On the Fossil Mammals of Australia. Part X. Family Macropodidae: Mandibular Dentition and Parts of the Skeleton of Palorchestes; Additional Evidences of Macropus Titan, Sthenurus, and Procoptodon

[ 197 ]

VIII, On the Fossil Mammals of Australia. — Part X. Family MacropodiDjE : Man-
dibular Dentition and Parts of the Skeleton of Palorchestes ; additional evidences
of Macropus Titan, Sthenurus, and Procoptodon. By Professor Owen, C.J?.,
F.P.S,, &c*

Keceived May 10, 1875,— Bead June 10, 1875.

§ 1. Introduction. — The evidences of these extinct Mammals which have been made
known through the 'Transactions' of the Royal Society have stimulated the search
and transmission of additional fossils, from which are selected for the present com-
munication those tending to complete the restoration of the gigantic kind of Kangaroo
indicated by the portion of skull described and figured in the volume for 1874*, and
others adding to the knowledge of the dental system and osteology of Sthenurus,
Procoptodon, and Macropus Titan.

To E. S. Hill, Esq., of Woollahra, Sydney, I am indebted, through his brother-in-
law Sir Daniel Cooper, Bart., for the portions of mandible adding to the dental
characters of Palorchestes ; and to Geoege Frederic Bennett, Esq., of Darling Downs,
Queensland, I chiefly owe, through his father Dr. Bennett, F.L.S., of Sydney, the parts
of the skeleton of the same extinct species about to be described, and the remaining
subjects of the present paper.

§ 2. Palorchestes Azael (Mandibular Characters and Dentition). — So much of the
dental characters of the genus and species as could be defined from the condition of
the maxillary teeth, described and figured in the above-cited volume, concurred with
the cranial characters in showing that such large extinct Kangaroo deviated less from
the type of the existing bilophodont Macropodidce than did the species of the genus
Procoptodon, some of which (Proc. Goliah, for example f) rivalled Palorchestes in
bulk.

This conclusion is sustained by the evidence afforded by the subjects of Plate 19.

The chief of these is a portion of the right mandibular ramus (fig. 1) with the teeth
symbolized as d 4, m i, m 2, and part of m 3. A smaller portion of the left ramus of the
same jaw (ib. fig. 5) contained the molar (mi) entire, a portion of m% and the sockets
of the teeth (d 4 and p 3).

The depth of the ramus at the interval between p 3 and d 4 is 3 inches 3 lines ; at
the socket of the last molar (m 3) it is 2 inches 8 lines. Such gain of depth as the

* Phil. Trans. 1874, p. 797, plates lxxxi.-lxxxiii.
f Tom. cit. p. 791,, plates Ixxix., Ixxx.

2e2

198 PKOFESSOE OWEN ON THE FOSSIL MAMMALS OE AUSTRALIA.

mandible extends from the back to the front of the molar series is a character of the
lower jaw of Macropus*, which contrasts with the uniformity of depth in that of
Sthenurusf and Protemnodon%.

The socket of the premolar (ib. fig. 5, p 3) shows the two roots of that tooth to have
been simple ; the hinder one, in section circular, with a diameter of 3 lines, was larger
than the fore root, which was compressed and elliptic in section. The fore-and-aft
length of the crown of this tooth is indicated to have been between 8 and 9 lines.
This accords with the same admeasurement of the upper premolar (p 3) of the fossil
from another and remote locality, figured in plate lxxxii. of the Phil. Trans, torn. cit.

The fore root of the next molar in the mandibular fossil from Darling Downs
(Plate 19. fig. 5, d*) is transversely elliptic, 5 lines in long diameter, with a mid groove
along its hind surface. The hind root, with a larger and deeper longitudinal groove
on its fore surface, is also transversely elliptical, with a long diameter of 6| lines. The
crown of this tooth, preserved in the right ramus (ib. figs. 1 & 2, d 4), shows a low, short
prebasal ridge (f), upon which the exposed dentine is continuous with that of the
much-worn surface of the fore lobe (a), indicative of a linked connexion therewith.
Along the mid link (r), also, a linear tract of dentine extends to the hind lobe (b).
The postbasal ridge is feebly indicated at this stage of wear by a narrow fold of enamel,
which extends from the dentinal tract at the back of the grinding-surface of the hind
lobe, downward and outward to near the base of the crown The fore-and-aft diameter
of d 4 is 10-J lines (23 millims.), the transverse diameter of the hind lobe is 7-§ lines
(15 millims.).

The chief difference in size of d a in the upper jaw§ is in the greater proportional
breadth of the crown — a characteristic of the upper molars in Macropodidw : so much
as remains of the prebasal ridge and of the two main lobes in the maxillary tooth
conforms to the character of the mandibular homotype.

The third molar (m 1) is well preserved in both rami of the present lower jaw. A
continuous tract of dentine is exposed, by wear, from the fore lobe along the mid link
to the postbasal ridge (g). The prebasal ridge is limited to the inner half of the fore
surface of the fore lobe, where, also, it is alone visible in m 1 of the upper jaw. The
postbasal ridge (g) is more developed on the inner side of the hind link, where it is
divided from the hind lobe by a fossa ; on the outer side of the link the ridge rapidly
sinks and subsides near the base of the crown ; the mid part of the postbasal ridge is
half an inch above the basal line of the enamel (fig. 8, g). The main valley expands
as it descends from each side of the mid link (r), and is closed by a low bar on both
the outer (ib. fig. 6, v) and inner (ib. fig. 7, v 1 ) sides of the tooth. The fore-and-aft
extent of the mandibular tooth (mi) is 1 inch, as in its homotype of the upper jaw;
the transverse diameter of the hind lobe is 8 lines, in the upper tooth 9 lines.

The less-w r orn crown of m 2 shows more of the true pattern of the grinding-surface

* See Phil. Trans. 1874, plate xxii. figs. 13, 15. J Tom. cit. plate xxv. figs. 7, 8.

f Ibid. figs. 5 j 6. § Tom. cit. plate lxxxii. fig. 1, d 4.

PEOFESSOE OWEN ON THE FOSSIL MAMMALS OF AUSTEALIA. 199

in the lower molars of the present species. The vertical indent (Plate 19. fig. 1, mz)
on the inner fore part of the crown, leading to the better developed inner part of the
prebasal ridge, is deeper than in m i, and causes a concave contour of the anterior
enamel-border of the fore lobe.

The mid link projects from the middle of the hind enamel-border of that lobe,
touches, but is not continuous with, the hind part of the link sent off from the outer
side of the fore part of the hind lobe. The vertical prominence from the corresponding
part of the crown of the lower molars in Nototherium is the homologue of this part of
the mid link in Palorchestes.

The inner concave part of the fore surface of the hind lobe shows two fine vertical
fissures in m 2. From the low link at the back of the hind lobe the postbasal ridge
sinks toward the base, thicker and shorter on the inner than on the outer side. The
fore-and-aft extent of m 2 is 1 inch 1^ line (29 millims.) ; the transverse breadth of the
hind lobe is 8 \ lines (18 millims.). Here the lower penultimate molar is longer in
the first diameter by 1 line than the corresponding tooth above, the breadth being
the same.

The breadth of the fore lobe ofms (ib. fig. 2) is 9 lines; the length of the crown, if
entire, would seem to have been, as in the perfect tooth in figure 4, 1 inch 2 lines, or
1 line longer than the corresponding molar above. The proportions of the upper and
lower last two molars in Macropus Titan are repeated in Palorchestes Azael.

The fracture of the molar (m 3, in situ) in figs. 1 & 2 shows the continuous part of
the mid link to be 3 lines in vertical extent, the whole height of the link being 5 lines
where it rises from the valley (fig. 3, r). The exposed hind root of this molar curves
outward as it descends, and is bifurcated at the extremity. Beneath it the fracture
shows the dental canal (fig. 3). The inner bar in m 3, as in m 2, is feebly marked at
the inner entry of the valley (fig. 4).

The enamel near the base of the crown and at parts of the outer surface shows, as in
the upper molars, the fine rugse and punctations like those in the molars of Nototherium
and Diprotodon; and the generic pattern of the molars of Palorchestes among the
Macropodidw indicates a transitional condition between the Kangaroos and those more
gigantic extinct Marsupials.

The outer surface of the mandibular ramus of the species, for which grounds are
above assigned to justify its reference to Palorchestes Azael, shows a vascular or
nervous outlet some way below the interval between m 1 and m 2 ; but this may be
an individual and inconstant character.

The depth of the horizontal ramus being given in fig. 1, I have restored the entire
skull in outline (Plate 20), with a finished view of the left side of the original and still
unique fossil, not figured in my former Memoir.

§ 3. Palorchestes (Sacrum). — From the same formation in Darling Downs, Queensland,
I have received portions of the pelvis and of the hind limbs with macropodal characters,
and of a size corresponding with that of the above-described portions of mandible.

200 PEOEESSOB OWEN ON THE EOSSIL MAMMALS OE AUSTEALIA.

The sacrum (Plate 21. fig. 1) consists of two vertebrae with the characters of those
of Macropm rufus*, but with a difference of size shown in the following admea-
surements : —

Macrojpus rufiis. Palorchestes.
in. lines. in. lines.

Length of sacrum (at zygapophyses) 3 2 4 10

Breadth of sacrum (across fore part) 3 5 5 6

Breadth of centrum of first vertebra 17 2 6

The Kangaroo yielding the recent bone compared was the largest example seen by
Mr. Gould in his travels in Australiaf , and no specimen of Macropm major has yet
been recorded of superior size.

The comparatively gigantic leaper yielding the fossil seems to have been an aged
individual, for so much anchylosis has taken place between the second sacral (c 2) and
first caudal (cd 1) as to have kept those vertebrae in natural connexion during the period
of petrifaction.

The transverse processes of the second sacral take a greater relative share in the
formation of the sacroiliac symphysis in Pdlorchestes than in the above-named recent
species, and the shape is rather more subquadrate than triangular. The joint between
the first and second sacral is not obliterated. The intervertebral foramina (ib. % i) are
rather smaller, relatively, than in Macropus rufus, and suggest that the hind limb may
not have predominated over the fore limb in so great a degree in the larger and
heavier Kangaroo. The pair of ridges on the haemal surface of the centrum of the
second sacral (ib. k, Jc) are better marked in the fossil. In this the neural arch of the
first sacral has been broken away.

The first caudal vertebra of Palorchestes (ib. figs. 2, 3) is 2 inches 11 lines in length,

1 inch 6 lines across the hinder articular end of the centrum. The base of the lamelli-
form depressed transverse process (ib. d, d) is 1 inch 8 lines in extent, reaching within

2 lines of each articular end of the centrum. The longitudinal extent of the base of
the neural spine is 1 inch. The characters of the corresponding vertebra in Macropus
rufus are closely repeated, with the difference of size and slight increase of breadth over
length, as in the sacrum; and these permit an inference that the tail in Palorchestes
corresponded in strength, if not quite in relative length, to that in Macropus.

§ 4. Palorchestes (Os innominatum).— The pelvis of the Kangaroo is characterized
by a long prismatic ilium, an oblong tuberosity above or anterior to the acetabulum, a
"pectineal" or "ileo-pubic" process, the articular surfaces for the marsupial bones,
the broad, compressed, subprismatic form of the ischium, the slenderness of the pubis,
and the great length of both those elements of the pelvic arch.

The length of the pelvis due to the great extension of the os innominatum both in
front and behind the acetabulum is a well-marked feature of resemblance to the same

* " Osteology of Marsupialia.-— Part V.," Trans. Zool. Soc. vol, ix. p. 429, pi. lxxvi.
f Phil. Trans. 1874, p. 253.

PEOEESSOE OWEN ON THE EOSSIL MAMMALS OF AFSTEALIA. 201

part of the skeleton in the Bird, and similarly relates to the faculty of station and pro-
gression on the hind pair of limbs — the movement being, in the Kangaroo, as in some
tribes of birds, by a series of hops, the fore limbs, however, in the mammal taking also
an occasional share in progression on land, which is not permitted to the volant
class.

The portion of pelvis (Plate 22) corresponding in size with the sacrum (Plate 21),
although reduced to the parts immediately surrounding the acetabulum, includes the
base of the ilium, showing its three-sided prismatic form, the precotylar tuberosity, and,
above all, the " pectineal process " (ib. e\ in size, shape, and position closely corre-
sponding with that in Macropus rufus f .

The anterior or hsemal facet of the ilium (ib. fig. 1, 62) is more deeply excavated than
in the recent Kangaroo ; the rounded angle (a) dividing that facet from the inner or
median facet (ib. fig. 2, 62') is relatively thicker. The precotylar tuberosity (d) is more
prominent, is relatively nearer the acetabulum ; and a ridge is continued from the fore
part of the tuberosity to the brim of that cavity, which more definitely or abruptly
there defines the hsemal from the neural facets of the ilium than in recent Kangaroos.

The contour of the acetabulum is more ovate, the smaller end, at the fore part of the
brim, being narrower in the larger fossil. The notch (y) at the hind part of the brim
is narrower, and is nearly bridged over by the extension of the ischial (es*) toward the
pubic (ed*) border of the notch; the depression (x) into which the notch expands at
the bottom of the cup is relatively narrower in Palorchestes than in Macropus.

These characters would have indicated specific distinction if the present fossil had
not exceeded in size the corresponding part in the largest living Kangaroos ; the degree
to which Palorchestes surpassed them is exemplified in the pelvis as in the skull.

Admeasurements of Pelvis.

Palorchestes. Macropus rufus.

Breadth of haemal surface of ilium above the precotylar in. lines. in. lines.

tuberosity 30 10

Breadth of neural surface of ilium above the precotylar

tuberosity 33 13

Breadth of median surface of ilium below symphysial

surface 20 10

Length of acetabulum 28 17

Breadth of acetabulum 26 14

Breadth of ischium behind the lower part of acetabulum .22 13

§ 5. Palorchestes (Femur). — The fossil thigh-bone, of which the two extremities are
figured in Plate 23. figs. 1 & 2, shows an articular head (a) fitting the acetabulum of
the pelvic fossil (§ 4). The height of the trochanter major (/), the length and back-
ward position of the narrow trochanter minor (n), the depth of the cavity (I) under-

f Mem. cit. Trans. Zool. Soc. vol. ix. plates lxxix., lxxx., e.

202 PKOFESSOR OWEN ON THE FOSSIL MAMMALS OE AUSTRALIA.

mining the hind extension of the great trochanter, and the ridge (p) at the back part
of the upper half of the shaft, bespeak the macropodal characters of the present fossil
in the upper portion of the bone, as the partial division of the outer condyle (fig. 2, v)
by the channel (w), and the deep rough oblong fossa (y) above that condyle, do in the
lower portion *. . .

Guided by the proportions of the femur in Macropus major and Macropus rufus, 1
estimate the subject of fig. 1, Plate 23, to include the proximal third of that bone in
Palorchestes Azael ; and suspect, as the upper portion of the great trochanter is still
epiphysial, or but partially united to the shaft, that this thigh-bone may have come
from a not fully mature individual.

As in Macropns rufus the extreme breadth of the proximal end of the femur exceeds
that of the distal end by 3 lines, I estimate the difference in those admeasurements of
the ends of the fossil femur in Plate 23 to be within the limits of individual character
in Palorchestes, the breadth of the shaft, where broken across, in both upper and
lower portions being the same ; and the circumference in both is 5 inches. The some-
what larger proportional proximal end, due to the development of the great trochanter,
may be taken as one of the differential characters of the present huge femur as com-
pared with that bone in the largest living Kangaroos.

The great trochanter is continued as a strong ridge (ib. fig. 1, g) 5-J inches along the
outer border of the bone ; but the trochanter itself (f) is, relatively, less raised above
the head (a) than in Macropns major. The articular surface of the head is less convex
than in Macropus rufus, and the anterior concavity between it and the trochanter is
less marked. The demarcation of the summit of the trochanter by the antero-internal
channel is more feebly given in Palorchestes. The upper surface of the neck of the
femur (c) is relatively broader in Palorchestes than in Macropus.

The lesser trochanterian ridge (n) is more posterior in position than in Macropus
rufus, in which its free margin just comes into view when the femur is seen from the
front f ; this is not the case in Palorchestes, in which the ridge descends to the parallel
of the beginning of the posterior ridge, which resembles at its most prominent part (p)
a third trochanter %. The trochanterian fossa (fig. 1, I) has the extreme generic or
family depth in Palorchestes ; it forms a long narrow cavity, undermining the hinder
basal part of the great trochanter.

From the summit of this trochanter (f) to the third trochanter (p) being 6^ inches,
the total length of the femur of Palorchestes may be reckoned, from the analogy of
Macropus rufus, to have been not less than 18 inches; in that recent species it is
10 J inches. The epiphysial line of the great trochanter is distinct, but confluence of
the central part has kept the process in place in the present fossil.

The upper end of the linea aspera is preserved, expanding to form the process^.

* See the characters of the femur of Macropus described and figured in the Mem. cit. in Trans. Zool. Soc.
vol. ix. p. 437, plate lxxxi.

f Mem. cit. torn. cit. plate lxxxi. fig. l,n. ' t See ibid. fig. 2, p, in Osphranter rufus.

PEOFESSOK OWEN ON THE FOSSIL MAMMALS OF AUSTRALIA. 203

At the distal end of the femur (ib. fig. 2) the chief distinction of Palorchestes from
Macropus is in the relatively narrower postintercondylar fissure (u) and its minor ex-
pansion, where it is closed anteriorly*. The depression (y) is strongly marked in
Palorchestes. The epiphysial line is traceable in the fossil ; a wedge-shaped process at
both the outer (z) and the inner (z f ) borders rises as if to clamp more securely the
epiphysis to the shaft.

The following admeasurements exemplify the difference of size between Palorchestes
Azael and Macropus rufus, the measured femur of the latter being of a full-grown
male : —

Palorchestes Azael. Macropus rufus.
in. lines. in. lines.

Extreme breadth of proximal end of femur ... 4 9 29

Extreme breadth of middle of shaft of femur ..18 10

Extreme breadth of distal end of femur . . . . 3„ 9 2 5

§ 6. Palorchestes (Tibia). — If the fine fragment of this bone figured in Plate 24 be
compared with the corresponding views of the entire tibia of the large male Red
Kangaroo figured in vol. ix. of the ; Zoological Transactions/ plate lxxxii., the associa-
tion of the peculiar characteristics of the macropodal tibia with the grand proportions
of that bone in Palorchestes will be readily appreciated.

The length of the present fossil remnant from the proximal end of the bone to the
subsidence on the shaft of the procnemial plate is 7 J inches. The fore-and-aft diameter
of the tibia, at the upper part of the plate, is 3 inches 5 lines ; the span of the exca-
vation between the procnemial and ectocnemial plates or ridges is 2 inches 4 lines ;
the antero-posterior diameter of the head of the tibia is 3 inches 10 lines ; the breadth
of the back part of the tibia, at 5 inches below the articular head, is 1 inch 6 lines.
The head of the tibia is in a state of epiphysis; its undulatory course along the inner
side of the bone is shown in fig. 4, but partial confluence, as in the case of the epi-
physis of the femur of probably the same individual Palorchestes, has tended to retain
the epiphysis in place, notwithstanding the movements and shocks of alluvial transport
through which, seemingly, the fractures of the fossil are due.

The inner articular facet (ib. fig. 5), the only one preserved on the head, is relatively
more extensive and more concave transversely than in Macropus rufus. The hind
surface of the shaft, continued down from that articular surface, is thicker and more
convex across ; it contracts in the large recent Kangaroo to an angular ridge, sharply
dividing the hinder from the antero-internal surface of the shaft of the bone.

The following are a few comparative dimensions of the tibia : —

#

* Compare fig. 2 with fig. 2, u {Macropus rufus) in plate lxxxi. of Mem. cit. Zool. Trans, vol. ix.

MDCCCLXXVI. 2 F

204 PEOEESSOE OWEN ON THE FOSSIL MAMMALS OE AUSTRALIA.

Palorchestes Azael, OspJiranter rufus.
in. lines. in. lines.

Length from head to subsidence of procnemial ridge .76 53

Fore-and-aft diameter at upper part of procnemial ridge 3 5 2
Transvere diameter of shaft halfway down procnemial

ridge .. 17 1

Span of excavation between procnemial ridge and fibular

ridge 24 16

Fore-and-aft diameter of head of tibia 3 10 2

§ 7. Palorchestes (Calcaneum). — In Macropus (OspJiranter) rufus the length of the
calcaneum exceeds the transverse breadth of the distal end of the femur by two sevenths ;
the extreme breadth of the calcaneum is one third greater than that of the middle of
the shaft of the femur. By these proportions I am guided in the choice of the two
fossil calcanea (figs. 4 & 5, Plate 23), and refer the longer bone to Palorchestes.

The length of the subject of fig. 5, Plate 23 ? exceeds the transverse breadth of the
distal end of the femur (ib. fig. 2) by two sevenths. The length of the calcaneum (ib.
fig. 4) exceeds the transverse breadth of the distal end of the femur (ib. fig. 3) by one
fourth. The breadth of the calcaneal process of fig. 4 is equal to- that of the longer
calcaneum, fig. 5.

On the grounds subsequently to be adduced for concluding the leg and foot of
Procoptodon to have been shorter in proportion to its length than in Macropus, I
therefore assign the shorter and thicker calcaneum to that genus, together with the
portion of femur (fig. 3), which shows more generalized characters, or those less strictly
macropodal, than the femora assigned to Palorchestes, Macropus, and Sthenwrus.

§ 8. Macropus Titan (Skull). — Of this species I am now enabled to add to maxillary
and mandibulary evidences adduced in former Parts some instructive cranial characters.

The specimen yielding these was found by W. F. Tooth, jun., Esq., at King's Creek,
near Clifton, Darling Downs, at a part of the bed which Dr. Bennett, F.L.S., had
pointed out to his friend as being likely to yield fossils after a flood *. To these gentle-
men the British Museum is indebted for the specimen. On receiving it as much of
the adherent matrix was cleared away as could safely be meddled with ; and the present
state of the fossil is given in side and base views, of the natural size, in Plates 25
and 26.

It is a great part of the skull of a Kangaroo, wanting the lower jaw, but including
the cranium proper, the interorbital and the hinder part of the facial division of the
skull ; also great part of the left zygomatic arch, with the included orbit and temporal
fossa, the bony palate, and the molar dentition, of which the two hindmost teeth are
sufficiently entire to afford the means of specific determination.

Other projecting parts and processes have suffered fracture, and the region of the

* See " A Trip to Queensland in search of Possils," by Dr. George Bennett, F.L.S., in 'Annals and Maga-
zine of Natural History,' April 1872.

PBOEESSOB OWEN ON THE FOSSIL MAMMALS OE ATJSTEALIA. 205

large frontal sinuses has been obliquely crushed. The specimen is in the usual heavy
petrified condition of fossils from the freshwater drift; it shows the effects of trans-
port and attrition during the movements of this matrix before reaching the locality
where it was found.

The first attention being directed to the teeth, of which the three last left molars
and the four last right molars were in place, the characters of transverse lobes, links,
and prebasal ridge were seen to be those of the genus Macropus, while the size and
the sculpturing of the hind surface of the last molar (Plate 26. fig. 2) determined the
species.

In Macropus major that surface (ib. fig. 3) is moderately hollowed lengthwise and
thickly coated with cement, which partly fills the triangular transverse concavity, the
apex of which shallows to the ordinary level of the hind surface before reaching the
base of the crown. When the cement is removed the inner enamel boundary (fig. 3, g)
is sharper and more produced than the outer one (ib. h).

In Macropus Titan the enamel, after coating the inner border of the hinder lobe,
extends backward, downward, and outward, projecting as a sharp-edged ridge (ib.
fig. 2, g\ defining a deeper depression on the hinder surface of the tooth. There is
also a shallow vertical groove (Ji) continued from the hind part of the apex of the inner
border of the hind lobe downward toward the base of the crown, which groove seems
to define the inner limit of the oblique posterior ridge. One sees that this groove
repeats the deeper cleft that defines the mid link internally from the inner end or
border of the anterior lobe. The oblique hind ridge (g) is indeed a serial repetition of
the mid (r) and fore (s) links, but subsides with a more oblique course downward
toward the base of the outer border of the hind lobe, having no other division of the
molar to connect with such lobe. From the fore part of the base of the inner end of
the hind lobe a low ridge defines the anterior surface of that lobe to the inner side
between it and the mid link ; this defining ridge is not present in the upper molars of
Macropus major, but there is a small tubercle at the inner entry of the valley between
the two main lobes of the upper molars in Macropus major which is not present in
Macropus Titan.

The fossil skull, with the molars agreeing in the above characters and in size with
those of more fragmentary examples of Macropus Titan*, is of a mature and somewhat
aged individual. The summits of both lobes of the hindmost grinder are worn so as to
expose a linear tract of enamel, widest of course on the anterior lobe. In the penulti-
mate grinder a broad field of dentine is exposed on this lobe, extending backward by a
linear tract along the base of the mid link (r), but not so far as the transverse tract of
dentine exposed on the hind lobe. In the antepenultimate grinder (m i) both lobes are
so worn that the lozenge-shaped fields of exposed dentine touch and communicate at the
base of the worn-down link. The foremost grinder (d 4, Plate 26) is retained on the
right side, worn down to its base ; but this tooth has been shed on the left side, and

* Phil. Trans. 1874, plate xxi. figs. 8, 10, 15, 16 ; plate xxii. figs. 10, 12.

2f2

206 PEOFESSOE OWEN ON THE FOSSIL MAMMALS OE AUSTEALIA.

the grinding series reduced to the three true molars, as in old individuals of Macropus
major.

The skull of this large existing species of Kangaroo which I have to compare with
the present fossil retains the last deciduous molar (d 4) on both sides of the upper jaw,
without any trace of the socket of the premolar which had worked in advance of the
four retained grinders (d 1, m 1, 2, 3). In the maxillary fossil of Macropus Titan (figured
in plate xxi. fig. 10, Phil. Trans. 1874) the two roots of p 3 are retained, the crown
having been accidentally broken away. That of the last molar (m 3) had not come into
place, although the front lobe had pushed its way out of the formative cell. The two
lobes of m 2 in the same fossil show attrition of the enamel ridge, but not so as to reach
the dentine.

In the younger subject of figs. 6, 7, 8 of the same plate and volume, the four teeth
in place are d 3, d 4, m 1, and m 2 ; above the first of these is exposed the crown of p 3 in
its formative cell, and part of that of m 3 is shown behind m 2.

The series of changes of the upper molar dentition of the extinct Macropus Titan are
thus as instructively and almost as completely displayed in petrified specimens as in the
existing species (Macr. major), of which I have obtained specimens in number and
periods of age sufficient to exemplify these phases *. I have previously described a
mandibular specimen of Macropus Titan, with the molar dentition reduced to m 2 and
m 3 f , as in the latest phase hitherto observed in Macropus major.

The cranial specimen exemplifying the reduction of molars to three on the left and
four on the right side, lacks, unfortunately, the part of the upper jaw which supported
the incisor teeth. Nearly two inches of the diastemal tract, however, is preserved in
advance of d 4 on the right side. A fossil mandible of another individual of Macropus
Titan, with a similar stage of dentition as the right side in Plate 26, fortunately gives
the extent of the diastema between the molars and incisors % ; and guided by the pro-
portion which this part bears to the upper diastema in Macropus major ^ I have restored
in outline in Plates 25 and 26 what is wanting in the present fossil, together with an
outline of the mandible and mandibular teeth.

The length of the mandibular diastema in Macropus major is 1 inch 9 lines, that of
the maxillo-premaxillary one is 2 inches 6 lines. The length . of the mandibular
diastema in Macropus Titan being 2 inches 6 lines, that of the maxillo-premaxillary
diastema, according to the pattern of the recent species, should be 3 inches 6 lines.
There are indications, however, that the muzzle was relatively rather shorter in the
larger extinct Kangaroo, and I have restored it with an interval of 3 inches 3 lines
between the foremost molar and hindmost incisor.

Of this characteristic tooth fossil- specimens reveal two patterns of the outer surface
of the crown in examples indicative of species as large as Macropus Titan and Sthenurus

Atlas.

In one type (Plate 25. fig. 4) the outer surface of the . crown is divided into three

* See notes, p. 245, Phil. Trans. 1874. f Tom. tit. plate lxxvi. ftg. 3. J Tom. cit. plate xxvi. 'fig 11.

PKOFESSOB OWEN ON THE FOSSIL MAMMALS OE AUSTEALIA. 207

unequal convex tracts by two oblique grooves, of which the hinder one extends nearest
to the base or root of the tooth ; in the other type (ib. fig. 3) a deeper oblique fissure
subequally bisects the crown ; it marks off a more prominent fore part of the outer
surface from a lower and vertically shorter, but rather more longitudinally extended,
hind tract. As the first of these patterns is repeated in the third upper incisor of
Kangaroos with a small premolar [Maeropus major*), and the second pattern is found
in Kangaroos with a large trenchant premolar {Halmaturus ualabatus, H. rvficollis)^
I refer the fossils of the second pattern to Sthenurus Atlas and those of the first pattern
to Maeropus Titan.

The skull in this extinct species has the triangular form of occiput as in Maeropus
major ', the apex of which, forming the summit of the superoccipital ridge, is somewhat
rounded off. The upper and larger ends of the condyles subside more gradually into
the occipital surface, and are not defined by a depression there as in Maeropus major.
The channel or concavity between the condyle and paroccipital is relatively wider in
Maeropus Titan. In this species the foramen magnum seems as if it had been notched
at its upper border, where the exoccipitals may not have met, and where the foramen
may have been bounded by an intercalated portion of the superoccipital.

As in Maeropus major ^ also, a second inner ridge from the base of the paroccipital
converges towards its fellow as it rises, parallel with the outer ridge, from the mastoid,
but subsides before attaining the summit of the exterior ridge.

The crown of the superoccipital arch projects rather more backward in Maer. Titan
than in Maer. major ; it is not on a vertical plane with that of the occipital foramen,
nor does it slope, as in many recent Kangaroos, forward from that foramen. The
surface below the arch is traversed by a less prominent median vertical ridge in
Maeropus Titan than in Maer. major.

The upper border of the occipital foramen is mutilated in the fossil, but seems to
have been more arched, less regular, than in Maer. major.

The basioccipital (Plate 26. fig. 1, i) is carinate below, as in. Maer. major %; but there
is more tumefaction at its suture with the basisphenoid in Maer. Titan.

A low crest runs along the line of the sagittal suture in the fossil, which bifurcates
anteriorly, the divisions diverging to the postorbital prominences, which, as usual in the
genus, are feeble. In Maeropus mfus, at a similar phase of dentition with the fossil,
the sigittal suture persists, and the low ridges bounding above the crotaphyte surfaces
have not met at the mid line.

The fore part of the glenoid surface for the mandibular joint, in Maeropus Titan^ is
contributed by the malar as in other Kangaroos. The outer surface of the zygoma
seems not to have been so deeply impressed or concave as in Maeropus major and Maer.
rufus. The facial part of the skull anterior to the orbits is relatively broader in Maer.
Titan than in either of the above-named existing species. The antorbital foramen is

* Tom. tit plate xx. fig. 17. t Ibid. figs. 20, 21.

Also as in Maeropus rufus t see "Osteol. of Marsupialia. — Part V.," Zool. Trans, vol. ix. pi. lxxiv. fig. 8.

+

208

PBOFESSOE OWEN ON THE FOSSIL MAMMALS OE AUSTEALIA.

relatively further from the orbit in Macropus Titan than in Macr. major; and in this
character Macr. rufus more resembles the large fossil Kangaroo. The front pier of the
zygoma springs from the side of the skull more posteriorly in Macropus Titan than in
either of the large existing Kangaroos.

The bony palate (Plate 26) is extended further back, and the production of the
alveolar border of the maxillary behind the last molar is more convex transversely
than in Macr. major or Macr. rufus. The bony palate is entire in the fossil as in Macr.
major \ but is relatively wider than in that species ; it has not the reticularly disposed
small perforations shown by the specimen of Macropus rufus described in the ' Zoolo-
gical Transactions'*. The interspace between the right and left ultimate molars in
Macr. Titan is twice and two thirds of the fore-and-aft diameter of that tooth ; in Macr.
major that interspace equals two diameters and one fifth of the last molar.

The lower area or outlet of the zygomatic arch (Plate 26. fig. l,z) is relatively larger
in Macr. Titan than in Macr. major ; it exceeds the length of the molar series of four
teeth (d 4, m i, m 2, m 3) by the length of m 3 in Macr. Titan, while in Macr. major the
outlet does not equal in length the same series of teeth by one half of the anterior
molar (d 4).

The diastemal border is less obtusely rounded in Macr. Titan than in Macr. major ;
it resembles more that border in Macr. rufus f and in some of the smaller existing
Kangaroos (Pliascolagus erubescms, e. g.). The extent of the diastemal interval in the
upper jaw of the fractured fossil has been approximately estimated on the grounds
above defined.

Subjoined are admeasurements of the fossil skull above described, and of that of a
male at the same phase of dentition of a large existing Kangaroo.

Macropus major. Macropus Titan.

Length of skull from superoccipital ridge to fore part in. lines. in. lines.

01 d 4 ..49 66

Breadth of skull at widest span of zygomata ....

Breadth of cranium at postorbital constriction . . .

Breadth of cranium at the mastoids

Breadth of foramen magnum

Length of cranium from superoccipital ridge to post-
orbital constriction c

Length of cranium from superoccipital ridge to fore
border of orbit

Length of orbito-temporal vacuity, lower opening . ,

Breadth of orbito-temporal vacuity, lower opening . .

From hind border of palate to fore border of d 4 . . .

Breadth of palate between the right and left molars (m 2)

Prom alveolus of m 3 to hind border of palate ....

4 3

5 9

1

1 6

2 11

3 8

10

11

2 7

3 2

4 3

5 8

1 9

2 6

1 4

1 8

2 4

3

1 4

1 10

6

1 1

* YoL ix. plate lxxiv. fig. 8.

Tom. tit. plate lxxiv. %. 3,

PBOHSSSOB OWEN ON THE FOSSIL MAMMALS OF AITSTKALIA. 209

Macropus major. Macropus Titan.
in. lines. in. lines.

From lower border of orbit to alveolar border at m i . 1 16

Length of series of four molars (d 4 to m 3) ..... 1 10 2 2

Length of m 2 and ms 11 13

Breadth of fore lobe of m 3 .......... 4^ 06

§ 9. Macropus Titan (Femur). — An almost entire thigh-bone, in the same petrified
condition as the skull above described, and from the same freshwater drift in King's
Greek, offers the same proportions to that skull and to the mandible and teeth of
Macropus Titan as the femur of Macropus major does to the same parts in that species.
It is of the right side, in length 11 inches 6 lines; but would equal, if not exceed, a
foot in length were the summit of the great trochanter entire. The bone is figured, of
the natural size, in Plate 27, 2J inches of the middle of the shaft being omitted in
figs. 1 and 2 to bring them into the quarto form. The macropodal characters of this
fine fossil femur, and the deviations, besides size, from the femur of the largest existing
Kangaroos, will be readily appreciated if Plate 27 be compared with plate Ixxxi. Zool.
Trans. * of the femur of Macropus (Osphranter) rufus.

A trace of the antero-internal groove, defining in that recent species the super-
trochanterian tuberosity, is plain in the fossil at e, figs. 1 & 2, Plate 27, where
that tuberosity has been broken away. Compared with the femur of Macropus rufus
that of Macropus Titan shows a relatively wider and shallower concavity (ib. fig. 1, d)
between the fore part of the great trochanter and the head (a) of the bone. The
" cervix femoris " (b) is relatively thicker. The transverse diameter below the head is
relatively greater, mainly through the greater extent of the bone internal to the
" small trochanter" (ib. fig. 2,%), whereby that outstanding ridge-like process does
not appear in a direct front view (ib. fig. 1). The same relative position of n in the
femur of Palorchestes Azael is also due to the inward extension of the support of the
neck and head of the bone.

The femoral shaft in Macropihs Titan is relatively thicker, especially from before
backward, than in Macropus major and Macr. rufus. The rough depression (ib.
fig. 4, y) above the outer condyle is relatively larger, deeper, more sharply defined*
The inner condyle (ib. figs. 1 & 2, t) has its inner (tibial) border better defined and
produced so as to give a slight concavity, transversely, to that half of the back part
of the condyle. This character is more marked in Palorchestes (Plate 23. fig. 2, i);
but there is no trace of it in the inner femoral condyle of the large existing Kangaroos.
The intercondylar notch (u^ fig. 2, Plate 27) is narrower and deeper in Macropus Titan
than in Macropus rufus, again repeating a femoral character of Palorchestes, but not
in so marked a degree* The ectocondylar pit (ib. fig. 4, v) is equally well marked.

The broad shallow vertical groove at the back part of the outer condyle, which in
Macropodidw offers so interesting an approach to the characteristic structure of that

* Yol. ix.

210 PBOFESSOE OWEN ON THE EOSSIL MAMMALS OE AUSTRALIA.

part of the thigh-bone in Birds, is well shown in the femur of Macropus Titan (ib.
fig. 2, w) as in that of Palorchestes (Plate 23. fig. 2, id).

The epiphyses are confluent with the shaft at both ends of the femur, but the line of
separation is traceable in the fossil as in the figured femur of Macropus rufus above
referred to.

I may here refer to portions of fossil femora which depart from the type of the two
already described by deviating further from the characters of the femur in the existing
species of Macropus. The chief difference is in the smaller and shallower depression
(y) above the outer condyle, such depression being filled up, as it were, by a rough and
thick ascending process of the distal epiphyses, of which a rudiment only exists in the
femur of Macropus Titan (Plate 27. fig. 4, z) and of Palorchestes Azael (Plate 23.
fig. 2, z). The femora with the larger and longer " clamping" process are thicker in
proportion to their length than in the above-cited fossils, and still more so than in the
recent Kangaroos. This stronger type is manifested by full-sized or mature femora of
three dimensions, of which the distal end of the largest is figured in Plate 23. fig. 3. I
shall at the conclusion of the present " Part " adduce evidence which leads me to deem
these fossils to belong to the genus Procoptodon ; and I. provisionally, refer the portion
of femur figured and the shorter type of calcaneum in the same Plate (fig. 5) to Pro-
coptodon Goliah.

§10. Sthenurus Atlas (Eestoration of the teeth and part of the skull). — Confirmation of
the ascription of the second type of upper third incisor to an extinct species of Kangaroo
with a large premolar tooth has been had by the reception of a specimen of that part
of the skull and dental system which, as a rule, is wanting in cranial fossils of these
extinct Marsupialia.

This specimen consists of the facial part of the skull, from the anterior halves of the
orbits to the ends of the premaxillaries, with their incisor teeth (Plate 25. fig. 2,
Plate 26. fig. 4).

The molar dentition is represented by an anterior tooth of trenchant character (d 3),
followed by three double-ridged molars on the left (Plate 25. d 4, m 1, m 2) and two on
the right side (Plate 26. fig. 4, d! 4, mi). The third on the left (m 2) is emerging from
its socket with the ridged summits of the lobes narrow ; a portion of a formative cavity
of a larger molar is preserved behind that tooth. This evidence of immaturity is
supported by the incomplete exclusion of the crown of the third incisor (ib. ib. i 3) ;
and the correspondence of the stage of dentition with the second (b) of the series in
Macropus major, figured in my 'Anatomy of Vertebrates'*, was demonstrated by the
usual test, viz. the exposure of the crown of the replacing tooth (Plate 26. fig. 4, p 3) in
its formative alveolus above the deciduous teeth (d 3, d 4) in place and use. The third
bilophodont tooth (m 2) on the left side is not so far advanced as its homologue in the
jaw showing the third stage (op. cit. ib. c) of the dentition of Macropus major.

The germ of the premolar and the crowns of the deciduous teeth in place (d 3, d 4,

* 8vo, 1868, vol. iii. p. 380 ? fig. 296.

PKOFESSOK" OWEN ON THE POSSIL MAMMALS OF AUSTRALIA. 211

fig. 2, Plate 25, and fig. 4, Plate 26) accord with the characters shown in more frag-
mentary specimens of Sthenurus Atlas, Consequently can be added by means of the
present fossil the characters of the first and third upper incisors to the previous know-
ledge of the dentition of that large extinct species.

The fossil evidence of this young individual of Sthenurus Atlas presents a condition
which significantly points to the nature of its violent death, and to the operation of the
powerful jaws and teeth of its carnivorous destroyer.

The upper jaw, anterior to the orbits, has been nipped in by a cross bite ; another
grip in a vertical or obliquely vertical direction in the orbital region has crushed the
right half in the course of the interfrontal and internasal sutures to a lower level than
the left half, with a similar degree of forward dislocation. The skull has been subject
to this violence in its fresh state, and the matrix has subsequently become petrified
about it, and has preserved the dislocations.

If they had been due to movements of the matrix after fossilization, the petrified
head would show fracture corresponding to the bone ; but no such evidence of post-
humous crushing of matrix and fossil being present, 1 presume that the skull, if it
had been imbedded uninjured, would have retained its form when petrified, and con-
clude that the actual state of the fossil was that in which it was interred before
petrifaction began.

The anterior incisor (Plate 25. fig. 2, i\) is curved, as in most existing Kangaroos;
but besides its superiority of size to that in the largest kind, as shown by the breadth
of the crown*, the exserted and enamelled portion is both absolutely and relatively
longer, and thus makes a nearer approach to the character of the first upper incisor in
JDiprotodonf. The convex or fore surface of the crown of i i in Sthenurus Atlas is
traversed longitudinally by a shallow and rather wide groove behind the mid line of
that surface, which groove deepens near the cutting-edge, and thus marks it with a
feeble notch. The enamel also shows some fine longitudinal striations. This wrap of
the tooth is uninterrupted, but becomes much thinner at the back part. A transverse
section of the crown would give a long, narrow oval, rather broader at the outer and
hinder end.

The breadth of the tooth, or length of the oval, is 10 millims., or 4§ lines ; the
thickness or antero-posterior extent is 4 millims., or 2 \ lines. The hind margin of the
tooth, near the cutting-edge, shows the shallow indent caused by the crown of the
second incisor ; but this tooth in both premaxillaries has been displaced by the lateral
crushing of these bones in the recent state, and was probably lost prior to the imbedding
of the skull. The second incisor is the smallest and least deeply implanted in most

Kangaroos.

The third incisor (Plate 25. fig. 2, i*) had not been fully developed; its crown had
only partially emerged from the socket, whence its preservation. It is in the form of a

* Compare with Phil. Trans. 1874, plate xx. fig. 17 (Macropus major),
t Phil. Trans. 1870, plate xxxy. fig. l,ii.

MDCCCLXXVI. 2 a

212 PBOFESSOB OWEN ON THE. FOSSIL MAMMALS OF AUSTEALIA.

scalene triangle ; the shortest side is turned forward, the longest side forms the cutting-
edge, which is notched anteriorly by an oblique groove extending from near the middle
of the outside of the crown down to the cutting-edge, then inward and forward along
that edge to near the anteroinferior angle of the crown. The grooved part of the edge
in its present narrow condition thus presents two trenchant borders. The fore side or
border of the crown shows an anterior low convex ridge through the subsidence of the
enamel between this and the mid groove ; the enamel behind this groove is again at a
lower level, thus the anteroposterior lay of the outer enamel is undulated. The fore-
and-aft extent of the exposed crown is 6 lines, the front border or side of the triangle
measures 5 lines.

The indent caused by the missing second incisor is present on both right and left
anterior incisors ; the interspace between the first and third incisors, from which the
second has been pushed, is greater in the left than in the right premaxillary.

A detached fully developed third incisor of a full-grown Sthenurus Atlas is figured in
Plate 25. fig. 3.

In the lower jaw of this species* the interval between the molars and procumbent
incisor is less than in Macropus Titan f. The present specimen shows that a similar
character marks the upper jaw. The extent of the maxillo-premaxillary diastema is
here 10 lines, but would be of course greater in the full-grown Kangaroo.

The anterior molar (Plate 25. fig. 2, Plate 26. fig. 4, d a), with a crown 6 lines in
antero-posterior extent, shows a depressed middle tract of the outer surface traversed
by two vertical ridges. The inner surface, which forms posteriorly a prominent convex
lobe, sinks rapidly to a basal ridge as it extends forward to a low angle on the inner
side of the anterior division of the tooth. The second molar (d 4) has a prebasal ridge
without the fore link ; the mid link is small and low placed, in chief continuation with
the inner angle of the fore lobe. The hind surface of the hind lobe has a triangular
excavation.

These characters are repeated in m 1 and m 2 ; the crown of the latter is protruding
from the formative cell, and is unworn. The crown of the premolar, exposed in its
formative cavity (Plate 26. fig. 4,_p 3), is incomplete with the concomitant wide and deep
excavation at the basal part for the unexhausted pulp. The longitudinal grooves and
ridges of the trenchant apical border, part of which are visible in the worn premolar of
the subject of a former Part J, are well shown in the germ of the premolar in the fossil
now described. The fore-and-aft extent of the crown agrees with that of the fully
developed homologue, viz. 9 lines=18 millims.

The bony palate, so far as it is preserved, appears to have been entire.

§11. Sthenurus Brehus (Restoration of dentition and part of the skull). — This species
was founded on a maxillary portion of cranium with the left molar series, in part muti-
lated, and with the last two molars of the right side (Phil. Trans. 1874, plate xxvii.
figs. 5 & 6) ; also on a fragment of the left maxillary with the premolar and contiguous

* Phil. Trans. 1874, plate xxii. fig. 5. f lb. ib. &g. 13. J lb. plate xxiy. %. 4,jp 3.

PBOFESSOE OWEN ON THE FOSSIL MAMMALS OF AUSTEALIA. 213

molar entire (ib. figs. 7, 8, & 9) from a younger individual. Both fossils were from
Mitchell's Breccia-cave in Wellington Valley, New South Wales, and formed the first
evidence of this extinct species which came to my hands.

Since the publication of the paper (loc. cit.) illustrative of these fossils I have been
favoured, through the persevering and successful quest of George F. Bennett, Esq.,
with portions of the skull, and lower jaw of older and younger individuals of the same
species from the freshwater drift at Clifton, Queensland. One of these specimens
exhibits the entire molar series, left side upper jaw, with that of the right side, less the
last molar ; another fossil includes the premaxillaries and upper incisors ; and a third
consists of the fore part of the mandible with the lower incisor and with the first three
molars of the right side. All three specimens are parts of the same skull.

A fourth lot consisted of four portions of the upper jaw of a mature but younger
individual than the subject of plate xxvii. loc. cit., and included the incisors and pre-
molars of both sides, and the entire molar series of the right side.

A fifth specimen consisted of the fore part of the upper jaw of an aged individual
with much of the crowns of the incisors worn away, and the smaller ridges on the inner
side of the premolars rubbed smoothly down.

These specimens instructively exemplify the constancy of the maxillary molar
characters of the genus and species as shown in the type specimen, and add those of
the mandible and the characters of both upper and lower incisors.

Of the upper incisors the first or foremost (Plate 28. figs. 1, 2 9 & 3, i i) has a greater
relative superiority of size over the second and third than in any existing species of
Kangaroo that has come under my observation ; in this character Sthenurus approaches
the Koalas and Potoroos among existing, and the colossal Diprotodonts among extinct
Marsupials.

The transverse diameter of the crown of i i is 8 lines (17 millims.); it nearly equals
that of the two following incisors, of which the third (i 3) is broadest, viz. 5 lines
(10 millims.) along the oblique trenchant or working border; the thicker, triturant
surface of the second incisor is 4 lines (8 millims.) in longest diameter. The crowns
of the six incisors describe a semicircle (ib. fig. 3) ; those of the anterior pair, separated
by a line's breadth in the fossil, evidently touched each other in the living animal at
their median angles, which show the effects of mutual pressure.

Each incisor is curved lengthwise, with a strong outward or forward convexity ;
the exposed enamelled crown of the first measures in a straight line 1 inch 1 line
(27 millims.), that of the second incisor 7-| lines (15 millims.), that of the third
incisor the same; this, as usual, expands to the working surface; its outer enamel is
bisected by a feeble linear longitudinal groove. The transverse interval between the
two incisors of the third pair is 1^ inch.

The outer half of the fore surface of the crown of the second incisor (i 2) is prominent,
and is pressed into a corresponding channel of the hind surface of the first incisor;
the channelled part of the contiguous surface of the second incisor reciprocally receives

2 g2

214 PEOEESSOB OWEN ON THE FOSSIL MAMMALS OF AUSTKALXA.

the prominent part of the opposed first incisor ; the crown of the third incisor (i 3)
presses closely against that of the second : thus firmly interlocked the three incisors in
each premaxillary worked as one tooth. The enamel of the second and third incisors is
continued from the outer or fore part of the crown upon part of the hind or inner
surface ; but the enamel of the large anterior incisor is limited to the fore part. In
both proportion and curvature the large incisor resembles the homologous tooth in
Nototherium and Diprotodon ; but it is an incisor of limited growth, and its implanted
fang tapers to the end, as in the rest of the family Macropodidce .

From the back of the median border of the front incisor (Plate 28. fig. 3, i 1) to the
fore border of the u foramen incisivum" (ib. a) is 1 inch 6 lines; from this border a
groove is continued forward, shallowing, to near the tooth.

The breadth across the outsides of the last pair of incisors is 1 inch 11 lines ; the
breadth of the palatal part of the premaxillaries at the fore part of the prepalatal or
incisive foramina is 1 inch 10 lines. From the third incisor to the premolar is
2 inches 1 line ; in other words, this is the extent of the diastema or toothless space
(Plate 28. fig. 1, d, i 3) between the incisors and the molars. The breadth of the bony
palate anterior to the premolars is 2 inches 1 line.

To the objection that the species Atlas and Br elms of the genus Sthenurus might-
have been based, in Part VIIL, on parts derived from the female and male of the same
species, the reply is that, although the males exceed in size the females in most, if not
all, Kangaroos, the difference is chiefly shown in the bones of the trunk and limbs, less
so in the skull, and little, if at all, in the teeth.

Now the third incisor is, relatively to the first incisor, smaller in Sthenurus Brehus
than in Sthen. Atlas. In the larger species the length of the molar series (Plate 28. fig. 1,
p 3-m 3) is 3 inches 6 lines, in Sthenurus Atlas it is 2 inches 9 lines. The premolar
(p 3) exceeds the rest in fore-and-aft diameter, which, as in the cave specimen (loc, cit.
plate xxvii. fig. 7, p 3), is 9-J lines (20 millims.) ; the three low transverse ridges which
connect the inner with the outer wall are well marked in the present comparatively
young though full-grown individual. These ridges become less salient in the course
of the oblique wear of the crown of the premolar from the outer to the inner ridge,
and in old individuals they are polished off. But all the generic and specific characters
of the premolar of Sthenurus Br elms from Mitchell's Breccia-cave in New South Wales
are repeated in the present specimens from the fluviatile beds of Queensland. The
same may be said of all the succeeding molars which in the type specimen are suffi-
ciently complete for comparison. The last molar in the present example (Plate 28.
fig. 3, m 3) has not quite come to the grinding level, and its ridges are untouched.
The enamel-fold from the inner angle to the hind ridge, which defines by its oblique
tract along the hind surface the angular depression there, seems as if it were folded on
itself or notched at its basal termination.

The descending process of the zygoma is more perfectly preserved than in any other
fossil hitherto transmitted to me of the genus Sthenurus ; it terminates below the level

PEOEESSOB OWEN ON THE EOSSIL MAMMALS OE AUSTRALIA. 215

of the grinding-surface of m 2 ; ia older examples it would show the same relation to m 3,
as the grinders move, or seem to move, forward.

I next proceed to notice the portions of skull of a more aged individual of Sthenurus
Brehus from Clifton, Queensland. The laterally crushed maxillary part of the skull
includes, with the incisors, the entire molar series of the left side and the major part of
that of the right side. The premolar with a fore-and-aft length of 10 lines (20 millims.)
in the type specimen (loc\ tit. plate xxvii. fig. 7, p 3) is but half a millimetre less in the
present fluviatile fossil ; and this seems due to the wear of the anterior prominence.
But all the formal characters are closely repeated. I have had no evidence from the
spelaean haunt of the Thylacoleons of a giant Kangaroo which had attained the expe-
rienced age of the original of the present Queensland fossil. The molar contiguous to
jf) 3 contrasts, as usual, its great degree of wear with the fresher crown and higher level
of the antecedent subsectorial tooth ; the fore-and-aft diameter, 6 lines (12 millims.), is
the same in both fossils ; the minor transverse breadth in the Queensland specimen is
due to the wearing down of the outer angles of the transverse lobes or ridges, which
are prominent in the cave fossil. The superiority of size, slight as it is, in m 1 of the
type subject of plate xxvii. figs. 5 & 6 (loc. cit.) is chiefly due to the minor wear of the
crown of that tooth in the cave fossil. The last two molars occupy a longitudinal
extent of 1 inch 6 lines (37 millims.) in both specimens. The linkless prebasal ridge
is transverse, not curving at either end to be continuous with the corresponding angles
of the fore lobe ; the low,, short mid link is less distinctly continued to the inner angle
of the fore lobe than in Sthenurus Atlas; the depression on the hind surface, due
chiefly to the ridge curving from the inner and hinder angle of the hind lobe toward
the outer side of the base of the crown, with the lower and shorter ridge from the outer
angle, are all characters of Sthenurus Brehus, repeated in the present as in the preceding
Queensland sedimentary fossil.

The left molar series in this instructive specimen occupies a longitudinal extent of
3 inches 3 lines.

The dentine is exposed on the fore lobe of the last molar, and the fore part of the
enamel ridge of the hind lobe is nearly worn through ; the prebasal ridge also shows
abrasion. A hollow transverse field of dentine is exposed on both lobes of the penulti-
mate molar. With these indications of greater age the maxillary pier has retrograded
and projects on the transverse parallel chiefly of the last, instead of the penultimate,
molar as in the younger specimens (plate xxii. loc. cit.). In all the bony palate is
entire.

The fore part of the present skull shows a diastema 2 inches in extent. From the
back of the socket of the third incisor to the fore part of the crown of the first is 1 inch
3 lines. Of the third incisor, the seat of variety in existing Kangaroos, the left is lost,
and of the right one only the fang remains. The crown of the second right incisor is
worn nearly to its base. The first or front incisor is present in both premaxillaries,
with its fang exposed in the left one. The crown is worn to the level of the palate ;

216 PEOFESSOE OWEN ON THE EOSSIL MAMMALS OE AUSTEALIA.

the smooth and polished surface (Plate 28. fig. 5) presents an oval figure, the great end
outward; the long diameter is 7-§ lines (15 millims.), the short diameter, near the outer
side of the worn surface, is 5^ lines (11 millims.). The enamel, nearly 2 millims. in
thickness, is limited to the anterior surface, bending slightly back at the outer and
inner margins ; not more than 9 lines in length of the enamelled crown remains. The
whole incisor, as usual, is curved lengthwise, with the greater convexity anterior; the
root contracts to its implanted end ; the length of the incisor, in a straight line, as here
worn, is 1 inch 8 lines. The portion of cranium preserved, from the fore part of the
front incisor to the back of the last molar, measures 6 inches 8 lines. About 2 inches
of the facial part of the premaxillaries are preserved, bounding by a curved and obtuse
border, thinning as it rises, the anterior nostril (Plate 28. fig. 1, 22). According to the
proportions of some existing Wallabies, which retain the premolar with the last molar
in use, 5 inches may be added for the extent of skull behind the last molar, and the total
length of the skull in Sthenurus Brehus may be moderately estimated at 12 inches.

The skull, so far as it is shown in the present specimen, has been crushed sideways,
not partially as in the cave fossil, but by a pressure operating along a more extensive
plane, and which I deem to be due to movement of the matrix, rather than to the jaws
of a destroyer or devourer.

The lower jaw, which appears to have been imbedded originally in connexion with
the upper one, has suffered similar lateral compression. Only the fore part of the
mandible has been obtained or transmitted: it includes the pair of lower incisors with
the premolar and two following molars of the right side ; these are much worn. The
length of the diastema is 2 inches: in Sthenurus Atlas f it is 1 inch 3 lines. In the
course of the pressure to which this mandible of Sthenurus Br elms has been subject,
the attachment of the broad symphysis has been overcome, and the right ramus has
been moved a little in advance of the left.

The thin alveolar sheath does not extend to, or has been lost from, the end of the
cement-clad root next the enamelled crown. This expands as it extends forward, and
terminates in a polished worn surface, 10 lines in long diameter (obliquely transverse),
5 lines in fore-and-aft diameter. The length of the enamelled (under or outer) part of
the crown is 8 lines, that of the entire tooth is 2-J inches. The fang, as usual, tapers as
it recedes in its socket (Plate 28. fig. 1, i*). The narrow symphysial border sinks from
the premolar alveolus with a sharp curve before extending forward to expand upon the
terminal part of the socket of the incisor. The outlet of the dental canal (v) is nearer
the molars, and the diastema, with the symphysial part of the ramus, is longer, relatively,
than in Sthenurus Atlas.

A third example of the premaxillary part of the skull with the fore part of the right
maxillary, of a still older individual of Sthenurus Brehus from Queensland deposits,
shows the six incisors in situ, much worn, and the form and transverse extent of the
palate between those teeth and the molars. The length of the diastema is 2 inches

Phil. Trans. 1874, plate xxii. fig. 5.

PEOEESSOE OWEN ON THE FOSSIL MAMMALS OE AITSTEALIA. 217

I line, the breadth of the palate at the prepalatal foramina is 1 inch 9 lines. These
foramina are elongate, from 2 to 3 lines broad, about 6 lines long, and continued forward
by a groove (ib. #), which shallows out when parallel with the third incisors. These are
much worn ; the working-surface of the third is 7 millims. by 6 millims. The enamel,
which is continued from the outer upon the hinder surface, is impressed, as in the less-
worn tooth of the younger JBrehus^ by a longitudinal groove almost equally bisecting the
outer surface. The second or mid incisor is worn almost to the stump. The first pair
of incisors, being more worn than in the subject of figure 4, show a working surface
of similar shape but rather smaller dimensions ; the anterior coat of enamel is reduced
to a length of 4 lines (10 millims.).

In the collection of Marsupial fossils from Queensland in the Museum of the Natural-
History Society at Worcester, I noted, in 1853, the left lower incisor of a Kangaroo, of
which I made drawings of the under or outer side (Plate 28. fig. 4). This tooth best agrees
with the corresponding incisor of Sthenurus Brehus. It had preserved an extent of
enamelled crown of 8 lines, the breadth being that of the more worn incisor of Plate 28.
fig. 5. In the large existing Kangaroo (Meter opus major*) the breadth of the crown
of the lower incisor is 4-J lines, in Macropus (Osphranter) rufus it is barely 4 lines.

The portion of skull of a Sthenurus Atlas (Plate 25. fig. 2, Plate 26. fig. 4) permits
comparison to be made of the first and third incisors with those teeth in Sthenurus
Brehus. The first incisor is but half the size of that in the larger species, while the
third incisor presents a crown of equal size. The generic character of equal division of
the crown by the longitudinal groove is retained, but the anterior border of the groove
is produced at the lower part of the crown. Such character, however, may have
existed in the part of the crown worn away in the subject of Plate 23. fig. 1, id. But
the difference of proportion in the upper incisors is, at least, specific. In Macropus
Titan (Plate 25. fig. 4) the modification of the pattern of the third incisor is associated
with generic distinction in other parts of the dental system and in the skull itself.

§ 12. Macropus affinis (Metatarsus)- — The hind foot in Macropodidw exhibits an
extreme modification of its bony structure. The inner toe (i) is suppressed ; the meta-
tarsals of ii & in are long and filamentary, supporting a pair of small pendent fur-
cleansing claw-toes ; that of iv is both long and large, with characteristic modifications
of its proximal end ; that of v is much less, the shaft compressed, but supporting, like
that of iv, a toe with a quasi hoof for station and progression.

In Plate 29. fig. 4 is given a view of the fourth and fifth metatarsals, natural size, of
the right hind foot of the large male Macropus (Osphranter) rufus, already referred to.
Two other views of the same bone will be found in the pap'er " On the Osteology of
the Marsupialia " in the Zoological Transactions f.

In the task of determining the fossil specimens of the homologous bones, those best
agreeing in proportions with the corresponding metatarsals of existing Kangaroos w 7 ere
set apart from the fossils deviating in a marked degree from such proportions.

* Phil. Trans. 1874, plate xx. fig. 15. f Yol. ix. plate lxxxiii.

218 PBOFESSOE OWEN ON THE FOSSIL MAMMALS OF AUSTBALIA.

In the first group was the upper two thirds of a left fourth metatarsal (Plate 30.
figs. 7, 8, & 9) with a proximal articular surface of the same breadth as in Macropus
rufus and Macropus major, but with a markedly thicker shaft, being broader from side
to side in proportion to the depth from before backward. The smooth tendinal groove
answering to that marked n in figure 4, Plate 29, on the fore part of the shaft in Macro-
pus rufus, was bounded in the fossil (which I have noted, for convenience, as from a
Macropus affinis) by stronger ridge-like risings, and the groove does not reach so far
down the fibular half of the anterior surface of the bone. On the tibial side of the
tibial ridge of this groove the fore part of the shaft shows a slight concavity in Macropus
affinis, whereas in both cited species of the large existing Kangaroos the answering part
of the metatarsal is transversely convex. At the back part of the proximal third of the
shaft the rough surface or ridge for muscular attachment is more prominent, better
defined, yet less extended longitudinally, in the fossil. This specimen is from King's
Creek, Darling Downs, Queensland.

§ 13. Phascolagus alius (Metatarsus). — A right fourth metatarsal (Plate 30. figs. 1-5)
from the same formation and locality has come from a larger kind of Kangaroo than
Macropus affinis. It is an inch longer than the subject of figure 4, Plate 29 (Macropus
rufus), is relatively thicker, and, like the previous fossil, differs in presenting a stronger
ridge bounding the fibular side of the anterior surface of the shaft. This part of the
bone is also more prominent, giving a convexity to the outline of a side view (ib. fig. 3)
not present in the fourth metatarsal of Macropus major or Maw. rufus*. The present
fossil likewise shows a relatively broader distal end (ib. fig. 6), which is barely 1 line
less in transverse diameter than is the proximal end (ib. fig. 5). On this surface, as in
Macr. rufus and Macr. affinis, the hinder prominence shows the oblique tendinal groove
(ib. fig. 4, g) and the flat inferior facet (h) for the articulation of the large tarsal sesamoidf .
The fibular or outer side of the proximal end, in both fossils, shows, as in the recent
Kangaroos, the antero-posteriorly prolonged, bilobed, articular surface (ib. fig. 2, m, n)
for the side of the head of the fifth metatarsal. Below this surface is the depressed
rough tract (ib. Jc), continued down nearly three fourths of the back third of the fibular
side of the shaft, for the ligamentous attachment of the smaller compressed shaft of the
fifth metatarsal. The posterior ridge in the present metatarsal, which I refer to a
Phascolagus alius, answering to that marked d in Plate 29. fig. 6 (Macropus affinis), is
continued lower down, nearly to the end of the shaft. The hinder half of the distal
articular surface (Plate 30. fig. 6) is fashioned by a mid rising into a double trochlea,
as in existing Kangaroos.

§ 14. Palorchestes (Metatarsus). — The breadth of the middle of the shaft of the
fourth metatarsal of Macropus rufus is two thirds that of the same part of the femur.
The breadth of both fourth and fifth metatarsals, naturally united one third down, is
nearly equal to three fourths of the breadth of that part of the femur.

The breadth of the middle of the shaft of the fourth metatarsal, the subject of

* See figs. 1 & 2 of plate lxxxiii. Zool. Trans, torn. eit. f See Cut, fig. 2, p. 441, Zool. Trans, torn. cit.

PEOEESSOE OWEN ON THE EOSSIL MAMMALS OE AUSTEALIA. 219

Plate 29. fig. 1, is two thirds that of the same part of the femur, as shown at the
broken ends in figs. 1 & 2 of Plate 23. The breadth of the naturally united fourth and
fifth metatarsals (Plate 29. figs. 1 & 2), one third from their proximal ends, equals four
fifths of the same part of the femur of Palorchestes Azael.

To this species, therefore, I provisionally refer the fossil subjects of figs. 1, 2, and 3 of
Plate 29. The length of the fourth metatarsal in those figures is restored in outline
from an entire homologous bone of the same extinct species which had become detached
from the contiguous metatarsals, and which I have not thought necessary to figure.

In the fossil under description, to the great fourth metatarsal lacking the distal end
there remained attached the fifth metatarsal (ib. figs. 1 & 2, v), wanting only a portion
of the plantar or posterior wall of the proximal end; and, similarly attached, the
proximal half of the third metatarsal (ib. ib. in), exhibiting the characteristic macro-
podal slenderness. It is interesting to note, however, that the fifth metatarsal in the
huge extinct Kangaroo shows more nearly the normal proportions of the bone than in
the existing species, even the largest, as, e. g., Macrojpus rufus (Plate 29. fig. 4, v); and
the more slender third metatarsal bone of Palorchestes (in, fig. 2, Plate 29) is relatively
less atrophied than in existing Kangaroos. Its proximal end (ib. fig. 3, in) presents an
oblong, subquadrate, almost flat surface for the ectocuneiform, and a triangular surface
on the inner side (ib. fig. 2, b) of the proximal end for part of the head of the second
metatarsal, which bone is wanting in the fossil. It articulates with the tibial side of the
proximal end of the fourth metatarsal by a surface which is extended by the backwardly
directed process (ib. fig. 2, c). Below the articular head (ib. a-c) the shaft narrows and
becomes compressed in its upper third, below which it assumes a subtrihedral figure,
with a trenchant margin both before and behind along its upper fourth.

The shaft does not exceed 2 lines in breadth at the point of fracture (d). The bone
closely adheres, either by anchylosis or matrix, to the shaft of the great fourth meta-
tarsal, inclining from the inner (tibial) side to the plantar aspect of the shaft, as does
the corresponding metatarsal in Macropus rufus*. The proximal articular surface of
the fourth metatarsal (Plate 29. fig. 3, iv) presents an undulating tract adapted to the
distal surface of the cuboid ; its broadest rotular (anterior) half is gently convex trans-
versely at the fibular half, concave at the tibial half; the narrower plantar or posterior
subquadrate tract is feebly concave rotulo-plantad, almost level transversely ; this tract
extends plantad, or backward, so as to overhang the shaft ; it is grooved at e, fig. 3, by
the "peroneus" tendon, which runs across the back or under (plantar) aspect of the
tarsus to be inserted into the entocuneiform bone ; also, as in existing Kangaroos and
in Macropus affinis^ it has a flat articular surface at the underside of the overhanging
part for the large tarsal sesamoid. The inner or tibial side of the proximal end of the
fourth metatarsal shows the two small vertical facets for the ectocuneiform, and a well-
marked rough depression for the proximal ends of the third and second slender
metatarsals.

* Zool. Trans, torn. cit. plate lxxxiii. fig. 1, in.
MDCCCLXXVI. 2 H

220 PEOFESSOE OWEN OF THE FOSSIL MAMMALS OF AUSTEALIA.

The shaft of the great metatarsal of Palorchestes is subtrihedral ; the rotular surface
(ib. fig. 1, iv) is slightly concave transversely along its middle third as in Macropus
Titan, not prominent as in Macropus rufus (ib. fib. 4). The plantar side (ib. fig. 2) is
produced into a ridge, broad along the upper third (o), becoming sharper (h, h) as it
descends, and subsiding about one half of the length of the bone (</) from the distal
end. The corresponding portion of the metatarsal of Macropus rufus (ib. fig. 6) is
widely channelled where in the fossil it is angularly convex. The greatest rotulb-
plantar thickness of the shaft in Palorchestes is 1 inch 4 lines, the greatest transverse
thickness is 1 inch ; that of the proximal end is 1 inch 6 lines, its rotulo-plantar thick-
ness is 1 inch 5 lines.

The fifth metatarsal of Palorchestes Azael (Plate 29. figs. 1, 2, 3, v) is relatively much
stronger than in Macropus major or Macr. rufus (ib. fig. 4, v). The plantar part of the
proximal end, broken away in the fossil figured, is entire in a later acquired homologous
bone of Palorchestes. This presents a small, oval, fiat, vertical surface for the fourth
metatarsal, a broader sub triangular one for the backwardly extended process of the
cuboid*, and a larger horizontal facet for the surface, marked If, of the same tarsal
bone. The proximal articular surface of the fifth metatarsal is very small in propor-
tion to the bone in Palorchestes. External (fibulad) to that surface the bone rises above
the proximal end of the fourth metatarsal in the form of an antero-posteriorly extended
thick round edge.

The shaft of the fifth metatarsal is subcompressed along the proximal three fourths ;
it measures 13 lines in rotulo-plantar thickness, one third down; and here, near the
plantar side, its thickness (or tibio-fibular breadth) is 6 lines. The opposite (rotular)
border is not sharp, as in Macropus major and Macr. rufus ; but though thin, in com-
parison with the plantar surface, the border is rounded off. The shaft loses rotulo-
plantar thickness and gains transverse breadth as it approaches the trochlear articular
surface (v); this is 1 inch transversely, 10 lines where thickest from before backward.
The surface is not simply convex, as in Macropus rufus (Plate 29. figs. 4 & 5, v), but is
made trochlear by a plantar median ridge, on each side of which the surface, trans-
versely, is feebly concave. The outer (fibular) side of the shaft has a feeble median
longitudinal channel along the middle of the proximal two thirds. The upper half of
the shaft shows in fractured portions of homologous fossils a small medullary cavity.

Thus we learn that in the large extinct Kangaroo of the genus Palorchestes the
fourth and fifth digits were less unequal in strength, and the fifth took more share in
station and locomotion than in the largest existing kinds. The metacarpal segment
and the rest of the foot were proportionally broader; but the length of the fourth
metatarsal in Palorchestes indicates, nevertheless, that it was a powerful leaper.

§ 15. Procoptodon (Metatarsal and Femoral Characters). — In deprecation of the fore-
going details, more wearisome perhaps to the reader than the inditer, I may plead the

* " Osteology of Marsupialia," Zool. Trans, torn. cit. plate lxxxiii. fig. 11, k.
f Ibid. fig. 10.

PEOEESSOE OWEN ON THE EOSSIL MAMMALS OE AUSTEALIA. 221

great proportion of fragmentary evidences of the hind feet of large extinct Kangaroos
as compared with entire or nearly entire bones. The grateful aid which such rare
specimens have yielded has impressed me with the desirability, if not duty, of defining
and recording all characters which may help future collectors, especially in Australia,
in determining such fossil fragments which are likely to accumulate in the public and
private museums of that great colony.

I refrain from trespassing on the lime and means of the Society with the results of
comparisons by which metatarsals of Macropus Titan, of the two dentally determined
species of Sthenurus, and of some species of Protemnodon have been worked out. But,
in regard to the extinct form of Kangaroo which has most interest in relation to its
aberrant or transitional character, I cannot withhold evidences which give some clue
to the characters of the hind limbs, and I finally pass to the result of the present
researches which has yielded me perhaps most satisfaction.

I have alluded to the primary step in the survey of the vast series of metatarsal
fossils which led to setting apart those indicative of a hind foot shorter in proportion to
its breadth, and yet retaining unmistakable "macropodal characters.

In the specimen, for example, of the naturally united fourth and fifth metacarpals
figured in Plate 31. figs. 1-5, the fourth is thicker than, but is little more than two
thirds the length of, the homologous bone in Macropus rufus (Plate 29. fig. 4). The
fifth metatarsal (Plate 31. figs. 1, 2, 3, v) shows a greater degree of thickness, in pro-
portion to its length, than in Palorchestes Azael (Plate 29. figs. 1 & 2, v).

The proximal end of the fourth metatarsal (Plate 31. fig. 4, iv), though somewhat
mutilated, exhibits the characteristic modifications of the articular surface in the normal
Kangaroos ; and these characters are shown more plainly in the homologous bone of a
larger kind of Procoptodon (ib. fig. 8, iv), viz. the non-articular peninsula (/), the back-
ward or plantar production (e), with the terminal groove (g 1 ) for the tendon before
mentioned ; but this groove is less deep than in the type Kangaroos : the flat surface
beneath (fig. 6, h) indicates a larger proportional sesamoid than in the species of
Macropus.

The fore surface of the shaft of the fourth metatarsal (ib. fig. 2) is more even or
flattened than in Macropus affinis, Macr. Titan, and the great recent kinds of Kan-
garoo. The distal end is more expanded, surpassing the proximal end in breadth.
The ridge on the back part of the shaft (ib. fig. 1, o) is broader, less produced, and less
extended downward than in Macropus, Sthenurus, ol* Palorchestes.

The well-preserved proximal end of the fifth metatarsal shows its apophysial pro-
duction (ib. figs. 1 & 2, z) proximad of the articular surface of the fourth metatarsal.
This process is absolutely as well as relatively longer than in Palorchestes Azael As
in that species three articular facets are present at this end of the bone, two for the
cuboid and one for the contiguous (fourth) metatarsal, the cuboidal surfaces being
relatively larger than in Palorchestes, and the tubercle projecting tibiad beyond the
metatarsal surface is more developed in Procoptodon.

2h2

222 PBOFESSOB OWEN ON THE FOSSIL MAMMALS OE ATJSTEALIA.

At the distal articular surface the chief difference is seen in the greater production
of the tibial convexity of the trochlea. By reason of this prominence its preservation
is rare in the rolled fossils of the present form of metatarsal from the creek-beds of
Queensland.

Of the metatarsals of this shorter type three modifications are shown by the fossils
that have reached me, which are indicative of three species of Procoptodon. The
metatarsal (Plate 31. figs. 10-12), of equal length with that (ib. figs. 1. 2) of Procoptodon
Pusio, but more slender, I take to be from a female Kangaroo of that species.

Deeming it probable that the form (family or genus) of Macropodal Marsupials
which, by dental and mandibular characters, offered the nearest approach to the large
isopodal or gradatorial family (Diprotodontidw) would present a corresponding approach
thereto in the form and proportions of the hind foot, I refer the present type of meta-
tarsal bones to the genus Procoptodon.

In this genus the above-described representative of the smallest known species would
answer, as to size, to the evidences which have been given of the maxillary, mandibular,
and dental characters of Procoptodon Pusio*.

The specimens next in size (Plate 31. figs. 6-9) I refer, on similar grounds, to
Procoptodon Baphaf.

Parts of a hind foot of a still larger species similarly relate to Procoptodon Goliah%,
Save in size, the characters of the metatarsal about to be described so essentiallv
resemble those of the homologous bone in Plate 31, that I have not thought it reason-
able to devote to it an additional Plate.

This metatarsal, the fourth, is 5 inches 3 lines in length, with a proximal breadth of
1 inch h\ lines, the opposite dimension being 1 inch 3 lines. The fore part of the
cuboidal surface is relatively broader from before backward than in Procoptodon Pusio.
There is no indication of the proximo-tibial ridge (#, fig. 10, Plate 30), in which cha-
racter the present bone resembles its homologue in Procoptodon Baplia.

The posterior angle at the proximal half of the shaft is less marked and less produced
than in Procoptodon Pusio. The posterior depressions above the distal trochlea are
deeper and better defined than in Procoptodon Baplia or Procop. Pusio.

The fifth metatarsal of Procoptodon Goliah is 5 inches in length ; the greatest dia-
meter of the shaft is 1 inch, equalling that of the homologous bone in Palorchestes
Azael, which is more than one fourth longer. The tibial covexity of the distal trochlea
is less produced than in Procoptodon Bapha, and the whole hind surface of the joint is
less obliquely disposed than in that species or in Procoptodon Pusio.

Thus the resolution of these shorter and stouter metatarsals into three categories,
characterized by modifications of shape as well as by size, concurs with the previously
adduced evidences of jaws and teeth in showing that the procoptodont modification of
Macropodidm was of old manifested by Australian Kangaroos under three specific forms.

* Phil. Trans. 1874, p. 788, plate lxxvii. figs. 2-7. } Ib. p. 791, plates lxxix., lxxx.

f Ib. p. 788, plate lxxvii. figs. 8-12, plate Ixxviii.

PEOFESSOE OWEN ON THE FOSSIL MAMMALS OE AUSTEALIA. 223

But these, which my present materials have enabled me to define, may prove not to
have been the only links connecting the saltigrade with the gravigrade groups of
phy tiphagous Marsupials.

Description of the Plates.

PLATE 19.

Fig. 1. Portion of right mandibular ramus and teeth of Palorchestes Azael, Ow.

Fig. 2. Working-surface of the same teeth.

Fig. 3. Hind fractured surface of the same fossil.

Fig. 4. Inner side view of last lower molar of Palorchestes Azael.

Fig. 5. Upper view of part of left mandibular ramus with teeth and stumps of the

same species.
Fig. 6. Outside view of the molar (m i) of left mandibular ramus of the same.
Fig. 7. Inside view of the same molar.
Fig. 8. Back view of the same molar.

PLATE 20.

Left side view of the fore part of the skull and teeth of Palorchestes Azael : the entire
skull and dentition restored in outline.

X JuA.XJLi JiLm

Fig. 1. Haemal (under or front) view of the sacrum and first caudal vertebra of Palor

chestes Azael.
Fig. 2. Eight side view of first caudal vertebra of the same.
Fig. 3. Neural (upper or back) view of the same.

PLATE 22.

Fig. 1. Outer side view of part of right " os innominatum " of Palorchestes Azael

Fig. 2. Inner side view of the same.

Fig. 3. Transverse section of ischium at 63, fig. 1.

Fig. 4. Transverse section of pubis at 64, fig. 2.

Fig. 5. Acetabulum of Procojptodon Goliah(V).

224 PEOFESSOE OWEN ON THE FOSSIL MAMMALS OF AUSTEALIA.

PLATE 23.

Fig. 1. Back view of proximal portion of right femur of Palorchestes AzaeL

Fig. 2. Back view of distal portion of the left femur of the same.

Fig. 3. Back view of distal portion of right femur of Procoptodon Goliah.

Fig. 4. Upper view of right calcaneum of the same.

Fig. 5. Upper view of left calcaneum of Palorchestes AzaeL

PLATE 24.

Fig. 1. Fibular (outer) side view of proximal portion of left tibia of Palorchestes

AzaeL
Fig. 2. Eotular (front) view of the same.
Fig. 3. Popliteal (back) view of the same.

Fig. 4. Tibial (inner) side view of proximal end with epiphysis of the same.
Fig. 5. Tibial (inner) proximal articular surface of the same.

PLATE 25.

Fig. 1. Left side view of cranium, with skull restored in outline, of Macropns Titan.

Fig. 2. Left side view of fore part of cranium of a young Sthenurus Atlas.

Fig. 3. Outside view of left upper third incisor of Sthenurus Atlas.

Jig. 4. Outside view of crown of left upper third incisor of Macropus Titan.

PLATE 26.

Fig. 1. Under or base view of cranium, with skull restored in outline, of Macropns

Titan.
Fig. 2. Hind surface of last molar (m 3) of the same.
Fig. 3. Hind surface of last molar of Macropus major.
Fig. 4. Under view of fore part of right upper jaw and teeth of a young Sthenurus

Atlas.

PLATE 27.

Fig. 1. Front view of right femur (omitting 2 inches of shaft) of Macropus Titan,
Fig. 2. Back view of right femur (omitting 2 inches of shaft) of the same.
Fig. 3. Inner view of proximal end of right femur of the same.
Fig. 4, Outer view of distal end of right femur of the same.

PKOFESSOE OWEN ON THE FOSSIL MAMMALS OE ATJSTKALIA. 225

PLATE 28.

Fig. 1. Outside view of parts of right upper and lower jaws and teeth, with part of the

skull restored in outline, of Sthenurus Brelms.
Fig. 2. Upper view of premaxillaries and incisors of the same.
Fig. 3. Under view of the same fossil.

Fig. 4. Under view of right mandibular incisor of Sthenurus Brelms.
Fig. 5. Upper surface of worn end of right mandibular incisor of an aged Sthenurus

JBrehus.

PLATE 29.

Fig. 1. Front view of metatarsals iv and v and part of in of Palorchestes Azael.

Fig. 2. Back view of the same.

Fig. 3. Proximal (upper) articular end of the same.

Fig. 4. Front view of metatarsals iv and v, with part of cuboid, of Macropus rufus.

Fig. 5. Back view of distal portion of the same.

Fig. 6. Back view of upper two thirds of fourth metatarsal of Macropus affinis, Ow

Phascolagus alius.

Fig. 1. Front view of the right fourth metatarsal.

Fig. 2. Outer side view of the same.

Fig. 3. Inner side view of the same.

Fig. 4. Back view of the same.

Fig. 5. Proximal articular surface of the same.

Fig. 6. Distal articular surface of the same.

Macropus affinis.

Fig. 7. Outer side view of portion of the right fourth metatarsal.

Fig. 8. Back view of part of the same.

Fig. 9. Proximal articular surface of the same.

Undetermined Species.

Fig. 10. Front view of proximal phalanx of fourth toe of an extinct Kangaroo, of the
size of Phascolagus alius.

226 PBOEESSOR OWEN ON THE EOSSIL MAMMALS OE AUSTRALIA.

PLATE 31.

Fig. 1. Back view of metatarsals iv and v of Procojptodon Pusio, Ow.

Fig. 2. Front view of the same.

Fig. 3. Outer side view of the same.

Fig. 4. Proximal (upper) articular end of the same.

Fig. 5. Distal (lower) articular end of metatarsal iv of the same fossil.

Fig. 6. Back view of metatarsals iv and v of Procoptodon Bapha, Ow.

Fig. 7. Outer side view of metatarsal v of the same.

Fig. 8. Proximal (upper) articular ends of metatarsals iv and v of the same.

Fig. 9. Distal (lower) articular end of metatarsal iv of the same.

Fig. 10. Front view of metatarsal iv of Procojptodon Pusio, foem.

Fig. 11. Proximal articular end of the same.

Fig. 12. Back view of distal half of the same.

All the figures are of the natural size.

(hrmv.

Phil Trans 1876 . Plate 19.

g^> i W'"

%$ #

S *■• '«''>'SW: '.*»WjM, WjJ

<n 2-

m,1

?1

£&4

kmtf^tsK f*fv«^v^»w

P«3

>ts, **>^„. '

{tWaaSW^W*^^

*wnwnw«tM(8$$INaiWlfe

i i

1§S§§

fissKgs

^"^slj

7.

^^'- !; ^^. ^S^/ ^¥^" : ^'^
H # K'ZJW

!&*'*§«« Si I
w

GHForcl ScC.L,Grles"bacb.

Mxrtern Bros - imp

(hvi

en

,^«JJ«'!S^'*'" XK '' tei "* W *** to «»^»f»H^ w ^

»"*->.'*v S f~ ! ,, l .,...

•■■■' ; "!'»« Sa »

...^"

?5. ,

'*^'^

^■"

./'

,*#■*

S^-^" r '

,**>""

l**"*

«*#*""'

WMjjj^jIKW

Phil, Trains .1876 . Flcute 20.

*<»'»*:«**>■,...,

■ '""•*'-•»«

''~ fts *fe>.^

-■*»:>-

'"Hi

"'•^ } ,.

C.L.GriesWcK cLei. et nth,.

Mmtern Bros .imp

WPJIa

PMl Irons. 1876. Plate 21

sit ■(- i /

8 /•• a

From, nat.on stone lay J.Erxleben.

M&N.HaxxhaH imp.

OtireiV;

Hul.Thms. 1876. Plate 22.

2..

trom nut. on .oUvneljy J-Er-vU^x-Jii.

:.M,;[.Haj:l!,i!t irso.

TireiV.

Pi

PkOiTrans. 1876. Plate 23

3

■■■■• ■ ■.<■■ ■ Mmmwwm

«5 i " i

tmz*- ,wu x v iff

1 HI

ill-

ZT

isjSKsraS^i

GL.Grriestacli lith.

M&NIWiaii imp.

Give-its,

PJul. Trans. \81Q. Plate 24.

^»».

'^i*i

kWl£8$

«|^W^,;^^ '■'• ;i v:

inn

{ 1 ■ <**

■ " , .■••i;.k'-/: ,: V---:. v i'.-V- 1 --?;

pl§li§ll •

«

irasg

'Sire

MHSSm

If

TO

K»

$$?iji

«$3S3#i

ifi 1 l - :

tePS

SIP

ansa ... K 11 1 '.

gff!
«1 ■

.'«•■

■Mr

4l

^fi :

iw^^v

i»t

api

i i m mm

ffl^i

:«#'

^ >

sill

^^

^^^

sua

■ ■'"■"' "'■'■•c'-X-

.•-.••;■ rf /-: •■^^^■•••^'••'.^.■.■^••■■'' ..••'

,.A.;i-i:.t; I .,"1;^^.. i* •

waP

«&«,,

»

intt

! r& I

m

M f>

mm?

- _

'.i*<5

::^V;

:"'.'V '-,"■'■ '■■

•J - ,* , '

^'M''

! .''- : :

m

m

! -?.;--

MPS

^

M

; rV

i»

4i

'if

:-*xm ■■■.

mm

%■
I

i'?fev. 4 v:.. '

»*?*-

From nat. on stone ij J-Erxleten.

M&-N.Ha»ftart imp.

o,

7ire?U

Plul.TransASlQ.PloM 25,

<m

rg

£-4

H '

&

a>

Pi

o

to

i'-i

o

g

i — ^

CD

ft
O

o
<

Ovreri-.

Pkil.Traiis .1876. PlaJ^ 26

^^fi^t't].

I

m

feife #

llltiiilliii:

H.O-wert. d«=l. on. stone by J.*Jirjde,beJi,

MACROFUS TITAN-

Prints S. "by M & JT.Haxt.'ha.rt;.

Onr&v,

Phil. Trans. 187 6. Plate 21.

./'

/

/•">

tr

GL.G-riesback KtL

M&NHanbaH imp.

Owen,.

PM'.TramlQ16. Plate 28

2 .

J

'**"■* ^4*-,

■"■ '!

s .' ^., . i/., '*_. .'. ... ; '. i ." '*■■

si* 1

«««a&;

SfttS<fef «*»««■ :V-

life

~\

\

V

\

- ... V

On stone hy J.'ErxLe.ben.

M <3c 'N . Haxi.ha.rt imp .

Otvejh.

PMl. Trans. 18 76. Plate 29

If'

III

lis

IV

IV

w

From nai; on Stone by J .Eralebe

n

MkKThwharL imp

O wen,.

PMl. Trans. 1876. Plate 30

.9

*■""

b

HhPp

Mmm

IBtip

•■■rar$raf8«fe

«§§

v9iviSu9v%n

'. ate

.tea;

*S?3^

^W-°

■"Y

vs&m

&m

If I; •

■ .J?

ill

'■"■" *,

:*.. ...

kJnS

IJ'S '

fcf'.

'fli

§11

Sag

&/

£

£A ; *sSiM>

HOHH

is-' •■:;;.

hi a

•.■;;r"'.^ .

[-1

I

i»i

■*'.

mm

i

mm

I*:

m

i

WmBz?
illfcl?

HI

••fe

wmm$i

liilllll - J
El mm

Mm

IW0

m

'';' ; '';'. : >.:V/'.>'..-..

— e

l

-*''':" ;

.•<'-}•'•.••:•..'■.. ■

..- s

»j

S§K8;>

SSsSs

S»

KM

P

s'/V.

sit *

*:.i.' ; >---

^

fee

msm

!•.':'>;.

/'•.'•«'

,-§.^# '

**«isMs«^ r

iTom'riou Jn BLone by J .Eivxleoen ,

J;4 cI'lH Fiarliart wjp

Oh>

r/t

Pht'J: 7rtms. 1876. Pfate 31

^.-.^j.

XV

From »at on Blone ly J-ErxJefee?! ,

M&lT^nbar^ imp

.. 187(1. m,r,- 2I\.