the position of leached or weathered calcitic
nodules. It is thirty feet thick.

Geographic distribution and facies changes.—
The Perseverance Jaspilite is readily recogniz-
able throughout its length in the area mapped
in de'ail and beyond to the south-east. Below
is a Table showing variations in thickness of
this formation within a lateral extent of three-
quarters of a mile.

TABLE II

Locality Thickness

| (feet)
G.M.L. 1287M 24
G.M.L. 1323M 40
G.M.L. 12°2M 38
G.M.L. 1438M 30
G.M.L. 1435M 30
G.M.L. 15335M 26

Underground, the Perseverance Jaspilite may
{be seen in the Perseverance workings at a depth
»of 313’ and in the Hill 50 gold mine, 1,060’ level,
iat the east end of the new east crosscut.

Mars Greenstone

Mars is the name of G.M.L. 1438M, a lease
sapproximately 500 feet west of the Hill 50 main
» shaft.

Definition—Mars Greenstone is the name
igiven to the sequence of jaspilites and green-
‘stones lying between and bounded by the Per-
iseverance Jaspilite and the Hill 50 Jaspilite.

The type section is defined underground on
| the 313’ level of the Hill 50 gold mine and the
{Perseverance workings. However, due to poor
‘structural data exposed in the underground
iworkings, thicknesses of the strata are in-
,completely known and have been adopted from
|:surface exposures near the south end of G.M.L.
| 1323M,

The following is a description of the section
jin stratigraphically descending order.

(Hill 50 Jaspilite
|Mars Greenstone:

Thickness
Unit Feet
7—Magnetite- and quartz-bearing biotite-
caroonate-chert-chlorite rock with
distinct relict lamination and small-
scale cross-lamination ei 10
6—Jaspilite 33 Bes ab Bee ‘ 6
5S—(?) Bedded, magnetite- and quartz-
bearing, carbonate-chert-chlorite rock,
fine-grained ilmenite-bearing plagio-
clase-quartz-chlorite rock, and (?)
medium - grained chlorite - carbonate -
albite rock 5 tes ees ae ss 55
4—Jaspilite fe: oe ps as 8
3—Magnetite - carbonate - quartz - pla -
gioclase-chlorite rock (20’ +), mag-
netite-bearing plagioclase-quartz-car-
bonate-chlorite schist, (?) ilmenite-
plagioclase-chlorite rock and (?)
ilmenite - bearing quartz - chloritized
plagioclase-chlorite rock ae Gaiman 8)
2—Jaspilite a ne nak iO 6
1—Fine-grained cherty quartz-plagioclase-
chlorite rock Ai Ae ie ae 40
Total estimated thickness: 237

‘Perseverance Jaspilite

Distribution—The Mars Greenstone may be
itraced through the area mapped in detail, and to
ithe south-east; its extension to the north-west
iis not known. The stratigraphic sequence is
lbroadly similar from G.M.L. 1287M, through
iG.M.L. 1323M, G.M.L. 1282M, to G.M.L. 1438M.
On G.M.L.’s 1435M and 1536M, the formation
boundaries may be traced, but, due to heavy
lateritization and poor outcrop within the forma-
tion, little of the actual sequence is known.

89

_ Approximate local thicknesses of the forma-
tion are shown in Table III. They indicate
convergence from south-east to north-west.

TABLE III

j Thickness
Locality | (feet)
G.M.L. 1237M 200
G.M.L. 1323M 230
G.M.L. 1438M 220-180
G.M.L. 1435M 240
GML. 1536M | 130

Hill 50 Jaspilite
Hill 50 is the name of GML. 1282M, on

which the upper levels of the Hill 50 gold mine
are situated. At the time of writing, ore shoots
were being worked within the Hill 50 Jaspilite.

Definition.—Hill 50 Jaspilite is the name given
to that jaspilite lying between and bounded by
the Mars Greenstone and the Havelock Green-
stone. The type section is exposed on the 613/
level of the Hill 50 gold mine, where a true
thickness of 75’ is exposed.

Distribution and thickness variations—The
Hill 50 Jaspilite, because it is by far the thickest
jaspilite in the Boogardie Group, serves as a
useful marker bed. It may be traced from
fault block to fault block in the area mapped
in detail. Its ex‘ension, north-west and west
of this area is not known, but it may be traced
to the south-east for at least one mile.

Approximate thicknesses of the formation are
shown in Table IV.

TABLE IV
Locality Thickness (feet)
GM..L. 1287M 70
G.M.L. 1323M a
G.M.L. 1282M ? 50-75
G.M.L. 1438M 65-80
G.M.L. 1536M | 75 true width X-cut
within the Hill 50

Central workings.

Havelock Greenstone

Havelock is the name of G.M.L. 1287M on
which this formation outcrops.

Definition Havelock Greenstone is the name
given to the chloritized basalt and minor jaspi-
lite lying between and bounded by the Hill 50
Jaspilite and the Saturn Formation.

The type section is exposed on the 313’ level
of the Hill 50 gold mine at which location the
formation is between 25’ and 30’ thick.

The following is a description of the type
section.

Saturn Formation
Havelock Greenstone:

Thickness
Unit Feet
3—(Top) Jaspilite she 3
2—Chloritized basalt .... we 25-30
1—(Bottom) Chlorite schist i

Total estimated thickness = “approx. 30°

Hill 50 Jaspilite

Distribution and thickness variations.—The
Havelock Greenstone is traceable through
G.M.L.’s 1462M, 1287M, 1323M, 1282M (subsur-
face) to G.M.L. 1438M. Its continuation on
G.M.L. 1536M is obscured by extensive lateriti-
zation.

Approximate thicknesses for each locality are
given in Table V.

TABLE V
Locality Thickness (feet)
G.M.L. 1462M ops
GM..L. 1287M 24
G.M.L. 1323M 20
G.M.L. 1282M 30 (subsurface)
G.M.L. 1438M 19

Underground in the workings of the Hill 50
gold mine, the formation is exposed on all
levels—its thickness reaches a maximum of 38’
on the 1304’ level.

Saturn Formation

Saturn is the name of G.M.L. 1457M. An
open cut, known as the Saturn open cut, has
been worked in the jaspilite members of this
formation.

Definition.—Saturn Formation is the name
proposed herein for that sequence of jaspilite
and greenstone directly overlying the Havelock
Greenstone, and bounded above by a consider-
able though unknown thickness of greenstone,
apparently devoid of jaspilite members.

The type section lies between surface co-
ordinates 10512E 11190N and 11230E 11130N.
In the type section, the high degree of faulting
and porphyry intrusion, coupled with a gener-
ally poor outcrop of both jaspilite and green-
stone, render interpretations of the geology and
thickness estimations inexact.

The figure given below (702’) is far less than
that observed in other fault blocks, and must
be accepted with reservations.

The following is a description of the section
in stratigraphically descending order.

Saturn Formation:

Thickness

Unit Feet
14—Jaspilite 8
13—Greenstone 27
12—Jaspilite 20
11—Greenstone 6
10—Jaspilite 6
9—Greenstone 3
8—Jaspilite ZI
7—Greenstone 63
6—Jaspilite 3
5—Greenstone 85
4—Jaspilite 5
3—Greenstone 94
2—Jaspilite 4
1—Greenstone 340

Total estimated thickness = 702
Havelock Greenstone
Geographic distribution and thickness varia-

tions.—The Saturn Formation may be traced
from aerial photographs, throughout the entire

90

examined length of the Boogardie Group. It
appears to thin from approximately two thou-
sand feet thickness, several miles south-east
of Boogardie, to approximately seven hundred
feet in the type section. Further north-west
from Boogardie, it firstly thickens and then
slowly thins until the junction with granite
four miles west of the Hill 50 gold mine.

Thickness variations of individual members
may be observed in units 8 to 14, which have
thickened considerably on G.M.L. 1487M, (west
of the Mars lease G.M.L. 1438M) from their
thicknesses in the type area further south-
east.

Lithology.—The typical rock type is a sericite-
carbonate-chert-chlorite schist cut by innumer-
able carbonate-quartz veinlets which run sub-
parallel to the foliation. On the 820’ and 1,060’
levels of the Hill 50 gold mine the unit is in
part conglomeratic, especially near its base.
One detrital fragment, about six inches in
diameter, is roughly circular in cross-section
and is composed of a white kaolin-like material
regardless of being enclosed in unweathered
schist on the 1,060’ level. Other fragments
(e.g. of quartz) are smaller (about 0.5 inch
in diameter) and roughly circular in cross-
section. Clear, closely spaced bedding indicates
a sedimentary origin for this rock. It weathers
to a yellow-brown colour.

The lithologies of the overlying greenstone
units are uncertain but are believed to include
both schists containing variable amounts of
quartz, biotite, sericite, chlorite and carbonate
and a chloritized medium-grained porphyritic
and trachytic rock composed dominantly of
albite and chlorite with minor iron ore.

Petrology

Within the Boogardie Group, fine-grained
sedimentary and igneous rocks and jaspilites
eccur in a conformable eugeosynclinal sequence.
Metamorphism has so altered many of the rocks
that their origin is indeterminate. Others are
igneous in origin and probably lavas or sills;
others are sedimentary in origin. Working
with those rocks whose origin is least in doubt
this section will deal with petrological features
of the jaspilites and igneous and sedimentary
greenstones which help to construct a picture
of their depositional environment.

Flow Lavas

Units 1, 3, and 5 of the Mars Greenstone, the
Havelock Greenstone, and parts (specimen
39515*) of the Mount Magnet Greenstone are
probably flow lavas. They are generally massive
and jointed and have suffered chloritization,
sericitization and carbonation. Undoubted por-
phyritic texture is apparent in most specimens.
The feldspars vary from albite Ab,; to oligoclase
Ab;; indicating all stages of alteration from a
more calcic plagioclase. The refractive indices
of chlorite are constant at 1.622 + .002 (6 speci-
mens), identifying the variety as either diaban-
tite or ripidolite.

Alteration of plagioclase to sericite, chlorite,
quartz, carbonate, albite and rare epidote has
taken place in many cases. These changes are

*Specimen numbers refer to specimens held at
University of Western Australia. oS

Miles (1941, p. 15) suggests “that this term
jaspilite might be extended slightly in meaning,
to include certain black and white varieties of
similar banded rock which, in many of the
Western Australian goldfields, show close field
relations with, and occasionally grade into, the
red banded kind.”

At depths of 600’ and greater the jaspilites
of the Beogardie Group contain bedded calcite,
often in an amount sufficient to designate a
small specimen as a banded magnetite-calcite
rock. Furthermore, under conditions of green-
schist facies metamorphism, minor chlorite is
developed rather than amphibole. Evidence
is brought forward below to demonstrate that
banded calcite-magnetite-chert rocks become
Silicified and oxidized at the surface to banded

haematite and magnetite-chert rocks—true
jaspilites under the definition.
Petrologic Description—Specimen 39575

(11616E, 12433N, 613’ level) is a specimen of
poorly auriferous jaspilite from the type locality
of the Hill 50 Jaspilite.

In hand specimen it is a black and grey,
clearly banded rock composed of alternating
iron-rich, silica-rich, or carbonate-rich bands,
10 mm to 2 mm in width. Intraformational
brecciation thought to be due to slumping, is
evident in hand specimen. Minor sulphide
mineralization occurs in small cross-cutting
fractures and along favourable bands. The
specimen has a specific gravity of 3.27.

In thin section the banding is seen to be pro-
minent. Graded bedding or cross bedding is not
present, although there is a variation in grain
size of the microcrystalline quartz in different
bands.

The carbonate, magnetite and quartz are
bedded in a sedimentary manner. There is in
this specimen and in all others examined from
the underground workings a distinct tendency
for the carbonate to occur in the same bands
as magnetite. Minor fractures traverse the
slide without any continuity or apparent dis-
placement. They contain microcrystalline
quartz where they cross chert bands and car-
bonate where they cross carbonate- and mag-
netite-rich bands, although no fracture was
traceable from a chert band to a carbonate-
rich band.

Constituent minerals are microcrystalline
quartz (45%), iron ores (30%), carbonate
(25%), and minor chlorite (<1%).

Microcrystalline quartz is composed of anhed-
ral grains 0.040 mm average size in coarser
bands, down to 0.020 mm average size in finer
grained bands.

Magnetite occurs as a dense cloud of fine or
coarser particles arranged in thin laminae
parallel to the bedding. The laminae may be so
close as to form a single thick band of magnetite
or may be as far as 0.25 mm apart. As a
general rule, the further apart these laminae are
the more microcrystalline quartz and the less
calcite there is between them. Hence, closely
spaced thick bands of iron ore contain abundant
interlaminated calcite. The grain size of the
magnetite is variable, thick bands containing
the coarsest magnetite; thin bands in quartz

92

resemble trails of dust when seen under the
microscope. On the average the magnetite is
of slightly smaller size than the associated
microcrystalline quartz.

Pyrite and minor pyrrhotite are found near
the edges of the iron- and carbonate-rich bands,
associated with a slightly coarser grained (re-
crystallized) quartz. They occur as irregular
crystals of varying grain size, generally much
larger than associated magnetite.

The association of pyrite in bands of mag-
netite and carbonate with a coarser grained
quartz suggests that some of this coarser grained
quartz has originated by metasomatic replace-
ment of carbonate.

The variation in grain size of the microcry-
stalline quartz is most likely not primary, but
is controlled by either recrystallization or
secondary introduction.

Chlorite is associated with pyrite
coarser grained microcrystalline quartz.

The microscopic characters of other jaspi-
lites, or of other specimens in the same forma-
tion, do not differ essentially from those above
excepting:

(1) where the jaspilite has suffered meta-
somatism in the ore shoots of the Hill
50 gold mine;

and

(2) where the jaspilite has suffered meta-
somatism causing the development of
actinolitic amphibole;

(3) where the jaspilite has suffered meta-
morphism adjacent to the granite;

(4) where the jaspilite outcrops at the

surface, extending to an unknown
depth, possibly hundreds of feet. Sili-
cification of the carbonate in the jas-
pilite yields a banded microcrystal-
line quartz-iron ore rock.

Metasomatic changes in ore shoots.—39576
(11610E, 12905N, 820’ level) is ore known at the
mine as massive pyrrhotite. It is a heavy,
metallic-lustred, bronze-yellow rock with a
faint relict banding. Specific gravity — 3.86.

In thin section the rock appears massive and is
composed dominantly of pyrrhotite and minor
pyrite (45%), quartz (30%), and carbonate
(25%). Extensive recrystallization has taken
place in this rock, quartz occurring in all sizes
from that of a normal jaspilite to 0.2 mm across.
Calcite and pyrrhotite also occur in larger
grains, averaging about 0.1 mm across.

It is concluded that metasomatic replacement
of magnetite by pyrrhotite and minor pyrite has
taken place. Most of the magnetite and car-
bonate was concentrated into bands before
ore formation. The process of metasomatism
has attacked these bands. Magnetite was re-
placed by sulphides and the carbonate recry-
stallized. Depending on the intensity of the
metasomatism, the associated microcrystalline
quartz was either poorly recrystallized for a
short distance on either side of these iron- and
carbonate-rich bands, or experienced all stages
of recrystallization.

A. B. Edwards (1955, p. 35) reported: “The
opaque minerals observed in the ore are mag-
netite and a trace of ilmenite, which are com-

jponents of the jaspilite rock forming the “host”
rock of the mineralization and were not in-
‘itroduced by the mineralization, and pyrite,
|pyrrhotite, chalcopyrite, (?) galena and gold.
|The gangue introduced by the mineralization
comprises quartz and a carbonate mineral.” He
‘concludes the ‘although the gold shows a dis-
{tinct preference for association with quartz it
iis genetically related to the sulphides. This is
apparent from the occurrence of occasional
“veinlets” of gold, about 0.50 mm X 0.005 mm,
tforming parts of pyrrhotite veinlets.

It is apparent also from the consistent
association of the gold with areas in which sul-
jphides are present, and its almost complete
-absence from areas lacking sulphides.”

Surface silicification of jaspilites—The evi-
‘odence of surface silicification is:

(a) At depth, the jaspilites contain abundant
»earbonate, while at the surface they contain
ionly traces of carbonate.

(b) At depth, the Perseverance Jaspilite and
tthe (?) Jupiter Jaspilite contain primary nodules
sof calcite which are overlain directly by a band
10f calcite (approximately +” to 3%” thick).
‘Whereas, at the surface these nodules are re-
ipresented mainly by cavities in several cases,
‘aS In specimen 39514, they are composed of
‘coarser microcrystalline quartz. Overlying these
modules at the surface is a band of microcrystal-
line quartz.

The depth to which the silicification extends is
mot known. Jaspilite on the 313’ level appears
Ito be highly siliceous in hand specimen while
(jaspilite from the 613’ level downwards contains
sa high percentage of carbonate. In particular,
sone specimen is composed almost entirely of
‘bedded carbonate and magnetite crossed by a
rvein of coarser microcrystalline quartz. The
[Hill 50 Jaspilite on the 1304’ level appears to
ave a higher percentage of carbonate than at
higher levels.

Within the quartz of the nodules of the Jupiter
Jaspilite there are particularly numerous and
often large inclusions of carbonate. The micro-
crystalline quartz in the body of the rock con-
Hains inclusions of carbonate of a much smaller
size. These are harder to detect and their
abundance is difficult to estimate.

Every thin section of jaspilite examined from
he area, even underground on the 820’ and
11060’ levels, contained inclusions of carbonate
of varying, generally minute size, in microcry-
stalline quartz. If these indicate silica replace-
ent of carbonate, then a very large portion of
ach jaspilite must once have been composed of
carbonate.

Surface silicification of jaspilites may be taken
laS proven. The silica in these jaspilites below
tthe zone of surface silicification may have
roriginated by metasomatic replacement of strata
Woreviously richer in carbonate. However, no
rvalid criteria have been found to prove this
ihypothesis.

Environment of Deposition of the Boogardie
Group

| The great thicknesses of sediments, isoclinal

folding and granite intrusion as in the Mount

agnet Greenstone belt are typical of ortho-

geosynclinal sequences. The lithologies in the
Boogardie Group (flow lavas, bedded chemical
and pelitic sediments and pure chemical sedi-
ments (jaspilites)) are characteristic of the
eugeosynclinal suite of Kay (1951). Further-
more, tectonic instability is implied by the
marked convergence of formations and mem-
bers, by a buried hill structure and by slumping.

The provenance of the detritus is probably
from within the eugeosyncline itself, rather than
from without. That is, the clastic material (and
some of the chemical) is primarily volcanic
ejectamenta which has passed through the
modifying influences of the normal processes of
clastic and chemical sedimentation before deposi-
tion. No orthoquartzite, greywacke or arkose is
intercalated in the section and the sequence
cannot be ideally equated with the tectonic
cycle of Krynine (1943, Fig. 1, p. 3).

The presence of small-scale cross lamination,
ripple marking and scour-and-fill structures in
the Boogardie Group is indicative of shallow-
water depositional environments.

Structure

The eugeosyncline has been recognized as a
wide, probably elongate, subsiding belt in which
voleanic activity is dominant. Flow lavas and
volcanic sediments form part of the geosynclinal
filing. When they accumulated more rapidly
than subsidence could accommodate them,
volcanic islands were formed, contributed peli-
tic material to the geosyncline and then sank
below sea-level, once subsidence overcame
accumulation. As a consequence of sloping sea
floors, members and formations were thickened
or thinned and slumping of unlithified material
occurred. Characteristically, the sediments of
an eugeosyncline next become buried and suffer
load metamorphism, isoclinal folding, faulting,
intrusion by granite and metamorphism. ‘The
structures within the Boogardie Group described
below are such that they could logically he
explained by the above type of environment.

Primary Structures within the Boogardie
Group
The following primary structures have been
recognized within the Boogardie Group: buried
valley structure; buried hill structure; calcareous
nodules; ripple marking and associated scour

and fill structure; cross lamination; thickening
slump

and thinning of units; structures and

brecciation.

Fig. 5.—Initial stages in the development of a buried-
valley structure showing the cutting of a channel.

CONTORTED JASPILITE

Fig. 6.—Final stage in the development of a buried-

valley structure showing diagrammatically the way in

which the upper jaspilite has slumped dcwn onto the
lower jaspilite, followed by burial.

Buried valley structure-—This structure is
developed at 10880E, 14420N. Figs. 5 and 6
illustrate diagrammatically how the structure
shown at this locality is thought to have de-
veloped. Initially, it is believed the two jas-
pilite members were conformably interbedded
with greenstone (Fig. 5) and that subsequently
a valley structure was developed by either slump-
ing or scouring. Fig. 6 illustrates the manner
in which the upper jaspilite member has slumped
into the valley so formed, the whole being
covered with later deposits of sediments.

Buried hill structure.—This structure has been
mapped at 10120E, 14500N. Fig. 7A, B, C, D and
Figure 8 illustrate diagrammatically how the
structure at this locality is thought to have
developed. Tectonism contemporaneous with
sedimentation has been the cause of both thick-
ness variations of the members and large-scale
and small-scale slumping.

Calcareous nodules.—The Perseverance Jaspi-
lite and the Jupiter Jaspilite contain small
calcareous nodules. Fig. 9 is a thin section
across a nodule in specimen 39585 from the
Perseverance Jaspilite (1060’ level of the Hill
50 gold mine). Laminae of chert, carbonate and
magnetite underlying the nodule have been
squeezed out in a compression fold. Overlying
the nodule there is a band of carbonate and iron
ore in which cross-bedding and gravity differen-
tiation of the minerals may be seen. Above
this latter band are laminae of chert, carbonate
and magnetite. The author believes that the
nodule grew by a process of chemical accretion
upon the depositional interface and that the
weight of the nodule was sufficient to depress
the laminae beneath it, adjustment being a com-
pression fold in a lateral direction. The growth
of the nodule was halted by the coagulation in
the overlying waters of carbonate and iron ore.

These were precipitated and the iron ore (by
reason of its higher specific gravity) became
separated from the carbonate. On reaching

the depositional interface the first of the car-
bonate and magnetite to arrive was cross-
bedded off the side of the protuding nodules
(supratenously). Subsequent precipitation
smoothed the new depositional interface.

94

Ripple marking and associated scour and fill
structure-—Ripple marking has been found in
unit 7 of the Mars Greenstone (specimen 39546)
(see Fig. 4). The cross lamination of the ripple
ridges and the collection of magnetite grains in
the deepest part of the trough serves to dis-
tinguish these structures from the fill structures
in the ripple valleys (Shrock 1948, p. 104). The
fill structures possess cross-lamination in which
the foreset laminae are asymptotic with one
slope and are sharply truncated on the other
slope. Magnetite and biotite are distributed

at the base of the foreset laminae.

Fig. 7.—Initial stages in development of buried hill
structure at 10120E, 14500N.

Deposition of the Poverty Flat Formation and

portion of the Jupiter Jaspilite.

Monoclinal warping during deposition of the

remainder of the Jupiter Jaspilite and subse-

quent deposition of the first two units of the

Mount Magnet Greenstone.

Repetition of the warping to produce a hill

structure. Slumping commenced due to steepen-

ing of depositional dip.

Normal sedimentation established over the

slumped sediments, unit 3 of the Mount Mag-

net Greenstone was laid down on a slightly

sloping sea floor and subsequently slumped.

Ke

ey
bit Dp
biti Pee

mies =
JASPILITE eRe DS atk] & aha
eae
PEAY VAN Oa Sac etc eh
Bee] MOUNT MAGNET GREENSTONE || Sm Ss

[2] supirer JASPILITE AMAT Teer

fo) POVERTY FLAT FORMATION

1=S5

Fig. 8—The buried hill structure as mapped after
regional tectonic deformation has been superimposed
on the initial structure.

Cross lamination.—Small scale cross lamina-
tion is developed within unit 7 of the Mars
Greenstone.

Thickening and thinning of units—The rapid
thickening and thinning of formations and
members has been demonstrated earlier in the
section dealing with the stratigraphy of the
Boogardie Group.

Slump structures and brecciation.—The dif-
ferentiation of slump structures and tectonic
structures depends upon criteria related to the
different modes of origin of the two features.
The following criteria have been used where
possible to determine the origin of contortions
in the jaspilites:—

(1) A highly puckered contortion of small
amplitude, contained within relatively undis-
turbed beds, suggests gravity slumping.

(2) Broad folds of large amplitude, repeated
in the underlying and overlying strata indicate
tectonic folding.

(3) A distinct or probable lineation running
parallel to the axis of a fold or series of folds is
evidence for tectonic drag folding, except in the
limiting case where the impressed lineation
coincides with the axes of true slump folds.

(4) Strata are considered to be slumped
where the folded beds are overlain by a local
unconformity.

(5) Slump folds are overturned towards the
depression whereas tectonic folds are overturned
towards the anticlines.

Many of the incompetent folds in the Boo-
gardie Group are indeterminate in origin.
Slumping requires a suitable unlithified state
of the sediment and either an initial slope of the
depositional interface or an external source of
energy such as could be provided by outpourings
of basic lava. These conditions are thought
to have existed during the deposition of many of
the jaspilites.

95

Slumping has been used above to indicate top
and bottom. Another use is to be found in the
correlation of jaspilite members. Slumping
being a primary feature could be typical of
certain units. Attention to slump features dur-
ing mapping might enable the correlation of
sporadic outcrops and of strata on each side of
fault blocks.

Tectonic Structures
The tectonic structures are folds, faults and
joints which have developed after deposition
and compaction of the sediments.

Folds.—Two types of folds have been recog-
nized, namely regional folds and drag-folds.

Fig. 1 illustrates the regional structure. The
Boogardie Group and the Lennonville Beds are
folded into a tight, southerly plunging, inclined
syncline which has characteristics as follows:—

(1) The axial plane trends 207° (S.S.W.).

(2) The west limb of the syncline strikes
approximately 225° (S.W.) and dips
approximately 60° E.

(3) The east limb of the syncline strikes
approximately 150° (S.E.) and dips
between 75° E. and 75° W.

(4) The lineation in hornblende schists

adjacent to the granite on the western
limb plunges 40°-65° S. in the direc-
tion 155° to 190° (S.E.-S.).

The folding has been interpreted as the result
of granite intrusion (see General Stratigraphy).

Drag-folds related to larger fold structures
are recognizable where they are not disharmonic.
The typical example is on the 313’ level of the
Hill 50 gold mine at (11623E, 12180N). At this
locality the junction between the Saturn Forma-
tion and the underlying Havelock Greenstone is
drag-folded. The north-plunging drag-fold is
related to the main north-plunging anticlinal
drag-fold of the mine at that level. The ‘Main
Oreshoot’ of the Hill 50 gold mine is structurally
controlled by a drag-fold (probably related to
the ‘Boogardie Break’ system) and there are
numerous disharmonic folds which are unreli-
able for structural interpretation.

Fig. 9—Thin section of a calcite nodule in the Per-
severance Jaspilite. Ordinary light X2.

N
N

"S05

/ >< cANGLOMERATE

WESTERN
AUSTRALIA

ROADS AND TRACKS

BOUNDARY POSITION APPROXIMATE
DUMP

INTERPRETED BOUNDARY
FRACTURE CLEAVAGE

FAULT POSITION APPROXIMATE
OPEN CUT

INTERPRETED FAULT

—

7%
~
3

»

WA. STRIKE AND PLUNGE OF DRAGFOLD

55

A. DIP AND STRIKE
“x CREEK OR GULLY

<vT,Y

i

Yuu
Pe
a

_X. FORMATION BOUNDARY

og FAULT

SCALE IN MILES

— LEGEND —

GEOLOGICAL MAP
OF
HILL 50 GML 1282M

AND SURROUNDING LEASES
BOOGARDIE MURCHISON GF

SATURN FORMATION

HAVELOCK GREENSTONE

MARS GREENSTONE
PERSEVERANCE JASPILITE
MOUNT MAGNET GREENSTONE
JASPILITE MEMBER
PORPHYRY

Vv

v
wv

F<] POVERTY FLAT FORMATION
[>] JASPILITE MEMBER

[*,°] THREE BOYS FORMATION
ESS suprrer saspiuive

FE} HILL 50 JASPILITE

ZA.
ws
ES
=<

EEE EO
— OOO

SHAF

eTamice =v

Ve yy

y
vy &@ MAIN

200

.
v

Persea

100
SCALE OF FEET —

Whit ites >
tual) ease ©:
< <
\iWies #2
\Va

> =
>?

< Ae es

aat¥ig a 2
A RC

Hil 50

-— “ll
.
.

Was “i
(eA YY
AN \\

\\\a9
‘van

mi\\
nerve

Ains\yyynny
re:

e
tune

Nep

.
"

Be 1246.”

Osh cso

BHF ORQASRBTNARA SHE BHA

Ke}

ios}

RH

uo)

subdivision of this thick sequence and in the
collection of rock types not described here.
Moreover, the group has not been drilled (apart
from shallow water bores) and no subsurface
samples are available. Nevertheless, the 53
specimens studied are probably — sufficiently
representative to justify a preliminary discussion
of the petrology.

Arrowsmith Sandstone
Petrography

The Arrowsmith Sandstone is a uniform
sequence of well-bedded, medium- to coarse-
grained sandstone that contains abundant feld-
spar and lithic fragments. All specimens
examined in the laboratory are arkoses, accord-
ing to the tabular sandstone classification of
Pettijohn (1957, p. 291). The mineralogy and
texture of specimen Pf2 is described below in
some detail, and is illustrated in Fig. 3A. The
rock is a medium-grained arkose made up
mainly of sub-angular to well-rounded quartz
and feldspar grains, and lithic fragments that
are generally angular. Its coefficient of sorting
(So) is 1.39 and the median grain diameter is
0.33 mm.

The feldspar is plagioclase (sodic oligoclase to
sodic andesine), microcline, microcline micro-
perthite and orthoclase. Lithic fragments in-
clude volcanic rock, sericite-quartz schist, many
composite quartzo-feldspathic grains, and
myrmekite. The volcanic fragments consist of
micro-phenocrysts of plagioclase in a red-brown
iron-stained groundmass containing some un-
altered black iron minerals: the fragments are
too small for precise classification, but much of
the plagioclase is fairly sodic and appears to be
oligoclase-andesine. Many of the grains are
surrounded by a narrow rim of fibrous, pale
green, authigenic chlorite which cements the
rock. Authigenic outgrowths on some of the
quartz grains also assist in the cementation.
Despite the presence of volcanic material, no
vitroclastic textures have been recognized.

Approximate composition of the rock by
volume is:

Percent
Quartz a : } 40
Feldspar aie ; as : PAL
Volcanic fragments ie ; 3
Other lithic fragments... : 13
Cement ns Bae ree 6
Other minerals ; 1

Heavy minerals were separated in bromoform
from the fraction with grain diameters between
0.66 mm and 0.124 mm, and compose about 15
per cent. of that fraction. Volcanic fragments
(generally angular) and rounded black opaque
grains (ilmenite with a little magnetite) are
abundant. Some of the volcanic fragments are
attracted by ordinary bar magnet, and at least
part of their contained iron minerals is there-
fore magnetite. Pale pink, slightly magnetic,
angular, strongly etched garnet is common,
and much of it is partly altered, apparently to
a mixture of feldspar and white mica. Clear
garnet grains have a refractive index close to
1.810 and they are probably almandine. Other
common heavy minerals are sphene and zircon.

CUMULATIVE WEIGHT %

The sphene is cloudy and angular, and there are
three varieties of zircon: colourless to pale
brown, very well-rounded grains; colourless to
pale brown, zoned, well-rounded grains; and rare
lilac euhedra. Rare heavy minerals include
rounded apatite, brown and black, rounded and
angular tourmaline, and rounded rutile.

/00

75

ey
fo}

Pate n
‘jf [+-#fo0

25

DIAMETER IN MM

Fig. 2.—Cumulative weight per cent curves of 8 speci-

mens of Arrowsmith Sandstone. The abscissae are on a

logarithmic scale. Note the high degree of sorting of
most specimens.

Hight specimens of Arrowsmith Sandstone
(36914) Pi 2, Pi silver 325) Piss) ein59o sein OF
Pit 61) were disaggregated by maceration in
water, and after mechanical analysis, cumula-
tive weight per cent curves were constructed
(Fig. 2). The curves do not precisely represent
the original sorting of the rock, which has
changed after deposition by secondary enlarge-
ment of some quartz grains and transformation
of cement to fibrous chlorite. Nevertheless, the
curves probably indicate fairly closely the
original sorting. The coefficient of sorting
ranges between 1.18 and 1.50 indicating a well-
sorted sediment (Trask 1932) and falls within
the range of sorting of most near-shore marine
sediments of sand grade, according to Hough
(1942).

The roundness of the quartz grains was com-
pared with the standards set forth by Krumbein
(1941). No precise grain counts were attempted,
as an unknown number of grains have an in-
duced angularity due to authigenic growth, but
some significant features are evident. Quartz
grains smaller than 0.124 mm in diameter are
practically all angular (.2 to .3), some grains
in the 0.124-0.246 mm range are slightly
rounded (.4 to .5), many grains in the

0.246-0.495 mm range are moderately well-
/ rounded (.5 to .6) and most coarser grains are
well-rounded (.6 to .8). It would appear from
the investigations of Kuenen (1959) that the
grains must have been modelled by either wind
or surf, and the earlier work of Twenhofel
(1945) suggests that wind traction may have
peen a factor in rounding the grains of less
than 0.495 mm diameter. The grains seem to
have been rounded in one cycle, for no earlier
formation capable of yielding rounded grains
has so far been recognized in the area.

Environment of Deposition

The parameters discussed above are consistent
with a near-shore, marine environment. The
abundant labile grains, such as feldspar and
volcanic fragments, indicate the absence of
effective chemical weathering. This may have
resulted from rigorous climatic conditions, but
other factors discussed at the end of the paper
suggest that climate was not the main influence.
It is more likely that dominantly mechanical
disintegration, anc rapid transport to the site
of deposition, prevented soil formation, with
its attendant chemical transformation of un-
stable grains. The Arrowsmith Sandstone was
derived from a granitic, meta-sedimentary and
volcanic terrain that was therefore probably
fairly rugged, and it appears to have heen
deposited near the strand line of a subsiding
basin.

A.—Arrowsmith Sandstone
spar.

(specimen Pf2).

Cement is fibrous chlorite.
(specimen 36916).

field.

B.—Arrino Siltstone
dark grains are volcanic.

C.—Enokurra Sandstone (specimen 38725).
and lined grains are muscovite.

sericitic matrix (closely stippled).

Lithic fragments are volcanic rock (dark) and
(lower left centre) shows authigenic enlargement, and there are more a
Diameter of field 2.4 mm.
Clear grains are quartz, stippled grains with cleavage are feldspar and
Fine grains are commonly difficult to identify, but most are quartz.
is sparse and forms an iron-stained rim to most grains.
Clear grains are quartz, stippled grains with cleavage are feldspar
Cementation is by enlargement of quartz grains and by clay-sized and
Diameter of field 2.4 mm.

Arrino Siltstone
Petrography

The Arrino Siltstone is a uniform sequence of
poorly bedded, dark, reddish brown, micaceous
siltstone that rests conformably on the Arrow-
smith Sandstone. Four specimens (Pf 3, Pf 4,
36915, 36916) were examined microscopically.
The rocks are sandy siltstones, with quartz con-
tent ranging from 37-73 per cent and all con-
tain numerous lithic fragments which are
generally concentrated in the coarser fractions.
The content of volcanic fragments ranges from
7-15 per cent and that of other lithic frag-
ments from 10-43 per cent. The volcanic frag-
ments consist of micro-phenocrysts of plagio-
clase in a red-brown, iron-stained matrix that
also contains abundant black magnetite: the
other fragments include garnet granulite,
muscovite-bictite-quartz schist, chlorite-quartz
schist, (2) chert, epidote-quartz rock, and other
composite grains commonly containing quartz,
carbonate and chlorite. Feldspar (plagioclase
and microcline) comprises up to 5 per cent of
the siltstone. Both sand and silt grains are
generally angular, and are commonly cemented
by pale green, locally fibrous, chloritic cement
that constitutes from 3-8 per cent of the rock.
In places, the cement is stained red-brown.
Specimen 36916 is illustrated in Fig. 3B.

None of the rocks examined contains recog-
nizable vitreclastic texture, and there is no
evidence that they are tuffaceous.

()
iq)
4

V

<3

Fig. 3.
Clear grains are quartz and stippled grains with cleavage are feld-

interlocking quartz. One quartz grain

grains of } i
ngular grains than is usual in this

Cement
Diameter of field 2.4 mm.

intermediate andesine and most seems con-
siderably more sodic. Other fragments are
fine-grained quartzite, chert, (?) hornfels,
chlorite-quartz schist, chlorite-epidote-quartz
schist and granitic rock.

Other specimens include siltstone (38712,
pfs, Pf9, Pf{10) and very fine-grained sandstone
‘Pf7). Volcanic fragments make up part of
the rocks (13-45 per cent) and other consti-
tuents are quartz, plagioclase, epidote, chlorite,
muscovite, biotite and minor microcline. Grains
are generally angular. Heavy minerals from
Pf7 (apart from volcanic fragments) are sparse,
and include subrounded to well-rounded apatite,
cloudy epidote, colourless, angular, faintly
magnetic garnet, biotite and rare pyroxene.

No undoubted vitroclastic texture has been
observed in the fine-grained rocks described
above, Dut a few pale brown, weakly anisotropic
fragments that probably represent devitrified
glass shards are present in them. Moreover,
some volcanic fragments have unusual shapes,
as though deposited when plastic. These fine-
grained rocks are likely therefore to be partly
tuffaceous. Unfortunately, fine-grained, water-
laid, lithic tuffs and fine-grained sediments
derived from erosion of a volcanic landmass are
not easily distinguished, even if of fairly recent
origin. In neither sediment are grains likely
to be rounded, and proof of tuffaceous origin
is to be sought in the presence of glass shards,
pumiceous fragments and embayed crystal frag-

ments as described by Pirsson (1915). The glass
cf submarine tuffs may alter rather quickly
however (see for example Miiller (1958) on

Recent sediments in the Bay of Naples), and the
features listed by Pirsson tend to become
unrecognizable under the influence of compac-
tion and diagenesis in older rocks such as those
of the Yandanooka Group. It is clear, however,
that much of the material is of normal, epi-
clastic origin, for it has been derived from
granites and meta-sediments. It is also evident
that voleanic fragments in the conglomerates
and coarse-grained sandstones are epiclastic, for
they are rounded, and are associated with
rounded fragments of similar size which are
not of volcanic origin.

Environment of Deposition

The Mt. Scratch Siltstone is a thick (25,000-
30,000 ft.), generally fine-grained sequence
derived from a terrain of volcanic, granitic and
meta-sedimentary rocks. Volcanism was ap-
parently active during sedimentation: some
volcanic material was derived from erosion of
the landmass, whereas some was probably
blown directly from volcanoes to the basin of
deposition and incorporated in the accumulating
sediment. The fine-grained rocks probably
represent a mixture of epiclastic and pyroclastic
Ceposition.

The general fineness of the formation sug-
gests deeper waters than those prevailing dur-
ing deposition of the Arrowsmith Sandstone,
Beaconsfield Conglomerate and Enokurra Sand-
stone. Ripple marks have been observed locally
by Playford (pers. comm.), but without careful
study their value as a criterion of shallow water
deposition must be accepted cautiously. Cur-

102

rent ripple marks have now been observed on
seamounts at depths of over 4,000 feet in recent
oceanographical work (Dietz and Menard 1951,
p. 2004; Heezen, Thorp and Ewing 1959, p. 59).
Small scale cross bedding is a characteristic
feature of the formation, but its significance
with regard to depth of sedimentation is not
understood. The great thickness of the Mt.
Scratch Siltstone, and the fact that it is under-
lain by the shallow marine or continer:tal
Enokurra Sandstone shows that subsidence dur-
ing sedimentation was considerable.

Origin and Deposition of the Yandanooka
Group

The Yandanooka Group was derived from a
terrain of granitic, meta-sedimentary and vol-
canic rocks. Mest of the sediments are notable
for the high proportion of volcanic fragments
in them, and they have long been regarded as
tuffaceous (Simpson and Glauert in Camptkell
1910, p. 97: Woolnough and Somerville 1924:
Johnson et al. 1954). Baker (1951) however
was unable to find evidence of pyroclastic
derivation for rocks he examined in the
sequence stratigraphically below the Mt. Scratch
Siltstone, and evidence cited in this paper sup-
perts his conclusions. The possibility that closer
sampling of the Arrino Siltstone will reveal
tuffaceous bands cannot, of course, be eliminated
at present. The Mt. Scratch Siltstone is prob-
ably partly tuffaceous in origin. The volcanic
fragments are commonly too altered for effec-
tive determination for their feldspar is cloudy,
and ferromagnesian minerals have been re-
placed by iron ores and chlorite. Their com-
position ranges from microdiorite (andesite) to
spilitic varieties.

Formations in the Yancanooka Group repre-
sent successively, from the base upwards, shal-
low water, probably littoral faces (Arrowsmith

Sandstone); deeper water facies (Arrino Silt-
stone); shallow water or continental piedmont
facies (Beaconsfielau Conglomerate); shallow
water or continental facies (Enokurra Sand-
stone); and a thick, fine-grained, partly tuf-
faceous facies, apparently deposited in fairly
deep water (Mt. Scratch Siltstone).

The high proportion of labile constituents,
such as feldspar and volcanic fragments, shows
that the processes of mechanical disintegration
dominated over those of chemical decay. The
tectonic environment that best explains the
petrography of this thick sequence is that of a
speradically subsiding basin bounded by a fairly
rugged land surface. Some of the volcanism,
at least, was contemporaneous with sedimenta-
tion.

The Enokurra Sandstone, unlike formations
below and above it, contains no volcanic detritus.
It was eroded from a granitic and meta-
sedimentary provenance, and presumably came
from a different direction from that of the other
sediments. Its distinctive lithology, and the
unconformity at its base, are best explained by
sudden tectonism.

The areas from which the Yandanooka sedi-
ments were derived are not yet known. The
nearest volcanic rocks are east of the Darling
Fault near the Billeranga Hills, where intru-

13.—Rainfall and Soil Control of Tree Species Distribution around
Narrogin, Western Australia

IBY hve eaaneie!
Manuscript received—21st April, 1960

In an area around Narrogin, Western Aus-
tralia, soils relate to erosional and depositional
surfaces, and to rainfall. Tree species incidence
also relates to the surfaces and to rainfall.
Relationships between surfaces, soils, tree inci-
dence and rainfall are presented. Evidence is
also produced for species migrations and this
is discussed in terms of the climatic and geo-
morphic history of the area.

Introduction

This report is of an autecological study of the
relationships of twelve tree species to the rain-
falls and soils in an area of transition in Western
Australia, and concerns the extent to which
these environmental factors and their histories
control the tree distributions. The area of
study (Figs. 1 and 2) surrounds the township of
Infsageesbe, Clens B> By So iWopays, ily? il’? Te )S ite
constitutes a rectangular strip of about 700
square miles between Cuballing in the north and
Highbury in the south, from Williams in the
west to Toolibin in the east, and is extensively
cleared for agricultural purposes.

The Environment
Climate

The area occurs in a climatic zone of winter
rainfall and summer drought (Gentilli 1956),
and the rainfail distribution in the area is
illustrated in Fig. 1. The annual precipitation
decreases on a fairly even gradient from 25”
of annual rainfall west of Williams to 15” per
annum in the east of the area.

Geology

The parent rocks in the area are the Precam-
brian granites and gneisses of the West Aus-
tralian Shield, with occasional basic intrusions
(Wilson 1958).

Geomorphology and Soils

The relief and drainage in the area are shown
diagrammatically in Fig. 2, which shows both
drainage west to the sea, and east to the salt
lake systems, hence the area includes elements
of Jutson’s Swanland and Salinaland physio-
graphic divisions (Jutson 1955). According to
Jutson, the lateritic peneplain or ‘‘old plateau”
of Western Australia occurs in the former divi-
sion as the Darling Plateau, and in the latter
division as relicts, particularly in the form of
mesas and buttes. Normal erosion in the
stream-bearing Swanland division is not suffi-
ciently advanced to obscure the old plateau, but
Salinaland exhibits a new plateau of arid

*Institute of Agriculture, University of Western Aus-
tralia, Nedlands, Western Australia.

104

64 MILES

Fig. 1—Locality plan and rainfall distribution.

erosion, and the boundary between the two divi-
sions is the line separating rivers of the coast
from interior drainage.

In the study area, drainage west occurs in
relatively sharp valleys, while that to the east
occurs in broad flat-bottomed ones, constituting
western extensions of Jutson’s “new plateau”
of arid erosion. In the study area, the old
plateau is largely destroyed, and is preserved
only as residuals on the divides. At York,
Mulcahy (1959) studied the soils of the old
plateau, and the erosional and depositional
surfaces resulting from its breakdown, and
showed that the distribution of these surfaces
determines the distribution of soils. A similar
series of erosional and depositional surfaces
occur in the study area, and although some
have no described equivalents near York, most
are the equivalents of surfaces described by
Mulcahy.

The distribution of the surfaces themselves
is controlled by the geomorphic history of the
area, and the older surfaces are distinctly relicts
left after the destruction of a great deal of their
previous extensions. These latter have the bulk
of their present distributions west of the study
area. The oldest surface in the study area
(Quailing erosional) is laterite, of Jutson’s
lateritic peneplain or ‘‘old plateau,” but lateritic

@WILLIAWS::

+t +34 DIVIDE

<-DRAINAGE

Fig. 2.—Relief and drainage.

profiles are not confined to the much modified
told plateau preserved on the divides. They are
‘found also as residual spurs and terraces on
walley sides and floors respectively, representing
an early phase of erosion cutting into the old
plateau. Subsequent erosions and depositions
have given rise to younger surfaces which
marry soils developed on the truncated older
\lateritic surfaces, or, where more deeply cut,
soils developed on exposed fresh rock.

This situation contrasts markedly with that
wertaining in the high rainfall zone of the Darl-
ing Range, to the west of the study area. That
area, although much dissected, carries laterite
fetrital material on valley sides and floors and
shis is often recemented such that nearly all
soils are laterites. Thus laterites in the study
area can be regarced as outliers.

In the east of the study area some of the
vcunger laterites of valley sides were found
'O be calcareous. Considering the origin of
haterite (Prescott and Pendleton 1952), this
fime must be secondary, i.e. brought in by wind
or ground water. Laterite development fol-
rowed by the occurrence of secondary lime
‘learly indicates formation in a wet climate,
sollowed by relatively arid conditions.

In the study area, there is a close correla-
ition between the surfaces and their associated
hoil types, and surfaces encountered in the area
re listed below, against their topographical
-osition and soil characteristics.

1) Surfaces Iccated on divides.

Quailing erosional—Massive residual laterite
‘r+ heavy ironstone gravel.

Quailing depositional_—Deep yellow
eposit derived from laterite.
Degraded Quailing erosional.—Laterite resi-
uals reduced to a thin veneer of ironstone
hravel over pale reddish clays resembling those
tf the Balkuling surface (q.v.).
Kauring.—Grey sand over massive ironstone.
Monkopen.—Deep grey sand in depressions.
Granite outcrop und associated sandy de-
iosits—Shallow skeletal soils and associated
ep sandy and gritty soils.

sands.

105

(2) Surfaces located on valley slopes and inter-
fluves.

Belmunging.—Spurs and ridges carrying iron-
stone gravelly soils extending from the divides
down towards the drainage lines. May be cal-
careous in the east of the area.

Breakaway face.—Pediments below the break-
aways, carrying pale reddish and greyish clayey
soils, often with a thin scree of ironstone gravel.
May be calcareous in the east of the area.

Malebelling erosional.—Brownish gritty sands
over yellow and red mottled weathered rocks.

Malebelling depositional—Brown or greyish-
brown gritty sands, often with a prominent
bleached A» horizon over variously mottled
weathered rock.

York.—Brown loams and loamy sands over
reddish-brown clays. Close to drainage lines.

(3) Surfaces located on valley floors.
Avon.—Brown or grey clay at the surface.
Mortlock.—Lateritic valley terrace.
Truncated Mortlock.—Grey sands over domed

clay. Often calcareous in deep sub-soil.

Sandy ailuvium.—Brown sand over yellow clay.
Not calcareous.

Flood plain sands.—Sandy deposits of braided
stream pattern associated with the lake system
in the eastern part of the area.

Baandee—Resembie deposits at Baancacee
(Bettenay, priv. comm.). Fine textured cal-
careous aeolian deposits associated with lakes
in the south-eastern part of the area.

(4) Fresh complex.

The situation pertaining in principal water-
courses.

The outstanding edaphic discontinuity in the
area occurs at the 20” annual rainfall isohyet,
which marks the region of transition between
external and internal drainage. Avon, Trun-
cated Mortlock, Sandy Alluvium and Baandee
surfaces of arid erosion occur in valley ficors
to the east of that isohyet only, together with
all calcareous soils, and thus the 20” annual
rainfall ischyet marks a pedocal-pedalfer boun-
dary in the area. Calcareous soils occur to the

iof the Monkopen deposits. It is notable that
while such an effect may simulate conditions
tof high rainfall in a low rainfall area, there
is no situation in the study area to simulate
cw rainfall in high rainfall regions.

The pedalfer-pedocal transition is the princi-
jal soil discontinuity in the area, and this
‘mecurs in the region of the 20” isohyet and the
associated transition from internal to external
idrainage. This line marks the margins of con-
tinuous distributions of 4 species in the study
area. Within the areas of pedalfers or pedo-
2als, tree species exhibit specificities to surfaces
and their associated soils, controlling local pat-

cerns of tree incidence. This control of dis-
|| cribution was not revealed by the grid system

jused to reconstruct the geographic distributions,
and was studied by a point sampling procedure.

Surface determinations were made at each
of 140 points distributed evenly across a broad
median east-west transect of the area. Tree
rneidence on these surfaces was observed and
species were scored as represented or not-
sepresented at each point. Scores from the 140
joints were listed as west or east of the 20”
annual rainfall isohyet, to compare species in
| -he areas where their continuous distributions

overlap, and scores from the sites of disjunct
} stands were listed separately. Each of these
| ists was totalled, and tree incidence-surface
interaction was detectable, ie. certain species

loxophleba

marginati

EB.

BE. salmonophloea

E. calophylla

4
E. longicornis

N. floribunda

s
m

Jiho

Balkuling (ees
+ — — | Belmunging Vera
Quailingerosional + + —
Quailing deposi-

tional ;
Monkopen a

Sandy alluvium | + 4

never occurred on certain surfaces and certain
species always ocurred on certain surfaces. The
various species exhibited a range of surface-
Specificities and conversely, surfaces varied in
the numbers of species incident on them. These
basic interactions are summarized in Table I,
where an entry in the negative (-) signifies that
the species so qualified never occurred on the
surface referred to, at any of the sample points.
Positive entry (+) signifies incidence of the
species on the surface referred to, at one or
more sample points.

Knowing the associated soils of the different
surfaces, the relationships of the different
Species to the soils in the study area may also
be summarized.

Nuytsia floribunda.—Restricted to deep sands
derived from laterite and deposited in water-
collecting hollows.

Banksia grandis —RFestricted to massive later-
ite, heavy ferruginous gravel or deep sands
derived from laterite and deposited in water-
collecting hollows.

Eucalyptus rudis—Restricted to the situation
pertaining in principal watercourses and to wet
drainage lines in loams and gritty sands over
mottled weathered rock.

Eucalyptus marginata.—Restricted to massive
laterite or heavy ferruginous gravel, sometimes
adjoining sands deposited in a water-collecting
hollow.

TABLE I*

elata

huegeliana
redunca var.
acuminata
loxophleba

rudis

E.

BE. calophylla

E. astringens

(Of;
E.
A.

Avon

Baandee Fresh complex
Breakaway face Breakaway face
Floodplain sands Granite outcrop

Granite outcrop - -

Monkopen

Monkopen a Malebelling de-

positional
Quailing erosional -+- — = Granite sandy
soils
Granite sandy = = == | Ikayuisie

soils
York e 6 = ea Mortlock
Quailing deposi- - +
tional
Quailing residual | ~- - .
|

Quailing deposi-
tional
| Quailing erosional

York

Truncated Mort- | — 1
lock |
Mortlock 0 et | == Belmunging
Malebelling
| erosional
Balkuling

Belmunging

we

Balkuling

upply beyond the study area.

inhloea occurs On

Balkuling surfaces that are neither calcareous nor de

109

ik rfaces Gills eal | pen surfaces ¢ ree species incidence, described in this paper, do not necessarily
ha relationships between surfaces and soils and between surfaces and tree speci » described in th n arily
pane cotanionehe ? For example, on traverses between Y ork and Quairading where essentially similar surfaces occur, E. salmono-

rived from basic materials.

Eucalyptus astringens.—Restricted to degraded
Quailing, Balkuling and breakaway surfaces.
The soils are pink or white weathered rock
usually with a scree of ferruginous gravel.

Eucalyptus salmonophloea.—Occurs on vailey
clays, or sands over domed clays, calcareous or
calcareous at depth. It also occurs on those
lateritic surfaces which contain secondary lime.

Eucalyptus calophylla.—Occurs principally on
non-calcareous sandy soils and on laterite within
its continuous distribution. Its relict represen-

tation at Toolibin is on relict deep sand
deposited in a water-collecting hollow.
Eucalyptus loxophieba—Occurs on brown

gritty sands over mottled weathered rock, on
leams, and on calcareous clays and sandy soils.
It occurs on those lateritic surfaces which con-
tain secondary lime.

Eucalyptus longicornis.—Occurs principally on
calcareous soils associated with valley floors
and salt lakes, but occurs also on calcareous
laterites and on soils of the Balkuling surface
derived from basic parent materials.

Acacia acuminata.—Occurs on various soils,
particularly sandy alluvium, loam and granitic
skeletal soils. It does not occur on very cal-
careous clays and sands, or on massive laterite,
breakaways, heavy ferringinous gravel or deep
sands.

Casuarina huegeliana.—Occurs on _ various
soils, particularly granitic skeletal soils and
associated sands but not on very calcareous clays
and sandy soils, or on loams or breakaways.

Eucalyptus redunca var. elata.—Occurs on
many soils, but not on very calcareous sands and
clays, or on granitic skeletal soils.

Discussion.

The two environmental factors of rainfall and
soils have a marked influence on tree distribu-
tions in the study area. and the occurrence and
distribution of the different soils themselves
relates to rainfall, which has brought about
soil changes in geological time. In the area
studied, margins of continuous tree distributions,
the pedalfer-pedocal boundary and the 20” rain-
fall isohyet all relate closely. Within areas of
continuous tree distributions, incidence patterns
relate closely to locai soil patterns. Only dis-
junct tree distributions on soils of some relict
surfaces do not relate to present rainfall. This
is not altogether inconsistent, however, as a
high ground-water effect simulates high rain-
fall conditions in many of these instances, and
there is evidence that disjunct distributions re-
late to past rainfall distribution.

All detected disjunct tree distributions aye
of species now otherwise distributed west of, or
in the west of the study area, in regions of
relatively high rainfall. In the case of E.
marginata, disjunct stands complete a_ line
observed to extend from areas of continuous
distribution west of the study area to outliers
at Jilakin Rock and Hyden. Similarly, £.
calophylla and B. grandis exhibit eastern dis-
junct and western continuous distribution. Tree
species now distributed continuously east of the
20” isohyet do not exhibit western outliers.

Disjunct tree distributions are located on
old lateritic soils on or derived from Jutson’s

110

“old plateau.” This was considered to have
had broad distribution as a peneplain under
conditions wetter than the present and favour-
able to laterite formation, and to have under-
gone subsequent arid erosion, resulting in its
partial destruction in the interior (Jutson 1955).

Wet climate of the kind necessary for laterite
formation has certainly changed in nett effect
towards arid conditions, indicated by secondary
deposition of lime in relict lateritic surfaces in
the east of the study area.

With increase in aridity, the original rainfall
and edaphic conditions have retreated west-
wards, and the sites of disjunct vegetation
stands are relicts of the original conditions,
where high soil-water availability substitutes
for rainfall. Aridity permits accumulation of
lime, and calcareous soils have extended west-
wards on the surfaces of the new plateau. Even
some relict lateritic surfaces have become cal-
careous under the drier conditions, and cal-
careous surfaces have been occupied by species.
which have migrated westwards as a result:

According to Crocker (1959), who recently
summarized the known history of vegetation and
climate in Australia, peneplanation and ap-
parent humid climate of the Tertiary limited
habitat diversity in Australia, and relatively
humid times may have persisted to the early
Recent with some arid periods. Subsequent
severe aridities in the late Quaternary elimin-
ated many vegetational units and resulted in the
retraction of others to more favourable situa-
tions. Since that time there has been expan-
sion. The genera studied here, with the
exception of records for Nuytsia, were apparently
all established by the end of the Tertiary.

The simplest hypothesis accounting for the
disjunct distributions in the study area is that
the species involved had continuous distributions
over the area, under the rainfall and soil con-
ditions preceding the most recent aridity.

Acknowledgments

The writer would like to express his thanks to
Professor E. J. Underwood and Dr. C. A. Parker,
who made this work possible; to Mr. M. Mul-
cahy for his assistance in the field, and to Mr.
D. M. Churchill and Mr. G. M. Storr for the
data for Figs. 3-14.

References

Crocker, R. L. (1959).—Past climatic fluctuations and
their influence upon Australian vegetation.
In “Biogeography and Ecology in Australia.”
pp. 283-290 (W. Junk: Den Haag.)
Gardner, C. A. (1942).—The vegetation of Western Aus-
tralia with special reference to the climate
and soils. J. Roy. Soc. W. Aust. 28: xi-lxxxvii.
(1956).—‘*‘Weather and Climate in Western
Australia,” (W. Aust. Govt. Tourist Bureau.)
Jutson, J. T. (1955).—The physiography (geomorphology)
of Western Australia. Bull. Geol. Surv. W.
Aust. 95 4th Ed. (Govt. Printer: Perth.)
M. J. (1959).—Topographic relationships of
laterite near York, Western Australia. J.
Roy. Soc. W. Aust. 42: 44-48.
Prescott, J. A., and Pendleton, R. L.
and lateritic soils. Tech. Commun. Bur. Soil
Sci., Harpenden. No. 47.
Teakle, L. J. H. (1938).—Soil salinity in Western Austra-
lia. J. Agric. W. Aust. Ser. II. 15: 434-454.
Wilson, A. F. (1958).—Advances in the knowledge of the
structure and petrology of the Precambrian
rocks of south-western Australia. J. Roy.
Soc. W. Aust. 41: 57-83.

Gentilli, J.

Mulcahy,

(1952) .—Laterite

Ce

PLATE I
Fig. 1.—Dingo paintings near Burralumma Spring, in Unggumi tribal territory. Map locality 1.

ig, 2—Wandjina, yam, and (?) “lightning men” paintings, near the eastern entrance to Windjana Gorge. Note
the two small human figures painted beneath the Wandjina. Map locality 2.

113

Fig. 2.

7

eo? 4 eae & |

-..

PE AH. Le

Fig. 1.—Bark painting of Nura Nura, the “lightning man.’ Worora Tribe, Mowanjum Mission, Derby.

-Wandjina painting in Unggumi tribal territory, between Carpenters Gap and Windjana Gorge, on the north
side of the Napier Range. Note the small female Wandjina beside the main figure. Map locality 3.

114

PLATE IV

Fig. 1.—Paintings of ‘‘stick-men” or “lightning men,” on each side of a yam. Map locality 6.
Fig. 2.—Paintings of a Bungarra lizard and an eel, 44-mile south of map locality 6. Bunaba tribal territory.

PLATE VI
Fig. 1—Cave gallery near Elgie Cliffs homestead, in territory of the Gidya Tribe. Map locality 9. The Nurunguni
figures include snakes (probably the rainbow serpent), frogs and a wallaby.
near Dampier Downs homestead. Map locality 11. The designs

symbolize the journeyings of the All-Father during the primeval dream-time. Note also the small human figure
on the right.

Fig. 2.—Paintings in a cave in Mangala country,

118

The Rock Paintings

The rock paintings described in this paper
secur in the tribal territories of the Unggumi,
Bunaba, Kuniandi, Gidya, Nyigina, and Mangala
Tribes.

Unggumi Tribe

Paintings have been studied from three locali-
ties in Unggumi country, on the northern side
of the Napier Range near Windjana Gorge.
This gorge is one of the most impressive physio-
zraphic features of the Kimberley District. It
is formed where the Lennard River has cut
through the limestones of the Napier Range, 70
miles east of Derby. On the southern side of
the Napier Range the territories of the Unggumi
sand Bunaba Tribes met at Windjana Gorge, but
‘on the northern side of the range the Unggumi
extended several miles east of the gorge. The
Wunaba name for the gorge is “Talay,” and on
‘most maps of Western Australia it is called
“Devils Pass,’ a mame which is probably due
to some ‘‘devil-like’’ Aboriginal paintings in a
seave near the eastern entrance to the gorge.
Other paintings are known 43 miles east-north-
east of the gorge and between the gorge and
‘Carpenters Gap.

Burralumma Spring.—On the north side of
ithe Napier Range, 44 miles east-north-east of
‘Windjana Gorge, there is a permanent fresh-
ywater spring, known to the Aborigines as
(Burralumma (map locality 1). In a small lime-
sstone cave 25 yards west of the spring there
iis an impressive painting of two large dingoes
\(Plate I, Fig. 1). The paintings are 8 feet wide,
and are executed in white clay, red ochre, yel-
flow ochre, and charcoal. Both animals are
sshown defecating, and the anus of each is
‘strongly outlined in brilliant red ochre. Other
«small paintings are drawn around the dingoes,
iincluding stars (windinya), the bardi grub,
ssmall men, and a number of shapeless unidenti-
tied objects. The general name used by the
fUnggumi for such rock paintings is Manjimanji-
igardingi.

Unfortunately full details of the myth con-
tmected with this painting now appear to be
lost. Only two male members of the Unggumi
Tribe are alive today. I interviewed Paddy
(Dijul), the older of these two men, but he
had never visited this cave, though he knew of
its existence. He later confirmed the story
vabout the paintings given to me by an old
Ungarinyin man named Napier Paddy (Panda-
marra). This man has lived most of his life
en Napier Downs Station in Unggumi tribal
country, and he had visited the cave several
years before with old Unggumi men. It appears
that the two dingoes (called kia) came from
the Leopold Downs area, where they were
responsible for clearing all the trees from the
wide black-soil plains, during the far-off dream-
time (Ungud). They travelled along the north
side of the Napier Range to Burralumma, where
they dug a deep hole which they filled with
water to form the present spring. Following
this they went to the nearby cave and turned
into paintings. The reason why the dingoes
are shown Gcefecating was either not known by
my informant or he was unwilling to tell. He

- male Wandjina figure (map locality 3).

did not know the mythological significance of
the other smaller paintings in the cave.

Windjana Gorge.—In the limestone cliff about
300 yards south-east of the eastern entrance to
Windjana Gorge, there is a large cave follow-
ing a bedding plane in the limestone (map
locality 2). It is decorated by numerous paint-
ings of men and animals, the main feature being
a large male Wandjina figure 6 feet 9 inches
high, standing upright (Plate I, Fig. 2). He is
drawn in white clay and red ochre, and shows
most of the features of the Wandjina paintings
of the Worora and Ungarinyin, including the
horseshoe-shaped band around his head, and
the lack of a mouth. However unlike Wand-
jinas of these tribes his sex is clearly indicated.
Two small figures beneath his feet may be his
children. On his left-hand side are two devil-
like figures with long ears and upstretched
hands, painted in red ochre, on each side of
a large yam. Similar ‘devils’ are painted in
other parts of the cave. It seems very likely
that each of these figures is of a “lightning
man,” though I was unable to confirm this.
The Worora paint similar figures which they
call Nura Nura, “the lightning man,” the main
feature of this person being his long ears (see
Plate II, Fig. 1). Nura Nura is believed to be
responsible for the lightning during the closing
stages of a rain storm. Other paintings in the
cave include a dingo, further human figures,
crocodiles, and several stencilled hands.

I was unable to learn much about this gallery
from the natives, though several knew of its
existence, and one Bunaba man told me that
his hand was stencilled there. The Unggumi
use the name Wandjina to describe the main
figure in the cave, and I was told that his main
function was to bring the rain every wet season.
My Unggumi informant told me that in addi-
tion to touching up the painting it was neces-
sary to break a certain tree near the Wandjina
cave in order to bring rain.

Below the Wandjina cave is another cave with
a smoke-blackened roof and a clay floor, which
may serve as a good site for archaeological
excavations, both for Aboriginal and animal
remains. It is interesting to note that the bones
oft the giant extinct marsupial Diprotodon aus-
tralis were found in gravels near the western
entrance to Windjana Gorge by the first geolo-
gist to visit the area (Hardman 1887).

Between Windjana Gorge and Carpenters
Gap.—About 2 miies east-south-east from
Windjana Gorge, on the north side of the Napier
Range, is another large cave containing a typical
The
cave is about 1 mile from Carpenters Gap, and
is in Unggumi tribal territory.

The Wandjina is shown lying on his side and
is 5 feet 9 inches long (Plate II, Fig. 2). He is
painted in orange and red ochre, white clay, and
charcoal, and has eyes but no nose or mouth.
Beside his head is a small female Wandjina. A
few other poorly executed paintings occur in
this cave, and there are numerous grooves cut
in the limestone rocks, the purpose of which is
not known.

I could not locate any natives who knew this
cave, though I was assured that it must be in

iA)

original Unggumi country. Further paintings in
the area once occupied by this tribe along the
Napier Range have been noted by other geolo-
gists, but I have not been able to visit them.

Bunaba Tribe

The Bunaba Tribe occupied the country west
of the Fitzroy River around Fitzroy Crossing,
the Oscar Range, the northern end of the Napier
Range, and the southern part of Leopold Downs
Station.

Numerous caves and rock shelters occur in
the limestones of the Napier and Oscar Ranges
and many are decorated with paintings. Th<e
Bunaba deny that these are actually paintings,
and refer to them as Nurunguni or, less often.
Djeralii. The Nurunguni are said to be men
and animals who inhabited the world during the
far-off dreamtime and who left their images in
caves (Nowani) after their journeyings were
over. Nurunguni is also a name for the dream-
lume, other expressions for this primeval period
in the Bunaba language being Ungud and
Djurda, terms which are also common to the
Unggumi tongue. Bunaba men told me that
even prior to white settlement the paintings
were never touched up by natives, in fact several
ot them told me severely that to touch any of
the Nurunguni would result in the crippling of
the offending limb.

Typical Wandjina-type paintings occur in
Bunaba territory, and these are referred to as
Nurunguni-Nowungoo. Nowungoo is the Bunaba
word for father, and the figure represented in
the paintings is supposed to be the All-Father,
who journeyed through the tribal territory dur-
ing the far-off dreamtime creating the physical
features of the country. With him were the first
representatives of the various animal and plant
species. At the end of their journeying they
went to the caves and rock shelters, leaving
their images in the rock. I was told that there
was only cne All-Father, but he left his image
in a number of different localities. There does
not seem to be any clear connection between the
Nurunguni-Nowungoo and rainmaking, even
though he is shown as a typical Wandjina. Move-
over I was told that his horseshoe-shaped head-
dress is the rainbow. Rain-making was achieved
in the Bunaba Tribe by means of “rain stones.”

The rainbow serpent forms an important part
of Bunaba mythology. He is generally referred
to as Ungeroo, but is Sometimes called Galery
or Ungud. He is believed to have made the
water-holes, and his present function is to keep
up the supply of spirit children in these water-
holes, from which they can be incarnated
through the “dreams” of their fathers.

Numerous paintings have been found in the
Oscar Range, and only the more interesting of
these will be described here.

4-mile south-east of Elimberrie Spring. —
Elimberrie is a well-known permanent spring on
the north side of the Oscar Range. A large cave
is present in the limestone just to the south-east
of the spring (map locality 4), and this is
decorated with numerous Aboriginal paintings.
The main figure is that of a man about 5 feet
tall, painted entirely in red ochre, and having
long ears. He is probably a “lightning man,”

120

equivalent to Nura Nura of the Worora. Another
smaller “lightning man” is present in another
part of the cave, and is overpainted by a rain-
bow-serpent (Ungeroo), in red ochre and white
clay, which is 5 feet 9 inches long (Plate III,
Fig. 1). Another smaller rainbow serpent is pre-
sent above it, and there are several other
paintings nearby, including a small figure which:
is probably that of a spirit child.

2; miles west-north-west of Elimberrie:
Spring.—At this place (map locality 5) there is:
a small cave with numerous paintings of
Ungerco, the rainbow serpent, together with
small human figures. They are painted in white
clay and red and yellow ochre. Some of the
small human figures show a “rainbow” head-
dress. Several stencilled hands occur in the
cave, and there is one stencilled boomerang.

33 miles west-north-west of Elimberrie
Spring.—The cliffs which mark the north side
of the Oscar Range in this area are almost verti-
cal, and are 250 to 300 feet high (map locality
6). There are a number of rock shelters at the
foot of these cliffs, and some of them are decor-
ated with Nurunguni paintings. The best
locality is 33 miles west-north-west of Elim-
berrie. It contains large numbers of paintings,
many of them well-executed, the most striking
figure being that of Nurunguni-Nowungoo,
drawn in white clay, charcoal, and red ochre
(Plate III, Fig. 2). The painting stands 4 feet
high, and closely resembles the Wandjina paint-
ings of the northern tribes. He has no mouth,
but there is a strange horizontal line drawn
between the eyes and nose, the meaning of which
I have not determined. A series of lines in
red ochre are shown radiating from his rain-
bow headdress. Similar lines in Wandjina
paintings of the Ungarinyin Tribe are regarded
as the hair. A further representation of
Nurunguni-Nowungoo is shown in the shelter,
drawn in white clay over an older figure in red
ochre. Numerous other human figures are pre-
sent, including two “stick-men,’”’ painted in
white clay, on each side of a yam (Plate IV, Fig.
1). The men have long ears and may be
“lightning men”: the representation strongly
resembles that in the cave near the eastern
entrance to Windjana Gorge in Unggumi coun-
try (Plate I, Fig. 2).

The rock shelter also contains representa-
tions of crocodiles, a kangaroo, and various
cther animals, together with numerous sten-
cilled hands.

In the rugged limestone country 4-mile south
of the rock shelter there are a number of other
paintings at the foot of a smooth Massive out-
crop. They are of a Bungarra lizard and an
eel (Plate IV, Fig. 2) together with small human
figures, all painted in red ochre.

Linesman Creek—An impressive series of
paintings occur in the small gorge cut by Lines-
man Creek, on the south side of the Oscar
Range, 4 miles north-west of 12-mile bore (map
locality 7). The paintings stare out over the
gorge from a cave 40 feet up in the eastern
limestone wall. The cliff-face in front of the
cave can only be ascended with great difficulty.
I managed to do so, but could not get down
again until a rope was thrown up. It seemed

PLATE I

Fig. 1.—Glaciated pavement in the Ripon Hills. View, lookin
ground, glacial sediments in west cliff face and lake sediment

Fig. 2.—Glaciated pavement in Ripon Hills. View looking south-east showing polished undulating surface of chert
breccia on west side of pavement.

g slightly west of north, showing: pavement in fore-
s in east cliff face. Both breakaway scarps may be seen,

124

pewnsso yybiay sndju09
Aomoyoasg
wWoaHsS

9OINS 0120/6 jo voNI91Ig

VIIWYLSNYV

\ooos—

dvW ALINYDON

auojspuos paysow
a/ddis puo 01990Ig {49Y)

auo[s{//s snoul6nssa,
pup snosayiuobuop

afosawo/buod
PUD BUO{S{IS PalsIIIIS

auojspuos COUT
Jos20/B01anjy pue aii, apiseoig

auoysyis Apuos 40/nNbvo speg f—
papsos Ajsood ‘ajosawo/6u07 4JaAoy0O L—

WNIANTO ‘105 ‘a/qqney

UHM

Aji wisojuooun

Aj iwuojyuosun

sp 7] AMVNYSLYNO

Abojouy!4 Pajpja4s0D

1D4U0 2140,
abunjd burmoys 340/990 vO1{Isod LURE ap
—ybno4 jouy2uKs paysyjqo1s 3 “a> Ug th
d. ot Sa
abunjd buimoys 34040220 vo! 1sod Sb! 0
— {S819 Jouljayuo paysijqoisz7 +t 5/0 40 po sh
Spjo4 D4yO44S yO dIP pUd 9y¥124S

C-6S VM _$822N0Say /o19UIW JO Noasng

BE
Cy

a
CALE Uf Ui iy

aaa
Ca

So ek
= Ge
Qa
CELI ET iy

DEG EEE BIEN
a
7
Ze

S26

‘
\
x
~

Yee 7
) Lie
ZO
~
t

\
XN

als
3
a3
ais
>
ee
qna
JUIW OAD. PajOlDD/O
is Wz
Ni
\
NN
JUBWAADd .
P9as0100) NS
In
A
K
(S
910Z0Y¥3L0ud §
Y3addn
NJ
NvIWuad
=| YSU eLLg ES

°
©.
S

°

VM STIIH NOdIY LNSWSAVd aalviovi9
SNIMOHS
dVW 1V91901039

dlow/rosede voisod Lae
—AJO0pUunog peysi/golsz wre

—Avopunog peysiqojs 7 9

$alJopunog j051b0);0a5

125