LATE PLEISTOCENE MAMMALS from the "KE1LOR CRANIUM SITE", SOUTHERN VICTORIA, AUSTRALIA by Larry G
https://doi.org/10.24199/j.mmv.1974.35.02 18 February 1974
LATE PLEISTOCENE MAMMALS FROM THE "KE1LOR CRANIUM SITE", SOUTHERN VICTORIA, AUSTRALIA By Larry G. Marshall*
(A Contribution from the University of California Museum of Paleontology) Department of Paleontology, University of California, Berkeley, California, U.S.A., 94720.
Abstract Two ate Pleistocene mammal faunas have _. „ i been recorded from the "Keilor Cranium U ictona Australia. y ' The older late Pleistocene Dry Greek Local Fauna from thek »rt° Clay"i" includes D Sarcophilus laniarius, Thylacinus cynocephalus, Perameles nasuta, Vombatus ursinus, Thylacoleo carnifex, Protemnodon anak, P. brehus, Macropus rufogriseus, M. agilis, M. titan, M. cf. ferragus, and Zygomaturus trilobus. Disconformably overlying the 'D Clay" is the slightly younger Doutta Galla Silt with a basal age of 18,000 yr B P The Manbyrnong Local Fauna of the Doutta Galla Silt is represented by Vombatus ursinus, Megaleia rufa, Macropus giganteus, Mastacomys fuscus, Pseudomys cf. australis and Pseudomys cf. gracilicaudatus. Late-Pleistocene dwarfing is demonstrated in four species in the Creek Dry Local Fauna and two species in the Maribyrnong Local Fauna, with one species common to both. Late Pleistocene extinctions and late-Pleistocene dwarfing were probably caused a factor. by common The term megafauna is defined to include (1) species now extinct, and (2) species which have undergone late-Pleistocene dwarfing.
Introduction 17, fig. 1 ). One mile N. of Keilor at confluence For nearly two decades fossil mammals of Dry Creek and Maribyrnong River, S. Vict., have been known from the Doutta Galla Silt Australia. Grid ref. 881495 on Sunbury Mili- in southern Victoria, Australia; these include tary Map. Rattus cf. assimilis, and species of kangaroos, Methodology wallabies, wombats and native cats similar to Linear tooth dimensions were measured those living today (Gill 1955a,b). A d 4 date with of 18,000 ± 500 yr B.P. (NZ-207), obtained a pair of vernier calipers to the nearest on a charcoal sample from the base of the 1 mm when possible. All measurements are in millimetres unless indicated Doutta Galla Silt, indicates that the fauna is otherwise. The following abbreviations of latest Pleistocene age (10,000 yr B.P. is are used: accepted here as the Pleistocene-Holocene a—approximate measurement boundary). AW—anterior width (protoloph and pro- Disconformably underlying the Doutta Galla tolophid width) distal Silt at the Keilor Cranium Site is an unnamed DB— breadth dark unit referred to by Gallus (1971) as the DD—distal depth "D Clay", from which are recorded species of DP—deciduous premolar Diprotodon and Thylacoleo (Gill 1967). L—length Over the past eight years Dr. A. Gallus and M—molar a team of workers from the Archaeological MW—maximum width Society of Victoria have made a large collec- P—premolar proximal tion of fossil mammals from the Keilor PB— breadth proximal Cranium Site, both from the Doutta Galla Silt PD— depth and "D Clay". These faunas are described PW—posterior width (metaloph and here as the Maribyrnong Local Fauna and Dry hypolophid width) Creek Local Fauna respectively. The specimens described here were de- posited in the fossil collections of the National Locality Museum of Victoria (NMV), Melbourne, "1940 Cranium Site" of Bowler (1970, p. Australia. The roman numerals following the
63 64 LARRY G. MARSHALL
specimen numbers in the tables refer to the resents the youngest, although this is only an level from where that specimen was collected. approximation and needs further clarification. A complete list of the identifiable specimens Until the stratigraphic and time relationships from each level of the is "D Clay" given in of these levels are worked out in detail it is Appendix 1. The specimens from the Doutta not possible to discuss differences in faunal Galla Silt are listed in Appendix 2. The composition within the different levels in any higher taxonomic categories are of Ride meaningful context. (1964).
1. "D Clay" (Dry Creek Local Fauna) The Fauna
Because there are no detailed studies of the The Dry Creek Local Fauna consists of 12 complex stratigraphy at the Kcilor Cranium species of marsupials representing three orders Site, I will restrict my comments to the general and seven families. Two of these families, relationships between the fossil bearing units. Diprotodontidae and Thylacoleonidae, are Three basic mapable units arc recognizable now extinct; the other five families, Thylac- at the Keilor Cranium Site: (1) the Arundel inidae, Dasyuridae, Peramelidae, Vombatidae, Formation, (2) an unnamed intermediate unit, and Macropodidae, are represented by extant and (3) the Doutta Galla Silt, from oldest to species. Table 1 lists the minimum number of youngest (formational names follow Gill individuals of each species necessary to 1962). All three of these units are separated account for all of the specimens recovered from by disconformities. The relationship of the each level (based on both dental and post- Arundel Formation and Doutta Galla Silt are cranial remains). discussed by Bowler (1970). The intermediate Macropods are the dominant group, com- unnamed unit has been referred to as the "D prising 76% of the total minimum number of Clay" by Gallus (1971), although it has not individuals and are represented by at least been formally named. The age of the "D Clay" five, and possibly as many as six, species has yet to be established, although it is cer- (Protemnodon anak, P. brehus, Marcopus tainly late Pleistocene in age and^ probably in rufogriseus, M. agilis, M. titan, M. cf. fer- the order of 25,000-40,000 yr B.P. (J. M. ragus). Macropus titan is the most abundant Bowler, pcrs. comm.). The material described species in most of the levels, followed by M. here as the Dry Creek Local Fauna was col- rufogriseus and M. agilis which appear in lected from the "D Clay". about equal numbers. The other species appear During initial excavation of the fossil mat- rather sporadically throughout and are not erials from the Doutta Galla Silt, Gallus abundant in any particular level. The ratio of organized his collections on the basis of their carnivores (Sarcophilus laniarius, Thylacinus relationship of one to the other (local concen- cynocephalus, Thylacoleo carnifex) to herbi- trations) and on superpositional relationships vores (all other species) is approximately where this was clearly defined. Differences in 1 : 10. the lithology of the sediment were also taken The species in the fauna can be placed into into account. For the most part the collection three basic groups: (1 ) species represented by is organized into specific collection sites (layers living forms indistinguishable from specimens or levels as sometimes used by Gallus) and in the fauna (Perameles nasuta, Vombatus the relationships of these sites to each other ursinus, Macropus rufogriseus); (2) the larger have not been firmly established. Tn the interest Pleistocene forms of living species (Sarcophilus of convenience and clarity T have given "level" laniarius, Thylacinus cynocephalus, Macropus numbers to each of Callus's collection sites agilis, M. titan), and (3) species now extinct (Appendix 1 ). These levels (I-XI) are roughly (Thylacoleo carnifex, Protemnodon organized anak, P. such that level I probably repre- brehus, Macropus cf. ferragus, and Zygo- sents the oldest and level XT probably rep- ma!urns trilobus). LATE PLEISTOCENE MAMMALS 63
TABLE 1
Under "levels" I-XI from the "D Clay" are _ listed the minimum number of individuals necessary to account for the specimens recovered from each level based on both dentitions and postcranial remains. The minimum number of individuals of each species in the total fauna based solely on dentitions is given in the right-hand column.
Species Level I II III IV V VI VII VIII IX X XI Total
Sarcophilus
laniarius 1 1 Thylacinus cynocephalus Perameles
nasuta 1 Vombatus ursinus 2 Thylacoleo
carnifex 1 1 Protemnodon
anak 1 2 Protemnodon
brehus 1
Protemnodon sp. 1 1 - Macropus
rufog ri seus 1 1 1 3 Macropus agilis 3 111
Macropus titan 13 2 2 3 3 1 3 13
Macropus cf. ferragus macropodid 2 12 2 1- 1 1 Zygomaturus trilobus diprotodontid Total 10 10 3 8 33 66 LARRY G. MARSHALL
Dry Creek Local Fauna Palaeoecology
(lass Mammalia The probable habitat preferences of the Infraclass Mctathcria species in the Dry Creek Local Fauna (Table Superordcr Marsupialia 2) are based on those of extant populations. Order Marsupicarnivora It is assumed that the larger late Pleistocene Family Dasyuridac forms (Sarcophilus laniarius, Macropus titan) Sarcophilus laniarius had habitat requirements similar to their living Family Thylacinidae descendants (S. harrisii and M. giganteus re- 1 hylacinus cynocephalus spectively). These data show that the region Order Peramclina in the Dry Creek area in late Pleistocene time Family Peramelidae was most probably covered by sclerophyll Perameles nasuta forest. It is possible that mesophytes lined the Order Diprotodonta river valleys which dissected the open rolling Family Vombatidae bushlands and grasslands much as occurs in Vombatus ursinus the area today. Family Thylacoleonidae Thylacoleo carnifex Faunal Correlation Family Macropodidae Lake Colongulac (= Lake Timboon) Protemnodon anak The Dry Creek Local Fauna compares well I'rotemnodon brehus with the late Pleistocene fauna from Lake Mac ropus rufogriseus Colongulac, N. of Camperdown, S. Victoria. Macropus agilis The following species are represented in the Mac ropus titan Lake Colongulac Local Fauna: Macropus el', ferragus Class Mammalia Family Diprotodontidae Superorder Marsupialia Zygomaturus trilobus Order Marsupicarnivora
TABLE 2
Basic habitat preferences of species in the Dry Creek Local Fauna based on living populations of these species (x indicates preferred habitat),
Rain- Species SclerophyllJ , • , , Woodland riorest forest Plains
Sarcophilu s laniarius (as S. ha rrj sii)
' T hylac inus cynocephalus X X
Perameles nasuta X X
Vombatus ursinus X
Macropus titan X X X
(as M. giganteus )
Macropus ag i I i s X X X
Macropus rufogriseus X X
t from Mariow (1958) * from. Ride (1970) LATE PLEISTOCENE MAMMALS 67
Family Dasyuridae cooperi) and dominated by the large browsing Sarcophilus lam at ins (P30218) macropod Procoptodon goliah. Family Thylacinidae The Lake Victoria Local Fauna (which is Thylacinus rostralis (see DeVis typical of the Murray River Basin faunas) is 1899) considerably different from that of the Dry Order Diprotodonta Creek Local Fauna. Five of the twelve species Family Vombatidae in the Dry Creek Local Fauna (Perameles Lasiorhinus sp.? nasuta, Vombatus ursinus, Macropus agilis, Vombatus ursinus (P30785) M. rufogriseus, and Zygomaturus trilobus) are Family Thylacoleonidae absent from the Lake Victoria Local Fauna Thylaco'leo carnijex (P24000) (Type whereas Perameles gunnii, Lasiorhinus Krefftii, locality) Lagorehestes leporides, Onychogalea fraenata Family Macropodidae and Diprotodon optatum occur in the Lake Sthenurus sp. (P29488) Victoria Local Fauna and not in the Dry Creek Procoplodon rapha (P26901) Local Fauna. These faunal differences are Thylogale billardierii (P23996) probably the result of the presence of an open Protemnodon anak (P30207) woodland-savannah-grassland in the Murray Protemnodon brehus (P30214) River Basin during late Pleistocene time. Macropus rufogriseus (P30217) The Collection Macropus agilis (M. sival) (P30215) The major part of the fossil materials Macropus titan (P28549) were Family Diprotodontidae collected in situ as isolated specimens. A few Diprotodon optatum (PI 5902) associated left and right rami were obtained, although these rinds are certainly exceptions. A large collection of fossil mammals from Most of the postcranial material was broken Lake Colongulac and the surrounding lake except for podial and metapodial bones. Be- region (generally referred to as the Camper- cause of the dearth of associations it is difficult down District) is in the National Museum of to assign postcranial elements to respective Victoria; but unfortunately this important dentitions with complete certainty. In the case fauna has not been adequately described. The of Protemnodon, all podial elements are diversity of this fauna is considerably greater referred to Protomnodon sp. as there were no than previously recognized as is seen in com- feet of P. anak or P. brehus available for com- paring the faunal list given by Gill (1953b, parison. All of the postcranial elements of a p. 35) with the revised and expanded list given large species of Macropus are referred to M. above. The fauna from Lake Colongulac and titan. Although associated rami (P30716) are that from Dry Creek are similar in age and referred to M. cf. ferragus, the presence of this share a close geographical proximity; differ- species in the fauna cannot be established with ences between these faunas are probably due complete certainty (sec below). The post- to sampling. cranial remains of smaller species of Macropus are potentially referable to M. agilis or M. Murray River Basin rufogriseus and some may even be referable Late Pleistocene local faunas from the to females? of M. titan; these arc all listed col- Lower Murray River Basin have been recov- lectively under the heading macropodid. The ered from Lake Menindee (Tedford 1967), postcranial remains of a diprotodontid arc Lake Tandou (Merrilees 1973), and Lake probably referable to Zygomaturus trilobus. Victoria (Marshall 1973). These faunas are Except for basing the identification of Thyla- typified by an abundance of grazers (Macropus cinus cynocephalus on a humerus, the identi- ferragus, M. titan) and browser-grazers fications of all other species in the fauna are (Sthenurus tindalei, S. andersoni, S. atlas, based on teeth. Nearly one third of the total Protemnodon brehus, P. anak, Osphranter bone sample consists of postcranials which are 68 LARRY G. MARSHALL
too fragmented to permit reasonably accurate No morphological differences were observed identification. These specimens are not dis- in the dentitions of the S. laniarius or S. harrisii cussed in this study although their presence is samples studied, and except for the superior recorded for completion. size of the former (approximately 15% based on tooth Systematics Review measurements) separation of the dentitions of these species was difficult. Class Mammalia S. laniarius is typically found in deposits of Infraclass Metatheria late Pleistocene age in direct association with Superorder Marsupialia extinct megafaunal species of Procoptodon, Order Marsupicarnivora Protemnodon, Sthenurus, and Zygomaturus, to Family Dasyuridae name just a few. S. harrisii, on the other hand, Sarcophilus laniarius (Owen, 1838) occurs in Specimens slightly younger deposits always in of Sarcophilus laniarius were col- association with a typically modern fauna lected from levels IV, V, (for and VI. As there is example no see Thorne 1972). Although the duplication in the elements represented association of specimens of they may in S. harrisii with theory be attributed to a single extinct megafaunal species are reported (Owen individual (Table 3). the 1877) contemporaneity of these species is The relationship of 2 2 L M and MW M of highly questionable (Frank 1971). The direct P29587 from Dry Creek is compared in Fig. 1 association and temporal overlap of these with an extant sample of Sarcophilus harrisii species have not been from Tasmania and substantiated. This evi- a sample of S. laniarius dence from suggests that S. laniarius represents a late Pleistocene deposits in Strathdownie larger, ancestral, late Pleistocene form Cave, Victoria (Merrilees of S. 1965). The Dry harrisii; a relationship first recognized by Creek specimen falls well within the range of Lydekker (1887). the S. laniarius sample.
12n TABLE 3 Dimensions & laniarius / of the cheek teeth of Sarcophilus laniarius
Specimen M M M M L, MW L L MW P29587(VI) - - 12. 9V 14.6 11.3 13.7 9.2 4.7 P2U631 - m (V) 6.8 6.0 10.0 - - _ _
S. harrisii Family Thylacinidae MW M 2 Thylacinus cynocephalus (Harris, 1808) A single left 10-1 • • •• humerus (P29588) collected from level VI is the only element referable • Tasmania to this species. A Strathdownie Cave Ride (1964) reviewed the status of the five proposed oDry Creek species of Thylacinus (T. major Owen 1877, T. breviceps Krefft 1871, T. —T— -1- spelaeus Owen 1845, T. cynocephalus 12 13 14 15 Harris LM- 1808, and T. rostralis DeVis 1894). Ride's study Fig. 1—Scatter of dental and cranial characters diagram showing relationship of of a L 1 and 2 living sample M MW M of a living sample of of Thylacinus from Tasmania, .Sarcophilus harrisii from Tasmania, a fossil and fossil samples from sample cave deposits in W.' of S. laniarius from late Pleistocene and E. Australia "do not deposits in Victoria, and a specimen (P29587) support separation referable to [of these S. laniarius from the Dry Creek samples] even at a subspecific level Local Fauna. at the present time". Ride recognized that these LATE PLEISTOCENE MAMMALS 69
samples represent a single variable species, Order Peramelina T. cynocephalus, which includes T. breviceps Family Peramclidae and T. spelaeus as junior synonyms (see foot- note of Ride 1964, p. 105 concerning the Perameles nasuta Geoffroy, 1 804 identity of T. major). Ride further noted, A right ramus fragment with Pi-M 3 and an "although recognition by is name not justified associated fragment of a left maxillary with there is no doubt that in the 3 Pleistocene of M-" (P29634) of this species were found in W. Australia there existed a population of level HI (Table 4). Thylacinus cynocephalus which on an average The upper molars of species of Isoodon contained smaller individuals than the modern possess a well developed hypocone which is form (and by inference its eastern Pleistocene subequal in size to the protocone. In species representative)". T. rostralis from the late of Perameles a hypocone is absent or only Pleistocene fluviatile deposits of the E. Darling incipiently developed. The Dry Creek specimen Downs, SE. Queensland was shown to be has only an incipiently developed hypocone larger than the specimens referred to T. cyno- and is readily referred to a species of Per- cephalus by Ride. Ride recommended that ameles. The dentitions of Macrotis, Chaeropus T. rostralis retain its specific identity, at least and Echimiptera are sufficiently distinct from for the time being. Perameles and Isoodon to dispense with a de- The Dry Creek specimen agrees well in tailed comparison with the Dry Creek morphology with the living specimens of T. specimen. cynocephalus from Tasmania (NMV C5742, The L M3 and AW M 3 of living samples of C5746, C5753) with which it was compared, Perameles bougainville (including P. fasciata, although the Tasmanian specimens are slightly P. myosura, P. arenaria; see Tate 1948, pp. smaller. No specimens of postcranial material 324-325), P. gunnii, and P. nasuta are com- referable to T. rostralis were available for pared with the Dry Creek specimen in Fig. 2. study and it may be that the Dry Creek speci- Values of the dental parameters of P. bougain- men is referable to this species. DeVis (1899) ville are significantly smaller than those of the reported T. rostralis from Lake Colongulac in other two species. Freedman (1967) reported S. Victoria. I was unable to relocate the speci- that the teeth of P. gunnii and P. nasuta "are men (s) upon which this identification was of approximately similar size" although those based. of P. gunnii were shown to be slightly smaller. The Dry Creek specimen agrees well in size He noted that the dentitions of these species and morphology with a nearly complete could be separated by differences in the mor- skeleton of T. cynocephalus (P26573) from phology of the upper incisors and canine. late Pleistocene deposits at Lake Victoria, Unfortunately, these teeth are not represented N.S.W. The cranial dimensions of P26573 fall in the collection from Dry Creek. I was unable well within the range of males of T. cyno- to find morphological differences in the molar cephalus studied by Ride (1964). teeth of these species which would allow them
TABLE 4
Dimensions of the cheek teeth of Perameles nasuta PI P2 Ml M2 Specimen M3 L MW L MW L AW/MW PW L AW/MWPW L AW/MW PW Upper P29634 (III) ----- 4.0 4.0 3.8 3.9 Lower P29634 (III) 2.9 0.9 3.3 1.3 3.9 2.2 2.6 4.1 2.6 3.0 4.2 2.5 2.7 —
70 LARRY G. MARSHALL
to be distinguished. As seen in Fig. 2, P. gunnii At the present time one late Pleistocene is slightly smaller than P. wowtfa. The Dry species, T. carnifex, is recognized and is (reek specimen lulls within the range of the virtually pan-Australian in distribution. Mcr- /'. nasuta sample and outside of the P, gunnii rilces (1 968) reports Thylacoleo sp. not T. sample. The Dry Creek specimen is tentatively carnifex from Mammoth Cave in SW. W. Aust. referred to P. nasuta based on this (admittedly This specimen (s) has not been described and minor) size difference. I do not know how it differs from T. carnifex.
Mylodon australis Krefft 1 870, Thylacoleo 2.5-i P nasuta owenl McCoy 1 876, and Thylacopardus australis Owen 1 888 are presently recognized as junior synonyms of T. carnifex (Anderson 2.3- '*--£ 1 929). A concise description of the cranium AW - and dentition of this species is given by Mo Woods (1956) and the species distribution is 2H P. bougainville outlined by Gill (1954). • * Living sample
1.9- x Dry Creek
' 3-2 Z.A ' ' ' ' 16 3.'a 4.0 42 44 LM3
lit,'. 2—Scalier diagram showing relationship of L M-, and AW M, for living samples of Perameles bougainville, P. gunnii, and P. nasuta, and a fossil specimen of Perameles from the Dry (reek Local Fauna.
Order Diprotodonta Family Vombatidae Vombatus ursinus (Shaw, L800) A minimum of two individuals of Vombatus ursinus are represented by two rami (P23026, P29519) and a fragment of a left ilium (P29548), collected from levels III and IV respectively. Remains of this species are com- monly found in late Pleistocene deposits throughoul Victoria, E. New South Wales and S. Queensland. The molars of V. ursinus typically have W- Shaped lobes Fig, J and sharp inlerlobe valleys as Protemnodon anak, right 1VL, NMV P29SS5 level V; a. lingual; opposed to the more U-shaped or rounded b, occlusal; and c, labial views; all x 11. lobes ami more open interlobe valleys of Lasiorhinus (Merrilees 1967, p. 407). Family Macropodidac Protemnodon Family Thylaeoleonidae anak Owen, 1873 Dentitions of Protemnodon Thylacoleo carnifex Owen, 1858 anak were found in levels V (P29554, P29555) and XI A single fragment from level XI of a right (P29604), representing a maxillary minimum of two (P29545) with the lower edgc^of individuals (Table 5). Postcranial remains the orbit and the anterior root of 1" is all that from levels III, VI, IX and is X may be referable known of I his species. to this species (Table 9). These specimens 1 8
LATE PLEISTOCENE MAMMALS 7!
agree in all respects with the holotype (BMNH mum of three individuals (Table 6). Some of figured Ml 895) by Stirton (1963, p. 136 the postcranials from levels II-VI, V1II-XI 13a,b) (Fig. fig. 3). listed under macropodid may be referable to TABLE 5 this species. In Fig. 4 the Dry Creek specimens are Dimensions of the check teeth of compared with specimens of M. rujogriseus Protemnodon anak and P. brehus from an extant population. There is complete overlap in I, PW L AW I'W I. AW PW the range of the two samples. M. P. brehus rujogriseus is found as living species the Upper a in
. . 111. P29522 (111) 13. 2t 1 2 .11 Dry Creek area; the living linns are indistin-
I .OWIT guishable in size and tooth morphology from P29628 (IX) - 16.5 11.5 11.5 17.7 - the specimens in the Dry Creek Local Fauna. P. anak Lower - P29604 (XI) 13.7 - - 15. G a. __ A (left) M. rufoariseus / ',
P29604 (XI) 15.D 5.5 lli.Sa - - 15.7 10.0 ,'%"' ... A (right) 5- xjj j \M. aj MIS
Protemnodon brehus (Owen, 1874) AW •• •
i . / \* Dentitions of this species were found in levels III (P29522) and IX (P29628). An !• */ Living sample isolated right lower incisor (P29586) from 4- X A Dry Creek level VI is probably referable to this species. \» x/ These specimens represent the first record of
1 1 1 P. brehus in Victoria. P29522 agrees in all respects with the holo- L Mf type (BMNH 43303a) figured by Stirton Fig. 4—Scatter diagram showing relationship of (1963, p. 140, fig. 15b). The lower molars L Mi and AW M, of a living sample of of P. brehus, as represented by P29628, differ Macropus agilis (squares) and M. rujogriseus (circle) and specimens referable to M. agilis in from those of P. anak (P29601) being (triangles) and M. rujogriseus (x) from the larger and relatively broader (Table 5). Both Dry Creek local Fauna. species cingula the have weakly developed on Macropus agilis (Gould, 1842) lower molars. Macropus agilis is the second most abundant Macripus rujogriseus (Desmarcst, 1817) species in the fauna and is represented by a Dentitions of this species were found in minimum of live individuals (Table 7, Figs. 5, levels IV-VII and XI, and represent a mini- 6). M. agilis, like M. rujogriseus, is repre-
TABLE 6 Dimensions of the cheek teeth of Macropus rujogriseus
DP3 P3 Ml M2 JV13 JV14 Specimen PW L AW PW L AW I'W L AW I'W I'W L AW Upper
8 . 7 P29578 (V) 6.96.9 3 9.0 . 0a Lower - P29577 (IV) - 4.3 7.15.0 7 . !)
P29574 (XI) - 4. 1 6 .7 5.0 5.4 7 . G 5.4 5.4 8.6
P30905 (VII) 4.2 - 7.0 4.3 4.7 7 . (i 5.3 5.3 P30725 (IV) G. 8 4. 72 LARRY G. MARSHALL
sented by dental remains in levels IV-V1I, although is not present in level XI. Some of a the postcranials from levels II-VI, VIII-XI may well prove to be referable to this species. DeVis (1895) erected M. siva on the basis of a partial ramus collected from the late Pleistocene fiuviatile deposits of the E. Darling Downs in SE. Queensland (see Bartholomai 1966, pp. 118-119). Specimens referable to b L|1^ M. siva have subsequently been recorded from WJi|\VlBlLli
' e est at>'isned; I would . estimate it 1 V by BartholomaidU"ulomai U^/i,MQ71 ™ 1 •* v^ *u- to be * p. 1 ). For this reason n the order nf miw a a -i a t ^
TABLE 7 Dimensions of the cheek teeth of Macropus agilis
2 pP3 Specimen — P P3 Mi 575 L Mw L aw '7 w= ™ """ ' '^^^ P295,! ^^^^^^^^" "V" ~ «•».. ...,.,,., . LowerI " - P29549 (IV) - ... _ ". 9-7 P29550 (IV) _ I " " - 6.2 10.5 6.7 6 7 (left) " " " - - 6.6 10.6 6.8a 6J 11.3 6~8 bl6~l P29550 (IV) _ fc . „ . 7.82.53.4g R - (right) - 5 . 6 -6.06.4 - 6.7 - P29575 (V)
9 ' P29576 (V) 6.6 3.5 4~7 ~- ' 8UM " 7.8 4.2 I *A\'\\-\ - 0.8 5.4 - P29603 (VII) ... 5.1 . . 10.6 6.7 6.4 — —
LATE PLEISTOCENE MAMMALS 73
Fig. 7 Macropus titan, left M3-4 of NMV P29524, level III; a, labial; b, occlusal; and c, lingual views; all x li.
samples. In linear dimensions of M3 , M. titan is intermediate in size between the smaller living species M. giganteus and the larger 3 Fig. 6 Macropus agilis, right P , NMV P29579, extinct late Pleistocene species M. ferragus. level VI; a, labial; b, occlusal; and c, lingual Of the four views; all x 3. complete metatarsal IVs listed in Table 9, two (P29534, P29535) are similar southern Victoria in late Pleistocene time. in size and are significantly smaller than the The absence of this species in Victoria today other two (P29556, P29589). I attribute these presents a problem because other macropods differences to sexual dimorphism with the in the Dry Creek Local Fauna which are smaller specimens representing females and represented by living forms (M. rufogriseus the larger the males (Fig. 8). and M. titan) axe, the dominant macropods in Macropus jaunus DeVis 1895 and Mac- the Dry Creek area today. The factor (s) ropus magister DeVis 1895 are recognized as responsible for the post-Pleistocene change in junior synonyms of M. titan (Tedford 1967). distribution of M. agilis is presently unknown. The specimen figured by Tedford (1967, Fig. 27a) and referred to Macropus birdselli is also Macropus titan Owen, 1838 referable to M. titan. Macropus titan, the most abundant species M. titan is the most widely distributed of the in the fauna, is represented by a minimum of late Pleistocene macropod species and occurs 13 individuals, accounting for 40% of the in almost every late Pleistocene deposit studied. individuals in the fauna (Table 8). Dentitions It is typically the most abundant species in late of this species were recovered at all levels Pleistocene deposits in Vict. (Lake Colon- except level I (Fig. 7). gulac), E. N.S.W. (Wellington Caves), and In Fig. 9 Dry Creek specimens are com- SE. Qd. (E. Darling Downs). At Lake Vic- pared with a sample of M. titan from Lake toria in SW. N.S.W., however, M. titan is less Colongulac, Vict., showing that there is nearly abundant than the larger grazing macropod complete overlap in the range of the two M. ferragus (Marshall 1973). M. titan was I 1 i 1 I l
74 LARRY G. MARSHALL
r- r- in
I I ' r-i • I o ' o o
CO CO <* r- o CO co LD CO CO O 03 03 1 < o i—i o 1 1 1 1—1 1 T— I 1 T— i— 05 T-i CJ3
CO o ^ o CO o CM a i— OS c- co w ^ cm 1 »' C- ' O l> co C3 J t- c- C- CO CO C- ' C- CO CO
' ' ' I o o . . o ' *-, r-, ' ' ' —i —i en a h ct>
M1 CM CD CO 11 1
J CO t- r- ' ' co ' in co r-
65
' ' ' ' * ' I I I b O O . I r-l ,-1 oi
cm o
I C- ' ' ' I I I I I O 2 cm" I I I I .1.11 I I CJJ CO 03 m co o CD C- O CO CO <* 3 I I I •s A I I I I I o CO CO CO
c 1 1 .If I I I I I < ' . . I I a>
£ CM CO (MO O
' I I * CO 0. .111111 I I I f 1 I . I CO ' I I I I I I I I I I ^ I I . . I CO CO > X > > X l> CO CO t-i co CD *mcocoHNcoOH CO CO O t-1 CM ^h Nco^ONcandirt a m in co co co •»-i Cl 05 CT3 Q O) o 1 CM CM CM CM CM 0) A 0. ft A A a g ^AAAAAAAAA ft AAAAAAAAa co LATE PLEISTOCENE MAMMALS 75 lcm Fig. Macropus titan, 8— a, left metatarsal IV (male?), NMV P29556, level V, dorsal view; b, left metatarsal IV (female?), NMV P29534, level IV, dorsal view. probably a woodland-savannah species and 13- occupied a niche which was probably identical 1 * -' to that of M. giganteus today (see below). 12- - M. titan appears to represent a 30% larger " P30716 late M. titan Pleistocene form of M. giganteus. The 11- only consistent difference between these species PW«j S ".-"^ is in oo o' the superior size of the former. M. titan 1CH , on D ," occurs in deposits of late Pleistocene age and l DD o vLaka Victoria • .-v • a Dry Crttk older, typically in association with extinct 9- o Lake Colongulac megafaunal species of Thylacoleo, Diprotodon, M. oiaintsu* 8- *Ory Craak Sthenurus, Procoptodon, and Protemnodon, to Victoria name just a few. M. giganteus is found in 1** **,-' 7- «**£ . _ late 16 17 younger Pleistocene deposits ( < 20,000 yr B.P.) and never occurs in direct association 9 with extinct megafaunal species or with M. Fig. —Scatter diagram showing relationship of L M3 and PW M3 of a living sample of titan. There is also no evidence of temporal Macropus giganteus from Victoria, a fossil overlap in the occurrence of these species. It sample of M. titan from Lake Colongulac, Vict., a fossil sample of M. jerragus thus appears most logical to regard M. titan as from Lake Victoria, N.S.W., and specimens refer- the larger late Pleistocene ancestor of M. able to M. titan and M. cf. jerragus from giganteus. This lineage represents the fourth Dry Creek. example of late-Pleistocene dwarfing in the cranials are assigned with relative certainty to fauna (see Discussion). M. titan. A large number of postcranials belonging to a large species of Macropus are represented Macropus cf. jerragus Owen, 1 874 in the Dry Creek collection. Except for the An associated left and right ramus (P30716) single questionable specimen referred to M. cf. from level III may be referable to this species jerragus from level III (P30716) (see below) (Table 10). The Dry Creek specimen is com- there is no other evidence that this species is pared in Fig. 9 with a sample of M. jerragus present in the fauna. For this reason the post- from Lake Victoria, N.S.W., M. titan from p 76 LARRY G. MARSHALL TABLE 9 Dimensions of macropodid metatarsals and phalanges Specimen PB PD DB DD Macropus titan Metatarsal IV P29533 (XI) - 35.0 32.5 P29534 (IV) 177.0 29.8 26.3 29.5 21.8 P29535(IV) 178.0 30.6 27.5 31.0 23.2 P29556 (V) 198.0 40.4 32.4 39.2 26.8 P29558 (IV) - 31.3 28.5 P29589 (VI) 207.0 39.0 28.0 39.6 28.3 P29591 (VI) - 31.0a 26.0 P29633 (X) - 33.0 27.8 — _ Proximal Phalanx Digit IV P29518 (III) 29.7 29.5 25.2 26.5 17.0 P29560 (V) - - 20.1 13.1 Medial Phalanx Di git IV P29520 (HI) 42.4 26.4 20.4 21.5al3.0a P29612 (VIII) 41.4 27.6 19.3 13.6 Metatarsal V P29557 (V) 156.0 14.8 18.4 18.6 18.7 P29605 (VII) 187.0 21.4 29.5 22.5 23.7 P29633 (X) - 17.6 20.9 Macropodid Proximal Phalanx Digit lv P29539 (IV) 38.3 18.4 13.7 14.6 10.3 Metatarsal IV P29618 (VIII) 80.0 13.6 12.0 13.1 10.0 P30722 (IV) 132.4 22.5 19.2 20.8 14.4 Metatarsal V P29544 (IV) 117.5 11.3 14.0 14.6 13.2 P29569 (V) 13.4 13.0 P29618 (VIII) 71.0 7.6 9.3 9.7 11.0 Protemnodon sp Metatarsal IV P29527 (HI) 136.5 35.8 30.8 40.2 P29625 24.0 (IX) 118.0 31.2 P29632 (X) 37.0 32.5 1 r1 il i ' I1(l l1I LATE PLEISTOCENE MAMMALS 77 Lake Colongulac, Vict., and an extant sample £ CO of M. giganteus from Victoria. In size of M3 , Ph i CM i P30716 is closest to the M. ferragus sample^ CO falls well outside the range <* of the M. titan CD sample and is i considerably larger than i M. < [> giganteus. In addition to size, P30716 agrees co best with the M. ferragus specimens in tooth CD morphology. As in M. ferragus there is i a large r- i pit on the posterior face of the hypolophid. tO 1 E. Darling Downs collections whereas M. fer- T— ragus is uncommon. Remains of M. ferragus Oi are abundant at Lake Menindee (Tedford 05 1967) and Lake Victoria (Marshall 1973) in t—1 i— SW. N.S.W. At Lake Victoria M. ferragus is ^_^ i c ^ s — l— (— l— more abundant than M. titan. M. ferragus was is 1 and the largest grazing species of macropod V tO 0) to -r to known. 0> OS — OT ^ OJ a CM CM CM Vi Ph Ph Ph 78 LARRY G. MARSHALL Family Diprotodontidae from the Doutta Galla Silt from the Keilor Zygomaturus trilobus Maclcay, 1857 Cranium Site, Braybrook, and Green Gully is collectively A well preserved mandible (P29514) found described here as the Maribyrnong in level I and an Local Fauna (see Appendix isolated M2 (P29515) from 2). level II, represent a minimum of two indivi- duals Age (Table 11). Postcranials of a diproto- dontid Bowler were found in levels VIII and XI (1970, p. 43) noted that "deposi- (Table tion of the 12). There is no duplication among Keilor Terrace silts [began] at these elements approximately suggesting that they may all 18,000 yr B.P. This situation is have come from a single individual. similar to that reported from the excavations The podial elements of A. Gallus at the cranium . were compared with site . . The inde- those of Diprotodon pendent optatum figured by Stir- radiocarbon dates from other sites ling and Zietz confirm the (1899), and it is evident that validity of the chronological they arc not sequence referable to that species. They are outlined above". An 18,000 yr B.P. probably referable date is accepted to Zygomaturus trilobus, here as the basal age of the the only Doutta diprotodontid so far known with Galla Silt. certainty from the fauna. Gill's (1967) record of Diprotodon The Fauna remains in this fauna was based Six species on these specimens. of mammals are represented in the fauna: Vombatus ursinus, Macropus giganteus, TABLE 12 Megaleia rufa, Mastacomys fuscus, Pseudomys cf. gracilicaudatus, Dimensions of metatarsals of and Pseudomys cf. australis, all represented by living Zygomaturus trilobus popu- lations. The fauna is modern in all respects. Specimen L PB PD DB DP Systematics P29619 Review (VIII) 90. 37.0 41.0 3U.5 32.9 P29620 Family Vombatidae (VIII) 90.5 28 3 43.2 30.8 33.6 Vombatus ursinus (Shaw, 1800) 2. Doutta Galla Vombatus ursinus Silt (Maribyrnong Local is represented by a single mandible Fauna) (P30724). This species has been discussed above under the Dry Creek Local Geology Fauna. Disconformably overlying the "D Clay" containing the Family Dry Creek Local Fauna at the Macropodidae Keilor Cranium Site is the Doutta Galla Silt Megaleia rufa (Desmarest, 1822) The Doutta Galla Silt consists of "basal This species is represented by a gravels three to four feet minimum of thick (basalt, sand- four individuals stone, and is the most abundant quartz, and mudstone) passed up species in the fauna (Table through two to three 13). The L M, feet of medium to fine and PW Ms of the Dry Creek quartz sands to specimens are 20 feet of yellowish-brown to compared in Fig. 1 1 with dark grey, very an extant sample of fine sands and silts which form Megaleia rufa, Osphranter robustus, and the main body of the terrace" Mac- (Bowler 1970 ropus giganteus. In size, the Dry Creek speci- P- 19). "Gill (1953a, 1957) has equated the men falls well within the range sediments of the of the M rufa Braybrook terrace with those sample. of the Keilor Terrace ... The sediments of On lower molars of M. these terraces have been rufa the posterior formally defined as lace of the hypolophid is plain and the the Doutta Galla Silt, for ento- which the type comd is set more posteriad from the locality is located at the metaeonid Dry Creek section than the hypoconid is from the (Gill 1962)" (Bowler ibid, protoconid p. 18). The fauna (Lig. In 10). O. robustus a faint diagonal LATE PLEISTOCENE MAMMALS 79 groove is present on the posterior face of the suggests that larger forms of M. rufa lived in hypolophid, whereas M. giganteus typically late Pleistocene time. This species apparently has a well developed (although usually small) underwent a late Pleistocene dimunition in vertical groove on the hypolophid which body size similar to that occurring in the Mac- appears as a well defined pit on the occlusal ropus titan-M. giganteus lineage. Average surface in worn teeth. In both O. robustus and individuals in living populations of M. rufa are M. giganteus the hypolophid and protolophid about 25-30% smaller than individuals in the are subparallel (Tedford 1967, pp. 113-114). late Pleistocene, i.e. >20,000 yr B.P. As a living species M. rufa has been re- corded in Victoria from Benetook and Ned's Corner in the NW. corner of the State (Wake- field 1966, p. 632). 10 4P30216 ' ~*\ Msgilail 9- * * A, Laka £ PWMj /A A * Waeranganuch / A ' 8- Osphrant.r ,-'.';>.*- o Dry Craak robustus,-' »*"^^»? * Living sample '' Macropus 7- giganteus • Living umpl* .- - Living sampl* i 11 12 13 14 15 L Mj Fig. 11 —Scatter diagram showing relationship of L M3 and PW M 3 of living samples of Megaleia rufa, Osphranter robustus, and Macropus giganteus, a fossil sample of Megaleia rufa from the Doutta Galla Silt, and a fossil specimen of M. rufa from Lake Weeranganuck, W. Victoria. Fig. - 10—Megaleia rufa, left Ms 4 , NMV P30750a. Macropus giganteus (Shaw, 1790) Doutta Galla Silt; a, labial, b, occlusal, and Two specimens represent a minimum of one c, lingual views; all x li. individual of this species (Table 13). The Fossil remains of M. rufa have been recorded relationship of M. giganteus to M. titan was in Victoria from the outlet of Lake Gnapurt discussed under the Dry Creek Local Fauna. to Lake Corangamite. A Cu date of 4,550 The Keilor specimens of M. giganteus are in- ±120 (GaK-2518) yr B.P. was obtained from distinguishable in size and tooth morphology associated Coxiella shells (Gill 1971). There from extant specimens. is also a specimen of M. rufa (P30216) from As a living species M. giganteus is abundant a bone bed on the E. shore of Lake Werran- in sclerophyll forest, common in woodland ganuck. Gill (1971, p. 76) reports that a C14 and rare in plains habitat (Marlow 1958). date of 25,300± 1,200 (GaK-986) yr B.P. That the species is presently found in such was obtained on Coxiella shells from this bone diverse habitats reduces its usefulness as an bed making the M. rufa specimen (P30216) indicator of paleohabitats. late Pleistocene in age and roughly equating Order Rodentia this Clay". P30216 is deposit with the "D Family Muridae compared in Fig. 1 1 with an extant sample of Mastocomys fuscus Thomas, 1882 M. rufa and specimens referable to M. rufa from the Maribyrnong Local Fauna. The This species is represented by a single spe- larger size of the Lake Werranganuck specimen cimen (P30906) consisting of the greater part . 80 LARRY G. MARSHALL of a skull and associated fragment of the left in P30906. P30906 is the specimen recorded ramus with M,- (Table 2 14). by Gill (1955a,b) as Rattus cf. assimilis. Approximately 4% of the specimens of At lower altitudes on the Australian main- M. fuscus studied by Wakefield (1972) had land this species lives in "wet sclerophyll forest supernumerary cusps on the lingual side of the with dense undergrowth containing ferns, upper molars. There is no trace of these cusps shrubs, and grasses" (Ride 1970, p. 152). TABLE 13 Dimensions of the lower cheek teeth of Macropus giganteus and Megaleia rufa from the Doutta Galla Silt DP o M M ° 1 Q M Specimen -1 o 4 L AW PW L AW PW L PW L AW PW Macropus giganteus P30750b 15.8 9.7 8.6 Megaleia rufa P30750a - _ 13.6 9.0 P30730 _ 14.5 9.0 P30751 9-0 4.7 5.7 11.2 6.5 6.9 TABLE 14 Cheek teeth dimensions of the rodents from the Doutta Galla Silt Specimen Mi M2 M3 Ml-3 L MW L MW L MW L Upper cheek teeth Mastacomys fuscus P30906 (left) 3.8 3.2 2.6 3.1 2.9 2.4 9.9 P30906 (right) 3.8 3.3 2.6 3.1 2. 9 2.4 9.8 Pseudomys cf. australis P15773 (left) 2.7 2.0 1.8 1.8 1.4 1.3 5.9 P15773 (right) 2.7 2.0 1.7 1.8 1.4 1.4 5.9 Lower cheek teeth Mastacomys fuscus P30906 (left) 3.9 2.8 2.6 2.7 - - - Pseudomys cf gracilicaudata P30907 2.3 1.7 1.8 1.7 13 12 5 P30908 5 2.5 1-7 19 1 ? ^ . . Pseudomys cf. australis P15773(left) 2.8 1.7 1.7 1.8 1.21.45.9 LATE PLEISTOCENE MAMMALS 81 That this animal was considered to be extinct associated with the now extinct megafaunal only four decades ago (Wood-Jones 1923-25) species underwent a reduction in body size indicates that there is still much to be learned giving rise to smaller living forms. Some of about its biology before palaeontologists can these smaller living forms are presently recog- use it in palaeoecological considerations. nized as distinct species (i.e. Macropus gigan~ teus represents Pseudomys cf. australis Gray, 1832 a smaller living form of M. titan; Sarcophilus harrisii represents a smaller and living form of S. laniarius). Pseudomys cf. gracilicaudatus (Gould, 1845) It is a well established fact that many living The genus Pseudomys is represented by species of both carnivorous and herbivorous three specimens, representing two species mammals are represented in late Pleistocene (Table 14). These have been identified by Mr. deposits by populations which are on the whole J. A. Mahoney as Pseudomys cf. gracilicau- larger in absolute body size. This reduction in datus (P30907, P30908) and P. cf. australis body size has been discussed by various (P15773). workers under the heading of "post-Pleistocene dwarfing" and has been shown to occur on a "Native cats" world wide scale (Hooijer 1950, Kurten 1959, Gill (1955a,b) reported the presence of Wen-Chung 1963). Hooijer (1950) has re- "native cats" in the Doutta Galla Silt, but I ferred to this dwarfing as "a sort of general have not been able to relocate the specimens evolutionary trend that is going on in the upon which this was based. "Native cat" is the Quaternary". Kurten (1959, 1968 and refer- colloquial name generally applied to one of ences) has established the present of post- the four species of Dasyurus (D. hallucatus, Pleistocene dwarfing in many European and D. geoffroyi, D. maculatus, D. viverrinus) Asian mammals with many of the lineages which live today on the Australian mainland showing a 25-30% reduction in body size. As and Tasmania. D. maculatus and D. viverrinus seen in the present study this dimunition in occur in the Dry Creek area today and are body size is not restricted to post-Pleistocene abundant in Holocene and late Pleistocene time but occurs also in the late Pleistocene. I deposits in S. Victoria. The presence of either have, therefore, used the term "late-Pleistocene or both of these species in the Maribyrnong dwarfing" throughout the text in reference to Local Fauna would not be unexpected. this phenomena in the Australian marsupials. The presence of late-Pleistocene dwarfing Discussion has not been well established for Australian The major reservation which must be imme- marsupials although I have found it to occur diately recognized in comparing the Dry Creek in a large number of species. These include Local Fauna with the Maribyrnong Local the following species or lineages as the case Fauna is that the sample size from the latter is may be (where the late Pleistocene forms are extremely small. For this reason larger sample recognized as distinct species they are listed sizes from the Doutta Galla Silt will surely first) : Macropus titan-M. giganteus, Macropus result in expansion of the present study; the siva-M. agilis (probably), Megaleia rufa, basic theme of late Pleistocene faunal succession Osphranter cooperi-O. robustus, Wallabia as outlined below should, however, remain un- vishnu-W. bicolor, Thylacinus cynocephalus, changed. Sarcophilus laniarius- S. harrisii, and possibly In the time interval between deposition of Dasyurus maculatus. Macropus rufogriseus has the "D Clay" and the Doutta Galla Silt two remained unchanged in size and tooth mor- phology from late major faunal changes occurred : ( 1 ) many of the Pleistocene to the pre- the megafaunal species became extinct (i.e. sent. M. rufogriseus is also the smallest Thylacoleo camifex, Protemnodon anak, P. macropod species in the Dry Creek Local brehus, Macropus ferragus and Zygomaturus Fauna and it is interesting and probably sig- trilobus), and (2) some megafaunal species nificant that larger macropod species experi- 82 LARRY G. MARSHALL enced cither a late-Pleistocene dimunition in In late Pleistocene faunas body size >20,000 yr B.P. (probably in the lineage M. siva- and M. represented by large sample sizes, both agilts and definitely in the lineage M. titan- megafauna! groups occur together. M. giganteus}, There are while still larger body species no faunas to my knowledge which are com- became extinct (i.e. Protemnodon anak, P. posed of only one brehus. of these groups. Based on Macropus ferragus). It would thus this evidence it may be induced appear that that where the late-Pleistocene fate of a extinct mcgafaunal species are species was determined found there to a large extent by its will also be found larger forms of living absolute body size, as reflected in tooth species; where living dimensions. species occur which are of their present size there The immediate will be no extinct problem arising from recog- megafaunal species found. I know of no excep- nition of late-Pleistocene dwarfing is how to tions to this generalization treat although it must be the populations of each species through kept in mind that time. within the time interval Are the larger late Pleistocene forms between extinction of specifically part of the megafauna distinct, subspccifically distinct, or and dwarfing of part of the megafauna these taxonomically indistinct from the smaller livim* groups may not show forms? Some of complete overlap. the lineages show size differ- Deposits showing this transition ences of 5%, others are presently 30%. Can all of these unknown. lineages be treated collectively or is it best to At Lake treat Menindee, N.S.W., a C date each lineage separately depending upon 14 of 26,300 ± 1,500 yr B.P. the amount of dwarfing was obtained on involved? An attempt charcoal from an Aboriginal oven associated to answer these questions must be based on with the remains more extensive of the extinct macropod collections and larger sample Macropus sizes ferragus (Tedford 1967). At Lake than those constituting the faunas Victoria, N.S.W., an described here. 18,000 yr B.P. date is tentatively Kurten recognized as representing a pos- (1959) considered the populations sible terminal date he studied which showed for extinction of late post-Pleistocene Pleistocene megafaunal dwarfing to be species (Marshall distinct at the subspecific level 1973). These studies complement These species differed by a magnitude the age and equal to faunal composition the differences of the Dry Creek and which occur in the Australian Manbyrnong Local species. If Faunas respectively Kurten's example is followed then Thorne (1972) described such late Pleistocene species the late Pleisto- as Macropus than cene would be fauna from Koonalda Cave, regarded as subspecies of the living SA collected from red, water-laid M. giganteus (i.e. Macrop'us deposits "which giganteus than). are Such approximately 20,000 yr B.P. old". taxonomic changes are not proposed All of the species present here although the in the fauna are represented option is made available for by living forms; future studies. no extinct megafaunal species or larger forms of Hie term megafauna as used living species are present. in this study The includes those fauna is modern in all large body sized species of respects. Thome's mammal study upholds the which occur in late Pleistocene conclusions reached by deposits Jones (1968, p. > 20,000 yr B.P. As shown above, 203) that "where fauna is the megafauna found in archaeological is composed of two very dis- sites spanning the last tinct groups: 20,000 yr it is (1) species now extinct, and (2) modern". species which have undergone On the bases a late-Pleisto- of these data it appears that cene dimunition in body extinction size. It is thus of late Pleistocene megafaunal necessary to indicate which of species and these groups is attainment of a fauna of being modern considered when using the aspect term mega- occurred in the time interval between fauna; if no distinction is made it must °be 20,000-25,000 yr B.P. In addition, it is con- assumed that both groups are beina considered cluded that late Pleistocene collectively. extinctions and late-Pleistocene dwarfing are correlative and LATE PLEISTOCENE MAMMALS 83 that these phenomena may possibly have been DeVis, C. W., 1895. A review of the fossil jaws of Macropodidae in the Queensland Museum. caused by a common, cryptic, factor(s). the Proc. Linn. Sue. N.S.W. 10: 75-133. , 1899. On some remains of marsupials Acknowledgements from Lake Colongulac, Victoria; with intro- ductory remarks on the locality by T. S. Hall. I am foremost indebted to Dr. A. Gallus for Proc. R. Soc. Vict. 12: 107-111. Frank, R., 1971. The clastic sediments of the Well- allowing me to study the Keilor collections. ington Caves, N.S.W. Hclictite 9 (1): 3-24. Thanks are due to Mr. T. A. Darragh, Curator Freedman, L., 1967. Skull and tooth variation in Anatomical of Fossils, and Mr. Ken Bell, Assistant Curator the genus Perameles. Part 1: features. Rec. Aust. Mus. 27: 147-166. of Fossils, of the National Museum of Vic- Gallus, A„ 1971. Excavations at Keilor. Rep. 1. toria for loan of materials and for placing Artefact. 24: 1-12. 1953a. Fluorine tests relative to the comparative fossil collections at my disposal; Gill. E. D., Keilor skull. Am. Jl pliys. Anthrop. 11: 229- Mr. J. A. Mahoney, University of Sydney, for 231. identification of the rodents; and Mrs Barbara , 1953b. Geological evidence in western antiquity of the Aus- Waters, Department of Palaeontology, U.C. Victoria relative to the tralian Aborigines. Mem. nam. Mus. Vict. 18: Berkeley for confirming identification of the 25-92. diprotodontid postcranials. The manuscript 1954. Ecology and distribution of the extinct giant marsupial, "Thylacoleo". Vict. was greatly improved upon by the constructive Naturalist 71: 18-35. comments and criticisms of Dr. W. A. Clemens, 1955a. Fluorine-phosphate ratios in rela- Department of Palaeontology, U.C. Berkeley; tion to the age of the Keilor skull, a Tertiary marsupial, and other fossils from western Vic- Bowler, Prehistory Department, Aus- Dr. Jim toria. Mem. natn. Mus. Vict. 19: 106-125. tralian National University, Canberra; Dr. -, 1955b. Radiocarbon dates for Australian Aust. Jl Sci. 18: 49-52. Michael O. Woodburne, Geology Department, archaeological samples. 1957. Current Quaternary Studies in Gill, Deputy U.C. Riverside; and Mr. E. D. Victoria, Australia. 1NQUA. V Internat. Cong. Director, National Museum of Victoria. The Madrid 1-7. 1962. In ANZAS. Committee for the In- following institutions provided working space vestigation of the Quaternary Strandline Clayton, for the study: Monash University, Changes. Aust. Jl Sci. 25: 203-205. Vict., Australia; National Museum of Victoria, 1967. Melbourne Before History Began. Australian Broadcasting Commission, Sydney. Australia; and the Department of Palaeon- -, 1971. Applications of Radiocarbon Dating tology, University of California, Berkeley. in Victoria, Australia. Proc. R. Soc. Vict. 84: This study was initially supported by a Monash 71-85. Hooijer, D. A.. 1950. The study of subspecific Graduate Scholarship (Monash University advance in the Quaternary. Evolution. Lancas- University) and completed under a National ter. Pa. 4: 360-361. Background to Science Foundation Traineeship Award (U.C. Jones, R., 1968. The Geographical the Arrival of Man in Australia and Tasmania. illustrations are by Mrs. Pat Berkeley). The Archaeol. phxs. Anthrop. Oceania 3 (3): 186- Lufkin, staff artist, Department of Palaeon- 215. Krefft, G., 1870. Guide to the Australian fossil tology, U.C. Berkeley, who was supported by by the trustees the Aus- Endowment. remains exhibited of the Annie Alexander tralian Museum, etc. Sydney. Kurten, B.. 1959. Rates of evolution in fossil mam- References mals. Cold Spring Harbor Symposium on Quan- titative Biology (Baltimore) 34: 205-215. , Pleistocene Mammals Europe. Anderson, C, 1929. Palaeontological Notes No. 1. 1968. of Chicago. Pp. 317. Macropus titan Owen and Thylacoleo carmjex fossil mammals Owen. Rec. Aust. Mus. 17: 35-49. Lydlkker, R.. 1887. Catalogue of British Museum (Nat. Hist.). Pt. 5, Mar- Bartholomai, A., 1966. The type specimens of some in the London. Pp. 146-295. of DeVis species of fossil Macropodidae. Mem. supialia. Marshall. L. G., 1973. Fossil vertebrate faunas Qd Mus. 14 (5): 115-126. > variation ot the from the Lake Victoria region SW. New South , 1971. Morphology and Wales, Australia. Mem. natn. Mus. Vict. 34: cheek teeth in Macropus giganteus Shaw and 16 151-171. Macropus agilis (Gould). Mem. Qd Mus. Mariow, B. J., 1958. A survey of the marsupials of New South Wales. CSIRO Wildl. Res. 3: 71-114. Bowler. L, 1970. Alluvial terraces in the Maribyr- 1876. Prodromus the Palaeontology nong Valley near Keilor, Victoria. Mem. rutin, McCoy, F., of Victoria. Dec. 3. Geol. Surv. Vict. pp. 7-12. Mus. Vict. 30: 15-58. of 84 LARRY G. MARSHALL Merrilees, D., 1965. Two new species of the ex- tinct genus Appendix 1 Sthenurus Owen (Marsupialia, Macropodidae) from South-western Australia including Dry Creek Local Fauna Sthenurus gilli sp. nov. R. Soc ' W Aust. 48 (1): 22-32. Level - , 1967. Cranial and mandibular characters I. Gravel layer 1, D excavation rn a ' nland wombats (Marsupialia, a) zysomaturus trilobus Vn t H f P29514, mandible. Vombatidae)m? /from tt t , a palaeontological viewl 11. X clay under L gravel point, and their bearing on the fossils called (a) Zygomaturus trilobus P29515, Phascolomys parvus by isolated Owen (1872) Rec M2 . S. Aust. Mus. 15 (3): 399-418. (b) Macropus titan P29644, left ramus- , 1968. Man the Destroyer: late Quater- P29531, right ramus nary changes with Ms-.. in the Australian marsupial fauna! (c) macropodid R. P29516, fragment of right /. Soc. W. Aust. 51 (1): 1-24. ramus of medium-sized macropod, with 1973 - Fos • siliferous deposits at roots ~a T Lake Tan- of M2_4?; P29517, distal end dou, New South Wales, of Australia. Mem. natn. humerus of large macropod. Mus. Vict. 34: 177-182. ill. D excavation, D clay Owen, R., 1877. Researches on the fossil remains (a) Perameles the of nasuta P29634, associated extinct mammals of Australia with a notice right ramus fragment with Pj-M, e Cl and PialS f E"g 'and L°nd0n associated left ° - maxillary Zy'!l^o?"2 ZT" fragment with M . —-, 1888. Description of the skull of an ex- fb) a'u tinct ursinm P23026, mandible; carnivorous mammal of the Z^P29519,n l leopard size of a fragment of right ramus. (Thylacopardus australis Owen) from a (c) Protemnodon brehus P29522, fragment recently opened cave near of the Wellington c^ve right maxillary with 2 "*. eW WaleS M Pr°C R Soc * S: (d) Protemnodon sp. ' - - P29527, left 99 (absto - metatarsal IV with proximal end of metatarsal Ride W. D. L.,' 1964. III A review of Australian fossil and left tibia. marsupials. /. R. Soc. W. Aust. 47: 97-131 (e) Macropus titan P29518, proximal 970 gUide to phalanx "te native mammals dl l IV; ~*Z7f lJ iAu of P29520, medial phalanx Australia. Melbourne, pp 149 f digit Stirling, 5 21, fragment of right maxillary E. C and A. H. C.~Zietz. •i }?, i 1899. Fossil with M->; P29523, right °f ramus; P29524 CalIabonna I. - Description left ramus \tTZl ^ o with Mi-,; P29525, right ramus DiProtod°» witn Fa, MemMem. R.R J" °! «£«£ Mw (associated with P29524)- Soc.l™ metatarsal V; P30722, left metatarsal IV; ated fragments of upper dentition; P29563, fragment of right scapula; P29612, medial phalanx of digit IV; P29566, fragment of right humerus; P29613, associated left and right innom- P29598, distal end of right femur. inate, V. A clay, level 3 (c) macropodid P29614, fragment of right (a) Sarcophilus laniarius P29631, right maxil- scapula; P29615, fragment of right in- lary fragment with P:, -M'. nominate; P29616, caudal vertebra; (b) Protemnodon anak P29554, left ramus P29617, right innominate; P29618, right with P3 , M,_«; P29555, right M a . metatarsal IV and V (both large and (c) Macropus titan P29552, fragment of right small species of macropod are included ramus with M 4 ; P29553, left ramus with in this group). Ms-,.; P29556, left metatarsal IV; P29557, IX. Excavation Y, various levels left metatarsal V; P29564, right cal- (a) Protemnodon brehus P29628, right ramus. caneum. (b) Protemnodon sp. P29625, left metatarsal (d) Macropus agilis P29575, left ramus; IV. P29576, right ramus. (c) Macropus titan P29624, right ramus; (e) Macropus rufogriseus P29578, right P29626, right maxillary fragment; maxillary fragment with M 2~\ P29627, left maxillary fragment; P29629, (f) macropodid P29559, acetabular region of fragment of right tibia. left innominate; P29561, left cuboid; (d) In addition to the above there are frag- P29562, right scapula fragment; P29567, ments of a large pelvis, a medium-sized caudal vertebrae; P29569, distal end of pelvis, a medial phalanx of digit IV. and left metatarsal V; P29570, proximal end several vertebrae of macropods. These of right ulna; P29571, distal end of right were not catalogued because of their humerus; P29572, right calcaneum; fragmentary condition. P29573, left acetabular portion of in- X. KA Excavation 1963 nominate; P30727, proximal end of left (a) Protemnodon sp. P29632, right metatarsal metatarsal IV; P30728, fragment of right IV. innominate. (b) Macropus titan P29633, right metatarsal VI. A clav, level 2 IV and V. (a) fhylacinus cynocephalus P29588, left (c) macropodid (small species) P29630, iso- humerus. lated right lower incisor. (b) Sarcophilus laniarius P29587, left maxil- XI. KAA, lowest level 1 "4 (a) Thylacoleo camifex P29545, right maxil- lary fragment with M . (c) Protemnodon sp. (probably P. brehus) lary fragment with lower edge of orbit, P29586, isolated right lower incisor. and anterior root of P*. (d) Macropus titan P29580, left ramus; (b) Macropus titan P29528. right ramus with P29581, right and left ramus; P29582, M 2 -,; P29529. left ramus with Ms_a isolated right M*; P29583. left ramus; broken, M, complete; P29530. left ramus P29584, right ramus; P29589, right with Ms-*; P29533, proximal end of right metatarsal IV; P29590, left metatarsal V; metatarsal IV; P29538, left maxillary 1 "*; right astra- P29591, left metatarsal IV. fragment with M P29540, 1 "2 galus. (e) Macropus agilis P29579, right P\ M . diprotodontid (probably referable to (f) Marcoput rufogriseus P29585, series of (c) upper incisors. Zvgomaturus trilobus) P30723, right calcaneum. (g) macropodid P29592, lumbar vertebra; P29593, vertebra; P29594, caudal ver- KAA, middle level of left tebra; P29595, left acetabular portion of (a) diprotodontid P29599, shaft innominate; P29596, portion of right humerus. ilium; P29597, portion of right scapula. (b) Macropus rufogriseus P29574, right ramus. VII. A clav, level 1 fragment of right (a) Macropus titan P29600, left ramus; (c) macropodid P29565, humerus. P29602. left ramus; P29605. left meta- highest level tarsal V; P29606, right maxillary frag- KAA, anak P29604, associated ment: P29607, left calcaneum; P29608, (a) Protemnodon left and right ramus. associated left calcaneum and astragalus. (b) Macropus titan P2960I, left ramus. (b) Macropus agilis P29603, left ramus with M*. Appendix 2 (c) Macropus rufogriseus P3O905, left ramus. VIII. A clav, level la Maribj'rnong Local Fauna (a) diprotodontid (probably referable to Zvgomaturus trilohus) P29619. right (a) Vombatus ursinus P30724. mandible. rttfa P30730. left maxillary fragment metatarsal V; P29620, right metatarsal (b) Megaleia with P30749, two lower molar fragments. IV; P29621, right cuboid; P29622, right M"?; was collected as float from the ectocuneiform; P29623, right navicular. This specimen soil quarry in Green Gully downstream from (b) Macropus titan P29609. right ramus; associ- St. Albans Road, Keilor, Victoria. P30749 and P29610, isolated left M 4 ; P29611, 86 LARRY G. MARSHALL other bones were found washed out of a bull- in Green Gully, dozed slope in the Doutta downstream from Keilor Galla Silt where it Terrace-St. abuts against the Albans Road. hillside at the southern end of (e) the quarry Pseudomys cf. P. australis P15773, (slope nearest shed). P30750a upper associated skull and postcranial fragments molar and lower molar series representing a (left) (M, ,) and single individual. isolated left This specimen was collected lower molar; P30751, associated by Mr. E. D. Gill seven with Cu date W169, Braybrook. feet from surface and above diastema (c) Macropus giganteus P30752, in east wall of Keilor cranium left ramus with quarry. e : P30750b isolated > lower molar, probablyu u, *fc* (f) Pseudomys cf. P. gracilicaudatus M 4 . P30907 left ramus with incisor, (d) Mastacomys fuscus P30906, M«; P30908, left ramus partial skull with with incisor, M,_ 2 . These left and right M« associated specimens were col- left ramus frag lected by ment with Mr H. E. Wilkinson 13 feet M ,. This specimen below was collected by the surface of Mr. H. b. high level terrace, in soil Wilkinson, in situ, in the soil quarry pit at mouth of Green Gully, Keilor : F 9 solid (l™r > • • • triangles represent extant sample off Macropus^ giganteus from Victoria solid circles represent fossil sample of Macropus titan from ury Creek (opposite as appears in the figure).