Records of the Western Australian Museum Supplement No. 57: 191-200 (1999).

Is Agrosaurus macgillivrayi Australia's oldest dinosaur?

Patricia Vickers-Rich1, Thomas H. Rich2

, Gregory C. McNamara3 and Angela MilnerA

1 Monash Science Centre, Earth Sciences Department, Monash University,Clayton, Vic 3168; email: prich@orion.earth.monash.edu.au

2 Museum Victoria, PO Box 666E, Melbourne, Vic 3001}Albert Kersten Geocentre, PO Box 448, Broken Hill, NSW 2880

4 Department of Palaeontology, The Natural History Museum, Cromwell Road, London SW7 5BD, UK

Abstract - The holotype and only known specimen of the prosauropodAgrosaunls macgillivrayi Seeley, 1891 probably comes from the Late TriassicDurdham Down locality near Bristol, England. Originally it was reported asbeing from the northeast coast of Australia. Firsthand examination of the mostplausible northeast Australian source for such a fossil, outcrops of the JurassicHelby Beds exposed on the east coast of Cape York Peninsula, havedemonstrated that the rock in that unit was quite unlike that associated withthe holotype. Gross and trace element comparisons between possible fossilbone fragments from the Helby Beds and the holotype also showed them to bequite different. On the other hand, similar comparisons between the holotypeand fossil bone from Durdham Down showed them to be quite comparable, aswere the rocks and microvertebrates there and the ones associated with theholotype. Furthermore, A. macgillivrayi is probably a junior synonym ofTlzecodontosaurus antiquus Morris, 1843 from Durdham Down.

INTRODUCTIONIn 1844, four British ships headed by H.M.s. Fly

were sent to construct a beacon on Raine Island offthe eastern side of Cape York Peninsula, Australia(Figures 1 and 2). They were in the area for fourand a half months, and while the beacon was underconstruction, the ships' crews explored in detail theeastern side of Cape York Peninsula, north of CapeGrenville Oukes 1847).

Nearly half a century later, Seeley (1891)published a description of a few limb bones andphalanges of the dinosaur Agrosaurus macgillivrayi

FLYHM Sloop Fly was launched at Pembroke Dock on25 August IS31.

(Figure 3). The only locality information availableto Seeley was a label, "in a small, delicatehandwriting, 'Fly, 1844. In. Macgillivray, from theN.E. coast of Australia.'" (Seeley 1891) (cf. Figure 4).

John MacGillivray was a naturalist attached to theexpedition whose job it was to collect animals forEdward Smith Stanley, the 13th Earl of Derby(Ralph 1993). His notebooks for the expeditionhave never been found, and he never mentionedanything about having discovered such a specimenin the book he wrote about a subsequentexpedition to the same area a few years later, when

j, ,

Arrived Sydney October 1842. Surveyed Great BarrierReef, visited Papua. Sailed for home December 1845.Reduced to coal hulk 1855. Broken up 1903.

TonnageLengthBeamArmament

485 BM114116 in (349 m)3211 (9.7 m)Eighteen 6 pndrs

(with permission).

192 P. Vickers-Rich, T.H. Rich, G.c. McNamara, A. Milner

14"

10'

146"

144

c=> Saibai I

TORRES STRAIT

1 l'

:CAPE YORK'PENINSULA

Raine Island

Cape Grenvi/le

NEWGUINEA

'0

..

30

,..140 130

14[ (1 Cl Wednesday Island :143

Thursday ISlanDdDOHom Island (> Uttle Adolphus IslandPossession 'V Mount Adolphus Island

Island

Prince of 0 jJ f I'Wales Island

Figure 2 Map of Cape York Peninsula, northern Queensland, and environs, with localities mentioned in text.

Australia's oldest dinosaur?

he sailed on H.M.S. Rattlesnake (MacGillivray 1852).A geologist, J. Bette Jukes, did publish a book onthe results of the 1844 voyage of H.M.5. Fly Uukes1847), but he, too, did not mention anything about

Figure 3 Postcranial elements of holotype ofAgrosaurus macgil/ivray Seeley, 1891, BMNH49984. From left: left tibia, ungual phalanx,distal caudal vertebra, proximal part of righttibia, distal end of right radius.

193

Figure 4 Original hand written label associated withholotype of Agrosaurus macgillivrayi Seeley,1891, BMNH 49984. Label reads, By In.[Mr?]Macgillivray. From N.E. coast of Australia."Fly" 1844

the discovery of fossil vertebrate bones during thecourse of that trip either there or in otherpublications Uukes 1850; Brown 1871).

Seeley (1891) regarded these fossils as mostsimilar to dinosaurs known to him from the LateTriassic and Early Jurassic. Modern opinionconcurs, placing Agrosaurus macgillivrayi in theprosauropods, a group of early dinosaurs confinedto that time (Galton 1990; Molnar 1991).

Institutional AbbreviationsBMNH = The Natural History Museum, London,

V.K. (formerly British Museum (Natural History));QM =Queensland Museum, Brisbane, Queensland,Australia.

EARLY MESOZOIC GEOLOGY OFCAPE YORK PENSINULA

Late Triassic and Early Jurassic continentaldeposits are not rare in eastern Australia.However, for the most part, they are deeplyweathered rock units, unfossiliferous except fortrackways in some cases; e.g. the aeolian PrecipiceSandstone. Therefore, it is not surprising thatadditional dinosaurian remains of this age have notbeen forthcoming there. The one possible exceptionis the partial skeleton of the primitive sauropodRhoetosaurus brownei from southeastern Queenslandwhich might be Early rather than Middle Jurassic.

Knowing where H.M.5. Fly was in 1844 andexamining the currently available geologicalinformation about the northeast coast of Australia,one rock unit stands out as being the most likelysource for the holotype of Agrosaurus macgillivrayi.This is the Helby Beds which occur on the eastcoast of Cape York Peninsula, due west of RaineIsland. The ships supporting the originalconstruction of the beacon on Raine Islandrepeatedly journeyed to the mainland to obtainwood and freshwater. Both MacGillivray and Jukeswould have taken whatever opportunities therewere to go ashore to collect specimens and make

194

observations. The easiest place for a block of rockwith fossil bones in it to have been collected bypeople operating out of longboats was along thatcoast west of Raine Island.

Not only are the Helby Beds exposed on the coastwhere voyagers might have easily reached them,but their age as indicated by palynological evidenceis appropriate: Jurassic (Powell et al. 1976).

RECONNAISSANCE OF THE HELBY BEDS

A reconnaissance flight carried out in September1993 over the northeastern coast of Australia fromCairns to the tip of Cape York Peninsula, revealedthat the only Mesozoic outcrops that from the air donot appear to be deeply weathered, are some of thecoastal exposures of the Helby Beds north of CapeGrenville (Powell and Smart 1977). This is the samearea that H.M.5. Fly visited in 1844.

The outcrops were observed to occur as coastalexposures reminiscent of those in Victoria, whichhave yielded a significant collection of AustralianEarly Cretaceous tetrapods (Rich et al. 1988; Richand Rich 1989). The apparently unweatheredoutcrops were seen over a distance of 80 km andhave a combined extent of 3-5 km along the coast.

A ship, El Torito, was engaged in July 1995 for sixdays of reconnaissance of the outcrops of the HelbyBeds between Captain Billy Landing at 11°40'5.,142°53'E. and the Tern Cliffs at 11°01 '5., 142°45'E.

During the winter on the far North Queenslandcoast, temperatures are commonly a pleasant 25°Cwith a cool onshore breeze, whereas duringsummer temperatures can climb to 35°C or morewith less breeze and oppressive humidity. Winterwas therefore chosen as the time of year to carryout this reconnaissance because it is easiest to keepa clear head searching for elusive fossils when it isnot too hot. This decision was made despite thefact that winter is also the time when steadysoutheasterly winds commonly cause high seas inthe area. Fortunately, unusually calm conditionsprevailed during most days of the work, whichmade it possible to examine all the outcrops of theHelby Beds for as long desired. Had normal seaconditions been the case during the work, athorough examination of all the outcrops duringthe time available most likely would not haveoccurred.

The people searching for the fossils had variedbackgrounds. This was important because thecontext in which fossils might be found wasunknown. The hardest fossil to find is the first one.Once a single specimen has been located in an area,the fossil hunter has an idea not only of what thefossils themselves will look like in the context ofwhere she/he is working but also the type of rockthey are likely to occur in. Unfortunately, theholotype of Agrosaurus macgillivrayi has been

P. Vickers-Rich, T.H. Rich, G.c. McNamara, A. Milner

completely removed from the rock it was found in,and all that associated rock has been chemicallydisaggregated. Thus, in a practical sense, the firstfossil bone from the Helby Beds had yet to befound. In bringing their different experiences to thesearch, this group made up of fossil hunters whohad collected in many parts of the worldmaximized the likelihood of finding a fossil. Inaddition to three of the authors (P.V.-R., T.H.R.,G.C.McN.), the searchers included TimothyHamley (Department of Zoology, University ofQueensland, St Lucia), Leaellyn Rich (OrmondCollege, University of Melbourne), Lucinda Hann(Wesley College, Glen Waverley, Victoria), LesleyKool (Earth Sciences Department, MonashUniversity), Alan Fraser (Johns HopkinsUniversity, Maryland, U.S.A.), Pablo Puerta(Museo Paleontologico Egidio Feruglio, Trelew,Argentina), and Fan Jun hang (Institute ofVertebrate Palaeontology and Palaeoanthropology,Beijing, China).

The field party departed on 10 July 1995 fromCairns on the Gulf Express, a coastal freighter thatmakes a weekly return voyage to Thursday andHorn Islands off Cape York. On the evening of 11July, the field party transferred to El Torito atPortland Roads.

A landing was made the following morning atCaptain Billy Landing, where the prospectingbegan. Most encouraging was the fact that the firstrocks seen were reminiscent of the coastal outcropsof the dinosaur-bearing units well known to severalof the party in Victoria. Working northward, alloutcrops of the Helby Beds near Captain BillyLanding were examined by the end of the secondday without the discovery of a single fossil bone,despite many rock exposures being remarkablysimilar to those that produce dinosaurs in Victoria.

On the third and first half of the fourth day, therocks of the Helby Beds further north, which fromthe air had not seemed remarkably different fromthose at Captain Billy Landing, proved on closeinspection to have been extensively chemicallyaltered. They were, for the most part, lateriteboulders and kaolinized clays. Searching in detailfor small areas that were not chemically altereddid turn up some places where the rocks weresimilar to those near Captain Billy Landing, butagain no fossil bones of any kind were found.

With no fossils found anywhere by noon on thefourth day, it was decided to return to CaptainBilly Landing and search there a second time tomaximize the chances of finding evidence of fossilbones. El Torito spent the afternoon retracing itsnorthward path and arrived off the southern mostexposures of the Helby Beds less than an hourbefore sunset. Going ashore, the field party spreadout and in quick succession two objects thatappeared to be bone, (QM F35957, Figure 5, and

Australia's oldest dinosaur?

Figure 5 QM F35957, one of two possible fossil bonefragments from 11°40'28"5., 142°51 '34"E.coastal outcrops of the Jurassic Helby Bedsjust south of Captain Billy Landing, easternside of the Cape York Peninsula, Queensland.Scale bar = 10 mm.

QM F35958), were found by Pat Vickers-Rich atleast 50 m apart. Passed amongst us, in the failinglight of day and again on board the ship that night,we could not convince ourselves one way or theother whether, in fact, the objects were bone ­although gradually the doubts became weaker.

The following day cracking rock at that site bya number of the field party did not revealanother object like either of the two possiblefossil bone fragments. After these two objectshad been found so quickly in a few minutes oftwilight one day, and then nothing similar tothem could be located searching all thefollowing day until the light conditions weresimilar to the previous one, we were greatlypuzzled as to what they were. Disappointed,some of us reprospected other outcrops of theHelby Beds near Captain Billy Landing for abouthalf the day, again to no avail.

The final day on El Torito was spent reachingThursday Island. While under way, the opportunitywas taken to examine the possible fossil bonefragments in full sunlight for the first time. There,with the knowledge that nothing more like themhad been found, it was harder to convince ourselvesthat they were, in fact, small pieces of fossil bone.To decide for sure, they were carefully rewrappedto be transported to Monash University where theycould be examined under a microscope andchemical tests carried out if necessary for a moredefinitive analysis.

195

LABORATORY RESULTS AND DISCUSSIONThe two possible fossil fragments found in the

Helby Beds were examined at length at MonashUniversity. If they were water worn rock pebbles,there were two things odd about them. First, onboth specimens, one smooth surface was a partialcylinder in shape. This smooth surface wastruncated by a smooth flat surface. Such acombination is what might be expected in afragment of bone but unlikely in a water-worn rockpebble. Second, again on both specimens, oppositethe smooth cylindrical surface was a very roughsurface, as if broken. Although pebbles with rough,broken surfaces are found, usually they are rare infine-grained sandstones, and in this case th~.onlytwo objects discovered had this pattern more.common in a bone fragment than an ordinary rockpebble.

Visual comparison of a bone of the holotype ofAgrosaurus macgillivrayi with the two possible fossilfragments found near Captain Billy Landing showsthey are quite different in preservation. TheCaptain Billy Landing specimens are solid,whereas the spongiosa of the holotype bone isopen. The rock in which the holotype of Agrosaurusmacgillivrayi was found has been disaggregatedwith acetic acid, suggesting it was originallycemented with calcium carbonate. The leachingquite evident in all outcrops of the Helby Bedsvisited had removed any carbonate that might oncehave been present in those rocks.

The difference in preservation between theholotype of Agrosaurus macgillivrayi and the twopossible bone fragments from Captain BillyLanding, coupled with the absence of carbonatesfrom the Helby Beds, strongly suggests that theholotype specimen came from elsewhere than theHelby Beds of Cape York Peninsula. The lack ofother coastal outcrops in "N.E. Australia" of theright geological age to have yielded a prosauropodto a ship's landing party operating inland on foot in1844, makes it unlikely that A. macgillivrayi camefrom that area at all.

Von Huene (1906) described the rock in whichthe bones of the holotype of Agrosaurusmacgillivrayi occurred as, " einer grauen Breccievoll von Knochensplittern [a grey breccia full ofbone fragments]." He then noted, "Das Gesteinerinnert in h6chstem Grade an die Knochen breccievon Durdham Down in Bristol [The rock isextremely reminiscent of the bone breccia atDurdham Down near Bristol]." This reference wasto the Late Triassic Magnesian Conglomeratefissure site in the Carboniferous Limestone fromDurdham Down near Bristol, England (Figure 6)(Locality 18, Avon, England in Weishampel 1990),dated as Carnian to Rhaetian (Benton and Spencer1995). Because of the subsequent chemicaldisaggregation of the rock surrounding the

196

ENGLAND

Figure 6 Location of Durdham Down near Bristol, thelikely actual source of the holotype ofAgrosaunls rnacgilIivrayi Seeley, 1891.

holotype, von Huene's observation cannot now beconfirmed, although there is no reason to doubt it.Microvertebrate remains recovered from theresidue remaining after the acid preparation includejaw fragments of a sphenodontid reptile and the tipof a prosauropod tooth crown bearing denticulatedmargins.

The most complete small reptilian element (andthe only one that can be identified with certainty)is a 3 mm long posterior fragment of a rightmaxilla of a sphenodontid, BMNH R14111. Thefragment includes ten teeth. The eight mostposterior ones, situated on the posterior process ofthe maxilla, are acrodont (firmly fused to the jawbone) with large and small teeth alternating, thesmall ones positioned on the lingual side of thejaw. The two most anterior teeth, in line with theback of the nasal process of the maxilla, arepleurodont (one side of the tooth base fused to thejaw bone) and similar in size. This dual type oftooth implantation is found only in thesphenodontian Diphyodontosaurus avonis Whiteside,1986. Diphyodontosaurus is recorded from severalTriassic fissure sites in the CarboniferousLimestones of the Bristol area: TytheringtonQuarry (Whiteside 1986), Cromhall Quarry (Fraserand Walkden 1983, 1984) and Durdham Down(Benton and Spencer 1995; Fraser 1994). Remains of

P. Vickers-Rich, T.H. Rich, G.c. McNamara, A. Milner

the prosauropod Thecodontosaurus also occur atthese fissure sites (Benton and Spencer 1995). Ofthese sites where both Thecodontosaurus and D.avonis have been found, only Durdham Down wasa known fossil locality in 1879, the year theholotype of Agrosaurus macgillivrayi was acquiredby the Natural History Museum, London. Thus,the occurrence of Diphyodontosaunls in the sameblock as elements of the holotype of A.macgillivrayi, BMNH 49984, together with thesimilarities in bone preservation and matrix,provide very strong evidence that the holotype isfrom Durdham Down.

This still leaves unanswered how the specimencame to be labelled as Seeley found it to be whenhe wrote his paper in 1891. However, the materialhad already passed through several hands; it waspurchased by the Natural History Museum in 1879from Edward Charlesworth after the dispersal of acollection belonging to Mr Samuel Long WaringEsq. of 'The Oaks', Norwood, following his deaththat year (Seeley 1891). It is clear that the label wasassociated with the specimen when the NaturalHistory Museum purchased it from Charlesworthin 1879 since the register entry detail for July 1879is an exact transcription in quotation marks of thatinformation, except that it reads "Mr.[.??]Macgillivray" instead of ''In.[.??] Macgillivray" asSeeley (1891) quoted it. So, any specimenmislabelling happened prior to that date. No earlierrecord of the history of the material can be traced.

Von Huene (1906) regarded Agrosaurusmacgillivrayi Seeley, 1891 as quite close toThecodontosaunls antiquus Morris, 1843 from theDurdham Down locality, noting that they differedonly in that the radius and ulna of A. macgillivrayiwere more slender. The similarity was so close thathe transferred A. macgillivrayi to Thecodontosaunlsmacgillivrayi. Despite this degree of similarity andthe likeness of the rock enclosing the holotype of A.macgillivrayi, nowhere did he express doubt as tothe Australian provenance of the specimen. Givenwhat was known of the geology of the Cape YorkPeninsula at the time, there was no reason for himto have done so.

One of the bones of the holotype of Agrosaunlsmacgillivrayi, together with a sample of fossil bonesknown to be from the Magnesian Conglomerate,Durdham Down, BMNH R1548, and the twopossible bone fragments from Cape York Peninsula,QM F35957 and QM F35958, were analysed using amass spectrometer in the VIEPS Trace AnalysisLaboratory at Monash University by Ms LouiseFrick, assisted by Ms Donna Korke and Mrs LesleyKool. The detailed results of this work are presentedin the Appendix. The data show that the fossil bonesample of the holotype of A. macgillivrayi and thatfrom Durdam Down, England have much higher U/Th ratios, and much lower Th/Pb ratios than the

Australia's oldest dinosaur? 197

• Editors footnote added in proof: Long and Molnar (1998), recentlydescribed a Middle Jurassic theropod, Ozraptor subotaii, fromWestern Australia.

Injune Creek Group where Rhoefosallrlls browllei wasfound, nowhere else in Australia are fossilterrestrial vertebrates of the age of the Helby Bedsknown*. Therefore, although the trail followed wasa century and a half old, and in the end may haveled nowhere, the effort to try and decide whetherthe holotype of Agrosallrlls macgillivrayi came fromthe Helby Beds was worthwhile.

A negative scientific result can never beestablished with complete certainty. It is notpossible, for example, to find the non-thylacine todemonstrate that there are no Tasmanian tigersalive today. The best one can hope to do is toconduct a thorough search. The one satisfaction ofthis result, although a negative one, is that it isrobust. Because the weather and sea conditionswere so favourable, the field party was able toexamine all the outcrops of the Helby Beds for aslong as necessary to establish with the highestprobability possible that fossil bones similar inpreservation to those of Agrosallrlls macgillivrayieither do not occur at all in that unit or if so, areextremely rare.

A caveat about this negative result is that theamount of outcrop examined was necessarilylimited. In Victoria, Early Cretaceous dinosaurshave been found in circumstances rather similar tothe outcrops of the Helby Beds. In both cases, theexposures are on shore platforms frequentlywashed by waves from the sea. If one were to berestricted at random to the same amount ofoutcrop in Victoria as was available to us in theHelby Beds, there is every chance that not a singlefossil bone fragment would be found there. It,therefore, cannot be completely ruled out that asingle block of rock containing the holotype ofAgrosallrlls macgillivrayi Seeley, 1891 was found byJohn MacGillivray in 1844 and no additional fossilswere found by us. But that is probably not whatoccurred.

Certainly, this thorough examination of theHelby Beds shows that it has a low potential ofeventually proving to be a rich source of fossilvertebrates. Given its age, however, as being froma time interval in Australia from which fossiltetrapods are unrepresented unless the two objectsfound at the end of the fourth day are, in fact,fossil tetrapod bones, if opportunities to revisit theoutcrops of the Helby Beds near Captain BillyLanding occur from time to time, an effort to findmore of whatever those objects are is justified.From its form alone, a third such specimen mightmake it clear whether such objects are in fact bonesor a rare rock type.

0.5

uThPbCeLaBa

Cape York OM F35958

Cape York OM F35957 j~o ,f'"\

OL...-~--l._~--'--~.....L-~':::"""'.J.U-~---'

0.0 0.1 0.2 0.3 0.4

Th/Pb

5

Cs

possible fossil bone samples from Cape YorkPeninsula (Figures 7 and 8). This further supportsthe contention that the holotype of A. macgillivrayi isfrom the Durdhanl Down locality rather than fromCape York Peninsula.

With the possible exception of the poorly exposed

Figure 8 Plot of the uranium-thorium and thorium­lead ratios in four samples. Note how theratios in the fossil bone from the holotype ofAgrosallrIls macgillivrayi and the sample fromDurdham Down are close to one another andquite distinct from the two objects found onCape York Peninsula which also are close toone another.

20 r-~--.--~-,--~-~~-,.-~....,

~o~\ Durdham Down

15 Agrosaurus macgillivrayi

10°..-------------------.

UfTh 10

Figure 7 Logarithmic plot of the abundances in partsper billion of seven trace elements in foursamples. Note how the abundances in thefossil bone from the holotype of AgrosaurusmaLxil/ivrayi and the sample from DurdhamDown closely track one another, while thetwo objects found on Cape York Peninsulahave a quite different trace elementabundance signature.

10 3--fr- Agrosaurus macgil/ivrayi--- Cape York OM F35957-0- Cape York OM F35958-0- Durdham Down

198

On the days the outcrops of the Helby Beds inthe vicinity of Captain Billy Landing werevisited, there was much mud on the shoreplatform. It is, therefore, well to keep in mindthe circumstances in which one of the mostproductive Cretaceous dinosaur localities inVictoria was discovered. Despite repeated, priorvisits to the site, nothing had been found there.On the day of discovery, a recent storm hadstripped away all the sand that ordinarilycovered the lens where the fossil bones occur,making them readily visible.

CONCLUSION

The brief answer to the question posed in thetitle of this paper is "No". The presentlyavailable evidence strongly suggests that theholotype of Agrosaurus macgillivrayi Seeley, 1891did not, in fact, come from northeasternAustralia despite the label associated with thespecimen. Rather, it is quite likely that it camefrom the Magnesian Conglomerate at DurdhamDown near Bristol, a site already well knownwhen the holotype was purchased by TheNatural History Museum in 1879 (Seeley 1891).Given the likelihood that the holotype ofAgrosaurus macgillivrayi Seeley 1891 did comefrom the Durdham Down locality, the species isquite probably a junior synonym ofThecodontosaurus antiquus Morris, 1843, aconclusion foreshadowed by von Huene (1906).

Following von Huene (1906), Galton and Cluver(1976) referred Agrosaurus macgillivrayi toThecodontosaurus macgillivrayi (Seeley) and listed itunder the family Anchisauridae as a nomen dubium,as they regarded the material as generically andspecifically indeterminate. More recently, Galton(1990) showed Thecodontosaurus to be a basalsauropodomorph taxon and assigned it to its ownfamily.

ACKNOWLEDGEMENTS

The Committee for Research and Exploration ofthe National Geographic Society provided thefunding to carry out this study: grants 4938-92,5257-94, and 5835-96. Andrew Kos and DavidCantrill hand-couriered the type specimen toAustralia, and back to England, respectively. PhilHurst of the Natural History Museum, London,took the photographs in Figures 3 and 4. Val Hoganof the Museum of Victoria Library, WendyCawthorne of the Geological Society Library,London, and John Thackray, Archivist for theNatural History Museum, London, and theGeological Society, London, helped gather historicalinformation.

P. Vickers-Rich, T.H. Rich, G.c. McNamara, A. Milner

REFERENCES

Bastock, J. (1988). Ships on the Australia Station: 1-175,Child and Associates Publishing, Frenchs Forest,N.5.W.

Benton, M.J. and Spencer, P.5. (1995). Fossil reptiles ofGreat Britain: 1-386, Chapman and Hall, London, U.K.

Browne, CA. (1871). Letters and Extracts from the Addressesand Occasional Writings of J.B. Jukes: 1-596, Chapmanand Hall, London, U.K.

Fraser, N.C (1994). Assemblages of small tetrapods fromBritish Late Triassic fissure deposits. In Fraser, N.Cand Sues, H.-D. (eds), In the shadow of the dinosaurs:214-226, Cambridge University Press, Cambridge,U.K

Fraser, N.C and Walkden, G.M. (1983). The ecology of alate Triassic reptile assemblage from Gloucestershire,England. Palaeogeography, Palaeoclimatology,Palaeoecology 42: 341-363.

Fraser, N.C and Walkden, G.M. (1984). Two Late Triassicterrestrial ecosystems from south west England. InReif, W.E. and Westphal, F. (eds), Third Symposium onMesozoic Terrestrial Ecosystems: 87-92, Tubingen, ShortPapers.

Galton, P.M. (1990). Basal Sauropodomorpha­Prosauropoda. In Weishampel, D.B., Dodson, P andOsmolska, H. (eds), The Dinosauria: 320-344,University of California Press, Berkeley, Calif., U.5.A.

Galton, P.M. and Cluver, M.A. (1976). Anchisauruscapensis (Broom) and a revision of the Anchisauridae(Reptilia, Saurischia). Annals of the South AfricanMuseum 69: 121-159.

Huene, F. von (1906). Ueber die Dinosaurier deraussereuropaischen Trias. Geologische undPalaeontologische Abhandlungen 8: 97-156.

Jukes, J.B. (1847). Narrative of the surveying voyage of theH.M.S. Fly: vo!. 1, 1-423, vo!. 2, 1-362, Boone,London, V.K

Jukes, J.B. (1850). A sketch of the physical structure ofAustralia as far as it is at present known: 1-95, Boone,London, U.K

Long, J.A. and Molnar, R.E. (1998). A new Jurassictheropod dinosaur from Western Australia. Records ofthe Western Australian Museum 19: 221-229.

MacGillivray, J. (1852). Narrative of the voyage of the H.M.S.Rattlesnake: vo!. 1, 1-402, vo!. 2, 1-395, Boone,London, U.K [Libraries Board of South Australiafacsimile edition, 1967.]

Molnar, R.E. (1991). Fossil reptiles in Australia. InVickers-Rich, P., Monaghan, J.M., Baird, R.F. andRich, T.H. (eds), Vertebrate palaeontology of Australasia:605-702, Pioneer Design Studio, Lilydale, andMonash University Publications Committee,Melbourne, Victoria.

Morris, J. (1843). A Catalogue of British fossils: 1-222, TheBritish Museum, London, U.K. [2nd ed. 1854].

Powell, B.5., Gibson, D.L., Smart, J., Grimes, KG. andDoutch, H.F. (1976). New and revised stratigraphicnomenclature, Cape York Peninsula. QueenslandGovernment Mining Journal April 1976: 179-188.

Powell, BS. and Smart, J. (1977). Jardine River-Grford Bay,Queensland, 1:250,000 Geological Series - map and

Australia's oldest dinosaur?

explanatory notes. Bureau of Mineral Resources, Geology andGeophySICS. Geological Survey of Queensland. AustralianGovernment Publishing Service, Canberra, A.c.T.

Ralph, R. (1993). John MacGillivray his life and work.Archives of Natllral History 20: 185-195.

Rich, P.V., Rich, T.H., Wagstaff, B., McEwen Mason, J.,Oouthitt, R.T., Gregory, R.T. and Felton, A. (1988).Evidence for low temperatures and biologic diversityin Cretaceous high latitudes of Australia. Science 242:1403-1406.

Rich, T.H. and Rich, P.V. (1989). Polar dinosaurs andbiotas of the Early Cretaceous of southeasternAustralia. National Geographic Research 5: 15-53.

Seeley, H.G. (1891). On AgrosaurIls m£lcgillivrayi (Seeley),

199

a saurischian reptile from the N.E. coast of Australia.Quarterly Journal of the Geological Society 47: 164-165.

Weishampel, O.B. (1990). Oinosaurian distribution. InWeishampel, O.B., Oodson, P. and Osm\'ilska, H.(eds), The Dinosauria: 63-139, University of CaliforniaPress, Berkeley, Calif., USA.

Whiteside, O.I. (1986). The head skeleton of the Rhaetiansphenodontid Diphyodontosaurlts avonis gen. et sp.novo and the modernizing of a living fossil.Philosophical Transactions of the Royal Society of LondonB 312: 379-430.

Manuscript received 9 March 1998; accepted 9 August 1998.

APPENDIX

Determination of trace elements in bone samples

Louise FrickVIEPS Trace Analysis Laboratory, Monash University, Clayton, Vic 3168

ChemistryAll wet-chemistry was performed in "class 350"

clean air cabinets. 50-100 mg samples of rockpowder were dissolved in HF-HN0

3mixture then

evaporated to dryness. They were refluxed twicein concentrated HNOy and once in concentratedHCl to remove fluorides, then dissolved in 2%HN0

3• This solution was then diluted to 50 ml of

2% HN03

(made with sub-boiled distilledconcentrated HN0

3and 18.2 M-ohm Millipore

water). Aliquots of this solution were then takenand diluted to an appropriate dilution factor forpresentation to the mass spectrometer(approximately 1:10,000). One analytical blank wasprepared with the run. A multi-element solutionstandard was used as a check standard to monitoraccuracy during the analysis procedure.

Mass SpectrometryDiluted rock and acid samples were analysed by

Table 1 Resolved concentrations of trace element isotopes in ppb.

I33CS 135Ba n9ta 140Ce 208Pb 232Th 238U

Acetone blank 304 633 390 235 299 335 N/OAgrosaurus macgillivrayi 1,077 393,931 126,264 315,082 449,818 1,405 25,987

Cape York Peninsula QM F35957 1,988 156,111 4,438 8,763 4,289 1,709 877Cape York Peninsula QM F35957 repeat 1,949 156,970 4,548 9,426 4,888 1,647 926

Cape York Peninsula QM F35958 837 23,344 2,404 4,843 1,770 729 140Ourdham Down 410 342,570 282,371 649,463 157,208 4,371 82,864

Reproducability of the standard datadum 1 9.837 9.719 9.825 9.588 9.938 9.863 10.101dum2 9791 9.687 9.829 9.493 9.944 9.895 10.109dum3 9.885 9752 9822 9.695 9.933 9.832 10.093

std 5 ppb 9.924 9.780 9.819 9.783 9.929 9.803 10.086mean 9.859 9.734 9.824 9.639 9.936 9.848 10.097

standard deviation (SO) 0.058 0040 0.004 0.126 0006 0.039 0010relative SO (= SO/mean) 0.59% 0.42% 0.05% 1.31% 0.06% 0.40% 0.10%

Accuracy of the 1 ppb dummy standarddum 1 -2.68'10 -3.85°1<, -2.79% -2.93% -1.67% -0.14% -0.07%dum2 -3.13% -4.17% -2.76% -3.89% -1.62% -0.18% -0.01%dum3 -2.20% -3.52% -2.83% -1.84% -1.72% -0.45% -0.14%

std 5 ppb -1.82% -3.25% -2.86% -0.95% -1.77% -0.75% -0.21 '/'o

Reproducability of the repeated sampleCape York Peninsula QM F35957 -1.93'>\, 0.55% 2.49% 7.56% 13.98% -3.63% 5.51%

200

inductively-coupled plasma mass spectrometry (ICP­MS) using a high resolution Finnigan MATELEMENTI'M magnetic sector ICP-MS. The followingisotopes were chosen for each element of interest, onthe basis of no or minimal isobaric interference: 133(:s,135Ba, 13"La, 140Ce, 208Ib, 23% and 23SV. Twenty-fiveruns of one mass spectrometer sweep through themass spectrum were performed. The runspecifications resulted in approximately 10 samplepoints per peak.

Data ReductionIntegrated count data from the instrument were

reduced using a custom, off-line, spreadsheet-basedprogram. During each analysis a series of dummysamples was run to provide a check on instrumentdrift. The count data were then corrected by a

P. Vickers-Rich, T.H. Rich, G.c. McNamara, A. Milner

polynomial in time for each element analysed toprovide constant counting efficiency in the dummysamples. Next, calibration curves were constructedfor each element using a simple, least squares linearregression of the multi-element solution standarddata points. Using these calibration curves, the datawere resolved into ppb in solution (Table 1; Figures7-8).

Estimate of Precision and AccuracyPrecision (in-run statistics quoted as relative

standard deviation = standard deviation / mean)is better than 5% for each analysis. As describedabove, determination of analytical accuracy isbased on analysis of the solution standards, and istypically better than 5% for most elements at the95% confidence level.