PROCEEDINGS OF THE AUSTRALIAN BICENTENNIAL HERPETOLOGICAL SOCIETY BRISBANE MEMOIRS OF THE VOLUME 29 20 SEPTEMBER, 1990 QUEENSLAND MUSEUM PART 2 Preface Most of these papers and notes were presented at the Australian Bicentennial Herpetological Conference held at the Queensland Museum, Brisbane, on 17-20 August, 1988. A handful were received subsequently and judged suitable for inclusion in the proceedings. Publication costs have been borne by the conference participants, the Australian Society of Herpetologists and the Board of Trustees of the Queensland Museum. The Queensland Museum generously supported the conference from concept to conclusion, so it is especially appropriate that the conference papers be published in a special issue of the Memoirs of the Queensland Museum. Production headaches and the attendant effort, and anxiety were cheerfully and willingly borne by Glen Ingram, Editor, and Neale Hall, Typesetter. Their diligence and patience have ensured the highest standards of presentation throughout. Peter Jell, Wayne Longmore, Peter Davie, Patrick Couper, Kim Easterbrook and Liza Hug also assisted in either running the conference successfully, or in the production of this memoir, or both. This is a fine collection of papers, clearly and concisely presented. A wealth of interesting and important new information is contained in these pages, all devoted to what are, without any doubt, the most fascinating elements of the Australian fauna. Without wishing to reveal my further bias, may I commend it to you? Jeanette Covacevich, President, Australian Society of Herpetologists Inc., Senior Curator (Vertebrates), Queensland Museum. 22 August, 1990. Delegates to the Australian Bicentennial He with Bette the giant Cane Toad. rpetological Conference, 17-20 August, 1988, assembled in the Dinosaur Garden of the Queensland Museum MEMOIRS OF THE (QUEENSLAND MUSEUM BRISBANE © Queensland Museum PO Box 3300, South Brisbane 4101, Australia Phone 06 7 3840 7555 Fax 06 7 3846 1226 Email qmlib@qm.qld.gov.au Website www.qm.qld.gov.au National Library of Australia card number ISSN 0079-8835 NOTE Papers published in this volume and in all previous volumes of the Afemoirs of the Queensland Museum maybe teproduced for scientific research, individual study or other educational purposes. Properly acknowledged quotations may be made but queries regarding the republication of any papers should be addressed to the Editor in Chief. Copies of the journal can be purchased from the Queensland Museum Shop. A Guide to Authors is displayed at the Queensland Museum web site A Queensland Government Project Typeset at the Queensland Museum ALTERNATIVE DIGITAL SCANSOR DESIGN IN THE NEW CALEDONIAN GEKKONID GENERA BAVAYIA AND EURYDACTYLODES AARON M. BAUER AND ANTHONY P, RUSSELL Bauer, A.M. and Russell, A.P. 1990 09 20: Alternative digital scansor design in the New Caledonian gekkonid genera Bavayia and Eurydactylodes, Memoirs of the Queensland Museum 29(2): 299-310. Brisbane. ISSN 0079-8835. Gekkonid feet are complex and highly integrated functional units. Convergence and parallelism are common themes in gecko digital design. Alternatively, closely relaled geckos may exhibit widely differing toe morphologies within a framework of phylogenetic constraint. The New Caledonian carphodactyline geckos Bavayia and Eurydactylodes are closely related lo one another yel they exhibil marked differences in external digital form. Bavayia possesses divided scansors and a highly arcuate penultimate phalanx. In Eurydactylodes the scansors are undivided and the penultimate phalanx is not raised. Internally Bavayia has a divided vascular sinus and a region of adipase tissue which helps to distribute forces laterally in conjunction with the divided pad. Proximally the digits of Bavayia are filled with adipose tissue which provides passive support and conformation forthe non-scansonal friction pads. In Furydactylodes a huge vascular sinus transduces forces directly from the penultimate phalanx. Differences in scansor morphology between [he taxa are related to the differences in control mechanisms of single versus divided pads. The functional significance of the alternative scansor designs is unclear, but the divided scansors of Bavayia may have played a role in the relative success of the genus in New Caledonia. L) Gekkonidae, Carphodactylinae, Bavayia, Eurydactylodes, digits, scansors, functional morphology, evolutionary con- straint. Aaron M, Bauer, Biology Department, Villanova University, Villanova, Pennsylvania 19085, U.S.A.; Anthony P, Russell, Department of Biological Sciences, The University of Calgary, 2500 University Drive N.W., Calgary, Alberta, Canada T2N 1N4; 16 August, 1988. The scansors of the gekkonid foot are extreme- ly complex and highly integrated functional units (Dellit, 1934; Russell, 1972, 1975, 1976, 1979, 1981, 1986). The adhesive ability of geckos is dependent not only upon the setal microstructures that interact with the substrate (Hiller, 1968, 1969, 1975), but also upon internal features of the scansors that transmit forces to the seta-bearing surfaces (Russell, 1975, 1981) and permit the distribution of forces associated with weight-bearing (Russell, 1986). The precise nature of the organisation of com- ponents of the musculo-skeletal and circulatory systems, as well as connective and adipose tis- sues varies among the taxa studied to date. Dis- tantly related taxa, however, often cope with similar locomotory demands (frequently sub- strate related) in near identical fashion. This has been well-documented in the case of ecalogical- ly equivalent members of the subfamilies Gek- koninae and Diplodactylinae (Russell, 1979) and has even resulted in paralle) radiations of geckos in these groups on the basis of key innovations (sensu Liem and Osse, 1975) in pedal structure and scansor architecture. Within a single sub- family, similar convergences have also been nated in genera occupying similar spatial niches or exploiting particular substrates (c.g. Russell, 1976), Indeed, a particular morphology, such as that characterising ‘leaf-toed’ geckos, may have been independently derived in many lineages (Russell and Bauer, 1989). Conversely, even closcly related taxa may ex- hibit a diversity of digital forms within the con- straints of shared descent (see Brundin, 1968). Russell (1976), for example, demonstrated morphotypic series. in digital design and com- plexity in the gekkonine genera Pachydactylus and Hemidactylus. Both of these genera are speciose and ecologically diverse and exhibit variations in both external and internal digital design. 300 The subfamily Diplodactylinae is less speciose than the Gekkoninae and much more geographi- cally circumscribed, with all taxa occurring in the Southwest Pacific (Australia, New Zealand and New Caledonia). None the less, there is great ecological diversity in the group, which includes burrowers, terrestrial forms and arboreal species. Like marsupials, the diplodactylines represent an ancient independent lineage which includes major radiations to some extent comparable to those of their more widespread relatives. The parallels between the Diplodactylinae and Gekkoninae are striking (Russell, 1979) but the analysis of the diplodactyline radiations is inter- esting in its own right. Unfortunately, few studies have examined diplodactyline morphol- ogy to date. Russell (1972, 1979) examined the pedal morphology of several species in the Diplodactylinae, but his data were derived al- most exclusively from gross dissection. Further, Russell lacked an explicit hypothesis of relation- ship among the diplodactylines upon which he could interpret the observed anatomy. Such an hypothesis is essential if the evolutionary and ecological significance of morphologies are to be evaluated in a phylogenetic context (Lauder, 1981, 1982). In this paper we examine aspects of digital FIG. 1. Proposed pattern of relationships among the New Caledonian carphodactyline geckos (including the Australian taxa previously recognised as Pseu- dothecadactylus), from Bauer (1986). MEMOIRS OF THE QUEENSLAND MUSEUM scansor morphology in representatives of the New Caledonian genera Bavayia and Eurydac- tylodes, two closely related diplodactyline geckos in the tribe Carphodactylini. Although Underwood (1954) initially placed the two genera in different subfamilies, he later (Under- wood, 1955) reconsidered the affinities of Eurydactylodes and placed both genera in his Diplodactylinae. Kluge (1965, 1967) accepted the affinities of these taxa and their close relationship to a third New Caledonian genus, Rhacodactylus and included all three in the tribe Carphodactylini. Bauer (1986), on the basis of a morphologically- based cladistic analysis, proposed a specific pattern of relationships among these three taxa (Fig. 1). Although a characteristic previously thought to be diagnos- tic of the Gekkoninae (sensu Kluge, 1987), the presence of extracranial endolymphatic calcium deposits, has since been identified in Eurydac- tylodes (Bauer, 1989), the overwhelming evidence of other characters suggests that the New Caledonian endemic geckos are indeed closely related. Despite this affinity, the external digital morphology of Bavayia and Eurydac- tylodes is markedly different. Bavayia is char- acterised by divided scansors and highly arcuate distal phalanges while the latter possesses single Bavayia cyclura Bavayia. sauvagi! Eurydactylodes symmetricus Eurydactylodes vieillardi Rhacodactylus (incl. Pseudothecadactylus) SCANSOR DESIGN IN BAVAYIA AND EURYDACTYLODES subdigital plates and a less markedly raised penultimate phalanx. Outward variation of this nature in the Gekkoninae is generally indicative of major design differences in internal anatomy and consequent functional differences (see Rus- sell, 1972, 1976, 1979). We here assess the specific anatomical differences exhibited by Bavayia and Eurydactylodes and evaluate sig- nificance (if any) of alternative digital designs within the well circumscribed New Caledonian carphodactyline lineage. MATERIALS AND METHODS Specimens examined in this study were col- lected by the senior author in New Caledonia under the authority of the Service des Eaux et Foréts and have been deposited in the California Academy of Sciences (CAS). In addition, specimens were also examined in the collections of several museums, most notably, the Australian Museum (AMS), the Naturhistoris- ches Museum Basel (NHMB), the Zoologisches Forschungsinstitut und Museum Alexander Koenig (ZFMK), and the British Museum (Natural History) (BMNH). Gross external observations and dissections were carried out on formalin-fixed, alcohol preserved museum specimens of Bavayia and Eurydactylodes. Radiographs of selected specimens were prepared using a self-contained x-ray unit. Cleared-and-stained preparations were made following a modification of the protocol of Was- sersug (1976). Specimens for light microscopy were decalcified, dehydrated, cleared and em- bedded in paraffin. Sections were cut on a rotary microtome at thicknesses of 8—12.m and stained according to the protocol for Mallory’s azan trichrome stain (Humason, 1979). Photomicrographs were prepared with a Wild compound microscope with 35 mm photo attach- ment. Specimens for scanning electron micros- copy were dehydrated through a graded alcohol series, critical point dried and sputter-coated to a thickness of 30nm with gold-palladium alloy before examination with an ISI-DS 130 micro- scope. RESULTS EXTERNAL ANATOMY OF THE DIGITS Eurydactylodes The digits of Eurydactylodes symmetricus and E. vieillardi are essentially identical in form. The 301 following description is based primarily upon specimens of the latter taxon. The digits are short and broadly dilated. Small fleshy webs are present between digits II and III and III and IV. The penultimate phalanges of digits II -V are mostly subsumed within their respective pads and the claw is carried only a short distance beyond the distal margin of the scansors. The ungual phalanx is firmly connected to the dor- sum of the pad by a fleshy sheath. Digits II -V of the manus fan out broadly whereas the first four of the pes are bound together at the level of the metatarsals, restricting the spread of the digits. The ventral surface of the pads of digits I-IV bear a series of broad scansorial plates that are generally straight proximally and somewhat chevron-shaped distally (Fig. 2A). Proximally, the scansors grade into enlarged subdigital scales that terminate at the level of the proximal portion of the first phalanx and are replaced by small non-setose scales similar to those of the palms. The proximalmost plates are generally non scan- sorial (sensu Russell, 1975) but do bear setae. There are typically 10 - 12 expanded plates under the fourth (longest) toe in both species. Digit one is small and bears a series of about five small friction pads proximal to the minute claw. The claw itself is sheathed and is bordered both laterally and medially by small terminal plates. Unlike all other New Caledonian and New Zealand carphodactylines (Bauer, 1986) these plates are completely separated from one another. The medial plate is substantially larger than the lateral (Fig. 2D). Bavayia The digits of Bavayia cyclura are moderately elongate and broadly dilated distally. As in Eurydactylodes the claws of digits II - V extend beyond their respective pads, but are firmly anchored to them by cutaneous sheaths. The penultimate phalanx is very strongly arcuate and tises well above the plantar surface of the pad. Small webs connect the bases of digits II and III, Tlf and IV and IV and V. As in all New Caledonian and New Zealand carphodactylines, metatarsals of digits I - IV are joined, reducing the digital spread of the pes (Fig. 3). All enlarged subdigital plates except the dis- talmost are divided and the pairs are strongly angled so that the medial ends meet at the mid- line far proximal to the lateral termini (Fig. 2B). At about the level of the antepenultimate phalanx the more distal true scansors (defined on their internal morphology - see below) give way to (a i=; th MEMOIRS OF THE QUEENSLAND MUSEUM FIG. 2.A. Eurydactylodes vieillardi (BMNH 1926.9.17.7), digit I], right pes. B. Bavayia sauvagii (BMNH 1926,9,17.25), digit IV, left pes. C. Bavayia sauvagii (CAS 159532), digit 1, right pes, dorsal view. D. Eurydactylodes vieillardi (BMNH 1926,9.17,7), digit I, right pes. E. Bavayia sauvagii (BMNH 1926.9.17.25), digit I, left pes. F. Bavayia cyelura (CAS 159550), digit 1, right pes. Note the architecture of the subdigital plates in A and B and the disposition of the terminal plates in D - F. simple friction pads. At the metapodial/phalan- geal joints the friction pads grade into irregular smaller scales which, in turn, grade into the palmar scales. Digit | of both manus and pes is reduced and carries a series of undivided friction plates, but no expanded pad. In contrast to those of the remaining digits, the claw of digit I is minute. It is bordered by a large, cleft terminal plate which is asymmetrical, bearing a larger medial pad (Fig. 2F). A diastema separates the terminal plates from the basal friction plates, which ex- tend well onto the palmar surface. The digits of Bavayia sauvagii are similar in most respects to those of their congeners but are somewhat more elongate and less broad, The scansor pairs of digits II - V are separated by a somewhat broader gap than are those of B. cyclura and break up into small scales somewhat more distally, Digital setae are longest at the free margins of the lamellae and appear to be better developed on the true scansors than on the fric- tion plates (Fig. 4). Most notably, the terminal plate of digit I is entirely medial to the claw (Fig. 2E). In dorsal view the claw of this digit appears completely sheathed (Fig. 2C). INTERNAL ANATOMY OF THE DIGITS Eurydactylodes The musculo-skeletal system of Eurvdac- tylodes and Bavayia are essentially identical to that described by Russell (1972, 1979) for the SCANSOR DESIGN IN BAVAYIA AND EURYDACTYLODES 30. td FIG, 3. Ventral view of the left pes of Bavayia sauvagii (CAS 159532) showing the metatarsal binding of the first four digits. Scale bar = Imm. 3 bar = 300um (right side shows 5 X enlargement of box on left). 304 Cp ee PN (to claw base) PA MEMOIRS OF THE QUEENSLAND MUSEUM FIG. 5. Diagrammatic representation of the musculo-tendinous system of digit IV of a generalised New Caledonian carphodactyline gecko. EDB = extensor digitorum brevis, FDB = flexor digitorum brevis, FDL = flexor digitorum longus, ID = interossei dorsales, LDT = lateral digital tendon, MT IV = fourth metatarsal, PA = plantar aponeurosis, PUP = penultimate phalanx, S = scansors. Arrows indicate points of attachment of muscles and tendons. closely related genus Rhacodactylus. The phalangeal formulae are unreduced (2-3- 4-5-3 manus, 2-3-4-5-4 pes). The penultimate phalanx is cylindrical in section and arches over the broadest part of the pad. The first phalanx is generally also cylindrical. Intermediate phalan- ges are strongly depressed and barbell-shaped, with broad epiphyseal surfaces. The dorsal extensor musculature consists of two asymmetrically developed bellies for each digit. Each pair of bellies sends a tendon to the middorsal region of its respective ungual phalanx. The dorsal interossei muscles reach only the level of the metapodial elements and do not send out tendons to the phalanges. Ventrally, lateral digital tendons extend from the metapodial- phalangeal joint cap- sule to the proximal borders of each of the true scansors. Digital flexor muscles run in parallel with the lateral digital tendons and insert on the distal ends of the phalanges (except the ungual and penul- timate phalanges; the long flexor muscle inserts on the lower part of the base of the ungual phalanx). In addition, the base of the digits receive slips from the femorotibial gastrocnemius (see Fig. 5 for a diagrammatic summary of the muscular and ten- dinous components of the digit). Internally, the scansor-bearing digits (II - V) of Eurydactylodes vieillardi are typified by the presence of an extensive digital sinus system consisting of a large central sinus and a reticular network (Figs. 6, 7) supplying both the medial and lateral portions of the scansors with vas- cularization in association with scansorial con- trol (see Russell, 1981). True scansors are present beneath the penultimate phalanx as well as the more proximal phalanges. There is a min- imal amount of loose connective tissue in the toe, and no adipose deposits were evident in his- tological sections. Bavayia The external differences between Bavayia cyclura and B. sauvagii are not manifested in the internal structure of the digits. In both taxa the proximal portion of the digit is largely filled with connective tissue invested with adipose cells (Fig. 8). This condition extends as far distally as the level of the penultimate phalanx where true scansors are located. In the region of the scansors lies a central blood sinus which has two lateral chambers, one on either side of the midline scan- sor cleft (Fig. 9). In all areas of the toe connective tissue, usually containing adipose deposits, oc- cupies the dorsal portion of the pad and the lateral portions of the scansorial plates themsel- ves. DISCUSSION PHYLOGENETIC CONSTRAINTS The observable morphologies of living or- ganisms are strongly influenced by the history of the taxa that possess them (Lauder, 1982). The influence of past environmental factors on form and function of organ systems should thus be reflected in descendent taxa, and taxa with shared descent should exhibit certain such his- torical features, or constraints, in common. Eurydactylodes and Bavayia appear to share a number of digital features as a result of common descent. In common with Rhacodactylus (in- cluding Pseudothecadactylus), Hoplodactylus and Naultinus, metatarsals I - IV, especially III and [V are parallel to one another, thus reducing digital spread (Russell, 1972; Bauer, 1986). These genera also share a ‘simplified’ muscular system in which the distal phalanges are free of the fleshy portions of the flexor and extensor muscles. Likewise, the digits receive no direct SCANSOR DESIGN IN BAVAYIA AND EURYDACTYLODES 305 FIG. 6. Cross-section through the fourth digit, right pes of Eurydactylodes vieillardi (ZFMK 16113) under the anterior portion of the penultimate phalanx. Note the many vascular lacunae (marked with X’s) that are components of the reticular network of vessels that regulates and transduces pressure within the scansors. Abbreviations as in Figure 5. Scale bar = 250y.m. 6. FIG. 7. Cross-section slightly proximal to Figure 6 showing the large central vascular sinus (VS) typical of Eurydactylodes. Note the absence of adipose tissue. Scale bar = 250m. tendinous slips from the dorsal interossei muscles. The presence of small, asymmetrical terminal plates on digit I is also a synapomorphy of the padded genera of carphodactylines (Bauer, 1986). At a more restrictive level, that of the broader padded carphodactylines, Bavayia and Eurydactylodes share the extreme flattening of intermediate phalanges. ALTERNATIVE DESIGNS The major differences in the digital morphol- ogy of Bavayia and Eurydactylodes are directly related to the scansors themselves, namely exter- nal scansor form and the internal support system of the scansors and friction pads. The autapomorphic condition of the completely divided apical plates of Eurydactylodes (Fig. 2D) seems unlikely to be of functional sig- nificance. Russell (1979) associated scansor division with the dissociation of the penultimate phalanx from the pad of the toe. In gekkonines scansor division is associated with the division of the blood sinus into two large lateral branches (Del- lit, 1934; Russell, 1976, 1979), ensuring intimate contact despite the lessening of the direct pres- Wh sure link through the penultimate phalanx. Rus- sell and Bauer (1988) have demonstrated that this pattern is also generally associated with the inception of some sort of paraphalangeal support for the lateral regions of the pad, probably as- sociated with the transmission of force to the sinus in the absence of a direct phalangeal trans- mission system. Scansors directly bencath the midline are subsequently lost with the absence ofan effective scansor control mechanism in that region, The highly arcuate penultimate phalanx of Bavayia spp. seems fo function to some degree in this manner, although here the sinus is only truly divided dorsal to the distalmost scansors, More proximally the central sinus sends two subdivisions out to the scansor pairs (Fig, 9). Although somewhat free of the pad, the penul- timate phalanges of Bavayia are much Jess inde- pendent than in some gekkonines such as Gehyra. Eurydactylodes, on the other and, with its undivided scansors possesses an undivided central sinus and the penultimate phalanx, al- though arcuate, is not strongly so. In addition to the subdivision of the scansors the «wo genera also differ in the distribution of athipose Hssue in the digit and the extent of the vascular network. In Eurydactylodes there is very liftle adipose tissue, and the posterior ex- pansion of large blood lacuna is suggestive of a posterior extension of the scansors (sensu Rus- sell, 1975). In Bavayia, on the other hand, the scansors are limijied to the area under the penul- timate phalanx and the remainder of {he ex- panded plates are filled with adipose tissue as is the entire dorsal surface of the toc. A subdigital adipose zone has been reported in the other New Caledonian genus, Rhacodactylus (Russell, 1972, 1979; Bauer, LY86) where il runs in (he midline of the tow, the area equivalent to the scansorial clefl in Bevayia, Adipose cells also fill the posterior portion of the pad in Rhacodac- tylus. In Rhacodacrlus, however, adipose tissue has not been associated with regions dorsal to the scansors in the lateral regions of the pad. The fat channel in Rhacodactylus appears to be a semi- controllable mechanism for the conformation of the pad to the substrate in the midline. The com- bination of vascular and adipose tissue as a means of control and/or mechanism of conform- ity of the scansors thus appears to be a general feature of the New Caledonian carphodactylines. (it may be independently derived in Bavayra and Rhacodactylus, or lost in Evrydactylodes). The combination of the two systems (vascular and adipose) deep to the scansors of Rhacodactylus MEMOIRS OF THE QUEENSLAND MUSELIM may be a solution to the problem of support in extremely wide (in both relative and absolute terms) digits. The problem of support in the much smaller digits of Bavayia and Eurydac- tylodes has been solved in different ways, InEun'dactylodes the digits, though wide rela- tive to the animal's size, are tiny in absolute terms. Here the sinus is so large in relation to the pad that it is able to provide active control of the scansors in the absence of adipose deposits. The less: arcuate penultimate phalanx of this taxon retains the ability to efficiently transduce forces onto a single, central vascular sinus (see Fig. 10A) which regulates pressure throughout the scansors by way of a reticular vascular network. In Bavayia pad control is achieved by pad division and concomitant vascular modification (Fig. 10B). The high arch of the penultimate phalanx probably precludes the direct use of a central adipose core as seen in Rhacodactylus, and the median scansors are Jost in favour of the dual scansor control mechanism. With the divisions of the subdigital plates and the elevated penullimate phalanx comes the requirement for lateral support of the pad, as the control mechanisms become at least partly restricted to one side of the digit or the other. Among divided- scansored gekkonines the position dorsal to the divided sinuses may be filled with adipose tissue, as in Thecadactylus or with incipient paraphalanges, as in Homepholis (Russell und Bauer, 1988). Clearly, a number of solutions to the problem of transducing pressure onto the lateral scansors are possible, Bavayia, starting with a diplodactyline or, more specifically, a carphodactyline heritage (and its coneomitant constraints) has solved the problem by maintain- ing a partially undivided sinus (thus allowing some central transduction of force from the penultimate phalanx) and by utilising a dorsal adipose zone to distribute pressure laterally. Basally in the digits, proximal to the scansors, extensive fut deposits are also present (Pig. 9) and Bavayia thus maintains at least some passive control of the mechanism of substrate conforma- tion in the series of friction plates. FUNCTION There is little information available about the biology of New Caledonian geckos, especially those of the genus Eurydactylodes. As fur as is known, both Eurydactylodes species are ex- clusively arboreal (Roux, 1913; Mcicr, 1979) and seem to prefer branches of small diameter, Bavayia cyclura dwells primarily on trees or in SCANSOR DESIGN IN BAVAYIA AND EURYDACTYLODES 307 FIG. 8. Cross-section through the proximal portion of digit IV, right pes of Bavayia sauvagii (author’s collection, AMB 267) showing the adipose zone (AZ) beneath the phalanx and above the friction plates (FP). Adipose invested connective tissue also fills much of the remainder of the digit. Scale bar = 250m. 8. FIG. 9. Cross-section through digit IV, right pes of Bavayia sauvagii (AMB 506) at the level of the penultimate phalanx. Note the division of the vascular sinus into two lateral branches on either side of the scansor cleft (SC) and the presence of loose connective tissue (LCT) containing adipose deposits above the branches of the sinus. Other abbreviations as in Figure 5. Scale bar = 250m. MEMOIRS OF THE QUEENSLAND MUSEUM Fic. 10. Diagrammatic séctions through the distal portions of the digits of A. Rhacodactylus/Eurydactylades (adipose zone would be lacking in the Jatter) type toe and B. Bavayia type toe. In A forces are transduced directly to the sinus or through the adipose zone to the sinus through the midline of the toe. An extensive reticular network of blood vessels (RN) is present. In B some transduction of pressure is direct while the remainder is spread through the adipose zone. FT = flexor tendon. Other abbreviations as in previous figures. logs and stumps (Roux, 1913; Meier, 1979; Bauer, 1986). Bavayia sauvagii, although generally perceived as arboreal, spends much of its time under rocks in terrestrial microhabitats, although animals may also live in tree holes or climb saplings at night to feed (Bauer and De- Vaney, 1987). Unfortunately our knowledge of the pedal performance requirements of scansors on different substrates is rudimentary. Further, the relative importance of claws versus scansors on substrates such as wood is unknown. Indeed, factors of safety (sensu Alexander, 1981) in digi- tal design (with respect to both claws and scan- sors) appear to vary greatly among even closely related geckos (Bauer and Good, 1986) and parts of the scansprial apparatus may be ‘overdesigned’ by more than an order of mag- nitude. Despite our ignorance, however, it is probable that divided scansors as seen in Bavayia do offer some advantage in terms of control, By passess- ing independently functioning halves of each subdigital plate, the animal is able to exert finer tendinaus and vascular control over the scansor and the scansor pairs themselves ure freer to deform to substrate irregularities, None the less, finer control of the scansors is also associated with a less direct transduction of forces onto the blood sinus and the ‘advantage’ (if any) of the divided scansor design of Bavayia overt the single scansor of Eurydactylodes or Rhacodac- tylus is difficult to determine. The anly valid assessment of the efficiency of these alternative designs would be one which was based upon performance of the morphologies in direct com- petition, Although many alternative solutions may suffice for a given problem of locomotor performance, some may be more effective than others under competitive circumstances (Rus- sell, 1976), Such situations are rare in nature but the exclusion of certain native geckos from human-commensal habitats by the introduced divided-scansored gekkonine Hemidactylus frenatus in both Hawaii (Hunsaker and Breese, 1967; McKeown, 1978) and New Caledonia (Bauer and Vindum, unpublished) may be in- dicative of the advantages of a particular digital design under certain circumstances. No such head-to-head competition appears to occur be- tween Bavayia and Eurydactylodes but it may be valid to evaluate the relative ‘success’ of the two forms by means of their geographic distribution and abundance. The species of Bavayia (actually species complexes, Ross Sadlier, pers. comm.) are distributed across all of New Caledonia and the Loyalty Islands and occupy habitats from houses and beach wrack to rainforest and savan- na, from sea level to over 1000m (Bauer, 1986). In contrast, the species of Eurydactylodes are known from scattered localities on the New Caledonian mainland and may be limited to regions of the edaphic vegetation of lateritic soils (Bauer, 1986; Bauer and Vindum, unpublished), Likewise, Bavayia is generally encountered in large numbers in the field, while Eurydactylodes is rarely found and has never been reported in high densities. Of course the patterns of species density and distribution are more than simple teflections of digital design. All other attributes of the animals’ biology, as well as the com- SCANSOR DESIGN IN BAVAYIA AND EURYDACTYLODES plexity of habitat type, and the vagaries of the search images of human collectors all combine to yield these patterns. None the less, the divided scansors of Bavayia may have played a role in the spread and habitat diversification of the genus. For the time being it seems prudent to regard the differences observed between Bavayia and Eurydactylodes as merely alternative designs for arboreal or semi-arboreal pedal function rather than specific adaptations for particular microhabitats or surface features. It is clear that both morphologies suffice for their possessors and it is likely that many other designs could also perform effectively in the same habitats. Op- timality may be a useful concept in theoretical considerations of biological phenomena but to our knowledge, animal morphologies are not, nor should they be expected to be, optimally constructed. Adaptation of the organism to its environment at this level is trivial (see Gould and Lewontin, 1979). The most (or only?) valid as- sessment of the ‘adaptation’ of alternative scan- sor designs is that which incorporates both biotic and abiotic features of the environment into the determination of selective value. Unfortunately, in the study of gekkonid morphology we are only at the stage that we can identify differences and suggest reasons (phylogenetic, functional, struc- tural or stochastic) for their existence. A com- plete analysis of the ‘meaning’ of alternative scansor design in Bavayia and Eurydactylodes must await a more detailed and fine-grained understanding of the biology of these taxa. ACKNOWLEDGEMENTS We thank Robert Drewes (CAS), Allen Greer (AMS), Eugen Kramer (NHMB), Wolfgang Bohme (ZFMK) and Nick Arnold (BMNH) for the loan of specimens in their care. Assistance in the laboratory was provided by Darcy Rae. Alain Renevier, Griff Blackmon, Kathy DeVaney, Larry Wishmeyer, Debbie Wadford and Katie Muir provided field assistance in New Caledonia. Raoul Wilson and Malcolm Swim- mer provided useful comments on earlier ver- sions of this paper. The manuscript was typed by Susan Stauffer. Funding for the completion of this work and travel to Australia to present the results was provided by a University of Calgary Postdoctoral Research Stipend to A.M.B. and a Natural Sciences and Engineering Research Council of Canada operating grant (No. A9745) to A.P.R. 309 LITERATURE CITED ALEXANDER, R. MCN. 1981. Factors of safety in the structure of animals. Sci. Prog., Oxf. 67: 109-130. BAUER, A.M. 1986. Systematics, biogeography and evolutionary morphology of the Carphodac- tylini (Reptilia: Gekkonidae). (Unpublished Ph.D. dissertation, University of California, Berkeley). 1989. Extracranial endolymphatic sacs in Eurydac- tylodes (Reptilia: Gekkonidae), with comments on endolymphatic function in lizards in general. J. Herpetol. 22: 172-175. BAUER, A.M. AND DEVANEY, K.D. 1987. Com- parative aspects of diet and habitat in some New Caledonian lizards. Amph.-Rept. 8: 349- 364. BAUER, A.M. AND GOOD, D.A. 1986. Scaling of scansorial surface area in the genus Gekko. pp. 363-366. In Rocek, Z. (ed.), ‘Studies in herpetology’. (Charles University: Prague). BRUNDIN, L. 1968. Application of phylogenetic principles in systematics and evolutionary theory. pp. 473-495. In Orvig, T. (ed.), ‘Current problems of lower vertebrate phylogeny’. (Inter- science: New York). DELLIT, W.-D. 1934. Zur Anatomie und Physiologie der Geckozehe. Jena Z. Naturwiss. 68: 613-656. GOULD, S.J. AND LEWONTIN, R.C. 1979. The spandrels of San Marco and the Panglossian paradigm: a critique of the adaptationist programme. Proc. Roy. Soc. London B205: 581- 598. HILLER, U. 1968. Untersuchungen zum Feinbau zur Funktion der Haftborsten von Reptilien. J. Mor- phol. Tiere 62; 307-362. 1969. Zusammenhang zwischen vorbehandeten Polyadthylen-Folien durch Korona-Entladung und dem Haftvermogen von Tarentola m. mauritanica (Rept.) Forma et Functio 1; 350- 352. 1975. Comparative studies on the functional mor- phology of two gekkonid lizards. J. Bombay Nat. Hist. Soc. 73: 278-282. HUMASON, G.L. 1979. ‘Animal tissue techniques.’ (W.H. Freeman and Co.: San Francisco). HUNSAKER, D. AND BREESE, P. 1967. Her- petofauna of the Hawaiian Islands. Pacific Sci. 21: 423-428. KLUGE, A.G. 1965. The Australian gekkonid lizard genus Diplodactylus Gray: an evolutionary and zoogeographical study. (Unpublished Ph.D. dis- sertation, University of Southern California, Los Angeles). 1967. Systematics, phylogeny and zoogeography 310 of the lizard genus Diplodactylus Gray (Gek- konidae). Aust, J. Zool. 15: 1007- 1108. 1987. Cladistic relationships in the Gekkonoidea (Squamata, Sauria). Misc. Publ. Mus. Zool., Univ. Michigan 173: 1-54. LAUDER, G.V. 1981. Form and function; structural analysis in evolutionary morphology. Paleobiol. 7: 430-442. 1982. Historical biology and the problem of design. J. Theor. Biol, 97; 57-67, LIEM, K.F. AND OSSE, J.W.M. 1975. Biological versatility, evolution and food resource utiliza- tion in African cichlid fishes. Amer. Zool. 15: 427-454. MCKEOWN, 8S. 1978. ‘Hawaiian reptiles and amphibians’. (Oriental Publishing Company). MEIER, H. 1979. Herpetologische Beobachtungen auf Neukaledonien. Salamandra 15: 113-139. ROUX, J. 1913. Les reptiles de la Nouvelle-Calédonie etdes iles Loyalty. pp. 79-160. /n Sarasin, F. and Roux, J. (eds). Nova Caledonia, Zoologie, Vol. 1, L. 2. (C.W. Kreidel’s Verlag: Wiesbaden). RUSSELL, A.P. 1972. The foot of the gekkonid lizards: a study in comparative and functional anatomy. (Unpublished Ph.D. dissertation, University of London). 1975. A contribution to the functional analysis of the foot of the tokay, Gekko gecko (Reptilia: Gekkonidae). J. Zool., London 176: 437-476. 1976. Some comments concerning interrelation- MEMOIRS OF THE QUEENSLAND MUSEUM ships amongst gekkonine geckos, pp. 217-244, In Bellairs, A, d’A. and Cox, C.B, (eds), ‘Mor- phology and biology of reptiles’. (Academic Press: London). 1979, Parallelism and integrated design in the foot structure of gekkonine and diplodactyline geckos. Copeia 1979; 1-21, 1981, Descriptive and functional anatomy of the digital vascular system of the tokay, Gekko gecko. J, Morphol. 169: 293-323, 1986. The morphological basis of weight-bearing in the scansors of the tokay gecko (Reptilia: Sauria). Can. J. Zool, 64: 948-955. RUSSELL, A.P. AND BAUER, A.M. 1988. Paraphalangeal elements of gekkonid lizards: a comparative survey. J. Morphol. 197: 221-240. 1989. The morphology of the digits of the golden gecko, Calodactylodes aureus (Reptilia: Gek- konidae) and the implications for the occupation of rupicolous habitats. Amph.- Rept. 10: 125- 140 UNDERWOOD, G. 1954. On the classification and evolution of geckos. Proc. Zool. Soc. London 124: 469-492. 1955. Classification of geckos. Nature 175: 1089- 1090. WASSERSUG, R.J. 1976. A procedure for differen- tial staining of cartilage and bone in whole for- malin-fixed vertebrates. Stain Technol. 54: 131-134, DENTITIONAL DIVERSITY IN RHACODACTYT.US (REPTILIA: GEKKONIDAE) AARON M. BAUER AND ANTHONY P. RUSSELL Bauer, A.M, and Russell, A,P. 1990 09 20; Dentitional diversily in Rhacodactylus (Reptilia: Gekkonidae). Memoirs of the Queensland Museum 29(2): 311-321. Brisbane. ISSN 0079- B&35, The teeth of gekkonid lizards have long (and erroneously) been considered to be simple, homodont, isodont and conical, Although some previously field beliefs about gecko dentition are largely true, the immense variability expressed among the approximately 900 species precludes the applicability of most generalities. A particularly wide range of tooth types is seen in the carphodactyline gecko genus Rhacodactylus (including those Australian forms previously assigned to the genus Pseudothecadaerylus), Rhacodactylus aurtculatis is characterised by huge caniniform teeth thatare few in number and widely-spaced. There is a steady increase with body size in the number of tooth loci in juveniles of this species (after the loss of the egg teeth). The pointed uni- or bicuspid teeth in this taxon are only slightly recurved and ure constricted at the crown base, In A, leachianus and R. trachyrhynchus the teeth are decidedly recurved and present a Jong, blade-like occlusal surface. The teeth of other members of the genus are smaller and much more numerous (up to 180+ marginal tooth loci). Teeth in these forms may be uni- or bicuspid and typically have short crowns, either conical or with moderately pronounced occlusal ridges. Dental anatomy in Rhacodactylus appears largely unrelated to phylogeny. Rather, the teeth correspond to dietary preferences. Many tooth designs are capable of processing insect prey but the caniniform teeth of R. auriculatus and the recurved teeth of some of its congeners seem to be specialisations for feeding on vertebrates and ather soft-bodied prey. Dietary data, though fragmentary, supports this interpretation. |’ Gekkonidae, Rhacodactylus, demtition, diet, phylogeny, functional morphology. Aaron M. Bauer, Bielogy Department, Villanova University, Villanova, Pennsylvania 19085, U.S.A.; Anthony P. Russell, Department of Biological Sciences, The University of Calgary, 2500 University Drive N.W,, Calgary, Alberta, Canada T2N 1N4; 16 August, 1988, The teeth of gekkonid lizards have been char- actetised as small, numerous, pleurodont, homodont, and pointed (Kluge, 1967). This generalisation has, however, not withstood the input of new data gained through more detailed analyses of dental morphology. Variation in tooth structure among the approximately 900 species of geckos is marked and has most recent- ly been evaluated by Sumida and Murphy (1987) who identified several morphologies, including both bicuspid and quadricuspid tooth crowns. Among the generalisations about gekkonid dentition that do hold true are that the teeth are pleuradont and, relative to most lizards, very numerous (see Edmund, 1969; Vorobyeva and Chugunova, 1986 for representative tooth counts for selected lizard groups). One species of gek- konine gecko, Uroplatus fimbriatus, may have more marginal teeth than any other living am- niote (Bauer and Russell, 1989), Tooth number in geckos increases with age and the concomitant increase in the length of the germinal tooth bed or dental lamina (Kluge, 1962; Edmund, 1969). Like most lizards with determinate growth, adult tooth number in geckos stabilises around a par- licular species mode (Owen, 1866) although variance may be quite high. All geckos bear teeth only on the maxillae, premaxillae and dentaries (Sumida and Murphy, 1987). Tooth size generally increases from posterior to anterior within a tooth-bearing cle- ment, but in at least one species, Teratoscincus seincus, the largest teeth are in the middle of the tooth rows (Edmund, 1969), Tooth replacement and addition of new loci is believed to proceed from back to front in waves affecting alternating positions. in accordance with Edmund's Zahnreithen theory (Edmund 1960, 1969; Os- born, 1973, 1975; Kline 1983; Kline and Cullum, 1984, 1985), although most evidence for this pattern comes from ather lizard groups- In addi- tion to the typical adult dentition, all oviparous geckos also exhibit embryonic egg teeth (Kluge, 1967) which drop out shortly after hatching (Ananjeva and Orlov, 1986), The paired condi- tion of gekkonid (including pygopodid) ces teeth appears to be a uniquely derived condition for the group (Kluge, 1967, 1987; Ananjeva and Orlov, 1986), During the course of revision of the carphodac- tyline geckos of the Southwest Pacific (Bauer, 1986) we noted a unique dental morphology characterising the New Caledonian forest gecko, Rhacodactylus auriculates - (he possession of enlarged, pointed caniniform ‘fangs’. This taxon is one of 44 in the tribe Carphodactylini, a monophyletic group of poorly known geckos endemic to Australia, New Zealand and New Caledonia. Rhacodactylus is of particular inter- est because the genus consists of moderately to very large species that are capable of taking prey types outside of the range normally available to other geckos. Further, Rhacodactylus (including (he Australian geckos previously assigned to the genus Pseudothecadactylus) occupies a fairly wide range of habitat types, from the Arnhem Land escarpment to the rainforests of eastern New Caledonia. In New Caledonia, in particular, members of the genus have few competitors and may be considered the primary non-volant predators. In light of the implications of Rhacodaetylus biology for dental form and a renewed interest in gekkonoid tooth form and function in general (e.g. Patchell and Shine 1986a,b; Sumida and Murphy, 1987) we take this oppertunity to present an analysis of the descriptive and functional anatomy of the teeth in the genus Rhacodactylus. MATERIALS AND METHODS Dentition was exammed in adult skeletal and cleared-and-stained specimens of five of the six recognised species of New Caledonian Rhacodacrylus as follows: R. auriculaius (10 specimens), &. chafroua (1 spectmen)_#. ciliatus {1 specimen), R. leachianus (1 specimen), R. MEMOIRS OF THE QUEENSLAND MUSEUM irachyriynchus (| specimen), In addition, ane cleated and stained pre-hatchling of RL auriculatus was also examined, Vascularization of the dental lamina was also observed in a number of cleared and stained adult R, auriculatus which had been injected with Microfil™ medium (see Russell et al., 1987, 1988). Radiographs of all of the above-men- tioned taxa, as well as R, sarasinorum and the Australian members of the genus, Rhacodae- tylus (Pseudothecadactylus) australis and R.(P.) lindneri, were examined and spirit-preseryed specimens of most taxa were alsa consulted. Specimens were borrowed from the collections of the California Academy of Sciences (CAS), the British Museum (Natural History) (BM(NH)), the Australian Museum (AMS) and the Zoologisches Forschungsinstitut und Museum Alexander Koenig, Bonn (ZFMK). Tooth counts for these taxa are expressed in terms of tooth loci, since teeth themselves may be lost or damaged post mortem. For compara- tive purposes reference was made to dry skeletal, cleared-and-stained and radiographic material representing all gekkonid genera except Micros- calabotes and Paragehyra. Skeletal material was examined under a binocular microscope and individual teeth were removed, sputter- coated with gold-palladium alloy and examined with an Hitachi S-450 scanning electron microscope. Photomicrographs were taken on a Wild stereo dissecting microscope with camera mount or With a 35mm camera With macro lens, RESULTS TOOTH MORPHOLOGY In Riiacodactylus chahoua, R. cilivtus, R. serasinorum and the Australian species R, (Pseudothecadactylus) australis and R. (P.) lindnert, the loath crowns are short, relatively blunt, and generally peg-like. All teeth are some- what labio- lingually compressed, with (he labial face of each tooth exhibiting a marked convex bowing. The lingual surface is somewhat more perpendicular with respect to the tooth-bearing bone, In R, eiliatus the crown is distinctly FIG. 1.4. Maxillary looth of Ritacodacolus chehewa (CAS 147764), occlusal view, Arrow in this and subsequent figures represents anterior direction, B. Same specimen as A, labial view showing asymmetrical crown height. C. Maxillary tooth of Rhacodactylus leachianus (CAS 167890), occlusal view showing elongate narrow culling ridge. D, Same specimen, labial yiew showing pronaunced recurvalure of the loath crown, E, Oblique view of maxillary tooth of Rhacodactylus auricttlatus (CAS 165891) illustrating pointed tooth crown and lateral extension of single cusp, F, Same specimen as E, occlusal view showing the dorsally directed single cusp and general smoothness of the tooth, Scale bars for all figures = 100m. DENTITION OF RHACODACTYLUS 3a bicuspid, with a shallow trough separating the higher and more pronounced labial cusp from its lingual counterpart. In R. chahoua the tooth crown is basically unicuspid (Fig. 1A), although older, more worn teeth may exhibit a weakly demarcated lingual ‘shoulder’. In the remaining taxa the teeth have weakly developed cusps or are conical with a rather blunt point. The latter condition is especially true of younger tecth at the posteriormost loci. In all taxa, but most notably in Rhacodactylus chahoua, the tooth crowns are directed somewhat posteriorly (Fig. 1B) and the posterior portion of the cusp(s) tise(s) slightly higher above the gum line. In R. chahoua the largest teeth are at the level of the posterior border of the nostril, in the anterior portion of the maxillae. The premaxillary teeth are uniformly smail, but similar in morphology to those of the maxillae and dentarics. Rhacodac- tylus sarasinorum, R. ciliatus and the Pseudo- thecadactylus species are characterised by isodonty. In the largest species of Rhacodactylus, R. leachtanus and R. trachyrhynchus, teeth at the posterior (younger) loci are generally similar in appearance to those of R. chahoua. Anteriorly, however, the teeth are elongate and their crowns ate decidedly angled. In R. Jeachianus the crown presents a very long and narrow occlusal blade (Fig. 1C). The highest (posteriormost) point of the crown projects backwards beyand the plane of the crown base as a rounded hook (Fig. 1D). There are no multicuspid tecth in this taxon. In R, trachyrhyachus the anterior teeth are only slightly curved and do exhibit a lingual ‘shoulder’ or weak cusp. As in R. chahoua the largest teeth in both &. frachyrhynchus and R. leachianus ate in the anterior part of the maxillae but in the former species, the premaxillary teeth are also quite large. A third type of morphology is exhibited by Rhacadactylus auriculatus, In this taxon all of the teeth are elongate and sharply pointed (Fig. 1E). The crown is only slightly inflected. Al- though most teeth appear to be unicuspid (Fig. 1F), a faint lingual cusp characterises some of the teeth in anterior loci of all tooth-bearing bones. Unlike other RAacodactylus species, some of the teeth of R. auriculatus are constricted at the junction of root and crown (Fig. 2). This condi- tion does not occur in the premaxillary teeth and is most evident in the teeth occupying older, anterior maxillary and dentary loci. No surface microstructures, such as grooves or serrations, Were located on the teeth of any MEMOIRS OF THE QUEENSLAND MUSEUM rm at} FIG, 2. Anterior portion of left maxilla of Rhacedac- tylus auriculatus (CAS 165891) showimg the can- striction (c) of the teeth at the junction of erown and Tool. species Rhacodactylus. Likewise, the teeth of upper and lower jaws do not normally contact one another in occlusion. TOOTH SIZE, NUMBER AND REPLACEMENT There is a two-fold difference in marginal tooth locus number in mature individuals of Rhacodactylus spp, Total locus number ranges from a mean of 105.5 in R. auriculatus to over [80 in A. sarasinerum, Figure 3 shows the dis- tribution of marginal tooth loci versus skull length for mature specimens for which total counts could be made umambiguously. Al- though sample sizes for all taxa except R. auriculatus are very small, qualitative differen- ces among species are evident. (Unfortunately the two specimens of R. frachyrhynchus ex- amined were sub- adults and thus the data for this taxon are not strictly comparable to those derived from other specimens.) Rhacodactylus sdrasinorum, R. ciliatus (Fig. 4A) and the twa Pseudothecadactylus have the smallest skull (and bady) size, yet they exhibit the greatest tooth number, The remaining, larger New Caledonian Rhacodactylus exhibit much lower tooth locus numbers (see Fig. 4B, C). {In interspecific comparisons tooth size is in- versely correlated with tooth number (see Table 1). This is.due in part to packing constraints, but also results from the relatively large diastimae that characterise the larger species, and Rhacodactylus. auriculatus in particular (Fig. 4D), DENTITION OF RHACODACTYLUS 31 _sarasinorun{1) 180: ciliatus(2) ¢ lindnen (4) 180 ott, ea - dustealig(2) pleachianus() 8 oa chahoua la) S . 6 140 & a c= 2 E uct} . 312 ¢ 0 _trachyrhyne| us| a co # auriqulalus 13) . 100 20 30 40 50 60 Mean skull length (mm) FIG. 3. Graph of mean marginal tooth loci versus mean skull length for eight Rhacodactylus species. Number of specimens of each taxon examined is listed parenthetically. Only adult specimens for which accurate tooth locus counts could be made are included (see text for exception of R. trachyr- hynchus), tn In all species, tooth replacement patterns leave at least some loci without functional marginal teeth. In these instances small replacement teeth of the next generation may be seen attached to the alveolar locus. Precise counts of the number of loci undergoing replacement at any time were not possible in the material available, as museum material, especially dry skeletal specimens, may lose teeth through the skeletonization process, storage or shipping. It is noteworthy, however, that premaxillary teeth were rarely absent and that maxillary and dentary tooth rows lacked approximately 20 - 25% of teeth in all taxa except Rhacodactylus auriculatus, in which 25 - 30% were lacking. Within a single species, such as Rhacodactylus auriculatus, tooth number increases ontogeneti- cally (Fig. 5) at least until sexual maturity. A more or less steady addition of tooth loci begins during prenatal life and continues until a body size of approximately 95mm SVL (skull length 26mm) is reached. This corresponds roughly to minimal breeding size. Other than absolute size, no differences in juvenile and adult dental mor- phology were noted in R. auriculatus. No sexual- ly related differences in either tooth size or number were noted in this, or any other Rhacodactylus species. ~ Wry tie FIG. 4, Right lateral views of the skulls of Rhacodactylus spp. illustrating variation in tooth size, number and shape. A. R. ciliatus (BM(NH) 86.3.11.4), B. R. leachianus (CAS 165890). C. R. trachyrhynchus (BM(NH) 85.11.16.7), D. R. auriculatus (CAS 165891). In D. the posterior end of the skull is somewhat distorted by the separation of the quadrate, squamosal and parietal bones. Scale bars for all figures = 10mm. 316 _ ! ° ° | r) A mee ie | Se juvenile of e nN ° oy ° ° va] 5 40 ° 3 | x ° E = g 2 30 me | 9 Q 2 ro) = & | adult 2 20 | | 0 T T T T T T T 1 10 20 30 40 Skull length (mm) FIG. 5. Graph of mean unilateral tooth loci (maxillary and dentary) versus skull length (measured from tip of snout to occipital condyle) for specimens of Rhacodactylus auriculatus. Premaxillary teeth were excluded because of the relative constancy of their number after hatching and subsequent loss of the egg teeth. Dashed line separates juveniles (including pre- hatchlings) (open circles) from adults (closed circles). EGG TEETH In all Rhacodactylus the number of premaxil- lary teeth is small and relatively stable. In the largest species and those species with high total tooth counts there are generally nine to eleven premaxillary loci while in R. auriculatus seven is the modal number. As in all oviparous geckos, embryos and hatchling Rhacodactylus possess only deciduous egg teeth in the premaxillae. These are shed within a few days of eclosion and replaced shortly thereafter by premaxillary teeth essentially identical to those of the adult. Egg teeth were examined in R. auriculatus and R. chahoua. These structures are about twice the size of the dentary or maxillary teeth. The egg teeth are broad and flattened and project anteriorly (Fig. 6). The presence of egg teeth in the viviparous species Rhacodactylus trachyr- hynchus has not been confirmed. TOOTH VASCULARIZATION A gross examination of vascularization of the mouth cavity of Rhacodactylus auriculatus MEMOIRS OF THE QUEENSLAND MUSEUM revealed that the entire oral mucosa is supplied by a dense network of blood vessels (Fig. 7). Individual tooth roots are invested with capillary cores which take their origin from the dental branches of the maxillary and mandibular arteries (vessel terminology follows O’Donoghue, 1921, and Oelrich, 1956). These arteries in turn arise from the stapedial artery which ultimately takes its origin from the inter- nal carotid. DISCUSSION PHYLOGENETIC IMPLICATIONS Gecko tooth form (Bauer, 1986) and cuspation (Grismer, 1986, 1987; Sumida and Murphy, 1987) have been employed to a limited extent in phylogenetic analysis. Unfortunately, many of the details of dental structure are difficult to examine without removing teeth for electron microscopy, and as a result only very few taxa have been adequately investigated. Relation- ships within Rhacodactylus proposed by Bauer (1986), largely on the basis of non-dental char- acters, are not strongly reflected in the evidence from tooth morphology. The sister group status of the Pseudothecadactylus species may be con- firmed by the similarity of their tooth morphol- ogy, but in the absence of comparable data from other carphodactyline geckos their shared fea- tures cannot be assumed to be apomorphic. The FIG. 6. Egg teeth of a cleared-and-stained pre-hatc- hling of Rhacodactylus auriculatus (AMB, personal collection). Note the orientation and large size of the egg teeth (e) relative to the maxillary and dentary dentition. Scale bar = Imm. DENTITION OF RHACODACTYLUS FIG. 7. Palatal view of a cleared-and-stained Microfil!™- injected Rhacodactylus auriculatus (CAS 165897) showing the dense vascularization of the oral mucosa. Tooth roots are served by dental branches of the maxillary and mandibular arteries. Scale bar = Smm. autapomorphic elongate caniniform teeth of Rhacodactylus auriculatus would appear to be unique among the Diplodactylinae, but provide no clues to affinity. Rhacodactylus leachianus and R. trachyrhynchus share some features, such as moderately great crown height and recurved crown tips, but differences in cuspation and a lack of corroboration from other characters tenders these features of little systematic value. Indeed, the diversity of tooth form among the eight taxa examined is unexpected and, if the monophyly of the Rhacodactylus/Pseudo- thecadactylus lineage is accepted, then function rather than phylogeny may be supposed to play the greatest role in the determination of tooth form. FUNCTION Dietary correlates of dental morphology have long been presented as evidence of a tightly coupled form-function complex. Mammals, in- cluding bats (Freeman, 1979, 1981, 1988) and sabre-toothed cats (Akersten, 1985), among many others, have been the subject of intensive analysis. Lizards, largely because of the mis- taken belief that their dentitions are simple, homodont and isodont have received less atten- tion (Edmund, 1969; Vorobyeva and Chugunova, 1986). None the less, iguanids (Hot- FIG, 8, Radiograph of an adult female Rhacodactylus auriculatus (AMS R 78121) illustrating the presence of a whole snail shell (s) as well as chitinous arthropod debris (d) in the digestive tract, Scale bar = 10mm. ton, 1955; Montanucci, 1968) macroteids (Presch, 1974; Dalrymple, 1979; Dessem, 1985), agamids (Cooper et al., 1970; Chugunova et al., 1987) and varanids (Lénnberg, 1903; Cowles, 1930) have received some attention in this regard, Among gekkotans functional dental mor- phology has only recently begun to be explored (Sumida and Murphy, 1987; Patchell and Shine 1986a,b - pygopodids). The diversity of dental form found within Rhacodactylus provides the opportunity to make predictions about dietary preferences on the basis of data from other lizard groups. The elon- gate caniniform teeth of R. auriculatus are typi- cal of tetrapods that feed primarily on soft-bodied prey items. Such pointed, elongate ‘fangs’ have been reported elsewhere in the teiid Callopistes flavipunctatus (Presch, 1974) and the iguanid Gambelia wislizenii (Hotton, 1955), both of which feed on vertebrate prey. Among other geckos Sumida and Murphy (1987) reported that Gekko and Chondrodactylus, again vertebrate predators, possessed upright, partly developed cusps implying that penetration rather than crushing or restraining functions are patamount in these forms. A similar morphology is seen in moth-feeding molossid bats (Freeman, 1979). The constricted ‘neck’ of Rhacodactylus auriculatus teeth (Fig. 2) may represent a preformed zone of weakness at which breakage may occur during excessive prey movement. Gecko teeth in general have been characterised as brittle (Vorobyeva and Chugunova, 1986) and high replacement rates (Edmund, 1969) have been interpreted as a compensation for this. In R. auriculatus facilitated crown loss may insure that the dental lamina itself is not damaged by trauma resulting from rapid root movement. Such breakage phenomena may be akin to the dietary-related ones evident in predatory mam- mals (Van Valkenburgh, 1988), where dental fracture is associated with occasional, unpre- dictable high stresses. Available information on diet confirms that Rhacodactylus auriculatus feeds, at least in part, on soft-bodied prey. Bauer and De Vaney (1987) recorded the remains of the small carphodac- tyline Bavayia sauvagii from the stomach of this species. The species is also known to eat insects (Bauer and DeVaney, 1987) and snails (Fig. 8) as well as flowers (Bavay, 1869) but there are no obvious dental correlates of these foods, suggest- ing that vertebrate prey (and perhaps large, soft- bodied insects) comprise the bulk of the diet. This prediction may be evaluated by analysing MEMOIRS OF THE QUEENSLAND MUSEUM the gut contents of the many specimens available in museum collections. Alternatively, however, a preformed zone of weakness may be associated with rare (Gretener 1984) but significant events and may represent an example of what may initially appear to be an ‘excessive construction’ (Gans, 1979). The recurved teeth of Rhacodactylus trachyr- hynchus and especially R. leachianus are rare among geckos. Sumida and Murphy (1987) did not acknowledge their existence, while Grismer (1986) reported them only in Hemidactylus flaviviridis, Hemitheconyx spp. Phelsuma sundbergi and Tarentola americana. Normally, recurved teeth are associated with grasping or holding prey during the ingestion cycle (see Wake and Wurst, 1977). However, such teeth are also believed to concentrate bite force at the recurved point (Gans 1974) and in non-geckos are frequently associated with multiple cuspa- tion (see Sumida and Murphy, 1987). In Rhacodactylus leachianus there is but a single cusp (Fig. 1C) and this is parallel to the tooth row. In this species the teeth appear best suited to piercing and shearing motions, again probably associated with soft-bodied prey. The large size of this gecko and its ability to overpower prey may obviate the need for the more piercing/hold- ing dentition of R. auriculatus, or alternatively it may permit finer intraoral processing of very large prey items. Little is known of the diet of this species in the wild, although small birds are known to be taken (Roux, 1913). In captivity mice (Bauer and DeVaney, 1987) and fruit (Mer- tens, 1964) are accepted. Small vertebrates, in- sects and fruit (Bauer, 1986) have been proposed as the food of Rhacodactylus trachyrhynchus but the native diet remains unconfirmed. Tooth mor- phology and the apparent ecological position of this, and other large Rhacodactylus species as dominant predators in New Caledonia, however, strongly suggest that this species could exploit saurian or avian prey. The teeth of the remaining species of Rhacodactylus, although varying in cuspation, are very small and tightly packed (Fig. 4A). Bicuspid teeth in particular, such as those of R. ciliatus, have been associated with a crushing function useful for hard bodied insects. Small blunt teeth, as in Pseudothecadactylus spp. have also been associated with insectivory, especially in regard to smaller, softer taxa (Hotton, 1955). A somewhat similar morphology characterises Coleonyx spp. (Kluge, 1962) which are known insectivores. Comparable analogues also occur DENTITION OF RHACODACTYLUS TABLE 1. Tooth number and size in adult Rhacodactylus. Maximum Maximum Crown height (mm) | Crown width (mm) Total tooth Taxon (n) Loci (x ) NEw CALEDONIAN R. auriculatus (13) R. chahoua (3) R. ciliatus (2) R. leachianus (1) R. sarasinorum (1)! R. trachyrhynchus (2y AUSTRALIAN R. (Pseudothecadactylus) australis (2) R. (P.) lindneri (4)' . . . . 2 - ' Tooth dimensions estimated from radiographs. ~ Data come from sub-adult specimens. in mammalian insectivores (see Freeman, 1979, 1988). Little is known of the diet of the smaller species of Rhacodactylus (including Pseudo- thecadactylus), but it is likely that all consume insects. In captivity Rhacodactylus chahoua ac- cepts both fruit and insects (Bauer, 1985). Rhacodactylus (Pseudothecadactylus) lindneri includes at least some vertebrates in its diet and has been observed to feed on Gehyra species in Kakadu National Park (Bauer, 1986). Although well-defined characteristic denti- tions appear to typify durophagous (Lénnberg, 1903; Dalrymple, 1979; Vorobyeva and Chugunova, 1986) and herbivorous (Hotton, 1955; Montanucci, 1968; Cooper et al., 1970; DeQueiroz, 1987) lizards, mechanical con- straints on the processing of other food types, including insects and vertebrates seem to be less testrictive. Thus dietary preference may be reflected in dentition, but, depending on size and phylogenetic limitations, several tooth mor- phologies may be equally efficacious in process- ing the same prey type (Chugunova et al., 1987). Much of the difficulty in understanding the cor- relation of diet and dentition stems from the fact that little is really known of the properties of different food types (Freeman, 1988) or of the dietary scope of most lizards. Indeed, most geckos, and certainly all Rhacodactylus, take a variety of prey types. Dentitions must not, there- fore, be expected to be optimally constructed to deal with single prey types. Rather they may be predicted to present a compromise solution to the problem of a varied diet, with an emphasis on those features most suited to the processing of food items that comprise the greatest bulk of the consumed prey, or those imposing the greatest mechanical stress (Gans, 1979; Van Valken- burgh and Ruff, 1987). Future additions to our knowledge of the diet of Rhacodactylus spp. may confirm or clarify the predictions stemming from the details of morphology discussed here. ACKNOWLEDGEMENTS We thank Drs Robert Drewes (CAS), E.N. Arnold (BM(NH)), Allen Greer (AMS) and Wolfgang B6hme (ZFMK) for the loan of skele- tal and spirit-preserved specimens used in this study. Curators of numerous other collections kindly loaned specimens for radiographic analysis. Scanning electron micrographs were prepared by Jim Alston (University of Calgary 320 Medical School). Drs B. Gueno and C, Perry provided useful comments on the manuscript. Financial support to travel to Brisbane to present this contribution at the Australian Herpetologi- cal Conference was provided by the University of Calgary. LITERATURE CITED AKERSTEN, W.A. !985. Canine function in Smilodon (Mammalia; Felidae; Machairodon- linae). Nat. Hist. Mus, Los Angeles County Con- trib. Sci. 356: 1-22. ANANIEVA, N_B. AND ORLOY, NLL. 1986, The anlage and development of the egg teeth in Squamata, pp. 319-322. Jn Rocek, Z, (ed,), ‘Studies in herpetology’, (Charles Universily: Prague). BAUER, A.M. 1985. Notes on the taxonomy, mor- phology and behavior af Rhacodacrylus chahoua (Bayay) 1869 (Reptilia: Gekkonidae). Bonn, Zoo}. Bietr, 36; 81-94. 1986. Systematics, biogeography and evolutionary morphology of the Carphodactylini (Reptilia: Gekkonidae). (Unpublished Ph.D. Dissertation, University of California, Berkeley). BAUER, A.M. AND DEVANEY, K.D. 1987. Com- parative aspects of dietand habitat in same New Caledonian lizards, Amph,-Rept. 8: 349- 364, BAUER, A.M. AND RUSSELL, A.P. 1959. A sys- tematic review of the genus U/roplatus (Reptilia: Gekkonidae), with comments on its biology. J. Nat. Hist. 23: 169-203. BAVAY, A. 1869, Catalogue des reptiles de la Nouvelle-Calédonie et description d’espéces nouvelles, Mem, Soc. Linn, Normandie 15; 1- 37. CHUGUNOVA, T. J, VOROBYEVA, E.I|. AND SEMENOV, D.¥V. 1987, Comparative analyses of the dental system and the feeding of the round- headed lizards of the genus Phrynocephalus (Reptilia, Agamidae), (In Russian]. Zool. Zhur. 66: 748-758. COOPER, J.S., POOLE, D.F.G. AND LAWSON, R. 1970, The dentition of agamid lizards with spe- cial reference to toath replacement. J, Zool., London 162: 85-98. COWLES, R.B. 1930. The life history of Varanus niloticus, as observed in Natal, South Africa. J, Entomol. Zool, 22: 3-31. DALRYMPLE, G.F. 1979. On the jaw mechanism of the snail-crushing lizards, Dracaena Daudin 1802 (Reptilia, Lacertilia, Teiidae). J, Herpetol, 13: 303-311. DE QUEIROZ, RK. 1987. Phylogenetic systematics of MEMOIRS OF THE QUEENSLAND MUSEUM iguanine lizards. A comparative osteological study. Univ. Calif. Pub, Zool, 118; 1-203. DESSEM, D. 1985. Ontogenetic changes in the den- tition and diet of Tupinambis (Lacertilia: Teiidae), Copeia 1985; 245-247. EDMUND, A.G. 1960. Tooth replacement phenomena in the lower vertebrates, Roy, On- (aria Mus., Life Sci. Diy. Contrib, $2: 1- 190. 1969. Dentition. pp. 117-200, Jn Gans, C., Bellairs, A.d’A. and Parsons, T.S. (eds), ‘Biology of the teptilia’, Vol. 1. (Academic Press: London). FREEMAN, P.W. 1979, Specialized insectivory: beetle-ealing and math-ealing molossid bats. J, Mammal. 60: 467-479, 1981. Correspondence of food habits and morpho]- ogy in insectivorous bals, J, Mammal. 62: 165- 173, 1988. Frugivorous and animalivorous bats (Microchiroptera); dental and cranial adapta~ tions, Biol, J, Linn, Sac, 33; 249-272, GANS, C. 1974, "Biomechanics, an approach to ver- tebrate biology’, (J,B, Lippincott Co: Philadel- phia). 1979. Momentarily excessive construction as the basis for protoadaptation. Evolution 33; 227- 233. GRETENER, P.E. 1984. Reflections on the “rare event" and related concepts in geology. pp. 77- 89. /n Berggren, W.A. and Van Couvering, J.A. (eds), ‘Catastrophes and earth history, the new Uniformitarianism'. (Princeton University Press: Princeton, New Jersey). GRISMER, L.L. 1986. The phylogeny, taxonomy, classification, and biogeography ot eublepharid geckos (Reptilia: Squamata). (Unpublished M.S. thesis, San Diego Stale University). 1987. Evidence for the resurrection of Goniurosaurus Barbour (Reptilia: Eubleph- aridae) with a discussion on geographic varia- tion in Gonturosaurus lichtenfelderi. Acta Herpetol, Sinica 6: 43-47, HOTTON, N. 1955. A survey of adaptive relation- ships of dentition to diet in North American iguanidae, Amer. Mid}. Nat. 53: 88-114. KLINE, L.W. 1983, The tooth replacement phenomenon and growth in the green iguana, Iguana iguana, Experientia 39: 595-596. KLINE, L.W. AND CULLUM, D. 1984. A long term study of the tooth replacement phenomenon in the young green iguana, /guana iguana. J, Her- petol. 18: 176-185. 1985. Tooth replacement and growth in the young green iguana, /gvana iguana, J. Morphol. 186: 265-269. KLUGE, A.G. 1962. Comparative osteology of the DENTITION OF RHACODACTYLUS 321 eublepharid lizard genus Coleanyx Gray. J. Mor- phol. 110: 299-332. 1967. Higher taxonomic categories of gekkonid lizards and their evolution. Bull. Amer, Mus. Nat. Hist. 135: 1-59. 1987. Cladistic relationships in the Gekkonoidea (Squamata, Sauria). Misc. Pub. Mus. Zool. Univ. Michigan 173: 1-54. LONNBERG, E. 1903, On the adaptations to a mol- luscivorous diet in Varanus niloticus, Atk. Zool. 1; 45-83, MERTENS, R. 1964. Neukaledonische Riesengeckos (Rhacadactylus). Zool, Garten, Leipzig, N.F. 29; 49-57. MONTANUCCL, R.R. 1968. Comparative dentition in four iguanid lizards. Herpetologica 24: 305- 315. O’ DONOGHUE, C.H. 1921. The blood vascular sys- tem of the Tuatara, Sphenodon punctatus, Phil. Trans. Rey. Soc, London Ser. B 210: 175-252. OELRICH, T. 1956. The anatomy of the head of Clenosaura pectinata (Iguanidae). Misc. Pub. Mus. Zool,, Univ. Michigan 94: 1-122, 57 figs. OSBORN, J,W. 1973. The evolution of dentitions. Amer. Sci, 61; 548-559, 1975. Tooth replacement: efficiency, patterns and evolution. Evolution 29; 180-186, OWEN, R. 1866. ‘On the anatomy of vertebrates’. Vol. 1. (Longman, Green & Co: London). PATCHELL, F.C. AND SHINE, R, 1986a. Hinged teeth for hard-bodied prey: a case of convergent evolution between snakes and legless lizards. J. Zool., London A 208: 269-275. 1986b. Feeding mechanisms in pygopodid lizards: how can Lialis swallow such large prey. J. Her- petol. 20: 59-64. PRESCH, W. 1974. A survey of the dentition of the macroteiid lizards (Teiidae: Lacertilia), Her- petologica 30; 344-349, ROUX, J. 1913. Les reptiles de la Nouvelle-Calédonie et des iles Loyalty. pp. 79-160. /n Sarasin, F. and Roux, J, (eds), Nova Caledonia, Zoologie. Vol. 1, L.2. (C.W. Kreidel’s Verlag: Wiesbaden). RUSSELL, A.P., BAUER, A.M., AND WALKER, R.L. 1987. Microfil injections of small ver- tebrates: an adjunct to clearing and staining. Collection Forum 3; 3-5. RUSSELL, A.P., WALKER, R.L. AND BAUER, A.M. 1988. A simple technique for visualization of the circulatory system in small lizards. Copeia 1988; 797-800. SUMIDA, S.S. AND MURPHY, R.W. 1987. Form and function of the tooth crown structure in gekkonid lizards (Reptilia, Squamata, Gek- konidae), Can. J. Zool, 65: 2886-2892. VAN VALKENBURGH, B, 1988, Incidence of tooth breakage among large, predatory mammals. Amer. Nat. 131: 291-302. VAN VALKENBURGH, B. AND RUFF, C.B. 1987. Canine tooth strength and killing behavior in large carnivores. J. Zool., London 212: 379-397, VOROBYEVA, E.I. AND CHUGUNOVA, T.J, 1986. The dental system in lizards. An integrated approach. pp. 315-318. Im Rocek, Z. (ed.), ‘Studies in herpetology’. (Charles University: Prague). WAKE, M.H. AND WURST, G.Z. 1977. Tooth crown morphology in caecilians (Amphibia: Gymnophiona). J. Morphol. 159: 331-342. a th tN WALUNARRA, BUNGARRA MALI AND THE GANGALIDDA AT OLD DOQMADGEE:-— Near Old Doomadgce (16°57’S, 138°49’E), the country of the Gan- galidda people, in the monsoon tropics, there are two species of freshwater turtles: walunarra, the ‘mud turtle’, is a species of Chelodina (possibly C, rugosa, possibly an undescribed species) and bungarra mali, the ‘stinking turtle’, is Chelodina novaeguineae. In the area, there is no permanent freshwater although lagoons provide plentiful surface water for much of the year. Al varying times during the dry season, which usually begins in June-July, the lagoons become rock-hard clay beds, The rate of drying depends on size, shape, depth, substrate, vegelation, previous rains and temperatures. Within a radius of 30km of Old Doomadgee, there are some ten lagoons. Both Walunarra and bungarra mali occur in all of these waters and they have been collected as a prized food source by the Gangalidda for as long as is remembered. As the waters of the lagoons recede during ‘the dry’, walunarra digs into the mud lo aestivate and bungarra mali migrates to other longer- lasting lagoons. Avan unnamed lagoon 3km SW of Old Doomadgee, three of us (JC, PC, KM) recently accompanied Major Walden, a senior Gangalidda, while he collected a specimen of wal- nunarra from its aestivation site. The site was well concealed both by the low, thick, sharp foliage and by the tangled root system of Melaleuca acactoides, which fringes many of the lagoons in the area. To the untrained eye, surface evidence ofan aestivating turtle is difficult to find. The mound af mud above the sile is some 60-70mm high, It is a whorl and resembles those made by freshwater crabs. However, they are much less common and they lack the central hole (diameter 2U-30mm) of those made by crabs, As well, al the base of the turtle’s mud-whorl is a small air hole (diameter about 3- Smm), used by the turtle for breathing. As Major Walden dug into the rock-hard mud below the whorl, the digging disturbed the aestivating turtle and it MEMOIRS OF THE QUEENSLAND MUSEUM emitted a sudden “whoosh’ of air; this is characteristic be- haviour, The turtle was located, head down, eyes-closed (they appeared ‘sealed’ against desiccation) at a depth, from rear of shell to ground surface, of approximately 120mm. The luttle was not vertically aligned. Rather, it rested on the diagonal al an angle of about 30°. A second specimen was located in the same way, under very similar condilions be- tween the surface roots of a large Melaleuca leucodendra close to the lagoon edge. When we visited the area (June, 1990) water levels were still high in the lagoons, although they were drying up quite rapidly, Inthe previous weeks, bungarra mali had been found, apparently moving lo new, temporary lagoons — something that has been observed in the area for many years (Alice Ned, pers. comm.) at roughly the same lime each year. Both species of lurtles are utilised as food by the Gangalid- da. Major Walden and others have supplied us with details of the method of cooking walunarra, (We have no data on whether the same procedure is followed for bungarra mali). The turtle is killed by ringing its neck. The neck is then cut lo expose its wind pipe into which air is blown. (In early times, the live turtle was held so air could be blown into its mouth and nose). The wind pipe is then tied to keep the air in soa ‘cushion’ exists to separate internal organs from the shell to prevent their becoming stuck to the shell during cooking. If air is not blown into the turtle, the carapace and plastron are cracked to ensure successful cooking. J. Covacevich, P. Couper, Queensland Museum, PO Box 300, South Brisbane, Queensland 4101, Australia; K. McDonald, National Parks and Wildlife Service, Townsville, Queensland 4810, Australia; D. Trigger, Department of Anthropology, University of Western Australia, Perth, Nedlands, Perth, Western Australia 6009, Ausiralia; 14 August, 1990. MOLECULAR EVOLUTION IN AUSTRALIAN DRAGONS AND SKINKS: A PROGRESS REPORT P.R. BAVERSTOCK AND S.C. DONNELLAN Baverstock, P.R. and Donnellan, S.C. 1990 09 20: Molecular evolution in Australian dragonsand skinks: a progress report. Memoirs of ihe Queensland Museunt 29(2): 323-331, Brisbane, ISSN 0079-8835. We have been using microcomplement fixation of albumin to assess the evolutionary téelationships of the dragons and skinks of Australia, and to provide approximate dates of divergence of extant taxa. The results are preliminary, bul sugges! the fallowing salient features. For the dragons: (1) The amphibolurid radiation is very recent, less than 20 MY BP; (2) Moloch is a part of the amphibolurid radiation; (3) the Australasian Gonocephalus are much more closely related to the amphibolurids and Physignarkus than to Asian Gonocephalus; (4) the divergence of the amphibolurids, Physignathus and Australasian Gonocephalus occured in the mid-Miocene; (5) The Australasian agamids (including Gonocephalus and Physignathus) are closer to the African Agama than apy Asian dragon so far tested, For the skinks: (1) The data are in accord with Greer’s (1979) recognition of ihree groups of skinks in Australia, diverging about 60 MYBP; (2) The genus Leiolopisma is paraphyletic with the genera Lamprapholis, Carlie, Menetia and Marethia; (3) The New Zealand Leiolopisma fall within the Australian Lefolopisma with a divergence time of about 20 MYBP. C) Dragon, skink, microcomplement fixution, molecular clock, biogeography, P. R. Baverstock and §. C. Dennellan, Evalutionary Biology Unit, South Australian Musexm, North Terrace, Adelaide, Souily Australia, 5000, Australia; P.R.B. present address: School of Resource Science and Management, University of New England, Northern Rivers, PO Box 159, Lismore, New South Wales, 2480, Australia; 17 August, 1988. Five families of lizards occur in Ausiralia - the Agamidae (dragons), Scincidae (skinks), Varanidae (goannas), Gekkonidae (geckos), and Pygopodidae (legless lizards). Of these, only the Pygopodidae are endemic to Australasia- The last 15 years have seen enormous changes in our understanding of the generic and specific limits of Australian lizards, as a comparison of Worrell’s (1963) book with Cogger’s. (1986) book reveals. Despite this work, the evolutionary origins and relationships among genera are offen poorly known, and subject to very varied opinions (¢,g. Tyler, 1979; Greer, 1979; Cogger and Heatwole, 1981; Witten, 1982). This uncer- tainty results from paucity of suitable mor- phological characters, high level of homoplasy in some groups, use of principally non-cladistic analyses, and paucity of fossils. It is in such areas that molecular genetic techniques can prove ex- tremely valuable. The molecular genetic approach to systematics and biogeography has two major contributions to make. Firstly, it provides a view of the evolu- tionary relationships of a group that is totally independent of that provided by morphology. This does not mean that it is the panacea for all problems in systematics. Rather, molecular genetic data should be seen as challenging estab- lished ideas about the evolution of a group. and highlighting areas of discrepancy. Secondly, there is mounting evidence that molecular genetic techniques can be used to provide a time-frame, albeit approximate, for the cladistic events in the evolution of a group (Wilson et al., 1977; Thorpe, 1982; Ayala, 1986), Over the past several years, we have been using the molecular genetic technique of microcomplement fixation (Champion et al., 1974) to assess molecular evolution in the Australian lizards, The study of the Varanidae with D. King and M. King is completed and will be published separately, while our work on the Gekkonidae and Pygopodidae has barely begun. However, our data on the Agamidae and Scin- cidae. although incomplete, are sufficiently ex- lensive to provide a rough picture of their evolution in Australasia. We have taken the op- portunity of the Bicentennial Herpetology Con- ference to present our preliminary data on these groups. Some aspects of the work we report here 324 on the Scincidae has involved S. Burgin, M. Hutchinson and C. Daugherty. MATERIALS AND METHODS Albumin was purified from plasma by disc electrophoresis and injected into rabbits (three per antigen) over a period of three months ac- cording to the schedule of Champion et al. (1974). Purity of antisera was checked by im- munoelectrophoresis. The microcomplement fixation procedure followed the protocol of Champion et al., (1974). The results of cross- reactions are reported as Albumin Immunologic Distances (AIDs). One AID is roughly equivalent to one amino-acid substitution (Max- son and Wilson, 1974). RESULTS THE AGAMIDAE Antisera were raised to six species of Australian agamids - Ctenophorus vadnappa, Pogona barbata, Lophognathus gilberti, Moloch horridus, Gonocephalus bruynii and Physig- nathus lesueurii. The full reciprocal matrix for these six taxa was corrected for reciprocity by the method of Cronin and Sarich (1975). The standard deviation for reciprocity (Maxson and Wilson, 1974) was 21.8% before correction and 8.2% after correction. The corrected reciprocal matrix is shown in Table 1. Also shown in Table 1 are the results of the one-way reactions to a range of other agamids from Australia, New Guinea, Asia, and Africa and two iguanids from North America. The reciprocal data were used to produce an unrooted tree by the Fitch-Margoliash method (Fitch and Margoliash, 1967), using the PHYLIP 2.7 package written and kindly supplied by J. Felsenstein. To root this tree, an outgroup is needed. The outgroup must be close enough to be able to detect differential rates of evolution in the ingroup, but far enough away to be sure that it is an outgroup. The taxa tested for suitability as Outgroups wereA gama aculeata, Calotes tym- panostriga, Dipsosaurus dorsalis and Iguana iguana (Table 1). Of these, only Agama aculeata was close enough to be useful as an outgroup. Because we do not have immunological dis- tances of all antisera to A. aculeata, it was not possible to produce a rooted tree for the Australian agamids using the Fitch-Margoliash criterion. However, we added A. aculeata to the tree by optimising the four distances available MEMOIRS OF THE QUEENSLAND MUSEUM (Table 1). The resulting rooted tree for the Australasian agamids is shown in Fig. 1. This tree should be treated as provisional since it is based on incomplete data for A. aculeata, and has not been tested for robustness by jackknifing (Lanyon, 1985) . On the tree in Fig. 1, Moloch stands apart from the amphibolurids represented (Pogona, Ctenophorus and Lophognathus). However, the one-way reactions to other am- phibolurids (Chlamydosaurus and Diporiphora) suggest that these genera fall outside a Moloch/Pogona/Ctenophorus/Lophognathus clade (Table 1). If this is true (and it needs to be tested by antisera to Chlamydosaurus and Diporiphora), then Moloch may in fact be part of the amphibolurid radiation. Moreover, again based on the one-way distance to Chlamydosaurus and Diporiphora, Physig- nathus lesueurii may be closely related to this clade. A second feature of the one-way cross-reac- tions shown in Table 2 are the albumin distances to non-Australasian taxa. Of all the taxa tested, the African Agama is much closer to the Australasian agamids than are the Asian agamids, including, significantly, Gonocephalus kuhili. THE SCINCIDAE Antisera have been raised to 10 species of Australian skinks, A partial reciprocal matrix for these 10 species is shown in Table 2. Table 2 also shows the results of cross-reactions of these 10 antisera to a range of other skinks, Because the reciprocal matrix is as yet incomplete, it is not possible to correct for reciprocity by the method of Cronin and Sarich (1975), nor to construct phylogenetic trees by the Fitch-Margoliash method. Nevertheless, a number of perhaps un- expected features emerge from the limited data available. They are: (1) The genus Lampropholis is highly diverse at the albumin level. AIDs among members of the genus range up to 29, which is as high as that characterising the entire amphibolurid radiation (see Table 2). Indeed, the species separated by 29 AIDs are La. basiliscus and La. challengert, which are sibling species. (2) The genus Leiolopisma is even more diverse at the molecular level, with AIDs up to 40! Indeed it is clear that the genus is not monophyletic. Some species of Leiolopisma (entrecasteauxil, pretiosum, palfreymani and metallicum) are closer to Lampropholis and Car- lia than to other Leiolopisma, while Le. duper- MO1ECULAR EVOLUTION IN AUSTRALIAN DRAGONS AND SKINKS 325 TABLE 1, Albumin immunologic distances (corrected for reciprocity) of antisera to six species of Australian agamids to a range of other agamids and iguanids. The standard deviation for reciprocity was 21.8% before correction and 8.2% after correction, CF is the correction factor. Antibody Antigen Ctenophorus vadnappa (Cv) Pogona barbata (Pb) Lophognathus gilberti (Lg) Moloch horridus (Mh) Gonocephalus bruynii (Gb) Physignathus lesuenrii (P|) Antigens only Tympanoeryptis intima Chlamydosaurus kingii Diporiphora bilineata Gonocephalus modestus Gonacephalus kuhli Caloies tympanostriga Agama aculeata Dipsasaurus dorsalis Iguana iguana reyi is Closer to Menetia and Morethia than to other Leiolopisma. The New Zealand Le. grande forms a third group. (3) The Eugongylus group of Greer (1979), here represented by Eugongylus, Carlia, Lampropholis, Leiolopisma, Menetia, Morethia, Cryptoblepharus and Emoia, appears to form a monophyletic group to the exclusion of Egernia, Tiliqua, Sphenomorphus, Ctenotus, Mabuya, Lamprolepis, Tribolonotus and, perhaps, Mabuya. (4) Of the non-Eugongylus group species, Egernia and Tiliqua are close, but we have no data yet on possible relationships among other species. DISCUSSION THE AGAMIDAE Current views of the biogeographical history of Australian and New Guinean agamids are highly disparate in some areas (cf. Tyler, 1979; Cogger and Heatwole, 1981; Witten, 1982). Briefly summarised, all schemes agree that there ig an endemic component which is referred to as the amphibolurid radiation but whose composi- tion varies between authors, and a group of genera (Physignathus, Gonocephalus and Geographic origin Australia Australia Australia Australia New Guinea Australia Australia Australia Australia New Guinea Asia Asia Africa North America North America Chelosonia) which arose from Asian ancestors and have entered Australia recently from New Guinea. The phylogenetic relationships of Moloch are not known with certainty, due to its highly autapomorphic morphology. Moloch has been considered as either the first agamid to have entered Australia and hence phylogenetically outside the amphibolurid radiation (Cogger and Heatwole, 1981), or as an embedded member of the endemic radiation (Witten, 19&2). The al- bumin data support the latter, and moreover sug- gest that Moloch is well embedded in the amphibolurid radiation. Thus the hypothesis of a separate entry into Australia by Moloch 1s not supported by our data. The origin of the supposedly Asian-derived species of Gonocephalus and Physignathus is also questioned by the albumin data, Most proposals in this area appear to have been strong- ly influenced by the current taxonomy. The al- bumin data suggest that the current taxonomy does not reflect the phylogenetic relationships of species in these genera. The New Guinean Gonocephalus available to us ate much more closely related to the amphibolurids than to the Asian Gonocephalus kuhli. Similarly, Physig- nathus lesueurti is much mare closely related to —___| 10 units 40 30 20 MEMOIRS OF THE QUEENSLAND MUSEUM Pogana barbata Ctenophorus yadnappa Lophognathus gilberti Moloch horridus Physignathus lesueurti Gonocephalus bruynii Agama aculeata 10 0 Approximate Age (Millions of years) FIG.1. Phylogenetic tree, constructed by the Fitch-Margoliash method, among the six species of agamids to which antisera were raised. The tree was rooted using A gama acileata as an outgroup. Branch lengths shown are proportional to the proposed amount of albumin change along each branch. An approximate time scale is given assuming T=0.6D. the amphibolurids than to the available Asian genera. The only other member of the genus, P. cocincinus, is found in Indochina, and may not be very closely related to P. lesueurii (Witten, 1982). Hence the proposed recent Asian origin for these genera must be questioned critically in the light of the albumin data. Taken at face value the tree in Fig.1 shows clear evidence that rates of albumin evolution within the Australasian agamids have been reasonably uniform among lineages. From the node common to all Australasian agamids, the range in amounts of albumin evolution vary from 12 units to Moloch horridus to 22 units to Ctenophorus vadnappa, a \ess than two-fold range. It is therefore appropriate to use a molecular clock for this data set. However, we need to calibrate the clock for agamids. Usually, such a calibration relies on obtaining from fossils an estimate of the age of at least one and preferably two cladogenic events in the history of the group. In order to date cladogenic events from fossil data, three requirements must be met. Firstly, the fossil must be well-dated, secondly, the fossil must be sufficiently well-preserved to be placed in a phylogenetic framework; thirdly, and most importantly, the systematics of extant forms must be well-established. Unfortunately, none of these requirements can be met for Australian agamids (Molnar, 1984). The relationship T=0.6D (where T=time in millions of years and D=albumin immunologic distance) has been used frequently in the litera- ture for a wide range of vertebrates including eutherians (Sarich, 1985), marsupials (Maxson et al., 1975), lizards and crocodiles (Gorman et al., 1971) and snakes (Cadle and Sarich, 1981), although usually without specifically calibrating the clock for the group in question. In the majority of cases, such a calibration has proved to be compatible with what limited available data there are for the group in question. Recently, however, Sarich (1985) has suggested that a relationship of T=0.37D is more appropriate for MOLECULAR EVOLUTION IN AUSTRALIAN DRAGONS AND SKINKS 327 TABLE 2. Albumin immunologic distances of antisera to 10 species of Australian skinks cross-reacted to a range of other skink species. The data are uncorrected. Antibody Geographic Lac Lag Lab Lad Ma _ Lep Lea Lee Led Ef origin Antigen Lampropholis challengeri (Lac) Lampropholis guichenoti (Lag) Lampropholis basiliscus (Lab) Lampropholis cf. delicata (Lad) Morethia adelaidensis (Ma) Leiolopisma pretiosum (Lep) Leiolopisma palfreymani (Lea) Leiolopisma entrecasteauxii (Lee) Leiolopisma duperreyi (Led) Egernia frerei (Ef) Australia Australia Australia Australia Australia Australia Australia Australia Australia Australia 2 2 1 = RBAONnNowW p | tN an Antigens only Leiolopisma metallicum Leiolopisma zia Leiolopisma grande Cryptoblepharus plagiocephalus Carlia rostralis Menetia greyi Emoia longicauda Australia Australia New Zealand Australia Australia Australia Australia/ New Guinea New Guinea Australia Australia Indonesia New Guinea New Guinea Australia Australia Eugongylus rufescens Sphenomorphus murrayi Ctenotus grandis Mabuya multifasciata Lamprolepis smaragdina Tribolonotus gracilis Egernia kingit Tiliqua rugosa eutherian mammals, although Baverstock et al. phibolurid radiation (including Moloch) and the (1989) have shown that such arelationshipis not second to New Guinean Gonocephalus and appropriate for marsupials. We herein use Physignathus, which recently entered Australia. T=0.6D, although this relationship may need to The morphological similarity of Asian and New be adjusted if and when relevant fossil datacome Guinean Gonocephalus and Physignathus is to hand. then seen to be due to convergence. Using Fig. 1. shows an approximate time-scale for T=0.6D gives a divergence time of Agama from the Australasian agamid radiation, using Australasian agamids of 40MY. This is much too T=0.6D. On this analysis, the three am-_— short fora Gondwanan connection, butit is based phibolurids represented form a monophyletic on only one species of African agamid, one-way group, radiating in the late Miocene-early cross- reactions, and an untested calibration of Pliocene. However, the one-way data (Table 1) T=0.6D. If the Australasian agamids do have an suggest that the radiation involving other am- Asian origin, then the possible sister taxa are not phibolurids (and Moloch) occurred a little ear- Asian Gonocephalus and Asian Physignathus, lier, perhaps mid-Miocene, and that the and have not been included in our analyses. Australian Physignathus and Gonocephalus also diverged about this time. THE SCINCIDAE Based on the phylogenetic relationships indi- Most inferences of the biogeographic history cated by Fig.1 and Table 1, it is tempting to of Australian skinks are based on the distribution speculate that the Australasian agamids do in- of extant forms (the fossil record is virtually deed have a Gondwanan origin. Onthisscenario, non-existent) and estimates of the time scale of the Australasian component gave rise to radia- cladistic events from comparisons of levels of tions in two land masses, Australia and what was faunal diversity (Greer, 1979; Tyler, 1979; Cog- to become New Guinea. The first gave the am- ger and Heatwole, 1981). It has been amply 328 MEMOIRS OF THE QUEENSLAND MUSEUM Carlia rostralis FIG.2. Inferred phylogenetic relationships among some Australasian skinks based on the raw data in Table 2. A very approximate time scale is given Lampropholis challengeri assuming T=0.6D. Lampropholis guichenoti Lampropholis basiliscus Lampropholis sp. Leiolopisma entrecasteauxii Leiolopisma pretiosum Leiolopisma palfreymani Leiolopisma metallicum Leiolopisma grande Leiolopisma duperreyi Menetia greyii Morethia adelaidensis \. \ Emoia longicauda \ \ Eugongylus rufescens Tiliqua rugosa Egernia frerei 60 40 20 0 Approximate Age (Millions of years) MOLECULAR EVOLUTION IN AUSTRALIAN DRAGONS AND SKINKS demonstrated that speciation and morphological clocks do not exist (Baverstock and Adams, 1987) and hence estimates of time based on these are purely speculative, Additionally, if the cur- rent systematics does not accurately: reflect the phylogenetic relationships then inferences based on the distribution of such groups can be er- ron¢ous, Greer (1979), Tyler (1979) and Cogger and Heatwole (1981) concur that skinks arose north of Australia’s present day position and Greer (1979) and Cogger and Heatwole (1981) propose that the ancestors of the scincid radia- tion entered Australia at least twice. Cogger and Heatwole (1981) suggest that the earliest in- vaders were here by at [cast {he mid-Tertiary. The finding of fossil cranial elements. from the mid- Miocene referable to the extant genus Eger- nia (Estes, 1984) at least gives a minimum age of entry which is compatible with this view, The microcomplement fixation data suggest some anomalies in the current systematics und provide a rough estimate of the timing of evolu- tionary events, However, our data are as yel not extensive cnough at the suprageneric level to provide information relevant to the evolutionary origins of the skink fauna of Australia. Fig.2 is summary cladogram of the relationships among some Australian skinks that seem to be indicated by the data in Table 2. We do stress however that these proposed relationships are very tentative, and will undoubtedly be refined as additional antigens and antisera are added to the data set, We have also added a very approximate lime scale assuming T=0.6D is an appropriate calibra- tion for the Australian skinks. While our data provide strong support tor a monophyletic Lugongy/us group, they are af odds with Greer’s (1979) concept of two sub- groups within the Exgangylus group. I indeed there are two subgroups present then theit com- positions are vastly different from those con- ceived by Greer (1979), Several authors concur that the genus Leiolepisma is composite (Rawlinson, 1974; Greer, 1982). Qur data demonstrate that this is so, but the groups delineated do not agree with previous schemes. £. duperreyi is more closely related to Moretiia and Menetia than to other Lejelopisma, Greer (1980) had previously suggested such a relation- ship but later included Lb. duperreyi in bis Lb, baudine species proup which included L. entrecasteauxié and £, metallicum, species not especially related by the microcomplement fixa- tion data. 349 Hutchinson (1980) from immunoelectro- phoretic comparisons and a reappraisal of Greer's (1979) morphological assessment sug- gested that the spincy skinks of the genus Tribolonotus are probably closest to the Eugon- gylus group. While our data on Tribolonotus are based at this stage on one-way comparisons, they do not provide strong support for such a view, and instead suggest that Tribolonotus is at least as divergent from the Eugongylus group as Eger- nia and Lamprolepis. While the present study shows that the genus Leijolopisma is at least paraphyletic, nevertheless the New Zealand representative of the genus available (o us (Le. grande) is clearly a member of the Eugongylus group, with a divergence time {rom its nearest Australian relatives of about 20 MYBP. Thus a Gondwanan origin for New Zealand Lefolepisma is clearly rejected by the albumin data, which suppon Hardy's (1977) view of a more recent invasion of New Zealand from Australia. CONCLUDING REMARKS We have heen struck by the highly disparate pattern of morphological and molecular genetic evolution in the Australian skinks and agamids, In the skinks. morphologically similar species are nevertheless highly divergent at the molecular level. This feature is emphasised in the genus Lampropholis, where sibling species have albumins that differ by up to 20 amino- acids. By contrast, the agamids show mor- phological diversity in the facc of relative unifor- mity at the albumin level. Species as diverse at the morphological level as bearded dragons (Pogona), \horny devil (Moloch), and frilled- neck lizard (C/amydeasaurus) are nevertheless as similar al the molecular level as sibling spectes of Lampropholts, These contrasts highlight the vast disparity between morphological evolution and molecular evolution, a feature which has been noted in other groups (e.g. Maxson and Wilson, 1974; Wilson ct al,, 1977; Bayerstock and Adams, 1987). While rates of molecular evolution may vary a little between different groups (perhaps two- or threcfold: see Brownell, 1983; Wu and Li, 1¥85), itis apparent that cates of morphologi- cal evolution can vary enormously between groups. Thus estimates of divergence times and biogeographic reconstructions that rely upon considerations of morphological diversity alone are unlikely to be valid. tod wey i=) ACKNOWLEDGEMENTS We thank the numerous people who have con- tributed to this study by supplying specimens or blood samples, especially W. Branch, D. Broad- ley, G. Johnston, G. Shea, R. Sadlier, A. Greer, G. Mengden, C. Moritz, M. Hutchinson, N. Brothers, T. Schwaner, S. Burgin, C. Daugherty, R. Jenkins and S. Morton. Various people have been involved in the rather boring task of per- forming the MC’F cross-reaction reported here including M. Hutchinson, S. Burgin, G. Sims, J. Nancarrow, M. Cotsios and C. Hefford. M. Krieg cheerfully prepared the antisera. We thank P. Kidd for typing the manuscript, and J. Riede for preparing the figures. J. Felsenstein supplied the PHYLIP programme, A. Gunjko helped set the programs up. M. Adams wrote the programme for correcting reciprocity. This project is sup- ported by an Australian Research Grants Scheme Grant (No. D18416251). LITERATURE CITED AYALA, F.J. 1986. On the virtues and pitfalls of the molecular evolutionary clock. J. Hered, 77: 226- 235. BAVERSTOCK, P.R. AND ADAMS, M. 1987. Comparative rates of molecular, chromosomal and morphological evolution in some Australian vertebrates, pp. 175-188. Jn Campbell, K., and Day, M. (eds), ‘Rates of evolution.’ (Allen and Unwin: London). BAVERSTOCK, P.R. RICHARDSON, B.J., BIR- RELL, J, AND KRIEG, M, 1989. Albumin im- munologic relationships of the Macropodidae (Marsupialia). Systematic Zoology. 38: 38-50. BROWNELL, E. 1983. DNA/DNA hybridization studies of murid rodents: symmetry and rates of molecular evolution. Evolution 37: 1034-1051. CADLE, J.E. AND SARICH, V.M. 1981. An im- munological assessment of the phylogenetic position of New World coral snakes. J. Zool., Lond. 195; 157-167. CHAMPION, A.B., PRAGER, E.M., WACHTER, D. AND WILSON, A.C, 1974. Microcomplement fixation. pp. 397-416. In Wright, C. (ed.), ‘Biochemical and immunological taxonomy of animals.’ (Academic Press: New York). COGGER, H. 1986. ‘Reptiles and amphibians of Australia.’ 4th edit. (Reed: Sydney). COGGER, H. AND HEATWOLE, H. 1981. The Australian reptiles: origins, biogeography, dis- tribution patterns and island evolution. pp. 1333- MEMOIRS OF THE QUEENSLAND MUSEUM 1373. Jn Keast, A. (ed.), ‘Ecological biogeog- raphy of Australia.’ (W. Junk: The Hague). CRONIN, J.E., AND SARICH, V.M. 1975. Molecular systematics of the New World monkeys. J. Hum. Evol. 4: 357-375. ESTES, R. 1984. Fish, amphibians and reptiles from the Etadunna formation, Miocene of South Australia. Aust. Zool. 21: 335-343. FITCH, W.M. AND MARGOLIASH, E. 1967. Con- struction of phylogenetic trees. Science 155: 279-284, GORMAN, G.C., WILSON, A.C. AND NAKANISHI, M. 1971. A biochemical ap- proach towards the study of reptilian phylogeny: evolution of serum albumin and lactic dehydrogenase. Syst. Zool. 20: 167-186. GREER, A. 1979. A phylogenetic subdivision of Australian skinks, Rec. Aust. Mus, 32: 339-371. 1980. A new species of Morethia (Lacettilia: Scin- cidae) from northern Australia, with comments on the biology and relationships of the genus. Rec. Aust. Mus. 33; 89-122. 1982. A new species of Leiolopisma (Lacertilia: Scincidae) from Western Australia, with notes on the biology and relationships of other Australian species. Rec. Aust. Mus. 34: 549- 573. HARDY, G.S, 1977. The New Zealand Scincidae (Reptilia: Lacertilia); a taxonomic and zoogeographic study. N.Z. J. Zool, 4:221-325. HUTCHINSON, M.N. 1980. The systematic relation- ships of the genera Egernia and Tiliqua (Lacer- tilia: Scincidae). A review and immunological reassessment. pp. 176-193. Jn Banks, C.B., and Martin, A.A. (eds), ‘Proceedings of the Mel- bourne Herpetological Symposium.’ (Zoologi- cal Board of Victoria: Melbourne). LANYON, S. 1985. Detecting internal inconsisten- cies in distance data. Syst. Zool. 34: 397-403. MOLNAR, R. 1984. Cainozoic reptiles from Australia (and some amphibians). pp. 337-341.Jn Archer, M. and Clayton, G. (eds), ‘Vertebrate zoogeog- raphy and evolution in Australasia.’ (Hesperian Press: Perth). MAXSON, L.M., SARICH, V.M. AND WILSON, A.C. 1975. Continental drift and the use of al- bumin as an evolutionary clock. Science 187: 66-68. MAXSON, L.M. AND WILSON, A.C. 1974, Conver- gent morphological evolution detected by study- ing proteins of tree frogs in the Hyla eximia species group. Science 187; 66-68. RAWLINSON, P. 1974. Revision of the endemic southeastern Australian lizard genus MOLECULAR EVOLUTION IN AUSTRALIAN DRAGONS AND SKINKS Pseudemoia (Scincidae: Lygosominae). Mem. Nat. Mus. Vict. 35: 87-96. SARICH, V.M. 1985. Rodent macromolecular sys- tematics. pp. 423-452. In Luckett, W.P., Harten- berger, J-L. (eds), ‘Evolutionary relationships among rodents. A multidisciplinary analysis.’ Series A: Life Sciences, Vol. 92. (Plenum Press: New York and London). THORPE, J.P. 1982. The molecular clock hypothesis: biochemical evolution, genetic differentiation and systematics. Ann. Rev. Ecol. Syst. 13: 139- 168. TYLER, M.J. 1979. Herpetofaunal relationships of South America with Australia. pp. 73-106. In Duellman, W.E. (ed.), “The South American herpetofauna: Its origin, evolution and disper- 331 sal.’ Monogr. Mus. Nat. Hist. Univ, Kansas No. 7. WILSON, A.C., CARLSON, S.S., AND WHITE, T.J. 1977. Biochemical evolution. Ann. Rev. Biochem. 46: 573-639. WITTEN, G.J. 1982. Phyletic groups within the fami- ly Agamidae (Reptilia: Lacertilia) in Australia. pp. 225-228. In Barker, W.R. and Greenslade, P.J.M. (eds), “Evolution of the flora and fauna of arid Australia.’ (Peacock Publications: Adelaide), WORRELL, E. 1963. ‘Reptiles of Australia.’ 1st edit. (Angus and Robertson: Sydney). WU, C.-I. AND LI, W.-H. 1985. Evidence for higher rates of nucleotide substitution in rodents than in man. Proc. Natl. Acad. Sci. USA. 82: 1741- 1745. FURTHER EVIDENCE OF OPHIOPHAGY IN AN AUSTRALIAN FALCON: Australia, despite the presence of arich terrestrial snake fauna, does nol supporta specialised predator of snakes among its raptors. One species, the Brown Falcon Falco berigora Vigors and Horstield, does possess certain morphological features - densely feathered breast, thick toes, and coarsely scaled legs and feet - that are usually associated with specialised snake-eating raptors such as the Old World snake eagles, Circaetus and Spilornis, and New World Laughing Falcon, Herpetotheres cachinnans. Although there are a number of records of Brown Falcons preying on snakes in the older literature (Shea, 1987), these do not provide identifications of the snakes involved, Recent Teports (See Sonter and Debus, 1985; Shea, 1987 and citations therein) provide evidence thal these raptors may be important predators of snakes in some areas (if not the greater part of the falcon's range). The following contribution provides evidence based on field observations of Brown Falcons as predators of both elapid and colubrid snakes. The observa- tions reported were made during general surveys of raptors, Each record is summarised in the anecdotal list below. Lengths of snakes given are estimates and are only provided for complete specimens. Brown Tree Snake Boiga irregularis: single record of 1.0m specimen being curried in flight, Maleny, SEQ (October 1971). Green Tree Snake Dendrelaphis punctulatus: one partially eaten specimen brought to nest, Woodford, SEO (November 1986); one being carried in flight, 0.8-1.0m, over Bruce Highway near Nambour, SEQ (August 1990). Keelback Tropidenotus mairii: two partially eaten specimens observed in nest, Woodford, SEQ (November 1986). Yellow-faced Whip Snake Demansia psammophis: one 0.5m specimen being carried in flight, near Townsville, NEQ (July 1979); one specimen being eaten on roadside post, Bundaberg, SEQ (August 1979); one specimen, U.6-0.8m, being carried in fight, Yarraman, SEQ (April 1987). MEMOIRS OF THE QUEENSLAND MUSEUM Whip Snake Demansia sp.: partially ealen specimen being carried in flight, Richmond, CQ (May 1984) Marsh Snake Hemiaspis signata: one specimen being eaten at perch, near Kenilworth, SEQ (December 1976). Red-bellied Black Snake Pseudechis porphryiacus: single partially eaten specimen being carried in flight, Maleny, SEQ (June 1978), Eastern Brown Snake Pseudonaja textilis: one specimen carried in flight, 0,9-1.1m, near Rockhampton, CEQ (May 1985); one specimen being carried in flight, 1.1-1.2m, near Gladstone, SEQ (April 1981); one partially eaten specimen being carried in flight near West Wyalong, SCNSW (Decem- ber 1984). Brown Snake Pseudonaja sp.: (possibly P.guttata) being carried in flight, 0.5-0.6m, near Richmond, CQ (August 1983), These observations provide further evidence that the Brown Falcon, although a generalist predator, is an accomplished predator of snakes. Not only are non-yenomous and mildly venomous species taken, highly dangerous species of the genera Austrelaps, Notechis, Pseudechis and Pseudanaja are also successfully preyed upon (see also Sonter and Debus, 1985; Shea (1987), It would also appear from the above observations and the records published to date, that there is no tendency on the part of these raptors to select smaller sized elapids relative to colubrids. Literature Cited Shea, G.M. 1987. Bibliography of herpetological References in Australian ornithological journals. Smithsonian Herpetological Information Service Series 73; 16-45 Sonter, C. and Debus. S.J.S, 1985. The Brown Falcon Falco berigora as a predator of snakes. Australian Bird Watcher 11; 92-93, G.V. Czechura, Queensland Museum, PO Rox 300, South Brisbane, Queensalnd 4101, Australia; 17 August, 1990. COURTSHIP AND MATING IN WILD VARANUS VARIUS(VARANIDAE: AUSTRALIA) D.B. CARTER Carter, D.B. 1990 09 20: Courtship and mating in wild Varanus varius (Varanidae: Australia). Memoirs of the Queensland Museum 29(2): 333-338.Brisbane. ISSN 0079-8835 Observations of wild Varanus varius in southern New South Wales indicate that mating takes place between mid-November and early January; that communication between individuals is by means of olfactory, visual and tactile cues; that females mate with several males, including subordinates; and that pairs mate frequently using hemipenes alternate- ly. Varanidae, mating, reproduction, behaviour, Varanus varius. D.B. Carter, Department of Zoology, Australian National University, GPO Box 4, Can- berra 2601, Australia; present address: Australian National Parks and Wildlife Service, Uluru National Park, PO Box 119, Yulara 0872,Northern Territory, Australia; 17 August, 1988. Goannas are wary lizards. Their behaviour in the wild is rarely documented although there are a number of published accounts of reproductive behaviour in captive varanids (Auffenberg, 1983; Moehn, 1984). The published records of mating in wild varanids include King and Green (1979), Tasoulis (1983), Wilson (1987) and Auf- fenberg (1978, 1981) but apart from Auffenberg’s studies of Varanus komodoensis these provide little detail. This paper reports the results of observations on courtship and mating obtained as part of a wider study of reproduction in wild lace monitors. MATERIALS AND METHODS This work was undertaken in the valley of the Deua River, Deua National Park in southeastern New South Wales (35°46’ S, 149°56’ E). The country is steep and rugged with rocky soils covered by an open forest dominated by Eucalyptus globoidea. The study site is in the southern temperate part of the range of Varanus varius, which is the only varanid known to occur in the area (Cogger, 1986). Goannas were cap- tured with a pole and noose, measured, weighed and marked and released at the place of capture usually within 24h. They were individually marked, temporarily with bands of acrylic paint around the tail and permanently by excising a combination of scales from the prominent fringe of scales under the rear fourth toes. The latter method was preferred to toe clipping because it did not deprive these arboreal lizards of one or more of their strongly clawed digits. Males were sexed when seen mating, when they fully everted their hemipenes during cap- ture or handling, or by radioxerography of the base of the tail showing the hemipenes which are partly ossified (Shea and Reddacliff, 1986). Animals were classified as females only when they were seen mating or by the absence of hemipene ossification in radioxerographs. Prob- ing at the base of the tail was considered an unreliable method of sexing (Weavers, 1983; King and Green, 1979). A selection of animals were fitted with radiotransmitters (tracking only), which were attached with glue and stain- less steel sutures to the side of the tail just behind the rear leg (method of Weavers, 1983). Com- plete transmitter packages weighed between 19 and 22g ( <1.5 % body weight). On the evening preceding a planned period of observation a telemetered animal was located in its roost. The next morning before 0730h I would take up a suitable position about 5 -15m from the roost before the animal emerged. I was equipped with binoculars and radio receiver and was con- cealed under a frameless canvas hide which al- lowed me to move to keep animals in view (Carter, 1988). Although these lizards are usual- ly very wary of humans and typically respond by climbing the nearest large tree, I was ignored by most animals whenever I wore the hide. A total of 37 hours of observations from within the hide were made between 17 November, 1987 and 8 January, 1988. During the period from September, 1987 to March, 1988 a total of 81 days were spent at the study site and information on location and activity was recorded whenever individuals, pairs and groups of goannas were encountered. Events were timed to the nearest 334 minute, Eastern Standard Time. Because it was impracticable to use video or still cameras, diagrams of behaviour were drawn from photographs of dead goannas arranged in ac- cordance with detailed field notes RESULTS Ten marked animals and another five un- marked were observed either mating or in what were presumed to be mating groups. Table 1 summarises data about the marked animals ob- served; Table 2 provides details of the groups seen. MATING SEASON Throughout most of the year lace monitors were found to be solitary (174 solitary in- dividuals were recorded). However between 18 November and 2 January, I saw 11 groups of goannas; some of which were mating, some fighting, some roosting together, and some that were surprised on the ground and climbed into SVL (cm) Sex (method) Weight (kg) 10.09.87 3* 18.09.87 7 06.10.87 11* 21.10.87 14* 22.10.87 18* 18.11.87 19* 18.11.87 22* 18.11.87 26 24.11.87 20.12.87 TABLE 1. Data concerning marked individuals found in mating/courting groups. Sex is followed in brack- ets by method of sexing, U = unknown sex, h = everted hemipenes, x = radioxerograph, m = mating, = probably male because of behaviour towards known males and females. Asterix indicates animals which were carrying a transmitter during observa- tions. MEMOIRS OF THE QUEENSLAND MUSEUM TABLE 2. Dates, activities and composition of mating groups observed November 87 to January 88. F=female, M=male, n=not marked, U=unknownsex, a= unmarked but recognisable individual. Date No of Activity | Individuals goannas 18.11.87 on ground ar 18(M), 19(M) 20.11.87 roosting in mound (F), 18(M) 21.11.87 mating 14(F), n(M)a 22.11.87 roosting 14(F), 11(M), n(U) 23.11.87 n(M)a, 11(M) 23.11.87 11(M), 14(F), 18(M), n(M), n(M), n(F) 24.11.87 7(U), n(U), n(U) 20.12.87 32(F), 22(M) 23.12.87 32(F), n(M) 26.12.87 32(F), 26(M) 02.01.88 3(F), 1(M) trees before I could see what they were doing.These observations indicate that the mating scason extends from about mid-Novem- ber to early January. MATING SYSTEM Females mated with several males over a period of days. Female 14 was seen to mate with two different males and over a period of 6 days associated with at least six different males. Female 32 was found on three separate days with three different males and was observed mating with one of them. Mating/courting groups varied from just the pair to as many as six individuals. If females were accompanied by more than one male the largest would chase other males that approached the females. COMMUNICATION AND COURTSHIP Lace monitors appear to use three forms of GOANNA MATING BEHAVIOUR 335 communication - olfactory, visual and tactile. Scent trails and marking of particular sites are used to communicate over long distances or periods of time whereas body position and head movements are used to communicate directly when animals can see each other. Tactile cues become important as the male comes close to the female and attempts to straddle and mount her. Male 11 wasseen to follow, for about 20m, the exact pathway taken by another male 10min earlier. Male 11 moved slowly with wide sweeps of the head and rapid flicking of the tongue consistent with detecting and following a scent trail (Auffenberg, 1982). On another occasion an unmarked male approached a tree where male 18 had basked 80min previously. After flicking his tongue around the base of the tree for about ]min the unmarked male carefully wiped both sides of his head on the tree trunk and then repeated this action before moving away. Another unmarked male was observed for th as he located the overnight roost (hollow log) af female 32 and then followed her exact path to the nearby tree where she had moved to bask 2h 18min earlier. While he was locating her he stopped twice to vigorously rub his cloacal region against the ground and twice to rub his head, neck and throat on a tree and a log near her roost. During his approach to her roost, he moved very slowly over a distance of about 100m and appeared to be carefully scenting the ground with rapid flicks of his tongue. He also seemed to be pressing his cloaca to the ground as he walked. One large termite mound appeared to be an important focus of breeding activity for animals 11, 14, 18, 19 and at least 5 unmarked goannas. The mound had a large hole dug into the side. Over one week of observations several of these goannas roosted in the hale overnight and during the day there were frequent skirmishes between males that tried to approach the mound, Two males, which were seen mating or attempting to male with females, had both visited the mound less than thirty minutes before they were seen with the females and may have picked up the female scent trail at the mound. Goannas, which were clase enough to see each other, adopted distinctive posture and head movements according to their sex (Fig. 1.). Females remained still or moved slowly, always with their bodies flat on the ground and necks extended along the ground. Females were incon- spicuous. In contrast, the moyements of males were rapid and exaggerated, making them con- spicuous. Males held their bodies raised from the ground, the neck usually vertical and the head —| = FIG. 1. Demeanour of male (top) and female (below) when they can see each other The female remains still or crawls slowly but always with body and neck Matlened to the ground. The male moves rapidity, with head raised and offen makes spasmodic shuddering movements of the head as he upprouwches the female. His behaviour towards other mules is similar. A36 held high and frequently jerked from side to side in a spasmodic, shuddering fashion. This head shuddering became more frequent and pronounced when closely approaching a female. The unmarked male which spent at least an hour locating female 32 used this head shuddering signal when he looked up from the base of her tree and first saw her looking down at him from a height of about Sm. She immediately descend- ed and they mated several times during the next hour. On another occasion when an unmarked female in a mating group was alarmed by my approach and climbed a tree, male 11 who was trying to mate with her climbed up after her and repeatedly gave the head shuddering signal. She ignored him and remained in the tree for the next 2h while | continued to watch the remaining animals, which appeared unconcerned by my presence. Head shuddering was also employed by males approaching each other - most vigorously by the largest of the males. MATING Twenty four acts of mating were observed, Female 14 and an unmarked male, with no other animals present, mated 16 times over a period of 3h. Another pair consisting of fernale 32 and an unmarked male mated 7 times over a period of Lh. In both these cases the unmarked males were relatively small - about 60cm snoul-vent length - and certainly smaller than other males seen in the vicinity. Female 14 and the large male 11 were ina group with another female and 3 other males and, over a period of 2h, mated only once although malel] spent much of the time circling, hissing and chasing other males. From these observations, mating behaviour can be summarised as follows. The female lies still and flat on the ground with neck fully ex- tended and head on the ground (Fig. 1). The male MEMOIRS OF THE QUEENSLAND MUSEUM approaches from the rear and to one side with body raised clear of the ground. The head is raised with the snout markedly tilted down towards the female and darts rapidly from side to side as he flicks his tongue aver her back and neck (Fig 2). Sometimes the male's head move- ments appear to be an involuntary spasm or shudder. He brings his head up to the right of hers. his body lying diagonally across hers and his vent adjacent to the left side of her tail, He then reaches over the base of her tail with his right hind foot and scrabbles at the right side of her tail with his claws apparently to stimulate her to raise the base of her tail, She recurves her back, lifts her hindquarters off the ground and raises her tail ina high arch. With his right hind foot still over the top of her tail and holding it firmly, he curves the base of his tail under hers to insert the right hemipenis (Fig. 3). Events so far have taken about half a minute and the pair may lie still in this position for a further half minute or so, Both may look around; the female with neck extended horizontally and head tilted up, the male with neck vertical and head horizontal. He may open his mouth slightly and pump the gular region presumably to cool. Then he begins vigorous thrusting, powered primarily by the left hind leg. Twenty or 30 thrusts may be made over about one minute with the lust one or Lwo being slower and much less vigorous. The muscle tension in both animals then subsides, the arched tails relax and they uncouple. The time taken from the initial ap- proach by the male to uncoupling is two to three minutes. The next copulation may commence within two or three minules and in most cases the male approaches from the other side and uses the other hemipenis. After several copulations the male may retire to the shade to cool for a few minutes FIG. 2. Courtship. The male begins to straddle the female, rapidly flicking his longue over her back, neck and head. He may rake the claws of one front leg down her hack. She remains flatlened. GOANNA MATING BEHAVIOUR 337 before returning to mate again, The female may terminate a period of mating by climbing a tree or may prevent the male from mounting by con- tinually crawling forwards, During the last few matings between 14 and the unmarked male there was no Vigorous thrusting as the male appeared to tire. On one occasion during mating between female 32 and the unmarked male, he raked his front claws several times along her back. DISCUSSION A comparison of my observations with those of Moehn (1984) for captive V. timorensis, Auf- fenberg (1978, 1981) for wild V. komodoensis and Auffenberg (1983) for captive V. bengalen- sis reveals differences in behaviour and timing of courtship events. None of these papers report head shuddering or jerking behaviour by males in courtship although Davis et al, (1986) report such behaviour by male V. dwmerili prior to combat, Neither Mochn nor Auffenberg em- phasise distinctive demeanours for males and females early in courtship although Auffenberg ey reports ‘do nothing’ behaviour by female , bengalensis which seems to encourage courtship by males and Moehn mentions that female V. timorensis are ‘passive’ during courtship. Mating of V. timorensis was markedly prolonged compared to V. varius. Whereas V. timorensis were coupled for up to 47min and pelvic thrusting occurred at intervals of 5 to 22s, V, varius coupled for no more than four minutes and executed 20 or 30 pelvic thrusts within about one minute. V. bengalensis completed courtship and intromission in a maximum of 123s. In both V. komodvensis and V. bengalensis Auffenberg emphasises the aggressive nature of females Which he regards as a danger to males attempting to mate. He regards the pacifying and immobhilising of females as an important part of courtship behaviour in these species and he in- terprets the mating success of large males as largely due to their ability to restrain females. [ observed no aggressive behaviour by V, varius females towards males or other females. In general, females appeared to be cooperative until they terminated a period of mating by climbing a tree. During mating males did not restrain the forelegs of the female as Auffenberg reports for V. bengalensis although male V. varius main- tained a firm grip on the female’s tail with his hind leg during intromission. [n this species males grow much larger than females (Carter, unpubl. obs.) and in all cases I observed the males were at least twice the mass of their partners. The initial approach by amale V. varius with raised body and conspicuous head move- ments. may advertise his size and strength, as well as communicate his sex, and inhibit aggres- sive behaviour in females, A prolonged period during which a pair mate frequently, as in V, varius, has not been reported for other species of varanids, Several communication and courtship acts of V, varius coincide with the behaviour of other varanids. Tongue flicking over the female’s back and neck, scratching upwards with the hind leg at the base of the female’s tail, scratching the female's back with the foreleg and the position for mating are similat in V. komodoensis, V. bengalensis, V. iimorensis and V. varius. Also, Auffenberg (1981) describes the importance of FIG. 3. Mating. The male has reached over her (ail with his hind leg, scrabbling upwards on the side. of her taal with his claws. She has responded by recurving her back, raising her hindquarters and arching her tail. With his hind foot gripping her tail he has then been able to curve his tai) under hers and insert his nght hemipenis. ty w oa scent marking in the behaviour of V. komodoen- sis and reports behaviour such as ‘head scraping’, ‘cloaca scraping’ and deposition of faeces. That females are so readily able to mate with smaller males even though a dominant male may be in the vicinity is puzzling. The fierce fighting between males during the mating season, often resulting in large wounds (Carter, unpub. obs.), suggests thal there is strong competition for mates. Yet females are left unaltended by dominant males for long periods when they may mate with subordinate males. There may be mechanisms related to sperm competition or timing of fertilisation which would confer some advantage to dominant males. ACKNOWLEDGEMENTS I thank Dean Ward and my wife, Margrit for assistance in the field. This work was supported by the Department of Zoology, Australian Na- tional University and was undertaken while I was receiving a Commonwealth Postgraduate Research Award, Permits for scientific research were provided by New South Wales National Parks and Wildlife Service. LITERATURE CITED AUFFENBERG, W. 1978. Social and feeding be- haviour in Varanus komodoensis. pp. 301-331. In Greenberg, N. and MacLean, P.D. (eds), ‘Be- haviour and neurology of lizards’, (Government Printing Office: Washington DC), MEMOIRS OF THE QUEENSLAND MUSEUM 1981. “The behavioural ecology of the Komodo Monitor’, (University of Florida: Gainsville). 1982. Noles on feeding behaviour of Varanus ben- galensis (Sauria: Varanidae). J. Bombay Nat. Hist. Soc. 80:286-302. 1983. Courtship behaviour in Varanus bengalensis (Sauria: Varanidae). pp. 535-551. Jn Rhodin, A.G.J. and Miyata, K. (eds), “Advances in Her- pelology and Evolutionary Biology: Essays in Honour of Ernest E. Williams’. (Museum of Comparative Zoology: Cambridge, Mas- sachusetts). CARTER, D. 1988. A simple hide for observing wary animals. Australian Zoologist. 25:19-20. COGGER, H.G. 1986. ‘Reptiles and Amphibians of Australia’. (A. H. and A. W. Reed: Sydney). DAVIS, R., DARLING, R. AND DARLINGTON, A. 1986. Ritualised combat in captive dumeril’s monitors, Varanus dumerili, Herp. Review 17:85-88. KING, D. AND GREEN, B. 1979. Notes on diet and reproduction of the sand goanna, Varanus goul- dii rosenbergi, Copeia 1979:64-70. MOEHN, L.D. 1984. Courtship and copulation in the timor monitor, Varanus timorensis. Herp. Review. 15:14-16. SHEA, G.M. AND REDDACLIFF, G.L. 1986, Os- sifications in the hemipenes of varanids. J, Herp. 20(4):566-568. TASOULIS, T. 1983, Observations on the lace monitor, Varanus varius. Herpetofauna 15:25. WEAVERS, B-W. 1983. Thermal ecology of the lace monitor Varanus varius Shaw. (Unpublished Ph.D, thesis, Australian National University). WILSON, S.K. 1987. Goanna! Geo 9(3):92-107. MIOCENE DRAGONS FROM RIVERSLEIGH : NEW DATA ON THE HISTORY OF THE FAMILY AGAMIDAE (REPTILIA: SQUAMATA) IN AUSTRALIA J. COVACEVICH, P. COUPER, R.E. MOLNAR, G. WITTEN AND W. YOUNG Covacevich, J., Couper, P., Molnar, R.E., Witten, G. and Young, W. 1990 09 20: Miocene dragons from Riversleigh: new data on the history of the family Agamidae (Reptilia: Squamata) in Australia. Memoirs of the Queensland Museum 29(2): 339-360. Brisbane. ISSN 0079-8835. Physignathus sp. and Sulcatidens quadratus gen. et. sp. nov. have been identified from a series of some 30 agamid jaw fragments recovered from the Riversleigh fossil beds. One specimen (dentary QM F18004) has a healed fracture. The presence of Physignathus confirms a strong Asian influence in the composition of the Australian lepidosaur fauna as early as the Miocene. It also suggests that the Riversleigh area was well-watered at that time. Of the eight Australian Miocene lepidosaur genera, six survive. Sulcatidens and Montypythonoides are extinct. Examination of the type specimens of the extant Physignathus cocincinus of Southeast Asia confirms that the Australia- 7Papua New Guinean P. lesueurii is properly assigned to the genus Physignathus. _] Sulcatidens, Physignathus, Riversleigh, Agamidae, Miocene. J. Covacevich, P. Couper, R. Molnar, Queensland Museum, PO Box 300, South Brisbane, Queensland 4101; G. Witten, Phillip Institute of Technology, Bundoora, Victoria 3083; W. Young, Department of Oral Biology, University of Queensland, St Lucia, Queensland 4067; 21 January, 1990, Fossils from freshwater limestones of Riversleigh Station, northwestern Queensland (c.19°02’S,138°45’E) were first reported near the turn of the century, but no serious work on them was undertaken until the 1960s. Intensive collection of this deposit began in the late 1970s with team work led by Dr Michael Archer (of the University of New South Wales; Archer, Hand and Godthelp, 1986). What is known of the Riversleigh stratigraphy and palaeoecology is summarized by Archer, Godthelp, Hand and Megirian (1989). The rich fossil fauna includes fish, frog, reptile (crocodile, turtle, lizard, python), bird, mar- supial (bandicoot, dasyurid, potoroid, phalangerid) and placental (bat, rodent) remains (Flannery, Archer and Plane, 1983; Hand, 1982, 1985). Studies of the fauna continue in many groups, including arthropods and snakes (typh- lopids, elapids). Reviews and descriptions of some of these groups have been published (e.g. Smith and Plane, 1985; Hand, 1985; Archer and Flannery, 1985; and references therein). From comparisons with other fossil faunas of known age, it has been suggested that the Riversleigh deposit yielding the material described herein is of Miocene age (Archer and Bartholomai, 1978; Archer, 1981; Godthelp, pers. comm.). Amongst material recovered are some 30 fragments of agamid skulls. Agamids are readily distinguished in having fixed, acrodont teeth which can be differentiated into ‘incisors, canines and molars’ (Boulenger, 1885). Agamid dentition and other distinctive characters are ably summarized by Boulenger (1885) and Estes (1983b). Modern Australian agamid maxillae and dentaries are illustrated in Fig. 1. The Riversleigh agamid material is small sized and in delicate condition. These fragments have been compared with small samples of all but one of the extant Australian genera of agamids. Not available to us is Cryptagama Witten 1984. (Cryptagama aurita, the type and only species, most closely resembles Ctenophorus clayi, another small cryptic dragon (Witten, un- published data)). The fossils have also been com- pared with a range of agamids from other continents, either through the literature (e.g. 340 MEMOIRS OF THE QUEENSLAND MUSEUM FIG. 1. Maxillae (left) and dentaries (right) from modern Australian agamids. (a) Physignathus lesueuriti x1.5 (QM J26671); (b) Diporiphora bilineata x3, (QM 311141); (c) Hypsilurus spinipes x1.5 (QM J45306); (d) Pogona barbata x1.5 (QM J23950); (e€) Chamydosaurus kingti x1.5 (QM J45307); (f) Chelosania brunnea x3 (WAM R41565); (g) Caimanops amphiboluroides x3 (WAM R15564); (h) Tympanocryptis tetraporophora x3 (QM 134580); (i) Ctenophorus caudicinctus x3 (QM J21654); (j) Lephognathus gilberti x1.5 (QM J39042); (k) Gemmatophora nobbi x3 (QM J38738); (1) Moloch horridus x3 (QM J11492), MIOCENE DRAGONS Cooper and Poole, 1973; Cooper, Poole and Lawson, 1970) or with specimens (e.g. Chamaelonidae, Sphenodontidae). Collection designations: AMNH, American Museum of Natural History, New York; MNHN, Musée Nationale d’Histoire Naturelle, Paris: QM, Queensland Museum, Brisbane, WAM, Western Australian Museum, Perth. COMPARATIVE MATERIAL EXAMINED RECENT AUSTRALIAN AGAMIDAE Amphibolurus nobbi &. QM J38748, Mt Windsor Tableland, NEQ. Caimanops amphiboluroides &. R14464, 34km S of Warratra. WA. Chelasania brunnea, WAM R41565, Mitchell Plateau, WA, Chlamydosaurus kingit. QM J3718, Zillmere, Brisbane, SEQ; QM J5707, Gulf Country; QM 319707, Rockhampton, MEQ. (dentary and max- illary fragments); QM J21929, Darwin area, NT; OM J45307, Cooktown area, NEQ: QM J47642, (juv) Wenlock, Pascoe River, Cape York, NEQ. Ctenophorus caudicincus ¢. QM J21654, Black Mountain, Warenda Station, WCQ. Dipariphora bilineaia 2. QM J11141, En- deavour R.. near Cooktown, NEQ. Lophognathus gilberti 2. QM J39042, 56.3km E of Camooweal, NWO. Gonocephalus beydii d. QM 117799, Mt Bel- Jenden Ker, NEQ. G. spinipes, QM J8330, Coomera Gorge, Lamington National Park, SEQ; ?. OM J45306.Richmond Range, via Bonalbo, NENSW, Moloch horridus 3 .QM J11492, Giles, WA. Physignathus lesueurli, QM J3865, Bellevue Station, via Coominya, SEQ; QM J5449, Bris- bane, SEQ; OM J26671, Caboolture, SEQ; OM 735270, Ferguson SF, Saltwater Ck tributary, via Maryborough, SEQ. (left dentary only); QM J38108, Cobble Ck. via Samford, SEQ; OM 347339, Mt Nebo area, SEQ. (right dentary anly); QM J47973, SEQ, Pogana barbata. QM J4141, Brookstead, Darling Downs, SEQ; OM 314402, Wacol, Brisbane, SEQ: QM 323950, Everton Park, Brisbane, SEQ; OM J45852, no locality: OM 347070 , Banyo, Brisbane. SEQ; OM J47077, Brisbane, SEO. Tympanocryptis tetraporophora 3. QM 334580, Cuddapan Station, via Windorah, SWO. WAM 34] RECENT FOREIGN AGAMIDAE Uramastyx aegypticus. AMNH 73160, Saudi Arabia; AMNH 74816, Saudi Arabia. Physignathus cocincinus. MNHN Ag&?, MNHN 2537, MNHN 1856, MNHN 2536. CHAMAELEONIDAE Chameleo (7) basiliscus. QM J45322, na dala. IGUANIDAE Feuana sp. QM J49263, no data. SPHENODONTIDAE Sphenadon punctatus. QM 11046, New Zealand. Fossil. AGAMIDAE Middle to late Miocene specimens, Gag Site. QM F18031-F18033. Early Miocene specimens, Godthelp Hills and Hal's Hill sequence. QM F18004-F18011, QM F18014-F 18030. 2, Inabayance Site. QM F18012-F 18013. FOSSIL AGAMIDS FROM RIVERSLEIGH Identification of this material has been con- founded by the small size and fragility of the fragments and by the fact that they are all from only dentaries or muxillae. Not one piece of cranium has yet been found. The studies by Cooper and Poole (1973) and Cooper et al. (1970) have been very useful in regard to cranial and dental anatomy comparisons und in recog- nising ontogenetic changes. Despite this, how- ever, distinguishing Juveniles from adults and making identifications from only a few charac- ters has been difficult. In the descriptions which follow, the extent of our material is shown in accompunying diagrams. The material present is shaded on stylized agamid skulls or jaws, based on those of Diperiphora bilineata, The descrip- tions concentrate on features of potential taxonuimic significance. In un effort ta refine the process of identifica- tion of these small fragments, epoxy resin teplicas of QM $5449, J47339 (Physignatiius lesueurii) and of fossil specimens QM F18024, FISOLL, FISUIS and FIS01S were made for comparisons of microwear patterns. The replicas were sputter coated with gold and examined under a Phillips 5U5 scanning electron micro- scope lo compare anatomical features with those produced by wear. However, there are few 342 microwear features on the specimens and, con- sequently, they are of no diagnostic significance. Family AGAMIDAE Gray, 1827 Physignathus Cuvier, 1829 Physignathus sp. QM F18004 LOCALITY Camel Sputum Site (Camel Sputum local fauna). DESCRIPTION A portion of left dentary which bears 12 acrodont teeth and a trace of the 13th acrodont tooth. Length 17.38mm, maximum depth (ex- cluding tooth row) 3.71mm. There is a healed fracture in the mid tooth row. The fracture has healed out of alignment, displacing the anterior half of the ig sentry towards the labial. ; The labial surface of the dentary >} is vertically notched between the tooth bases. Three foramina are present close to the ventral surface, below the third, fifth and seventh acrodont teeth. The anterior ventral surface of the dentary is badly weathered, so it is impossible to determine whether more foramina were present. On the lingual surface a longitudinal groove runs beneath the tooth bases. This groove shallows posteriorly. It is not continuous, having been displaced by the mid-dentary fracture. A large foramen is present in the roof of the Meckelian Groove, below the third posteriormost acrodont tooth. When viewed occlusally, the displace- ment of the tooth row is obvious. Both tooth surfaces are rounded, but the lingual is slightly flattened. An inward curve can be detected towards the anterior tip of the dentary. QM F18007 LOCALITY Camel Sputum Site (Camel Sputum local fauna). DESCRIPTION Right dentary; length 17.67mm; maximum depth (excluding teeth 4.96mm). No pleurodont teeth are present, although one empty ‘socket’ is visible in occlusal aspect. The first and third acrodont teeth are damaged. - In a labial view, only one foramen is visible, situated below the eighth acrodont tooth. MEMOIRS OF THE QUEENSLAND MUSEUM The labial surface is pitted and there are deep notches at the bases of the acrodont teeth. Viewed occlusally the acrodont teeth are rounded on both lingual and labial surfaces. The posterior three acrodont teeth are weakly cusped. The lingual faces are slightly flattened. The anterior tip of the dentary curves inward. On the lingual face of the dentary a deep groove runs below and parallel to the acrodont tooth row. The Meckelian Groove is clearly visible, but badly chipped anteriorly. QM F18008 LOCALITY Wayne's Wok (Wayne’s Wok Site local fauna). DESCRIPTION Right, worn maxilla bearing 11 acrodont teeth, with a flattened labial aspect. Length 16.25mm; maximum depth (excluding teeth) 4.01mm. The posterior three acrodont teeth are weakly cusped. On the labial surface, there is a ~w slight inward curve of the maxi- *< | lla, tilting the tooth row lingual- ly. High on the labial face, below the posterior margin of the nasal process, are two foramina. Further foramina are present in the groove parallelling the maxi- llary/jugal/lacrimal suture. The largest of the foramina lies directly behind the nasal process. Posterior to this are four small foramina. A small process projecting towards the rear of the maxi- lla is situated posteriorly on the maxillary/jugal suture. This is located above the third acrodont tooth from the posterior margin of the tooth row. There is a broad palatal process that reaches its widest point towards the middle of the tooth row, then rapidly tapers off, becoming almost non-ex- istent posteriorly. QM F18012 LOCALITY Wayne’s Wok (Wayne’s Wok Site local fauna). DESCRIPTION Portion of right maxilla bearing complete acrodont tooth row. Length 19.67mm, maximum depth (excluding teeth) 3.81mm (Fig. 2). The first four acrodont teeth are damaged, but the rest of the acrodont tooth row has a flattened labial aspect and the teeth are slightly cusped. The acrodont teeth are moderately ‘shouldered’. The maxilla has a MIOCENE DRAGONS FIG. 2. Occlusal view of QM FI8012 x3. slight inward curve tilting the acrodont teeth lingually. Three foramina are present on the labial surface above the third acrodont tooth; and a further two above the sixth and eighth acrodont teeth. Three foramina are present in the groove which runs parallel to the jugal/maxilla/lacrimal suture. A small process projecting towards the rear of the maxilla Is situated posteriorly on the maxillary/jugal suture, This is located above the third acrodont tooth from the posterior margin of the tooth row, On the lingual surtace, the palatal process is moderately developed anteriorly, broadening to its widest point half way along the acrodont tooth row, then narrowing posteriorly. QM FI8013 LOCALITY Inabayance Site. DESCRIPTION Almost complete left side of snout, including the premaxilla. Length 18.13mm; maximum depth (excluding teeth) 14.07mm. Perfect suture definilion is present between the premaxilla, nasal, maxilla and prefrontal (Fig. 3). The pleurodont teeth of the ~.| premaxilla and maxilla have nol been preserved. The anierior four acrodont teeth are reduced from wear, each tooth being dif- ficult to distinguish from those adjoining it. The posterior four acrodont teeth are intact, their broad triangular form being well preserved. QM FI8014 LOCALITY RSO Site. DESCRIPTION Specimen F18014 is a fragment of the left dentary, bearing cight acrodont teeth. Length 11.24mm, maximum depth (excluding teeth) 3.66mm. The labial surface of the dentary is deeply notched vertically between the tooth bases. Also clearly visible is a single foramen placed close to the ventral surface of the den- eae tary, below the third acrodont 343 tooth. The lingual surface bears a distinct lon- gitudinal groove immediately below the tooth row. This groove shallows posteriorly. In the roof of the Meckelian Groove is a large foramen below the fifth acrodont tooth. Occlusally viewed, both tooth surfaces are gently rounded, but slightly flattened labially. QM FI8016 LOCALITY Upper Site (Upper Site local fauna), DESCRIPTION Damaged right dentary, bearing two pleurodont and eight acrodont teeth. Behind the eighth acrodont tooth the fragment is damaged, revealing Lwo internal longitudinal cavities, one above, and the other below, the MO") Meckelian groove. Length 14.60mm, maximum depth (ex- cluding tecth) 3.28mm. The labial face of the dentary bears four foramina, one below each of the pleurodont teeth and one below the second and sixth acrodont teeth. Strong vertical notch- FIG. 3. Dorsal view of snout, illustrating variation in sulures between (a) Physignathus sp. x3 (QM F18013) and (b) Physignathus lesueurii x3 (QM 35449), 344 ing is evident at the bases of the acrodont teeth. The occlusal aspect shows a distinct inward curve of the.dentary anteriorly. Both lingual and labial faces of each acrodont tooth are rounded, the tooth margin being centrally positioned. On the lingual surface of the dentary a distinct groove runs below the tooth row. The Meckelian Groove its clearly defined and has a distinct foramen posteriorly on the upper surface, QM FIBO017 LOCALITY Upper Site (Upper Site local fauna). DESCRIPTION Specimen F18017 is the upper anterior portion of the left dentary. Eight acrodont teeth are present on this fragment, the first and third broken at the base, Length 9.64mm, maximum depth (excluding teeth) 2.17mm. The labial surface exhibits deep vertical notch- ing between the tooth bases and, on the lingual surface, the only distinctive fea- “>~~,| ture preserved is a longitudinal = groove below the tooth bases, Viewed occlusally, both tooth surfaces are rounded but the labial is more strongly convex. A slight curvature of the jaw line is evident. QM Fishis LOCALITY Upper Site (Upper Site local fauna). DESCRIPTION An almost complete left dentary with 14 acrodont teeth and three pleurodont teeth. Acrodont teeth two, three, four and seven. are ey PeyPE ETE EEE EP EP tebe bebe bab FIG. 4, Physignathus sp: (QM F18018). MEMOIRS OF THE QUEENSLAND MUSEUM damaged. Length 24.38mm; depth (including tooth row) 4.76mm (Fig. 4). Below the tooth row there is a distinct lon- gitudinal groove lingually =| which shallows posteriorly. QM F18019 LOCALITY Upper Site (Upper Site local fauna). DESCRIPTION Anterior half of left dentary, bearing three forwardly inclined pleuradont teeth and eight acrodont teeth; length 12.64mm, maximum depth (excluding teeth) 2.67mm. The second tooth is the larger. The acrodont teeth are sharply tipped and dis- >| tinctly grooved at their bases labially along most of the tooth row. There are six foramina on the labial surface; one below each of the pleurodont teeth, and one below the second, fourth and seventh acrodont teeth. When viewed occlusally the acradont teeth are rounded both lingually and labially and the dentary has a distinct inward curve anterior- ly. On the lingual surface beneath the tooth row is a distinct longitudinal groove, which shallows posteriorly. QM F18020 LOCALITY Upper Site (Upper Site local fauna). DESCRIPTION Specimen F18020 is a portion of mid to anterior right dentary, bearing eight acrodont teeth. Length 9,09mm: maximum depth (exclud- ing teeth) 3.27mm. The labial face of the dentary ‘| is deeply notched vertically at the tooth bases. Three foramina are present towards the ventral surface, one each below the second, third, and sixth acrodont tecth. On the lingual surface a longitudinal groove runs beneath the tooth bases. The Meckelian Groove is clearly defined. Occlusally viewed both tooth surfaces are rounded but the lingual face is less convex. Anteriorly the tooth row curves medial- ly. QM F18021 LOCALITY Upper Site (Upper Site local fauna). MIOCENE DRAGONS DESCRIPTION Tip of right maxilla. Length 9,10mm, maxi- mum depth (excluding teeth) 4.54imm. This frag- ment bears one complete acrodont tooth, followed by three damaged teeth. There are two pleurodont teeth, posteriorly inclined, and one empty socket, A ridge is present on the labial surface of the maxilla. This originates at the base of the third pleurodont tooth and runs paral- lel to the acrodont tooth row. Below this ridge the maxilla is flexed lingually. Six foramina are present, three on the labial surface, one in the ventral margin of the natis and twa in a groove where the maxilla joins the jugal behind the nasal process. The position of these foramina corresponds closely to those of F18024. QM F18022 LOCALITY Upper Site (Upper Site local fauna). DESCRIPTION Damaged left dentary. The anterior tip of the dentary is also damaged behind the second acrodont tooth. Two (possibly three) acrodont teeth are missing. All but the anterior six acrodont teeth are weakly cusped. Length 18.31mm; maximum depth (excluding teeth) 3.01mm (Fig. 5). - Anteriorly three pleurodont -|teeth, which incline slightly anteriorly, are present. Three prominent foramina are present labially, onc each below both the second and third pleurodont teeth, and one below the second acrodont tooth. FIG. 5. Physignathus sp. (QM F18022), 345 The body of the dentary bears nine acrodont teeth, the first of which is damaged. There is a foramen below this first damaged tooth. The acrodont teeth have rounded labial and lingual surfaces, and deep vertical grooves at the bases on the labial aspect. On the lingual surface a distinct longitudinal groove runs beneath the tooth row. This groove shallows posteriorly. The Meckelian Groove is distinct and contains a foramen, situated below the fourth acrodont tooth from the posterior margin of the tooth row, QM F18024 LOCALITY Upper Site (Upper Site local fauna). DESCRIPTION Left maxilla, bearing complete tooth row. Length 18.5mm, maximum depth (excluding teeth) 5.44mm. There are 13 acrodont teeth, and two pleurodont teeth. Anterior to these is an empty ‘socket’ of a third pleurodont tooth. ~— A distinct ridge is present on the ~>.| labial surface. This rises at the base of the third pleurodont tooth and runs parallel to the acrodont tooth row. Below this ridge, the maxilla inclines inwards. Five foramina are present on the labial face of the maxilla, beginning above the third pleurodont tooth and running back to the seventh acrodont tooth. Viewed from above the teeth are worn labially and rounded lingually. The acrodont teeth are weakly cusped. A distinct groove is present, running parallel to the maxill- ary/jugal/lacrimal suture just above the palatal process. This groove contains five foramina, two of which lie behind the nasal process. The second foramen is the largest. The remaining three are of equal size and are situated above the seventh and eighth acrodont teeth. On the lingual surface of the maxilla a distinct palatal process is present above the acrodont tooth row. This process is moderate anteriorly, broadening to its widest point half way along the acrodont tooth row. It then narrows and is only barely visible below the posterior acrodont teeth. QM FL&025 LOCALITY Upper Site (Upper site local fauna). DESCRIPTION Right maxilla bearing a complete acrodant tooth row of 14 teeth and 1 pleurodont tooth. Tan FIG, 6. Physignathus sp. (QM F18025). Length 19.31mm; maximum depth (excluding tecth) 4.93mm, Just posterior to the anterior tip a tidge of enamel is preserved, indicating the presence of at least one other pleurodont tooth. The nasal process, the palatal process and the posterior tip of the maxilla are slightly damaged Three foramina are present on = | the labial surface; one above the = | second acrodont tooth; and one each above the sixth and cighth acrodont teeth. Three additional foramina are present in the groove which paral- lels the jugal/maxilla/lacrimal suture; the anterior (below the nasal process) is the largest. The remaining twe foramina are slightly smaller and lie above the eighth and ninth acrodont teeth. There are also four smal! foramina in this groove and one lies on the floor of the naris. A small process projecting towards the rear of the maxi- lla is situated posteriorly on the maxillary/jugal suture. This is located above the third acrodont tooth from the posterior margin of the tooth row. The rear six acrodont teeth are weakly cusped (‘shouldered’). On the lingual face of the maxilla, there is a broad palatal process which reaches its widest point opposite the eighth and ninth acrodont teeth and is only barely discernable at the 11th acrodont tooth, A small secondary flange is present opposite the 13th and 14th acrodont teeth. The acrodont tooth row, occlusally viewed, has a flattened labial aspect. The pleurodont tooth is strongly inclined outwards. QM F18026 LOCALITY Upper Site (Upper Site local fauna). MEMOIRS OF THE QUEENSLAND MUSEUM DESCRIPTION Portion of the right maxilla bearing cight acrodont teeth. Three of these (the posterior two and the fifth from the posterior margin) are damaged. Length 9.98mm, maximum depth (¢x- cluding teeth) 3.93mm. The labial face of the maxilla has .. |a distinct longitudinal ridge tis, | about halfway down. Below this | ridge, the maxilla and tooth raw slope inwards. Two foramina are present, the first and larger is situated above the second anterior most tooth, and above the lon- gitudinal ridge. The second foramen is much smaller, situated above the third tooth, and below the longitudinal tidge. A further three foramina are present in the groove running parallel to the maxillary/jugal/lactimal suture. These lie close together and are situated anteriorly. The anteriormost foramen is the smallest und the posteriormost is the largest. Viewed laterally, the acrodont teeth are ‘shouldered’. On the lingual face of the maxilla is a broad palatal process which tapers posteriorly on the tooth row. Posteriorly, near the dorsal edge of the medial surface of the labial face there is a deep notch positioned posteriorly in the groove run- ning parallel to the maxillary/jugal suture. This notch is normally associated with the posterior process of the maxillary/jugal suture. On specimen F18026 the process has not been preserved, but the position of the groove indi- cates that the process was positioned anterior to the end of the tooth raw. Viewed occlusally, the teeth have a flattened labial face and the lingual surface is distinctly rounded. Posteriorly in the tooth row the teeth have a distinct wear facet on the anterior face, QM F18027 LOCALITY Upper Site (Upper Site local fauna). DESCRIPTION Left dentary bearing complete tooth row. The two anterjor-most acrodont teeth are damaged and the three pleurodont teeth are represented by broken stumps, Length 18.96mm, maximum depth (excluding teeth) 3.81mm. The labial face of the dentary ~=| bears four foramina; one below both the second and third pleurodont teeth, and one below both the second and fifth acrodont teeth. When viewed occlusal- ly the acrodont teeth are rounded on both the MIOCENE DRAGONS lingual and labial surfaces. and the dentary ex- hibits a distinct inward curve anteriorly, On the lingual surface beneath the tooth row is a distinct longitudinal groove which shallows posteriorly. QM F1i8028 LOCALITY Upper Site (Upper Site Jocal fauna). DESCRIPTION Right dentary bearing 12 acrodont teeth; length 16.38mm, maximum depth (excluding teeth) 355mm. The acrodont tooth row is almast complete. 7 The labial surface of the dentary ae a_i is pitted. Nonetheless two foramina are clearly visible below both the third and fifth anterior-most acrodont teeth. Deep notching is visible between the acrodont tooth bases, except on the two posterior acrodont teeth. On the lingual surface of the dentary, a distinct longitudinal groove shallowing posteriorly runs immediately below the tooth bases, There is.a large foramen on the roof of the Meckelian Groove, below the sixth and seventh anterior-most acrodont teeth. Oc- clusally viewed both surfaces of the teeth are rounded, but the lingual is slightly flattened. A slight inward curve of the dentary is detectable anteriorly. QM F18029 LOCALITY Upper Site (Upper Site local fauna). DESCRIPTION Anterior fragment of left dentary bearing one pleurodont and several acrodont teeth. There are two empty ‘sockets’ of pleurodont teeth, one on either side of the tooth. The first two of the acrodont teeth are in good condition. Length 9.45mm; maximum depth (excluding teeth) 2.52mm. Occusally viewed, the dentary ~| anteriorly displays a distinct in- ward curve. The pleurodont tooth is slightly procumbent. Four foramina are present on the labial surface of the dentary; one below both the second and third pleurodont teeth; one below both the third and sixth acrodont teeth. QM F15030 LOCALITY Upper Site (Upper Site local fauna). 347 FIG. 7. Occlusal view of dentaries illustrating the varying degrees of dentary curvature between (a) Physignathus lesueurit x3 (QM 326671). (b) Physig- nathus sp, x3 (QM F18018) and (c) Chlamydosaurus kingit x3 (QM 145307). DESCRIPTION Fragment of right anterior maxilla bearing seven acrodont teeth. The 3rd anteriormost is damaged, broken off at the base. Length 9.54mm; maximum depth (excluding teeth) 3.15mm, In a labial view, the lower edge of the maxilla curves inward. ..| Four well-spaced foramina are | present on the labial face, above the first, third, fifth and sixth actodont teeth. The acrodont teeth have well rounded tips. The teeth are not cusped. Viewed occlusally, the labial tooth surface is flattened. The lingual surface is very angular, with both the anterior and posterior faces. flattened, giving the teeth a three sided appearance. On the lingual face of the maxilla there is a broad palatal process. Four foramina are situated in the groove in the base of the nasal process, the anteriormost above the third acrodont tooth. The remaining three are above the fifth and sixth acrodont teeth. IDENTIFICATION The specimens F18004, F18007-8, F1S012- 14, F18016-22, F18024-30 are all referred to Physignathus sp. Dentary F18018 is one of the best preserved in the sample and, although the numbers of and wear pattern on agamid teeth vary greatly (Coaper and Poole, 1973), it shows similarity to the dentary of CAlamydosaurus, Lophognathus, Amphibolurus, Physignathus, Clenophorus, Pogona and Diporiphora, This specimen is easily distinguished from Chlamydosaurus kingit by number of acrodont teeth (19-20 in C. Aingii vs 14 in F18018); num- ber of pleurodont teeth (1-2 vs 3); shape and size of pleurodont teeth (one very large caniniform tooth, with one smaller tooth on right ramus vs three teeth, with the first the smallest, followed by two teeth of almost equal size). Further, the upper edge of Meckel’s Groove inC, kingit lacks the slight upward curve present in FI8018, and the mid-line (occlusally viewed) of the dentary in C. kingii is almost straight, whilst in F18018 this curves markedly inward anteriorly, The den- taries of Clenophorus, Pogona, Lophognathus, Amphibolurus, Dipariphora and Physignathus have the distinct medial anterior curve of specimen FI8018, However, Laphognathus, Amphibolurus, Ctéenophorus and Pogona specimens lack the enlarged forward projecting teeth of FI8018 and Pogona has only two pleurodont tecth (vs three in FIS018), Specimen FI8018 and the dentary OM J26671 (Physignathus lesueurii) are very close (Pig. 7). Table 1 summarizes these comparisons. F1S018 does not differ in any substantial de- gree from specimens of the extant P, lesueurti 'FIS8016 has only two pleurodont teeth, while three are present in other complete material in the sample. Examination of modern Physignatius contirms thal this character varies hetween two (QM J35270)) and three (other material examined). MEMOIRS OF THE QUEENSLAND MUSEUM FIG. 6, Lateral view of snout, illustrating the distinct “Roman nose’ profile shared by (a) Physignathus sp. x3(QM FIS8013) and (b) Physignathus lesueurti x1.5 (QM 35449). (Gray 1831), bul, from only this dentary we cannot assign it to this species with any certainty. The following characters, found only in Physignathus, have formed the basis of the remaining identifications: the form of the mmward curve on the dentary (FIS004, FI8007, FI8019, F}8022, P18027, FI8028, FIS029); the lingual longitudinal groave below the bases of the den- tary tecth (FISO04, FI8007, FISO14, FL8016, FISO19, FES022, FI8027, F18028): the number, shape, size and cusping of the acrodont dentary tecth (FI8004, FI8007, FISO14, FI8016, PISO19, FLS022, F1S027, F18028); the number, shape, and size of the pleurodont dentary teeth (Fi8016!, FI8019, F18022, FL8029); the degree MIOCENE DRAGONS TABLE 1. A comparison of the dentaries of F18018 (Riversleigh fossil agamid) with J26671 (Physignathus lesueurii). acrodont tooth number pleurodont tooth number angle of projection of pleurodont teeth curvature of midline of jaw (occlusal view) pattern of tooth wear of flattening of the labial aspect of the acrodont teeth (F18012, F18024, F18025, F18030); the number, shape, size and cusping of the acrodont maxillary teeth (F18008, F18012, F18025, F18026); the number, shape and size of the pleurodont teeth on the premaxilla-maxilla (F18021, F18024, F18025); proportions of the palatal process (F18008, F18012, F18024, F18025, F18026); the steeply-sloped posterior edge of the narial opening (F18012, F18024, F18025); the ‘Roman-nose’ profile (F18013) (Fig. 8); shape and position of the nares (F18013); the form of the small posterior maxi- llary process (F18008)(Fig. 9). Family Agamidae Gray 1827 Sulcatidens gen. nov. TYPE AND ONLY SPECIES Sulcatidens quadratus gen. nov. and sp. nov. (~~ ote WA Ak Reh FIG. 9. Maxilla from Physignathus lesueurii x1.5 (QM J26671), showing the position of the posterior maxillary process, Similar processes are present in many of the modern agamid maxillae examined (Fig. L) and in fossil maxillae where they are complete enough (e.g. F18008). This feature has not been used previously to help identify agamid fragments, 1, anteriorly; ,3 upward and anteriorly marked lingual face flattened; labial face rounded 1, anteriorly; 2,3 upward and anteriorly marked lingual face flattened; labial face rounded FIG. 10. Occlusal view of the holotype of Sulcatidens quadratus gen. et sp. noy. OM F18010 x3. GENERIC DIAGNOSIS Suleatidens is distinguished from all other agamids in having the distal margins of the posterior maxillary teeth set in a notch in the mesial margin of the adjacent posterior crown. ETYMOLOGY From the Latin sulcare, to furrow, in reference to the notches or furrows in the mesial margins of the posterior maxillary teeth, The gender is masculine. Sulcatidens quadratus sp. nov. HOLOTYFE QM F18010, an incomplete tight maxilla bear- ing 11 acrodont teeth (Fig. 10). TYPE LOCALITY Wayne's Wok Site, Riversleigh Station, NW Queensland. Miocene. SPECIFIC DIAGNOSIS As for the genus. ETYMOLOGY From the Latin quadratus, from quadrare, to 350 make square. Named in reference to the nearly quadrate profile of the posterior maxillary teeth. DESCRIPTION Right maxilla bearing 11 acrodont teeth, length 18.28mm, maximum depth (excluding teeth) 3.92mm. On the lingual face of the maxi- lla, some of the bases of the teeth are corroded. —_ In labial view the maxillary face ~ | is rounded, tilting the acrodont .|tooth row lingually. The ®) acrodont teeth are cusped, and their tips are directed posterior- ly. The four posteriormost acrodont teeth are unusual, distinct from the typical triangular agamid dentition. In profile, these teeth are al- most quadrate and are very closely spaced. Viewed occlusally the acrodont teeth are flat- tened on the labial face. The four posterior acrodont teeth are very distinct in form from those situated more anteriorly in the toothrow. In most agamids the posterior acrodont teeth slight- ly overlap, as a result of rotation about their vertical axes (Cooper and Poole, 1973). This is not the case with F18010. On this specimen the posterior five acrodont teeth tightly abut. The distal edge of each tooth is notched into the mesial edge of the posteriorly adjacent tooth. We have not seen this character in any other agamid. DISCUSSION Although anteriorly F18010 shows some similarity (a broad palatal process and the form of the acrodont teeth) to fossils assigned to the extant genus, Physignathus in this study, posteriorly the differences from Physignathus are marked. The tightly abutting posterior acrodont teeth are unique, and this state is not found in any other genera of Australian agamids examined. Even in genera where tooth wear plays an important role in shaping the posterior acrodont teeth (e.g. Pogona, Chlamydosaurus and Physignathus), the teeth remain broadly tri- angular in profile, rather than nearly quadrate as in F18010. REFERRED SPECIMEN QM F18015, fragment of right maxilla with four acrodont teeth (Fig. 11). LOCALITY 9 DESCRIPTION Fragment of right maxilla with four acrodont MEMOIRS OF THE QUEENSLAND MUSEUM FIG. 11. Sulcatidens quadratus (QM F18015). teeth. Maximum length 4.95mm; maximum depth (excluding teeth) 1.08mm. =F In both labial and lingual view, no distinguishing features have P| been preserved on the maxilla. RP However, the acrodont teeth have a distinct chisel-like form. Viewed occlusally, the acrodont teeth are some- what flattened on the labial face. The lingual face is gently rounded, and the teeth abut tightly. There is slight development of the ‘tongue-in- groove’ contact of the teeth seen in QM F18010. DISCUSSION F18015 shares with F18010 similar form of the posterior maxillary dentition. In F18010, each of the posterior five teeth is notched into the adjoin- ing tooth. This is evident in F18015, but less pronounced. It may be that F18015 preserves the more anterior of the teeth showing this feature, or it may have derived from a younger individual that had not yet developed the notching to the extent seen in F18010. F18015 is referred to Sulcatidens quadratus gen. et sp. nov. because it has the distinctive notching in the posterior five teeth. Family AGAMIDAE Gray, 1827 Unidentified Material QM F18005 LOCALITY Camel Sputum Site (Camel Sputum local fauna). DESCRIPTION Portion of left dentary bearing 6 posterior acrodont teeth. The anterior tooth row is badly MIOCENE DRAGONS t ‘ Sager Pipetite hd FIG. 12. Unidentified agamid (QM F18005). worn so several acrodont teeth are missing. Length 19.75mm; maximum depth (excluding teeth) 7.40mm. The labial face of the dentary is badly eroded and it bears little surface detail (Fig. 12). Posteriorly the acrodont teeth are weakly cusped, and there is deep vertical notching between the tooth bases labially. On the lingual surface a longitudinal groove, which shallows posteriorly, runs beneath the tooth bases. A large foramen is present in the roof of the Meckelian Groove below the acrodont tooth fourth from the posterior end. In occlusal view the tooth ridges are displaced towards the lingual surface. QM F18006 LOCALITY Camel Sputum Site (Camel Sputum local fauna). DESCRIPTION A small fragment from the middle of the right maxilla. It bears two undamaged acrodont teeth and anteriorly the bases of three / cS fragmentary acrodont teeth. d) Q “jp | Length 8.48mm, maximum ex! depth (excluding teeth) 2.58mm. On the labial face of the maxilla are three foramina, one below each of the broken acrodont teeth. The labial surface is curved in cross-section. On the lingual surface a groove runs above the acrodont tooth bases. The palatal process is broad. Viewed occlusally, the acrodont teeth have a flattened labial surface. ~~ The posterior acrodont tooth is very slightly tricuspid. Three foramina are present in a groove which parallels the maxilla/jugal/lacrimal suture medially. These foramina are tightly clustered, the posterior being the largest, and the anterior smallest. QM F18009 LOCALITY Wayne’s Wok (Wayne’s Wok local fauna). DESCRIPTION Posterior fragment of right dentary, bearing one damaged and five complete acrodont teeth, Length 12.82mm; maximum depth (excluding teeth) 4.01mm. =| Few distinguishing features are preserved on the labial surface. The posterior teeth are ‘shouldered’ and broadly triangular. Vertical notching is present at the tooth bases, except on the two posteriormost teeth. Viewed occlusally, both the lingual and labial tooth surfaces are well-rounded. The lin- gual surface of the dentary clearly displays a longitudinal groove running parallel to the tooth bases. At the posterior margin of the tooth row is a distinct depression in the surface, once car- trying an additional acrodont tooth. The Meck- elian Groove is clearly visible and contains a large foramen close to the anterior fracture. QM F18011 LOCALITY Wayne’s Wok (Wayne’s Wok site local fauna). FIG. 13. Unidentified agamid (QM F18011). 352 DESCRIPTION Damaged left dentary bearing seven acrodont teeth. The anterior one is damaged near the tip. Length 20.83mm, maximum depth (excluding teeth) 4.43mm. Viewed in labial aspect, the mar- gin of the dentary/coronoid suture is well preserved (Fig. 13). 7 Two foramina are present on the [agate —] Labial face. One lies on the anterior fracture line. The other is immediately behind the damaged anterior acrodont tooth. Both foramina are displaced towards the ventral surface. The acrodont teeth are sharply pointed and inclined slightly back. The two hind-most teeth have a small cusp on the posterior edge. Deep vertical grooving is present between most of the tooth bases on the labial face of the dentary. However, the four posterior teeth do not show this. Viewed oc- clusally, the tooth ridge is slightly off-set labially and both tooth surfaces are gently rounded. The teeth curve slightly inwards, toward the lingual face. On the lingual surface of the dentary a well defined groove runs parallel to the tooth row, immediately below the tooth bases. A shallow cavity behind the posteriormost tooth has held a tooth. The damaged ventral surface of the den- tary clearly exposes the Meckelian Groove in which there are two foramina, situated in the roof. The first and largest lies below the second acrodont tooth. The second is below the fourth acrodont tooth. QM F18023 LOCALITY Upper Site (Upper Site local fauna). DESCRIPTION Right maxilla bearing 12 acrodont teeth. Length 15.30mm, maximum depth (excluding teeth) 2.83mm. The posterior margin of the max- illa is well preserved. Labially two foramina are present above the third and fifth acrodont teeth. Anteriorly a dis- tinct ridge is present on the labial face, giving the lower labial sur- face a distinct inward curve. The acrodont teeth are not cusped and the tooth tips are well rounded. The two posteriormost teeth have a broader profile than the rest of the tooth row. Occlusally, most of the acrodont teeth are flat- tened on the labial face, and well rounded on the lingual face. The posteriormost tooth differs in being gently rounded on both faces and almost MEMOIRS OF THE QUEENSLAND MUSEUM FIG. 14. Occlusal view of QM F18031 x3. chisel-shaped with a distinct ridge at its tip. The second posteriormost tooth, although similar, has a flattened anterior face, giving the tooth three faces. The lingual face of the maxilla has a deep vertical notch along the tooth bases. A distinct groove runs parallel to the tooth row terminating before the posterior two acrodont teeth. The palatal process is widest towards the middle of the tooth row. From this point it slopes back towards the tooth row and reaches its nar- rowest point level with the fourth posteriormost acrodont tooth. From here, it broadens to form a distinct posterior flange level with the end of the acrodont tooth row. Above the palatal process is a deep groove at the base of the nasal process. This groove contains two closely-spaced, large foramina, which are situated above the fifth and sixth acrodont teeth. QM F18031 LOCALITY Gag Site (Dwornamor local fauna). DESCRIPTION Right maxilla bearing 11 acrodont teeth (Fig. 14). Length 13.97mm; maximum depth (exclud- ing teeth) 3.60mm. Labially three foramina are present, situated above the second, fourth and fifth acrodont teeth. The teeth are broadly triangular, and cusped particularly on the posterior edge. Occlusally for most of the tooth row the labial face is flattened and the lingual face well rounded. The three posterior acrodont teeth are gently rounded on both faces. Their broad form and pronounced central ridge, give an almost chisel-like appearance. The lingual face of the dentary shows broad vertical grooving at the tooth bases, each groove located between ad- jacent teeth. As in fossil QM F18023 the palatal process is widest towards the middle of the tooth tow. This tapers and then gives rise to a small posterior medial flange, level with the end of the tooth row. A deep groove at the base of the nasal process, contains three foramina, situated above the fourth and sixth acrodont teeth. MIOCENE DRAGONS OM F18032 LOCALITY Gag Site (Dwornamor local fauna), DESCRIPTION Left dentary fragment bearing three posterior acrodont teeth and a fragment of a fourth. Length §.02mm; maximum depth (excluding teeth) 1.25mm. The only distinguish- “| ing feature on this fragment is the nature of the acrodont teeth. These are broadly triangular and tricuspid, Both the lingual and labial surface of the acrodont teeth are rounded when viewed occlusally. The tooth margin is displaced slightly toward the labial side. Slight rotation of the vertical axes gives the acrodont teeth some degree of overlap. OM FI8033 LOCALITY Cag Site (DWornamor local fauna). DESCRIPTION Right maxillary fragment bearing four acrodont teeth, probably originating from the mid-posterior maxilla, Length 4.66mm; maximum depth (ex- cluding teeth) 2.15mm. The only ===) possibly distinguishing feature is one foramen on the labial sur- face. The teeth are broadly triangular and one is slightly shouldered. Viewed occlusally the teeth huve a flattened labial face. IDENTIFICATION The specimens FLS005-6, FISO09, FISO11, FI8023, FI8031-3 can easily be assigned to the Asumidac. All have fixed acrodont lecth, a major character of the family. However, the specimens are generally small and have worn or damaged teeth, or are otherwise damaged. They thus lack other characters useful in identifica- tion. One specimen (F18032) has distinctly tri- angular, tricuspid teeth, This is seen in Physignathus, Chelosania, Caimanops, Gonocephalus, and Diporiphora but, as no other useful characters are preserved on the specimen, iL is impossible to refine its identification. THE FAMILY AGAMIDAE, EXTINCT AND EXTANT Fossil agamids are known from China, Mon- golia, the Middle East, Belpium, France, southeastern Europe (including Crete), north- er tay ty eastern Africa, Wyoming (USA) and Australia (Estes 1983b). The oldest material is from the Cretaceous of China and Mongolia, Many taxa have been described from this region (Estes, 1983a, Holman, 1981; Hou, 1976). Estes (1983a) has reviewed the fossil record and early distribution patterns of lizards and concluded that agamids originated in the Late Cretaceous in southern ‘Eurasia’, the area which conforms today with South-east Asia and India, including part of Australia, This view is supported by both early and recent reports of Cretaceous agamids from Mongolia and China. Anhutsaurus hauinanensts, Tinosaurus doumuensis, and Agama sinensis (all of Hou, 1976) are agamids from the Palaeocene of China. Agamids are not present in North America (Estes, 1964; Holman, 1981), South America, or Europe. And fossils af appropriate age in Africa are so poorly known that it is impossible to comment on the presence of the group in Africa, The remaining material is more recent, from the Eocene, or, like that from Pulestine and Australia (excluding the Riversicigh material), from the Pleistocene. The Australian fossil record of agamids is extremely poor. Estes (1984) describes. briefly, ‘an actodont lizard (presumably an agamid) from localities in (the Wipajiri Formation ... Etadunna Formation ... South Australia.’ This Middle Miocene fragment presenis the first evidence of a possible agamid presence here. There is a gap in the record (prior to the discovery of the Riversletgh deposits) until the Pliocene. Archer and Wade (1976) report an agamid lizard similar to same species of Amphibolurus and Estes (1983a.b) reports part of a skull of Chlamydosauruy kingit. The only other agamid matcrial is from the Pleistocene (Benneit, 1876; Smith, 1976, 1982). The occurrence of modern agamids reflects this fossil record, Agamids are strongly tepre- sented in both Asia and Australia, which are widely regarded as centres of diversity for the group. The group 1s most diverse at both specific and peneric levels on the Indian subcontinent. Dragons occur in Africa, southern Europe, and some Pacific islands. Species from Africa and southern Europe are not numerous, and belong to genera from India or the Middle East. In Australiu, 60-70 species are probably present, although notall are yet described. There is general agreement about species boundaries, bul na such accord in regard to generic definition and allocation. Thus, there are several recent proposals regarding generic divisions (e.g. Wit- ten, 1982a: Cogger, et al., [983:; Storr, et al. 1983; Wilson and Knowles, 1988). Here, we follow Storr et al. (1983), and Wilson and Knowles (1988). with minor modification, THE AUSTRALIAN LEPIDOSAUR FOSSIL RECORD In order to provide a perspective on the ayamids from Riverslcigh, we briefly review the history of the lepidosaurs in Australia. Molnar (1982, 1984a,b, 1985) summarized what is known from the fossil! record here, The earliest remains are a Triassic (¢.240 my) specimen iden- tified as Kudnu mackinlay(. (Kadimakara ausiraliensis, previously regarded as the oldest Australian lepidosaur, is referred to the sister group of the archosaurs, Molnar. 1990), Estes (pers. comm.) notes further that this species is nota ‘lizard’ and that no Triassic forms have the squamate synapomorphies. Some lacertilian {ragments are known from the Lower Cretaceous (100my) of Victoria and a mosasaur has been recorded from the Upper Cretaceous of Western Australia. None of these have descendants amongst our modern fauna. Molnar (1985) describes “a yawning gap’ inthe record and notes elsewhere (1982), that the bulk of Australian fossil lepidosaur material is of Pleistocene age. The Pleistocene remains belong cither to essen- tially modern taxa or to extinct, related fossil taxa (e.g. Varanus vs Megalanta). PRE-RIVERSLEIGH HYPOTHESES Despite a scant fossil record, a plethora of theories about the evolution of agamids in Australia has emerged (Heatwole, 1987). 1. Harrison (1928) noted (hree elements in Australia’s fauna: A. Autochthonian which ‘must have had its origins at.a lime when Australia was in connection with other land mas- ses...'; B. Euronotian, “... which has reached Australia from ¢lsewhere, and undergone radia- tion ... a bone of contention for a long time .., derived chiefly from South America, by means of antarctic conneetions,,."; C, Papuan, *.,, not Well-named since it came from further aficld than New Guinea, through which is merely passed..,.’ From his discussion of the occurrence of 'The Agamidac’ (pp, 378-380) it can reasonably be inferred that he regarded the group as having fundamentally Asian origins (i.c. forming part of his Papuan clement), 2. Cogyer (1961) suggested *,., there were 4 MEMOIRS OF THE QUEENSLAND MUSEUM agamid invasions of Australia beginning some time in the carly to mid Tertiary, all of which entered Australia via New Guinea. The earliest invasion was by the ancestor of Moloch hor- ridus, the secand one arrived probably in the Pliocene and gave rise to the amphiboluroid radiation.... The final twa, Physignathus lesueurii followed by Gonocephalus species are little differentiated from their New Guinea rela- lives and are probably of recent origin. These two are found only in the wet, forested part of eastern Australia, whereas the older elements are primarily adapted to semi-arid regions.,,.” This approach has been reiterated by Cogger and Heatwole (1984), 3. Witten (1982a) observed two possible cx- planations for the occurrence of agamids in Australia. ‘Either (a) the family evolved in Asia and has spread into Africa and Australia, or (b) the family evolved in Gondwanaland, part of which now makes up Australia, Africa and the Indian subcontinent.,..’ He regards Phiysignathus as one of the Asian-derived agamids in Australia, ‘a more recent Australian arrival than Chelosanin...’ ic. more recent than 10-20mya when he suggested the first Asian-derived agamid species invaded. 4, Estes (1983a,b) has a world view of the agamid fossil record and has commented in detail on origin and carly distributions of the family, He deseribed a Middle-Late Jurassic (190-145my) *... more northern Gondwanan ucrodoent iguanian group (which) underwent Vicarianee as the Asia-Southeast Asia-Australia- India blocks separated from Africa ...° and resulted in the agamids (in Asia and Australia) and in the chameleanids (in Africa). He notes elsewhere (p. 392) the centre of origin for the agamids (along with geckos, skinks, and varanids) during the Early Cretaceous (i.e, about 120my) as “the conjoined area’ of what is today India, Australia and Southeast Asia. Tyler (1979) has also contributed to this interpretation, 5. Greer (1990) contends that *.. agamids evalved Initially on the northern landmass and entered the southern continents, including Australia, only relatively late in their history..." MODERN PAYSIGNATHUS SPECIES The affinities of the bulk of the agamid fossils from Riversleigh lic with the extant genus Physiznatias, the water dragons. Three species of Physignathus are currently recognised - 2. lesueurti, P. coctacinus and P. mentager, MIOCENE DRAGONS P. lesueurii occurs in coastal Australia from southern Victoria (Gippsland) to northeastern Queensland (Cooktown area), and in Papua New Guinea (Wilson and Knowles, 1988; de Rooij, 1915). That many species are shared between Papua New Guinea and Australia is the well documented result of several recent land links that have favoured the exchange of both open and rain forest faunas (de Rooij, 1915; Tyler, 1972; Storr, 1964; Covacevich and Ingram, 1980; Kikkawa, et al., 1981; Covacevich and McDonald, in press). That Physignathus lesueurii should have colonised Papua New Guinea from Australia (or vice versa) sometime in the last 100,000 years is not a matter for comment, as such colonisations are entirely con- sistent with patterns for other taxa. However, the Australo-New Guinean distribution of P. lesueurii presents a paradox. It seems reasonable to suggest that a non-specialised, reptile which could colonise an area that is now two separate land masses, should be rather ‘evenly’ dis- tributed throughout its range. Such is not the case. P. lesueurii is listed from only the Western and Gulf Provinces (Whitaker et al., 1982) and from the Arfak Mts of Irian Jaya (de Rooij, 1915). In Australia, it ranges from the banks of the Endeavour River, Cooktown, NEQ to Vic- toria, some 2500km. A distribution including New Guinea and most of coastal, eastern Australia, but excluding Cape York Peninsula north of Cooktown, is unique amongst the ver- tebrates. This appears anomalous in the light of the generalised habits of P. lesueurii, which oc- curs in a wide variety of riparian habitats and is a catholic feeder. There are two possible ex- planations. 1. P. lesweurii occurred, and has, become secondarily extinct, on Cape York Peninsula north of Cooktown. 2. P. lesueurii occurs only in Australia and ‘P. lesueurii’ from Papua New Guinea is in fact another taxon whose status and affinities are not known. P. cocincinus occurs on mainland southeast Asia, in Indochina and Thailand (Boulenger, 1885; specimens in the Musée National d’- Histoire Naturelle, Paris). Very little has been written about this species since Boulenger’s (1885) work. As the affinities of the bulk of the fossil material lie with Physignathus, any attempt at interpretation of the data presented by the fossils makes desirable an assessment of the relation- ship between the modern Australo-New Guinean representative of the group (i.e. P. lesueurii) and the Asian P. cocincinus. However, assessment of the status of the New Guinean taxon is not possible because of the lack of accessible material. Skulls of P. cocincinus (MNHN, two specimens Ag8, here termed Ag81 and Ag82) and of P. lesueurii (several specimens, see specimens examined) are available to us. P. cocincinus and P. lesueurti resemble each other in general skull form, form of cusps on the teeth, the labial aspect of the teeth, position and size of maxillary foramina, and numbers of both acrodont and pleurodont teeth. In both species the maxillae are inflexed just above the tooth row. There is no discernable groove at the max- illary-jugal-lacrimal suture. In both, also, there is a marked longitudinal groove shallowing posteriorly below the tooth row of the dentary, although the degree of grooving varies slightly from side to side in Ag82 and also between the two specimens. They differ in size, but there is only slight difference in proportion. The shape of the parietal and frontal bones is distinct. In P. cocincinus these bones are broadly flattened, while in P. lesueurii they are more gracile and the occipital processes of the parietal are narrower. The posterior maxillary process in Ag82 lies above the second last maxillary tooth, while in J47973 it is just posterior to the last maxillary tooth. Further, the snout profile of P. cocincinus is acute rather than ‘Roman’, almost truncate, like that of P. lesueurii. Notwithstanding the differences, the skulls of P. cocincinus and P. lesueuriti are closer to each other than either is to skulls of any of the other genera examined by us or illustrated by Cooper, Poole and Lawson (1970). Skull sizes, shapes, and dentition, particularly the pleurodont teeth, distinguish the genera. Table 2 summarizes the results of our examination and that by Cooper, Poole and Lawson (1970) of the skulls of a wide range of agamid genera. Physignathus (includ- ing P. cocincinus) has 15 or more pleurodont teeth in total, made up of a maximum of 6 on the maxillae, 6 on the dentaries and up to 5 on the premaxilla. The presence of 3 caniniform pleurodont teeth on each dentary and each max- illa sets Physignathus apart from all other Australian and Asian agamids examined for this study. The three spirit syntypes of P. cocincinus (MNHN 2537, 1856, 2536) have been examined by one of us (RM). They have no external fea- tures which suggest separation at the generic level from P. /esueurii. Because the specimens 356 MEMOIRS OF THE QUEENSLAND MUSEUM TABLE 2. Numbers and sizes of pleurodont teeth in selected agamid taxa. TAXA PLEURODONT TEETH PREMAXILLA * DENTARIES * TOTAL NUMBER SIZE# MAXILLARIES Physignathus lesueurit medium (QM specimens) + Physignathus lesueurii 2 ? ? Physignathus cocincinus medium + Agama ? : ? Amphibolurus large Caimanops large + Calotes ? ? ? Chelosania small Chlamydosaurus large Ctenophorus medium Diporiphora large + Draco 7 ’ ? Lophognathus large Hypsilurus small + Japalura + Liolepis Moloch + Phrynocephalus Pogona * Uromastix * based on actual teeth. # of largest teeth. + after Cooper, Poole and Lawson,(1970). MIOCENE DRAGONS are old and faded, colour and pattern can not be assessed confidently. However, Taylorand Elbel (1958) describe distinct banding on the tail of P. cocincinus specimens from Thailand. Tails of P. lesueurii are also distinctly banded. Differences in size, degree of nuchal cresting, colour pattern, and head scalation, along with skeletal differen- ces, warrant maintenance of the two species. The status of Physignathus mentager Giinther 1861 remains unresolved, but is inconsequential inastudy of the Miocene Physignathus and their broad relationships. It was described by Boulenger (1885) from ‘Siam’ (= Thailand). In- formation on this taxon is scant, but comparisons between descriptions of P. cocincinus and of P. meniager suggest the latter could be a junior synonym of P. cocincinus. The authors of recent reviews of the tepliles of southeast Asia have not used the name P. mientager. GENERA SHARED BETWEEN AUSTRALIA AND MAINLAND SOUTHEAST ASIA Affinities between the faunas and floras of south-east Asia (including the archipelagos), Australia and New Guinea have long been the subject of observation and discussion. Among the modern lizard genera there is an easily demonstrated affinity. The following genera occur in both Australia and south-east Asia: Cyrtodactylus, Hemidaciylus, Gehyra, Lepidedactylus, Gonocephalus, Physignathus, Cryptoblepharus, Emoiu, Sphenamorphus, Trapidophorus and Varanus. It seems reasonable to suggest thar this results fram a combination of past continetal connections and recent migrations, ¢ithcr across ihe sea, or via land bridges at times of lowered sea levels. (New Guinea and the islands between it, and the In- donesian archipelago, abound with endemic genera which complicate the clear pattern evi- dent at the Australia-Southeast Asian poles of the cominuum of the Australasian region). Notwithstanding the fact that close exumina- tion af the taxa shared between Australia ane South-cast Asia will undoubtedly bring about new allocations, and the recognition of new species and generic definitions, there are strong asspciitions between the two continents. DISCUSSION *.. The Water Dragons are so conspicuous that it cunnot be supposed (hal they have been overlooked, Since they are aquatic and freely enter the sea, their distribution becomes even more remarkable and beg- gars explanation...” Harrison 1928, p. 380. The following Miocene agamids are now known from Australia: Physignathus sp. (QM V18004, 18007-8, 1S8012-14. 18016-22, 18024-30); Sulcatidens quadrats gen. ct. sp. nov. (QM F18010 holotype; 18015); unidentified jon F18005-6, 18009, 18011, 18023, 18031-33). In assessing the significance of the Riversleigh apamid material, we know that: 1. Physignathus sp. was present in Australia in the Miocene. Physignathus remains dominate the sample which also includes a taxon quite distinct ftom any extant form, Sulcatidens quad- ratus gen, et, sp. nov, The Riversleigh material establishes that agamids have had a much longer history here than has generally been supposed. Agamids have been here for at least 15-20my and the Physignathus represented appear to have changed little in that time. 2. Australia was separated by water from lands to the north until the late Miocene. This has been illustrated by Archer et al. (1989) at 40, 30, and JOmya. Despite the controversy about the exact details of the timing of fracturing of the con- linents and fluctuating sea levels, it is apparent thatsea separated Australia {rom the archipelago to its north throughout the Eocene, and Oligocene and into the late Miocene. Thus, any dragons moving berween Southeast Asia and Australia would have required high salt tolerance. 3. P. lesueurii lives in large numbers on the banks of the Brisbane River where salinity is of the order of 20 parts per L000. The dragons, although terrestrial, use the waler as a refuge. They readily enter the water and are capable of spending long periods submerged. This suggests a devel of salt tolerance unusual in other Australo/New Guinean lizards observed. (Covacevich, pers. obs.) 4, Phy signathusyspecies occur caastally in bath Australia (2. lesucurii) und SE Asia (P. cacir- cinus), und in the intervening area, New Guinea. Companson of f lesueurit and P. cocincins indicates a close phylogenctic relationship. 5. Moderi Australian agamids can be placed in two groups. Piysignathus, Gonocephalus and possibly Chelosania form one group. The remaining genera all have a reduced number of microchromosomes, lack lacrimal bones, and are 358 adapted to arid or semi-arid conditions (Witten, 1983). lt is not certain that these both represent monophyletic groups, This evidence can be used lo support two hypotheses about the origins of agamids in Australia. Either they have evolved [rom Asian stock that entered Australia across the seas or via land bridges about 20mybp. (An obvious corol- lary of this hypothosis is that they may have originated in Australia and colonized Asia), or the agamids had earlier origins in Gondwanaland. They are, today, conspicuous in some of the the southern continents, Witten (1982) surmised that Physignaihus was Asian and that the first of the Asian-derived agamids ‘arrived’ in Australia between 10 and 20mybp. He postulated that Chelosania was the most likely direct descendant of such an invasion and that PAysignathus appeared *... to be a more recentarival....’ The evidence presented here, coupled with the well-documented, long history af the Agamidae (at least from the Cretaceous) in China and near regions, most strongly support the suggestion that Physignathus is Asian-derived and that P. fesueuru’ and P, cocinercus are the direct and similar descendants of the Asian ancestral Physrunathes. IL seems reasonable to infer from the agamid remains identified, that the Riversleigh environ- ment in the Miocene may have resembled (in climate and forest profile, at least) present day coastal Queensland. If present requirements for Piysignathus are relevant the urea must have been well-watered, As Suleatidens quadratus gen, ct. sp. nov. is based on an incomplete max- ila and appears to have no obvious close relationships to any modern agamid genus, it reveals nothing of the palaeoecology of Riversleigh. Lepidosaurs are poorly known from the Miocene in Australia. Egernia, Tiliqua, Varanus, Ramphotyphlops?, Morelia (= Python), and Montypythonoides, have been reported (Malnar. 1990). In addition, we here report Physivnathus and Suleatidens gen. nov. Six of these cight genera are represented in the modern fauna. The extinction of Sulcatidens and Montypythanoides does not conform to this pat- tern af general conservatism in the Australia lepidosaurs, NOTED ADDED IN PRESS Since the review work for this paper was com- pleted, 38 further agamid jaw fragments have MEMOIRS OF THE QUEENSLAND MUSEUM been extracted from the Riversleigh matrix. The specimens have been registered into the Queensland Museum reference collection (F18044-F 18081). There are no taxa represented in these fragments that differ from those already identified. All compare well with the initial sample, These specimens do not, therefore, alter the conclusions already drawn. ACKNOWLEDGEMENTS We wish to ucknowledyge, with thanks, the ussistance we have received with this work from Alan Bartholamai, Roger Bour, Bruce Cowell, Gary Cranitch, Bob Domrow, Richard Estes, Allen Greer, Peter Jell, Arnold Kluge, Marie- Louise Racine, Jean-Claude Rage, Philippe Ta- quet, Stephen Van Dyck, and Richard Zweifel. Michael Archer (The University of New South Wales) co-ordinates the Riversleigh Project. He, Henk Godthelp, and Suc Hand supervised the Held and preparatorial work involved in making ihese specimens available, and generously in- volved us in their project. They received finan- cial support from the following: Australian Research Grants Scheme: Department of Arts, Sport, the Environment, Tourism and Ter- ritories; National Estate Programme Grant Scheme: University of New South Wales: Wang Computer Ply Lid: Australian Geographic Ply Lid; Mount Isa Mines Pty Ltd; The Queensland Museum: The Australian Muscum; The Royal Zoological Society of N.S.W,; the Linnean Society of N.S.W.; Ansett/Wridgways Pty Ltd: Mount tsa Shire Council; The Riversleigh Society and Friends of Riversleigh. LITERATURE CITED ARCHER, M, 1981, A review of the origing and radiations of Australian mammals. pp. 1437-88, Jn Keast. A. (ed.) “Ecological Biogeography of Australia’. (W, Junk; The Hague). ARCHER, M. AND BARTHOLOMAL, A, 1978. Ter- liury mumtmals of Australia: a synoptic review. Alcheringa 2: 1-19. ARCHER, M. AND FLANNERY, T, 1985, Revision of theextinel gigantic rat kangarous (Polorvidae: Marsupialia), with description ufa new Miocene genus and species und anew Pleistocene species of Propleopus, Journal Paleontology 59(6): 1331-1349, ARCHER. M,, GODTHELP, H, AND HAND, SJ. AND MEGIRIAN, D. 1989. Fossil mammals of MIOCENE DRAGONS 3 Riversleigh, northwest Queensland: Preliminary overview of biostratigraphy, correlation and en- vironmental change. Australian Zoologist 25(2); 29-65, ARCHER, M., HAND, S. AND GODTHELP, H. 1986, Uncovering Australia’s Dreamtime. The Fossil Deposits of Riversleigh: Anextraordinary Australian Discovery, (Surrey Beatly and Sons; Chipping Norton). 32pp. ARCHER. M. AND WADE, M. 1976. Results of the Ray E. Lemley expeditions. Part 1. The Allin- gham Formation and.a new Pliocene vertebrate tauna from northern Queensland. Memoirs of the Queensland Museum 17: 379-97. BENNETT, G. 1576. Notes on the Chlamydosaurus or frilled lizard of Queensland (Chlamydosaurus kingii, Gray), and the discovery of a fossil species on the Darling Downs, Queensland. Pap. Prac. Roy. Soc. Tas. 1875: 699-700, BOULENGER. G.A. 1885, ‘Catalogue of the Lizards in the British Museum (Natural History)’, Volume |. (Trustees of the British Museum of Natural History: London). COGGER, H.G. 1961. An investigation of the Australian members of the family Agamidae (Lacertilia) and their phylogenetic relationships. (Unpublished MSc thesis, University of Sydney: Sydney). COGGER, H.G., CAMERON, E.E. AND COGGER, H.M. 1983. ‘Zoological Calalogue of Australia Volume 1. Amphibia and Reptilia’. (Australian Government Publishing Service: Canberra). COGGER. H, AND HEATWOLE, H. 1984. The Australian teptiles; origins, biogeography, dis- tribution patterns and island evolution. pp.343- 360. In Archer. M. and Clayton, G. (eds), ‘Vertebrate Zoogeography and Evolution in Australasia. Animals in Space and Time’. (Hesperian Press: Carlisle). COOPER, J.S, AND POOLE, D.F.G, 1973, The den- lition und dental tissues of the agamid lizard. Uromastyx. Journal of Zoology 169; 85-100. COOPER, J.S., POOLE, D.F.G, AND LAWSON, R. 1970. The dentition of agamid lizards with spe- cial references to tooth replacement. Journal af Zoology 162: 85-98. COVACEVICH, J, AND INGRAM, G. 1980. The endemic frogs and reptiles af Cape York Penin- sula. pp.49-57. Jn Stevens, N.C. and Bailey, A. (eds), ‘Contemporary Cape York Peninsula’. (Royal Society of Queensland: Brisbane). COVACEVICH, J. AND MCDONALD, K.R. In press. Frogs and reptiles of tropical and subtrapi- cal eastern Australian rainforests, Distribution patterns and conservation, J *The Rainforest wv o Legaey. Ausiralian Rainforests Study, Volume 2". Special Australian Heritage Publication Series, (Australian Government Publishing Ser- vice: Canberra), ESTES, R. 1964. Fossil vertebrates from the late Cretaceous Lance Formation east Wyoming, University of California Publications in Geological Sciences 49: 1-180, pls 1-5. 1983a. The fossil record and early distribution of lizards. pp.365-398. fm Rhodin, A.G.J. and Miyata, K. (eds), “Advances in herpetology and evolutionary biology. Essays in honor of Ernest E. Williams’, (Museum of Comparative Zool- ogy: Cambridge, Massachussetts), \9S3b. Sauria terrestria, Amphisbaenia. pp.52-56, In Kubn O. and Wellnhofer, P. (eds) ‘Handbuch der Paldeoherpetologie Encyclopaedia of Palaeoherpetologv.” Part JOA. (Gustav Fischer Verlag: Stutlgart). 1984. Fish, amphibians and reptiles from the Evadunna Formation, Miocene of South Australia. Australian Zoologist 21(4): 335-43. FLANNERY, T.F. 1958. Origins of the Australo- Pacific land mammal tauina. Australian Zoologi- cal Reyue 1; 15-24, FLANNERY, T.F.. ARCHER. M. AND PLANE, M. L983. Middle Miocene kangaroos (Macro- podoidea: Marsupialia) from three localities in northern Australia, wilh a description of twonew subtamilies. Bureau of Mineral Resources Jour- nal of Australian Geology and Geophysics 7: 287-302. GILMORE, C.W. 1943. Fossil lizards of Mongolia. Bulletin American Museum Natural History 8]: 361-88, GREER, A.E. 1990). ‘The biology and evolution of Australian lizards’. (Surrey Beatty and Sons: Chipping Norton). 264pp. [1989 on title page]. HAND, S. 1982. A biogeographically unique Miocene mammal deposit from northern Australia. [Abstract]. Bulletin Australian Mam- mal Society 7(2): 57. 1985. New Miocene megadermatids (Chiroptera = Megadermatidae) from Australia wilh com- ments on megadermatids phylogeny. Australian Mammialogist 8(1-2): 5-44. HARRISON, L. 1928. The composition and origins of the Australian fauna, with special reference to the Wegener hypothesis. Report of the eighteenth meeting of the Australasian Associa- lion for the Advancement of Science (8: 332- 396. HEATWOLE, H, 1987. Major components and drs- tributions of the terrestrial fauna. pp. 101-135. da Dyne, G.R. and Walton, D.W. (eds.), ‘Fauna of Australia. General articles.’ Vol. 1a. (Australian Government Publishing Service: Canberra). HOLMAN, J.A, 1981. A herpetofauna from an eastern extension of the Harrison Formation (early Miocene: Arikareean), Cherry County, Nebras- ka. Journal of Vertebrate Paleontology 1(1): 49- 56. HOU, LIANHAT 1974. Paleacene lizards fram Anhui, China, Vertebrata Palasiatica 12(3): 193-200. (english abstract p. 200). 1976. New materials of Palaeocene Jizards of Anhul. Vertebrata Palasiatica 14(1): 45-52. (english abstract p. 52). KIKKAWA, J., MONTEITH, G.B. AND INGRAM, GJ. 1981. Cape York Peninsula. Major region of faunal interchange, pp.1695-1742. Jn Keast, A, (ed.), ‘Ecological Biogeography of Australia’. (W. Junk: The Hague). MOLNAR, R.E. 1982. Cenozoic fossil reptiles in Australia. pp.227-234, Jam Rich, P.V. and Thompson, E.M. (eds), ‘The fossil vertebrate record of Australia’, (Monash University Offset Printing Unit: Clayton, Victoria). 19844, Cainozoic reptiles from Australia (and some amphibians). pp.337-341. Jn Archer, M. and Clayton, G. (eds), ‘Vertebrate zoogengraphy and evolution in Australia. Animals in Space and Time. (Hesperian Press: Carlisle). 19846, A checklist of Australian fossil reptiles. pp.405-406, in Archer, M. and Clayton, G. (eds), ‘Vertehrale zoogeography and evolution in Australasia. Animals in space and time.” (Hesperian Press: Carlisle). 1985, The history of lepidosaurs in Australia. pp.155-158. /a Grigg, G., Shine, R. and Ehmann, H. (eds), ‘Biology of Australasian frogs and Tepliles’. (Surrey Beatty and Sons, Royal Zoological Society of New South Wales: Syd- ney). \990), Fossil reptiles in Australia. fm Rich, P.¥.. Barid, R.F., Monaghan, J.M. and Rich, T.H. (eds), ‘Vertebrate palaeontology of Austral- asia’, (Chapman and Hall: London), ROOW, DE, N. 1915. ‘The reptiles of the Indo- Australian Archipelago I’, (E.J, Brill: Leiden). SIGE, B., HAND, 8S. AND ARCHER, M. 1982, An Australian Miocene Brachipposideros (Mam- malia, Chiroptera) felated to Miocene repre- sentatives from France, Palaeovertebrata |2(5); 149- 172. SMITH, M.J. 1976. Small fossil vertebrates from Vic- tona Cave, Naracoorte, South Australia. TV. Reptiles. Trans. Roy. Sac. S.A. 100(1): 39-51. 1982. Reptiles from late Pleistocene deposits on MEMOIRS OF THE QUEENSLAND MUSEUM Kangaroo Island, South Australia. Trans. Roy. Soc. S.A. 106(2): 61-66. SMITH, MJ, AND PLANE, M. 1985. Pythonine snakes (Boidae) trom the Miocene of Australia. Bureau of Mineral Resources Journal of Australian Geology and Geophysics 9: 191-3. STORR, G. 1964. Some aspects of the geography of Australian reptiles. Senckenbergiana biologica 45: 577-89. STORR, G,.M., SMITH, L.A. AND JOHNSTONE, R.E. 1983. ‘Lizards of Western Australia. IL. Dragons and monitors’, (Western Australian Museum: Perth), TAYLOR, E.H. AND ELBEL, R.E. 1958, Contribu- tion to the herpetology of Thailand. University of Kansas Science Bulletin 38(2), no.13, 1033- 1130. TYLER, M. 1972. An analysis of the lower vertebrate faunal relationships of Australia and New Guinea, pp. 231-256. In Walker, D. (ed.), “Bridge and barrier: the natural and cultural his- tory of the Torres Strait’. (Australian National University: Canberra). 1979. Herpetofaunal relationships of South America with Australia. pp. 73-106. In Duellman, W.E. (ed.), ‘The South American herpetofauna: Its origin, evolution, and disper- sal.’ University of Kansas Museiim of Natural History Monograph 7, (University of Kansas Press: Lawrence). WHITAKER, R., WHITAKER, Z. AND MILLS, D. 1982. Reptiles of Papua New Guinea. Wildlife in Papua New Guinea no. 82/2. (Division of Wildlife, Box 585, Konedabu, Papua New Guinea). 53pp. WILSON, S.K. AND KNOWLES, D.G. 1988. ‘Australia’s reptiles. A photographic reference to the terrestrial reptiles of Australia’. (William Collins: Sydney). WITTEN, G.J, 1982a, Comparative morphology and karyology of the Australian members of the Family Agamidae and their phylogenetic im- plications. (Unpublished Ph.D. thesis, Univer- sity of Sydney, Sydney). 1982b. Phyletic groups within the STamily Agamidae (Reptilia: Lacertilia) in Australia. pp.225-228. Jn Barker W.R, and Greenslade, P.J.M. (eds), ‘Evolution of the flora and fauna of arid Australia.” (Peacock Publications, Australian Systematic Botany Society, AN- ZAAS SA Division: Adelaide). 1983, Some karyotypes of Australian agamids (Reptilia: Lacertilia). Australian Journal of Zoology 31: 533-40. TAUDACTYLUS DIURNUS AND THE CASE OF THE DISAPPEARING FROGS GREGORY V. CZECHURA AND GLEN J. INGRAM Czechura, G.V, and Ingram, G. J. 1990 09 20: Taudacrylus diurnus and the case of the disappearing frags. Memoirs of the Queensland Museum 29(2): 361-365. Brisbane. ISSN 0079-8835. The southernmost representatives of the myobatrachid genus Tandactylus is the Southern Dayfrog (7. diurnus). This frog has been recorded from only the subcoastal ranges near Brisbane in southeastern Queensland. 7. diurnus inhabits creeks and their edges in rainforest and tall open forest communities at elevations in the 300-850m range. [tis diurnal, terrestrial, and easily observed and most commonly encountered during the summer months. Breeding is restricted to the spring-summer wet season. T. diurnus has not been recorded in the wild since early 1979. It is possible it is now extinct. T. diurnus und Rheobatrachus silus are apparently linked by similar fates. They were closely associated in the wild and both were last seen in 1979. As well, their related species, T. eungellensis and R, vitellinus have not been seen since 1985, If these missing frogs are exinct, their passing was not a slow process, Rather it appears to have been a catastrophic event thal could not have been anticipated or prevented, There may be reasons to fear for the other species of Taudactylus. These species should be studied and monitored, (] Taudactylus diurnus, biology, Rheobatrachus, extinction. Gregory V. Czechura and Glen J. Ingram, Queensland Museum, PO Box 300, South Brisbane, Queensland 4101, Australia; 20 June, 1990. Taudactylus diurnus and Rheobatrachus silus are two very different frogs that are linked by similar fates (Ingram, 1990). Both species ap- parently disappeared in the same year and neither have been seen in the wild since 1979. While &, silus is a well- studied frog, (see Tyler, 1983, for an overview, and papers therein ), no overview is available for 7. diurnus. Myobatrachid frogs of the genus Taudactylus occur in association with upland and montane rainforests of high rainfall in eastern Queensland (Ingram, 1980; Czechura, 1986a; Winter and McDonald, 1986). Of the six species of Taudac- tylus, only two occur south of the Tropic of Capricorn, These are the Kroombit Tops Dayfrog (7. pleione) and the Southern Dayfrog (T. diurnus). The former species is restricted fo a small area of Kroombit Tops near Gladstone (Czechura, 1986a). T. diurnus is found near Bris- bane in the south of the region (Ingram, 1980). No detailed field investigations of the biology of eilher species have been conducted. Czechura (1986b), however, summarised what was known of T. pleione. Aspects of the biology of T. diur- nus have been reported by Straughan and Lee (1966), Johnson (1971), Liem and Hosmer (1973), Czechura (1975), Corben (1977), In- gram (1980) and Czechura (in press). In this review, we detail field observations made during general and specific investigations of the ver- tebrates found in the Blackall-Conondale Ran- ges of southeastern Queensland, The general results of these investigations have been reported elsewhere (Czechura, 1974, 1975, 1976, 1978, 1983, 1985, in press; Roberts, 1977; Ingram, 1983). We also incorporate information about T. diurnus in the D' Aguilar Range. Unless other- wise stated, all reports are based on the authors” observations. Vegetation types follow Webb (1978) for rain- forests and Specht (1970) and Groves (1981) for other vegetation types, and Stanton and Morgan (1977) provide a broad view of the area, Young and McDonald (1987) describe rainforests of southern Queensland, Taudactylus diurnus GENERAL T. diurnus has been recorded from three sub- coastal mountain ranges (Blackall, Conondale, and D* Aguilar Ranges) near Brisbane (26° 30°- 27°23'S). The species occurs over a relatively narrow altitudinal range of 350-800m, with most records falling in the 500-800m range. The lowest altitudinal records were all attained on the southern and southeastern slopes of the Blackall Range near Maleny (Mooloola and Stanley Ab? MEMOIRS OF THE QUEENSLAND MUSELM River drainages), These frogs are found in association with per- manent and temporary watercourses in extensive forests or gallery forests of the following types: Notophyll Vine Forest (NVF), Tall Open Forest (TOF), NVF-TOF transitions, and Sclerophyll Fern Forest. In addition, they have been found along watercourses in pure stands of the palm Archontophoenix cunninghamiana, in exposed rocky areas, in gorges, in dense non-lorest riparian Vegetation (Lomandra longifolia, Carex neuroclamys, Elastostems reticulatum and Blechnum nudum with or without a shrub story of Callistemon sp.), and where these have been lightly infested with exotic Lantana camara, T. diurnus have not been found where such arcas have been heavily infested with Lantana, where weed infestations of Baccharis halimifilia and Eupatorium riparium occur, The frogs are also absent from streams that have very muddy waters and seem to prefer clear or ‘black-water’ conditions, Permanent streams with rocky substrates are favoured, but T. diurrus also occurs in per- manent and ephemeral streams on gravel, clay, sand and soil substrates. Active 7. diuraus are found in and along the margins of watercourses and Jeaf-litter within 10m of water. The greatest distance that any T. diurnus has been recorded beyond a watercourse was about 22m in wet Weather. Individuals are usually very active. moving or swimming about the stream and en- virons, but they will remain rather motionless at times on stones, debris or low vegetation in or near the walercourse, At night individuals have been located in rock crevices, under stones al ihe water's edge, under debris, in fallen palm fronds, in old burrows or clinging to broad-leafed riparian vegetation. These trogs avoid danger by diving into fast- flowing water and allowing themselves to be carried downstream, swiniming across still or slow-flowing water, or hiding among submerged stones, debris, in rock crevices or in Jeaf-litter (Straughan and Lec, 1966; Liem and Hosmer, 1973; Corben, 1977). In watercourses with muddy substrates, the frogs have been observed diving into the loose upper mud layers and then remaining motionless on the bottom partially covered with mud. DAILY ANB SEASONAL ACTIVITY PATTERNS Active T. diurnus have been observed throughout jhe year, alihough they were en- countered less frequently during the generally colder winter months (e.g. this is the period when minimum activity temperatures are likely to be reached; Johnson, 1971). During periods of prolonged inactivity individual frogs and small groups of frogs (6-13) have been found under stones, in boulder piles, deep stone beds. stream- side rocks and stones, fallen logs, deep soil crevices and rock shelves , often near waterfalls. Shelter sites in rock crevices and under stones are often shared with other species such as Litoria barringtonensis and L. lesueurii. T. diurnus is diurnal, Activity begins at or soon after sunrise, peaks at full light, and rapidly declines with the onset of evening, Most forag- ing occurs during the day, but may extend into the evening (Ingram, 1980). Captive individuals often remain active until late at mght, although this has not been observed in the wild. These [fogs appear to engage in basking be- haviour; individuals have been observed sitting motionless in patches of sunlight before moving away into Jeaf-litter oralong stream edges, Other individuals have been noted sitting on warm rocks for lengthy periods, possibly absorbing heat from the substrate, which in mid-summer remains warm until late at night. Regular move- ment between patches of sunlight and shade was also observed. Individuals regularly move between land and water by swimming from one point ta another, They also cling to rocks, debris or vegetation, with most of their bodies submerged or sit in shallow water. The thermal relations and water balance of T. diurnus Were studied by Johnson (1971), who found the body temperatures of T. diurnus ranged from 13.6-22.8°C (mean 18.4), The criti- cal thermal maximum was 31.1°C (range 28.4- 33.7). It was found that 7. diurnus shows little tolerance (o desiccation despile an ability Io rehydrate rapidly. Breeding occurs in warm weather afler or during heavy rain commencing in late October, November or early December, Straughan and Lee (1966) reported finding gravid females be- tween November and May, with a January- March peak. The eggs are deposited in gelatinous clumps under rocks in water. The tadpoles are bottom-dwellers that feed by scrap- ing food from the substrate with their umbrella- shaped lips. Liem and Hosmer (1973, fig.G) illustrated the tadpole. Males call despite the absence of vocal sacs (Liem and Hosmer, 1973; Corben, 1977; Ingram, L980), Calls are barely audible in the field, al- DISAPPEARING FROGS though they are obvious in captivity and are ulicred in response to disturbance (Corben, 1977), male-male interactions (Ingram, 1980) or advertisement calls. There is no evidence that breeding choruses are formed. Very little courtship and territorial behaviour has been reported, However, some interactions have been observed: male-male interactions involving head to head encounters while ‘vek-eek’ were uttered (Ingram, 1980); males making soft cluck- ing calls trom leaf-litter along stream margins or from rocks in stream; amplectant pairs (umplexus is inguinal) in water-filled rock crevice during day and under stone in afternoon! and male in amplexus with male 4. bar- ringtonensis during late morning (amplexus in- guinal). Foop Straughan and Lee (1966) analysed gol con- tents and showed these frogs to be opportunistic feeders of the forest floor. Amphipods, hymenopterans and lepidopteran larvae Were the most commonly recorded prey in their sample, Captive T. diurnus take a variety of small soft-bodied insects such as immature cock- tuaches, moths, winged termites and flies, In the wild, feeding frogs have been observed taking small insects along or nearstreams. There are no ohservations to suggest that prey ts taken from the waler. THE CASE OF THE DISAPPEARING FROGS For T. diurnus, there are no quantitative es- timates of population density. However, the general impression is (hal most observers thought 7. diuenus was an abundant frog where it becurred. For example, McEvoy etal. (1979) reported them (o be abundant along streams in NVF in the Kilcoy Creek drainage of Conondale Range. Nevertheless, surveys (G.V. Czechura and GP. Maywald, pers, obs.) of several streams in the Blackall Range (Obi Obi Creek drainage, Stanley River headwaters) and Conondale Runge (Booloumba Creek drainage) indicated that the density varied along a given water- course, In addition, changes in numbers some- limes occurred in a particular urea between Successive visits (i.e. weekly to monthly varia- tion) and was sometimes maintained across one or two seasons (e.g. streams in the Narrows National Park area, Blackall Range. 1973-1975). No one expeeted 7! diurrtus to disappear: the last sightings had to be reconstructed. [tscems it has nol been met in the field since possibly late 1975 on the D'Aguilar Range (C. Corben pers, comm.), early November [978 on the Blackall Range (G. Cveehura, pers, obs.) and January 1979 on the Conondale Range (K. McDonald pers. comm.), The significance of these last sightings of 7. diurnus have taken Sometime to emerge. However, at present all that can be stated with certainty is that during late 1978- early 1979 on the Blackall-Conondale Ranges and possibly as early as 1975 on the D' Aguilar Range (the time of disappearance here is un- known) the last encounters with this frog o¢- curred. and forthe last decade attempts to locate them have failed. Like T. dturnus, Rheobatrachus silus also ap- pears (0 have vanished. Both species were close- ly associated in the wild, although R. silus was never recorded from the D’ Aguilar Range. There are some data from Ingram’s (1983) study for the disappearance of R. silus. He worked with wild populations of the frog in the Conondale Range between October, 1976, and December, 1981. During his study, they disappeared and the study had to be placed in limbo, The last frogs captured by Ingram were two juveniles on 18 October, 1979. Apparently, the last wild frog seen was a juvenile at Ingram’s study site on 8 December, 1979, by Gregory Czechura. (However, Tyler und Davies (1985) reported thal the last known frog died in captivity on S November, 1983, in Adelaide, presumably in their laboratory). In- gram (loc. cit.) had noticed that R. silus had declined belore 1979 and said the number oi captures in 1978 had decreased. He presumed that this was related to late rains. His actual numbers forthe years 1976-1981 forthe number of individuals captured (not cou nting recaptures) were; 1976, 59; 1977, 35; 1978, 24; 1979, 2; 1980, 0; 1981, 0 (Ingram unpubl, data), R, silus was never a common frog, Ingram (1983) caleu- lated there were 1.11 frogs/are on his study site. Barinaga’s (1990) estimate of cammonness ap» pears to be journalistic licence. Both A, st/us and T, diurnus were last seen in 1979 - but what makes 1979 a special year for disappearing frogs in Australia? The report of Heyer et al, (198) of population reductions and extinctions of frogs im southeastern Brazil in 1979 adds fo the intrigue. Is this coincidence? Maybe the question should be rephrased to “What takes 1979 special for the disappearance of trogs on two large, bul remote southern con- tinents?” Ingram (1990) argued that the disap. 3b4 pearances in southeast Queensland were probab- ly due to late rains. falling in cooler months. Heyer et al. (1988) argued that unusually heavy frosts were responsible in southeast Brazil, If the disappearances. are related, it is likely that Os- borne (1989) is correct. He conjectured the cause to be deterioration in climate affecting both can- tinents. However, if climate did change, there is evidence that it might still be changing. Winter and McDonald (1986) noted the sudden disap- pearance of Eungella Gastric Brooding Frog (A. vitellinus) and Eungella Daytrog (7. eungellen- sis) in 1983, These species still had not been Jocated in early 1990 (K, McDonald, pers. comm). There disappearances illustrate another coincidence. Their sister species are respectively the vanished frogs, R. silus and T. diurnus (see Ingram, 1980; Mahony etal., 1984). For wildlife conservation, this subsequent disappearance of closely related species illustrates the wisdom of one of the Berne Criteria for CITES, that ‘...the listing of one Linnean species on Appendix [ of CITES means that other species in the genus. must be listed on Appendix If unless there is a reason for listing them otherwise’ (Holt, 1987, p.20), In fact, the warning signs are there for the safety of the other species of Taudactylus. A project should be commenced to monitor their numbers and study their bialogy. However, it is very difficult to decide whether or not the missing frogs are extinct. Extinction is not easy to prove. If the missing frogs are extinct, their passing was not a slow process. Rather it appears to have been a catastrophic event that could not have been anticipated or prevented, LITERATURE CITED BARINAGA, M. 1990, Where have all the froggies gone? Science 247: 1033-1034. CORBEN, C.J. 1977. A unique amphibian fauna. pp.t6-17. Jn Roberts (1977), CZECHURA, G.V. 1974. A new southeast locality for the skink Anomalopus reticulatus, Herpetofauna 7(1): 24-25, 1975. Notes on the frog fauna of Conondale Range, south east Queensland. Herpetofauna 7(2): 2-4. 1976, Additional notes on the Conondale Range herpetotsuna. Herpetofauna 6(2): 2-4, 1978, A new locality for Literia brevipalmata (Anura: Pelodryadidac) from south east Queensland, Victorian Naturalist 95; 150-151, 1983, The rails of the Blackall Conondale Runge region with additional comments on Latham’s MEMOIRS OF THE QUEENSLAND MUSEUM Snipe Gallinage hardwickii. Sunbird 13(2): 31- 5. 1985. The raptors of the Blackall-Conondale Ran- ges and adjoining lowlands. Corella 9(2): 49-54. 1986a. A new species of Taudactylus (Myobatrachidac) from south eastern Queensland. Memoirs of the Queensland Museum 22: 299-307, 19866, Kroombit Tops Torrent Frog Taudactylus pleione, with a key to the species of Taudaciylus, Queensland Naturalist 27; 68-71, in press. The Blackall-Conondale Ranges: lrogs, reptiles and fauna conservation, Ja ‘The rain- forest legacy. Australian rainforests study. Vol. 2.” Special Heritage Publication Series, (Australian Government Publishing Service: Canberra). GROVES. R.H. (ed.) 1981. ‘Australian vegetation.” (Cambridge University Press: Cambridge). HEYER, W.R., RAND, A.S., DA CRUZ, C.A.G. AND PEIXOTO, O,L, 1988, Decimations, ex- linctions, and colonizations of frog populations in southeast Brazil and their evolutionary im- plications. Biotropica 20(3): 230-235. HOLT, S.J. 1987, Categorization of threats to and status of wild populations. pp.19-30. /n Fitter, R. and Fitter. M. (eds), ‘The road to extinction’. (IUCN: Gland, Switzerland & Cambridge, U.K,). INGRAM, G.J, 1980. A new frog of the genus. Taudactylus (Myobatrachidae) from mid-east- em Queensland with notes on the other species of the genus, Memoirs of the Queensland Museum 20(1): 111-119. 1983, Natural history. pp. 16-35. /# Tyler (1983). $990, The mystery of the disappearing frogs. Wildlite Australia 27(3); 6-7. JOHNSON, C.R. 1971. Thermal relations and water balance in the Day Frog Taudacrylus diurnus, from an Australian rainforest, Australian Journal uf Zoology 19: 35-9. LIEM, D'S. AND HOSMER, W. 1973. Frogs of the genus Tauductylus with descriptions of two new species (Anura: Leptodactylidae). Memoirs of the Queensland Museum 16: 435-457, MAHONY, M., TYLER, MJ. AND DAVIES. M. 1984. A mew species of the genus Rheobatrachus (Anura: Leptodacylidae) from Queensland, Transactions of the Royal Society of South Australia 108(3): 155- 162. MCEVOY, J.S., MCDONALD, K.R. AND SEARLE, A.K. 1979. Mammals, birds, reptiles and am- phibians of the Kilcoy Shire, Queensland. Queensland Journal of Agriculture and Animal Science 36; [67- 180, DISAPPEARING FROGS OSBORNE, W.S. 1989. Distribution, relative abun- dance and conservation status of Corroboree Frogs, Pseudophryne corroboree Moore (Anura: Myobatrachidae). Australian Wildlife Research 16: 537-547. ROBERTS, G.J. (ed.) 1977. ‘The Conondale Range. The case for a National Park.’ (Queensland Con- servation Council, Brisbane). SPECHT, R.L. 1970. Vegetation. pp.44-67./n Leeper, G.W. (ed.), ‘The Australian environment.’ (CSIRO, Melbourne University Press: Mel- bourne). STANTON, J.P. AND MORGAN, M.P. 1977. ‘The rapid selection and appraisal of key and en- dangered sites: The Queensland case study.’ (University of New England School of Natural Resources: Armidale). STRAUGHAN, I.R. AND LEE, A.K. 1966. A new genus and species of leptodactylid frog from Queensland. Proceedings of the Royal Society of Queensland 77: 63-66. TYLER, M.J. (ed.) 1983. ‘The Gastric Brooding Frog’. (Croom Helm: London). 365 1990. “Declining amphibian populations: a global phenomenon?” An Australian perspective. A.C.T. Herpetological Association Newsletter, May 1990, pp.4-12. TYLER, M.J. AND DAVIES, M. 1985. The gastric brooding frog, Rheobatrachus silus. pp.469- 470. In Grigg, G., Shine, R. and Ehmann, H. (eds), ‘Biology of Australasian frogs and rep- tiles.’ (Royal Zoological Society of NSW: Syd- ney). YOUNG, P.A.R. AND MCDONALD, W.J.F. 1987. The distribution, composition and status of the rainforests of southern Queensland. pp.119-142. In Werren, G.L. and A.P. Kershaw, A.P (eds), ‘The rainforest legacy.’ Vol.1. (Australian Heritage Commission: Canberra). WEBB, L.J. 1978 A general classification of Australian rainforests. Australian Plants 9: 349- 63. WINTER, J. AND MCDONALD, K. 1986. Eungella: the land of cloud. Australian Natural History 22(1): 38-43. NOTES ON THE DISTRIBUTION AND OCCUR- RENCE OF SOME STRIPED SKINKS (GENUS CTENOTUS) IN QUEENSLAND: The following provides details on the occurrence and distribution of some species of Ctenotus in northern and western Queensland, Crenotus piankai Storr has been included in the Queensland herpetofauna (e.g. Cogger, 1986) on the basis of 4 specimen in the South Australian Museum (SAM 5387, Doomadgee Mission) reported in Storr (1970), G.M_ Storr (pers. comm.) has informed us that this specimen is referable to Crenoms striaticeps Storr, a species that has since been found throughout the area of north-western Queensland and adjacent parts of the Northern Territory. On this basis, Crenotus piankai should be deleted from lists of the Queensland herpetofauna, Ctenotus decaneurus Storr has been included in the Queensland herpetofauna by Wilson and Knowles (1988), who figure a specimen from ‘Muellers Range, Winton Dis- trict, Queensland’ (Photograph 343, p.264). This view of the distribution of Clenotus decaneurus contrasts with that provided by both Cogger (1986) and Rankin (1978) who indicate that this skink is more-or-less restricted to the area between the northern Kimberley region of Western Australia and western Amhem Land. Wilson and Knowles’ identifica- tion is correct and a relatively extensive distribution of the species through western Queensland is confirmed by refer- ence lo material in the Queensland Museum herpetological collection (QMJ), Three specimens, all clearly referable to Ctenotus decaneurus, indicate that this skink occurs through hummock grass habitats on extensive stony substrates in the arid western half of the State, (QMJ30430 ‘Cloncurry’; QMJ43244 ‘Mica Creek via Mt Isa’: QMJ43267 ‘8&km south west of Winton’, the latter specimen is the one figured by Wilson and Knowles, 1988). Ctenotus inornatus (Gray) has been reported to occur on eastern Cape York Peninsula in the vicinity of Iron Range (Wilson and Knowles, 1988). This population is apparently isolated from its conspecifics in subhumid and semi-arid MEMOIRS OF THE QUEENSLAND MUSEUM northern Australia (between Broome and the southern Gulf of Carpentaria), The occurrence of this skink in what is a humid, wel region appears problematical and requires some explanation. The presence of Crenotus inornatus on eastern Cape York is based on two specimens collected by Donald Thomson from ‘Near Lockhart River’ and held in the Museum of Victoria (MV DT-D273-4). Both specimens are referable lo Cteno(us inornatus (Gray). Although Thomson collected extensively in Arnhem Land, NT, and at Aurukun, W Cape York Peninsula (Dixon and Huxley, 1985), there appears to be no reason to doubt the provenance of these two specimens (AJ. Coventry, pers. comm.). However, the locality given is sufficiently vague in thal the specimens may have actually been taken some distance from Lockhart River, Further comment on the range of Crenotus inornatus in eastern Cape York must await collection of additional Specimens, Literature Cited Cogger, H.G. 1986, ‘Reptiles and amphibians of Australia.’ (Reed; Sydney). Dixon, J.M. und Huxley, L. (eds) 1985. ‘Donald Thomson's mammals and fishes of Northern Australia’, (Thomas Nelson: Melbourne), Rankin, P.R. 1978. A new species of lizard (Lacertilia: Scincidae) from the Northern Territory, closely allied to Clenotus decaneurus Storr, Records of (he Australian Museum 32: 395-407. Storr,G.M, 1970. The genus Crenones (Lacertilia: Scincidac) in the Northern Territory. Journal of the Royal Society of Western Australia 52: 97-108. Wilson, S.K. and Knowles, D.G. 1988. “Australia’s reptiles. A photographic reference to the terrestrial reptiles of Australia.” (William Collins: Sydney). G.V, Czechura und GJ. Ingram, Queensland Museum, PO Rox 300, South Brishane, Queensalnd 410], Australia; 17 August, 1990. THE RELATIVE IMPORTANCE OF HOST BEHAVIOUR, METHOD OF TRANSMISSION AND LONGEVITY ON THE ESTABLISHMENT OF AN ACANTHOCEPHALAN POPULATION IN TWO REPTILIAN HOSTS. CHRISTOPHER B. DANIELS Daniels, C.B. 1990 09 20: The relative importance of host behaviour, method of transmis- sion and longevity on the establishment of an acanthocephalan population in two reptilian hosts. Memoirs of the Queensland Museum 29(2): 367-374. Brisbane. ISSN 0079-8835. The acanthocephalan parasite Sphaerechinorhynchus rotundocapitatus occupies the rec- tum and large intestine of the riparian Australian snake, Pseudechis porphyriacus. Eggs are released into water to be consumed by an aquatic arthropod (intermediate host) which in turn is captured by the eastern water skink, Eulamprus quoyii (transport host). The parasite adopts a resting, encysted stage in both hosts until the lizard falls prey to the snake. Pseudechis porphyriacus exhibits a relatively high frequency of infection but EF. quoyii comprises only 2% of prey items. Aquatic prey also represent only a small proportion of the diet of E. quoyii. Unlike all other acanthocephalans so far examined, there is no evidence that the parasite alters intermediate host behaviour or physiology to increase the chance of capture by the next host in the life cycle. Rather, the operation of the food web appears to provide sufficient momentum to transfer the parasite from one stage to the next, provided both the hosts and the parasite are long lived. Transfer mechanisms involving parasite mediated alterations in host behaviour can be termed ‘active’ while those which do not significantly affect a host are termed ‘passive’. The advantages of passive transfer mechanisms are discussed. L] Parasite, lizard, snake, invertebrate, Acanthocephala, life cycle, passive transport, mathematical model. Christopher B. Daniels, Department of Physiology, School of Medicine, Flinders Univer- sity of South Australia, Bedford Park, South Australia 5042, Australia; 12 July, 1988. Almost all parasites with multistage lifecycles often rely for transmission on the predation of their intermediate host by their final host. How- ever, usually only a small proportion of inter- mediate hosts (infected or otherwise) are captured by predators. Thus, sometimes natural selection influences parasites to alter inter- mediate host behaviour to increase the chance of capture by the final host. Many such parasites produce extreme and often spectacular altera- tions in intermediate host behaviour to move the prey into the feeding niche of the final host and/or decrease the frequency of predation by other inappropriate carnivores (c.g. Holmes, 1976: Moore, 1984). Holmes (1976) suggested that if the final host is an efficient predator, the strategy of the parasite should be to make the prey more conspicuous, and when the predator is inefficient the parasite should make the prey more conspicuous and easier to catch. In both cases the parasite may cither institute novel be- haviour patterns or simply clicit pre-existing host behaviours under inappropriate conditions (Moore, 1984). In cases where the parasite interacts with, and influences, host behaviour to promote transmis- sion, the methods employed can be termed ‘active’. Mechanisms whereby parasites promote their own transmission without in- fluencing host behaviour or physiology can be termed ‘passive’. Thus, increasing parasite lon- gevity and reproductive output may ‘passively’ promote transmission by increasing the number of infective individuals which can contact hosts. Digeneans amplify their numbers in inter- mediate hosts by producing cercariae highly adapted for transmission. Active interactions be- tween parasites and their hosts have received considerable attention recently from be- havioural, physiological, evolutionary and genetic viewpoints (Holmes and Bethel, 1972; Bethel and Holmes, 1973,1974,1977; Clarke, 1979; Smith-Trail, 1980; Brassard et al. 1982; Rand et al. 1983, Schall, 1983) as well as in studies of population dynamics (Holmes, 1982). However, this study will show: (1) the impor- tance of passive forces in influencing a parasite life cycle; (2) present a simple model to illustrate how selection can act on passive mechanisms to increase the probability of parasite transmission 368 from one host to the next; (3) present an example where an apparently active interaction between a parasite and its host in fact represents a method of passive transmission; (4) discuss the ad- vantages and disadvantages of active and passive transmission techniques. The system examined involves an acan- thocephalan parasite of a snake. The parasite has two sequential intermediate hosts: an aquatic invertebrate and a riparian lizard. Because I have been unable to identify the invertebrate host and in order to test the ‘worst case scenario’, | will ignore any possible differences between the first and second intermediate hosts in the acan- thocephalan life cycle. Behavioural changes in the first intermediate host are well documented but are relatively uncommon in the second host, possibly because parasites are usually associated with the first host for a longer period. However, in this system it will be demonstrated that while behavioural transformations in a host may im- prove the probability of transmission, such chan- ges are not necessary for the successful establishment of the parasitic life cycle. MATERIALS AND METHODS This study applies aspects of the ecology of the acanthocephalan parasite Sphaerechinor- hynchus rotundocapitatus, the eastern water skink Eulamprus quoyii and the red-bellied black snake Pseudechis porphyriacus to a simple probability model to test whether the parasite is utilising active or passive transfer mechanisms. More complex ecological and physiological studies of these animals are documented else- where (Shine 1975; Daniels, 1984; Daniels and Simbotwe, 1984) and only the salient charac- teristics will be present here. Sphaerechinorhynchus rotundocapitatus Sphaerechinorhynchus belongs to the order Palaeacanthocephala and contains two species both of which probably utilise snakes as final hosts (Schmidt and Kunz, 1966; Morris and Crompton, 1982). Palaeacanthocephala occupy the intestine of aquatic or semiaquatic ver- tebrates and their intermediate hosts are usually aquatic crustaceans, especially ostracods, am- phipods or isopods (Crompton, 1970, 1975; Morris and Crompton, 1982). Palaeacan- thocephalans sometimes utilise a second inter- mediate host, often a vertebrate (Morris and MEMOIRS OF THE QUEENSLAND MUSEUM Crompton, 1982). The second intermediate host of S. rotundocapitatus is the Australian skink Eulamprus quoyii which consumes infected in- termediate hosts (currently unknown) and the lizards in turn are eaten by the final host (Daniels and Simbotwe, 1984). The fully embryonated eggs of S. rotun- docapitatus measure 0.07-0.09 x 0.025mm. These are released into water in the faeces of the snake, and are immediately infective and retain their infectivity for many months (Johnston and Deland, 1929a,b; Crompton, 1970, 1975). In most acanthocephalan life cycles the eggs are consumed by the correct arthropod host, hatch in the gut and the larval stage (acanthor) burrows through the intestinal wall to reach the haemocoel, The acanthor then develops into an acanthella and encysts. The encapsulated acan- thella is termed a cystacanth (Crompton, 1970,1975). Most of the cystacanths so far ex- amined alter intermediate host behaviour which increase the likelihood of consumption by the final host (Holmes, 1976; Moore, 1984). Infected arthropods are consumed by E. quoyii. The cystacanths hatch and the acanthel- lae again burrow into the peritoneum and encyst. Thirty four percent of 53 E. quoyii contained worms (X = 2.0, S.D. = 2.2, range 1-8), which measured up to 26mm long (Daniels and Sim- botwe, 1984). Cystacanths were removed from the peritoneal wall, liver, outer gut wall, and sperm ducts. These can survive in the lizard for at least 6 months (Daniels and Simbotwe, 1984). Only 8% of juvenile E. quoyii were infected compared to 41.5% of adults (Daniels and Sim- botwe, 1984). A few cystacanths have been recovered from other species of small lizard but not, to date, from frogs (Johnston, 1911, 1913; Johnston and Deland, 1929a; Daniels and Sim- botwe, 1984). Infected E. quoyii exhibited a mean voluntary diving time of nearly 8 minutes, while uninfected ones dived for an average of 4.5 minutes (Daniels, 1985a). Altering the voluntary diving time of E. quoyii may represent an active strategy promoting parasite transport because the red bel- lied black snake forages underwater and thus contacts more infected E. quoyii (Gilbert, 1935; Fleay, 1937; Shine, 1975). Twenty three percent of 22 P. porphyriacus contained adult S. rotun- docapitatus in their rectum and lower large in- testine (X = 2.3; S.D. = 2.3; range 1-7). Female worms measured up to 37mm while males ex- hibited a maximum length of 23mm (Johnston and Deland, 1929b). PARASITES IN REPTILES Eulamprus quoyii The eastern Water skink, Bulampras queyii, 1s # common Inhabilant of creck banks in castern Australia (Veron and Heatwole, 1970; Speller- berg, 1972; Daniels, 1984). This lizard is ter- ritorial and intraspecifically aggressive (Done and Heatwole, 1977) often existing in dense populations. Inthe New England region of north- om New South Wales, &, quayii is active from September to May, and hibernates during the winter (Veron 1968, 1969b). Water skinks ure viviparous, mate in October and the young are born in January/February (Veron, 1969b). Haichlings are 35mm SVL (snout to vent length) and occupy fossorial habitats until they reach 55-80mm SVL. The juveniles then emerge and occupy suboptimal habitats, which are more ex- pased regions often some distance from water. Adults measure 80-110mm SVL and prefer rocky regions near expanses of water. Water skinks become sexually mature in theirthird year and live for 6-10 years (Veron, 1968, 1969b; Daniels, 1984). Water skinks consume at least 25 taxa of prey including insects, worms, frogs, crustaceans, mammuals, spiders, myriapods, snails, lizards and fish (Veron 1968, 1969a; Daniels, 1987). Ap- proximately 25% of prey taxa are aquatic (Daniels, 1987), Of the aquatic prey items, the possible intermediate hosts of 8S. rotun- docapitatus could be; Coleoptra (7.1% of the prey of F. quoyii), Hemiptera (2.9%), Odonata (3.8%), Plecoptera (2.2%), Crustacea (2.2%), Gastropoda (0.7%) or perhaps Anura (0.9% of prey taken by water skinks). Thus, whatever the immediate host, it must comprise less than 7% of the prey items of the transport hosts, A more realistic estimate is probably 1-2%. The most important prey items for £. gueyii are terrestrial Coleoptra (15% of prey) and ants Which vary from 2% to 95% of the prey consumed depend- ing on the season (Veron, 196%a; Daniels, 1957). Water skinks can avoid predators by practising tail autotomy, with 49% of 110 New England lizards possessing regenerated tails (Daniels, 1985b), These lizards also exhibit a diverse range of escape tactics. Of 698 lizards chased by me around streams in New England, 32%% es- cuped by swimming across open water lo cover (tocks or reeds), 5% dived to the bottom of ponds and remained submerged and motionless for at least 2 minutes, while 61% ran fo cover and 2% remained motionless (Daniels, 1984: Daniels and Heatwole, 1990). Pseudechis porphyriacus The red-bellied black snake Pseudechis por- phyriacus is a large riparian elapid common in stream habjtats in eastern Australis. In New England the snake is active from September toa May with 5-8 young born alive in January/February (Shine, 1975, 1978), Hate- hlings measure 24em SVL and reach sexual maturity in their third year (Shine, 1978). Large snakes are 150¢m SVL and at least 10 years old (Shine, 1975, 1978; Daniels, 1984). Red-bellied black snakes consume 34 types of prey including lizards, frogs, mammals, snakes and fish. Inver- (ebrates are almost non-existent dietary items (Shine, 1977), Frogs comprise 82,4% of prey with Limnodynastes tasmantensis the most com- mon (1.9%), Water skinks are only 2% of lhe prey of P. porphyriacus (Shine, 1977). P, por- phyriacus 1s an active forager and can capture prey on land or in water. The snakes will swim underwater for considerable periods in search of tadpoles, fish and other animals hiding amongst litter on the pond bottom (Gilbert, 1935; Fleay, 1937), RESULTS AND DISCUSSION If intermediate hosts comprise less than 7% of the diet of FE. guovif but 34% of E. quoyii are infected with cysts of S. rotundocupttates and if E. guoyli comprise 2% of the diet of P. por- phyriacus but 23% of P. porphyriacus are in- fected, then how can infection occur? One alternative is for the parasite to employ an active transfer mechanism, Some aspects of (he be- haviour of the transport hast indicate this pas- sibility, Parasitised lizards possessed much longer voluntary diving times than unparasitised ones and may be more likely to be captored by P. porphyriacus foraging underwater (Danicls and Simbotwe, 1984). However, it is unlikely (hal the parasite is exerting an active effect on the behaviour or physiology of E. quoyii, fora number of reasons. Firstly, an enhanced volun- tary diving time may not necessarily indicate an increased tendency to use diving as the predominant escape method. Moreover, lizards rarely dived, with only 5% of individuals diving to avoid me (Duniels, 1984). Secondly, neither body mass nor swimming stamina were affected by parasitism (Daniels, 1985a). If the parasite exerted some physiological, behavioural or me- tabolic effect an the lizard to promote diving, it is surprising that the other parameters remained unatiected. Swimming was the predominant aquatic escape tactic employed by E. guoyil and therefore seems a much more suitible mechanism for the parasite to exploit. Acan- thocephalans alter the swimming behaviour of muny invertebrate intermediate hosts (Holmes and Bethel, 1972; Bethel and Holmes, 1973, 1974, 1977). Parasites also often interfere with the Stamina of many vertebrate hosts (Rau and Caron, 1979), Aberrant swimming behaviour may still confer protection from many terrestrial predators while reducing the ability of the lizard to escupe from P. perphyriacus, Diving is so infrequently practised that even if all the divers 1 observed were parasilised they represent a barely significant proportion of the total popula- tion, In addition, if 5% of the lizard population were divers and all Were parasitised, then 85% of the parasitised individuals did not dive, Thirdly, il S, rofundocapitatis actively in- fluences diving, which increases the likelihood of infected lizards being consumed by the snake, (hen the proportion of the dict of P. perphyriacus comprised of parasitised lizards must increase jrom that predicted by random collection, ic. from 34% of all E. quoyii captured, to a muxi- mum of 100% of the water skink component of the snake dict. However, infected lizards cay only increase from 0.68% to 2% of the dietary items of red-bellied black snukes. The snakes do most of their foraging on land, collecting frogs and lizards fram holes or undercover on the stream banks (Shine, 1975, 1977). Comparative- ly few of the prey speevies of the snake are com- pletely aquatic (Shine, 1977), Menee any parasitic alterations to skink behaviour which promote water ulilisation may menther increase nor decrease the likelihood of cansumption by the snake However, some selective advantages for an active strategy can sill exist if the increased ulilisajion of aquatic escape tactics cunfers protection from all predators except ?. por- phyriacas, This does not appear Irhely. Water skinks in the New England region #re potential prey items for 101 species of predater (Danicls and Heatwole, 1984). fifty species are known Lo consume fizards and include mummals. birds, reptiles, fish and invertebrates, An additonal 51 species have been reported to capture other similar sized Vertebrates, OF the total 101 poten- lial predators, 27.7% ure most likely to capture F gueve anly in water, 11.9% carn capture lizards on land or in water, 54.5% are purely terrestrial predators while 5.9% of predators. are MEMOIRS OF THE QUEENSLAND MUSEUM fossorial (Daniels and Heatwole, 1Y84). Thus the use of water as an escape medium still exposes E. quovii to attack from 39 predatory species, several of which are as important, if not more so, than P?. porphyriacus (e.g. kingfishers and herons}. It seems unlikely therefore that S. rojun- docapitaius is using active mechanisms to promote the consumption of the transport host by the final host. The only behavioural aberration so far observed in infected lizards is unlikely to affect capture frequency, particularly as the lizards are so rarcly eaten by snakes. In addition, as a Consequence of any active changes, natural selection may promote host resistance either to the parasite or to the behavioural change. There is always the risk That the response of the host may outweigh any advantage of the changed behaviour, to the detriment of the parasite. All my examinations to date have failed lo isolate the intermediate host. However, itis possible thatthe parasite docs. not employ any active transfer mechanisms in thal host either or thaLemploying active transport mechanisms may nol be neces- sary. I do not know that the parasite affects the first host inthe same mannerasthe second, Many olfer acanthocephalans influence the behaviour of (he first host but are benign in the second. However it is reasonable to hypothesise that passive strategies are the primitive ones from which time and natural selection develop more active methods in some species. As | wish to examine the primitive “worst case” situation | will assume thal in this system transition through the first host is also passive. Holmes (1976) observed that ance in an inter- Wediate host the normal operation of the food web will greatly enhance the probabilities of renching some potential tinal host, In the case of 5, rofuindocapitatus, itis possible (hat the opera- tion of the food web will passively support purasite transfer and enable the establishment of a viable parasite population, provided both the hosts and the worms are long lived, Utilising the knowledge of the ecology of the two vertebrate hosts and the parasite, it is possible to calculate the time required for a parasite population to become established. Assume an infected ( por- phyriaeus enters a riparian habitat previously Iree of the acwnthocephalan, Then S$. rotiai- docapitatuy eggs became abundant and can retain their infectivity for very long periods until consumed by the intermediate host, Assuming that the abundance of eyes result in the very rapid infection ofa substantial proportion of the inter- PARASITES IN REPTILES mediate host population then there are two cru- cial periods in the parasite lifecycle. Firstly, the infection of 34% of the E. quoyii with cys- tacanths and then the infection of 23% of the P. porphyriacus. To calculate the time for the passive transfer of the parasite from the intermediate host to the lizard, six assumptions are made: (1) As aquatic prey comprise 1-7% of prey items, we assume 1% of food items are the intermediate hosis; (2) 33% of intermediate hosts are infected; (3) Lizards eat every day; (4) Lizards cat two types of prey/day; (5) Lizards are active 6 months/year; (6) 20% of eystacanths hatch inthe lizard. and survive to again form a cystacanth. The predicted proportion of infected arthropods is unsubstantiated because the intermediate host is unknown. However, the proportion is similar to the levels of infection observed in the other hosts. Both Daniels (1984, 1987) and Veron (1969a) observed virtually all E. quoyii to con- tain fresh prey, Half of Veron’s lizards contained prey from more than five taxa while most of mine had more than two prey items. The value af 20% of cystacanth viability is an estimate, probably an underestimate. About 20% of cystucanths removed from E. quoyii and fed to P. por- phyriacus developed into adults (Daniels, pers. obs.), Thus the time for 34% of E, quayii to become infected with S, rofundocapitatus: = (% ol prey infected)x(% of prey in diet)x(no. prey items/day )x(% ol lizards infected)x(viability of cystucanths) = (100/33)x(1O0/1)x(1/2)x(34/100)x (100/20) = 258 days or approximately one year Because lizards are only active 6 months/year, ittakes about L.5-2 years for 34% of E. guoyii to become infecied. To calculate the time for 23% of P. por- phyriacus to become infected it is necessary to make five assumptions; (1) Snakes eat one prey item at a time; (2) Snakes eat once every three days; (3) 34% of E. quoyti contain cystacaiths; (4) 20% of cystacanths exsheath and survive to reproduce in P. porphyriacus; (5) Snakes are active 6 months/year, A third of 22 P. par- phyriacus | examined had prey in their stomachs, although Shine observed that a greater percent- age of his snakes had fed recently (Shine, 1977), Tusually found one prey item/snake although the average in Shines’ was nearly three (Shine, 1977). P. porphyriacus probably feed more often (hin assumed here. Thus the time for 23% of 2 porphyriacus to become infeeted is: = (no, of prey eaten/feed)s(frequency ol eating)x(% of snukes infected) a(viability al eystacanths)x(% of lizards infected) = (1)x(3)x(100/2)x(23/1008)x( 100/20) (100/34) = 507 days Snakes are active for only half of the year sa it takes approximately 3 years for 23% of the snake population to become infected, The total time for a parasite population to become established in all hosts is therefore 5-6 years, well within the lifespan of both the lizard and the snake, the more so considering the conservative nature of the calculations. However; can the parasiles maintain their numbers in two reptilian hosts in the face of natural mortality and the short activity period? Assuming lizards live 6 years then 17% of the population die cach year. On third of that, or 6% of the total lizard population die containing parasites. (This is probably an overestimate be- cause during any year, most mortality within the E&. queyii population occurs amongst the juveniles, which are mol parasitised (Veron, 1969a: Daniels, 1984). Thus, to maintain the population stability §, rotundecapitatus must in- fect 6% of the uninfected lizards/year; = (% of prey in diet)x(% of prey in- fected)x(prey items ealen/day) x(% of lizards infected)x(viabilily of cys- tucanths) = (100/1)8¢100/33)x(1/2)x(6/100) x(100/20) = 45 days But 28% of the surviving population ig already infected, Thus, the lime for infection: =45xc. 125/100 = 56 days or approximately 2 months Similarly if snakes live 10 years then the population turnover is 10%/year with up- proximately 2.5% of the population dying while containing parasites (again an overestimate he- cause the greatest mortality occurs amongst juveniles which are relatively unparasitised (Shine, 1978; Daniels, 1984). Therefore, for population stability, approximately 2.5% of the uninfected snakes must collect a parasite/yeur. Thus the time for infection: = (% of prey in dict)x(prey eaten at a lime)x(no. days between feeding)x(% of snakes infected)x(viability of cys- lacunths)x(% of lizards infected) = (100/2)x( 1 )x(3)x(2,5/100)x( 100/20) «( 100/34) = 55 days However, 20% of the surviving population is already infected. Therefore the tine for infec- tion: = 55x120/100 = 66 days or approximately 2 months Hence it takes approximately 4 months for the parasite to replace individuals lost when their hosts die. Four months is well within the yearly activity period of the reptiles, the more so con- sidering the conservative nature of the calcula- tions. This study supports the hypothesis that the operation of the food web will passively transport the parasite from one stage in its life cycle to the next. The time predicted is probably an overestimate. However, as acanthocephalans have separate sexes, it is crucial for at least 2 worms to reach each snake. The number of cys- tacanths per lizard is not important. The conser- vative basis within the calculations provides excess time which may be important for allow- ing extra parasites to infect hosts. Moreover, the great longevity of P. porphyriacus may compen- sate for the low rate of consumption of E. quoyii and enable populations of S. rotundocapitatus to become established in each infected snake. Perhaps the most convincing evidence for the utilised of the passive transfer mechanism invol- ves the very low numbers of parasite/host. It is possible that the low numbers of parasites per host represent a truncated negative binomial dis- tribution because host mortality may be as- sociated with increasing intensity of infection. However, no pathological effects were observed in either host, even in the skink with 8 parasites (Daniels and Simbotwe, 1984). Other animals appear to be capable of supporting large numbers of acanthocephalans without discernible effect (Holmes, 1982). Hence, parasites practising pas- sive strategies rely on time and the laws of prob- ability to promote infection. As each infection is an independent event, then the structure of the parasite population should be characterised by many infected hosts containing relatively few individuals. The distribution of S. rotun- docapitatus in both E. quoyii and P. por- phyriacus clearly falls into the passive type. Parasites can increased the probability of transfer and promote the success of the passive mechanism in three ways. Firstly, by prolonging the survival time and viability of the eggs, cys- tacanths and adults. In some acanthocephalan species the eggs retain their infectivity for more than 3 years (Crompton, 1975) while acanthellae MEMOIRS OF THE QUEENSLAND MUSEUM at least 6 months old have been removed from cystacanths in E. quoyii (Daniels and Simbotwe, 1984). Longevity is crucial because of the slow tate of transfer between hosts. However, a prolonged survival time is especially important for cystacanths and adult worms because of the long hibernation periods of their hosts. Host hibernation results in long periods of food un- availability for adult S. rotundocapitatus and teduces mating opportunities because of the delay in the rate of acquisition of more parasites. Hibernation also affects reproduction because it stops the release of eggs into the correct aquatic habitats. Adult worms must either live a long time and be capable of producing large numbers of eggs or live a short time and produce very large numbers of long lived eggs in order to maximise reproductive output and compensate for the high attrition rate in this type of life cycle. It is also crucial for cystacanths to be capable of exsheathing, transferring and establishing themselves in the next host when the opportunity arises. The value of 20% used in my calculations may be unrealistically low, A doubling of this percentage incurs a major reduction in the time necessary to establish a parasite population. Cys- tacanth viability is probably the most important variable in the population dynamics of the parasite because it is the most malleable. Without a high cystacanth viability it is unlikely that enough cystacanths would reach the final host in time to develop into adults, mate and maintain egg production. Secondly, the passive transfer mechanism could be promoted if the parasite was not specific to one intermediate or transport host. In situa- tions where both the transport host and the final host consume a wide range of prey items all at a low frequency, parasites will transfer more rapidly from one host to the next by using many host species rather than by increasing the level of infection within one specific host. Therefore, cystacanths should tolerate a relatively broad range of physiological conditions in order to survive in many different types of host. Cys- tacanths of S. rotundocapitatus have been ex- tracted from two other small skinks, Hemiergis decresiensis and Lampropholis guichenoti but in both cases the level of infection was less than 6% (Daniels and Simbotwe, 1984). Undescribed acanthocephalan cystacanths have also been ex- tracted from other snakes and lizards (Johnston and Deland, 1929a). However, somewhat surprisingly, cystacanths have not been collected from frogs. From a transfer viewpoint, frogs PARASITES IN REPTILES would be better transport hosts than lizards be- cause they are more important dietary items for P. porphyriacus. However, its appears either that S. rotundocapitatus cystacanths cannot survive in other animals, or the consumption of infected E. quoyii is sufficient to maintain the parasite population. The third mechanism available to S. rotun- docapitatus to promote the passive transfer mechanism involves exploiting behavioural variability within the host population. The presence of parasites in water skinks with ex- tended voluntary diving times may reflect such an exploitation. The tendency to dive may vary greatly between animals within a population. Long dives might be characteristic of lizards which use the water most frequently as an escape medium and also as a food source. These more aquatic lizards may be more likely to feed at the waters’ edge, capture aquatic prey and thus be- come infected with S. rotundocapitatus. When attacked, the more aquatic lizards may exhibit a greater tendency to swim or dive, conferring a degree of protection from terrestrial predators but not from P. porphyriacus (Daniels, 1985a). Thus the parasite can exploit a polytheism within the host population to increase the chance of reaching the final host. The presence of S. rotun- docapitatus correlates with, rather than causes, a behavioural or physiological difference and is therefore less likely to stimulate host resistance. Active transfer mechanisms can be disad- vantageous when the intermediate host repre- sents a small proportion of the diet of the final host, the final host consumes a wide variety of prey and the intermediate host is a prey item for a large range of different predators. This type of food web is common in Australian ecosystems and specialist systems involving one predator and one prey are rare (although there are a num- ber of vertebrates which specialise on ants or termites). The relative absence of simple food webs, especially amongst the vertebrates may be a result of the low vertebrate biomass. It may be impossible for one carnivore to specialise on one, or a few, prey species because the densities of the latter are too low. The optimal transport strategies for parasites in diffuse and complex food webs are often passive rather than active, provided reproductive output remains high enough to compensate for the attrition. ACKNOWLEDGEMENTS This paper grew from discussions with 373 Richard E. Lenski and Karen M. Calloway to whom I am particularly grateful. A.F. Bennett, T. Garland, J.J. Schall, J.C. Holmes and R. Shine also provided comments and criticisms. Julie Dodwell and Karen Rowe typed the manuscript. LITERATURE CITED BETHEL, W.M. AND HOLMES, J.C. 1973. Altered evasive behaviour and responses to light in am- phipods harbouring acanthocephalan cys- tacanths. J. Parasit. 59: 945-956. 1974. Correlation of development of altered evasive behaviour in Gammarus lacustris (Am- phipoda) harbouring cystacanths of Polymor- phus paradoxus (Acanthocephala) with the infectivity to the definitive host. J. Parasit. 60: 272-274. 1977. Increased vulnerability of amphipods to predation owing to altered behaviour induced by larval acanthocephalans. Can. J. Zool. 55: 110- 115. BRASSARD, P., RAU, M.E. AND CURTIS, M.A. 1982. Parasite-induced susceptibility to preda- tion in diplostomiasis. Parasitology 85: 495-501. CLARKE, B.D. 1979. The evolution of genetic diver- sity. Proc, R. Soc. Lond. B. 205: 453-474. CROMPTON, D.W.T. 1970. ‘An ecological approach to acanthocephalan physiology.” (Cambridge Univ. Press: Cambridge). 125 pp. 1975. Relationships between Acanthocephala and their hosts. Symp. Soc. Exp. Biol. 29: 467-504. DANIELS, C.B. 1984. The adaptations to a riparian habitat by the eastern water skink Sphenomor- phus quoyii. (Unpublished Ph.D. thesis., Univer- sity of New England). 1985a. The effect of infection by a parasitic worm on swimming and diving in the water skink Sphenomorphus quoyii. J. Herpet. 19: 160-162. 1985b. The effect of tail autotomy on the exercise capacity of the water skink Sphenomorphus quoyii. Copeia 1985: 1074-1077. 1987. Aspects of the aquatic feeding ecology of the riparian skink Sphenomorphus quoyii. Aust. J. Zool, 35; 253-258. DANIELS, C.B, AND HEATWOLE, H. 1984. The predators of the water skink Sphenomorphus quoyii. Herpetofauna. 16; 6-16. 1990, Factors affecting the escape behaviour of a riparian lizard. Mem. Qd Mus. 29(2): 375-387. DANIELS, C.B. AND SIMBOTWE, M.P, 1984. The biology of acanthocephalan parasites of Australian skinks. J. Herpet. 18: 211-213. DONE, B.S. AND HEATWOLE, H. 1977. Social behaviour of some Australian skinks. Copeia 374 1977: 420-430. FLEAY, D. 1937. Black snakes in combat. Proc. R. Zool. Soc. N.S.W. 1936/37: 40-42. GILBERT, P.A. 1935. The black snake (Pseudechis porphyriacus, Shaw). Proc. R. Zool. Soc. N.S.W. 1935/35: 35-37. HOLMES, J.C. 1976. Host selection and its conse- quences. pp. 21-39. In Kennedy, C.R. (ed.), ‘Ecological aspects of parasitology.’ (North Holland Publ. Co.: Amsterdam). 1982. Impact of infectious disease agents on the population growth and geographical distribution of animals. pp. 37-51. Jn Anderson, R.M. and May, R.M. (eds.), ‘Population biology of infec- tious diseases.’ (Springer-Verlag: Berlin). HOLMES, J.C. AND BETHEL, W.M. 1972. Modification of intermediate host behaviour by parasites. Zool. J. Linn. Soc. 51:123-149. Suppl. 1. JOHNSTON, T.H. 1911. A census of Australian rep- tilian entozoa. Proc. R. Soc. Qd 22: 232-249. 1913. Notes on some entozoa. Proc. R. Soc. Qd 24: 62-109. JOHNSTON, T.H. AND DELAND, E.W. 1929a. Australian Acanthocephala No. 1. Census of recorded hosts and parasites. Trans. R. Soc. S. Aust. 53: 146-152. 1929b. Australian Acanthocephala No. 2. Trans. R. Soc. Aust. 53: 153-166. MOORE, J. 1984. Altered behavioural responses in intermediate hosts. An Acanthocephalan strategy. Amer. Nat. 123: 572-577. MORRIS, S.C. AND CROMPTON, D.W.T. 1982. The origins and evolution of the Acan- thocephala. Biol. Rev. 57: 85-115. RAND, A.S., GUERRERO, S. AND ANDREWS, R.M. 1983. The ecological effects of malaria on populations of the lizard Anolis limifrons on Barro Colorado Island, Panama. pp. 455-471. /n Rhodin, A.G.J. and K. Miyata, K. (eds), ‘Advan- ces in herpetology and evolutionary biology.’ (Mus. Comp. Zoology: Cambridge, Mas- MEMOIRS OF THE QUEENSLAND MUSEUM sachusetts). RAU, M.E. AND CARON, F.R. 1979. Parasite-in- duced susceptibility of moose to hunting. Can. J. Zool. 57: 2466-2468. SCHALL, J.J. 1983. Lizard malaria: Parasite-Host ecology.pp. 84-100. /n Pianka, E.R., Huey, R.B. and Schoener, T.S. (eds.), ‘Lizard ecology: Studies of a model organism.’ (Harvard Univ. Press: Cambridge, Massachusetts). SCHMIDT, G.D. AND KUNZ, R.E. 1966. Sphaerechinorhynchus serpenticola sp. nov. (Acanthocephala: Sphaerechinorhynchinae) a parasite of the Asian cobra Naja naja (Cantor) in Borneo (Malaysia). J. Parasit. 52: 913-916. SHINE, R. 1975. Ecological studies on Australian elapid snakes. (Unpublished Ph.D. thesis, University of New England). 1977. Habitats diet and sympatry in snakes: A study from Australia. Can. J. Zool. 55: 1118-1128. 1978. Growth rates and sexual maturation in six species of Australian elapid snakes. Her- petologica 34: 73-78. SMITH-TRAIL, D.R. 1980. Behavioural interactions between parasites and hosts: Host suicide and the evolution of complex life cycles. Amer. Nat. 116: 77-91. SPELLERBERG, I.F. 1972. Thermal ecology of al- lopatric lizards (Sphenomorphus) in south-east Australia. III]. Behavioural aspects. Oecologica 11: 1-16. VERON, J.E.N. 1968. Aspects of the role of tempera- ture in the ecology of the water skink Sphenomorphus quoyii. (Unpublished M.Sc. thesis, University of New England). 1969a. An analysis of the stomach contents of the water skink Sphenomorphus quoyii. J. Herpet. 3: 187-189. 1969b. The reproductive cycle of the water skink Sphenomorphus quoyii. J. Herpet. 3: 55-63. VERON, J.E.N. AND HEATWOLE, H. 1970. Temperature relations of the water skink Sphenomorphus quoyit. J. Herpet. 4: 141-153. FACTORS AFFECTING THE ESCAPE BEHAVIOUR OF A RIPARIAN LIZARD CHRISTOPHER B. DANIELS AND HAROLD HEATWOLE Daniels, C.B. and Healwole, H. 1990 09 20) Factors affecting the escape behaviour of a riparian lizard. Memoirs of the Queensland Museum, 29(2): 375-387. Brisbane. ISSN 0078-8835. The escape tactics employed by water skinks, Eulamprus guoyit, are determined by their immediate location and anentation and by the physical characteristics of the habitat. Running was the most frequently practised type of escape, with rocks the preferred form of cover. Swimming and diving were employed lo a lesser extent. Juvenile water skinks did not differ in escape behaviour from adulls. Waiter skinks use only a small proportion of their maximal Jocamotor and diving abilities during escape. Short dives and/or short bursts of swimming or running enable escape from d predator and still allow the individual (a resume normal aclivities almost immediately (or execule more evasive manoeuvres). Juveniles are poorer swimmers and divers but only slightly poorerrunners than adults.) Eulamprus quayii, eseape diversity, running, swim- ming, diving, micohabitat, Christopher B. Dantels and Harold Heatwole, Department of Zoalogy, University of New England, Armidale, New South Wales 2351, Australia; C.B.D. present address: Depart- ment of Physialogy, School of Medicine, Flinders Ualversity of South Australia, Bedford Park, SA, 5042, Australia; 12 July, 1988, The selection of an appropriate cscape response is a critical factor in survival, and recently has been analysed using lizards as model animals (Jaksic and Nunez, 1979; Sim- botwe, 1983; Schall and Pianka, 1980), The lat- {er found that lizards alternated among various methods of eseape, and predicted that prey populations faced with higher per capita preda- tion pressure should evolve more diverse escape tactics than less heavily predated conspecilic populations. The present paper deals with the more proximate nature of diversity and escape tactics, by asking the following question: When a lizard is threatened, what factors determine which of the escape tactics in its repertory it will use? The subject of the study was the water skink Eulamprus quoyit (formerly Sphenomarphus quoyii), a tiparian lizard that eludes predators by running to terrestrial cover, and/or by swimming or diving (Veron and Heatwole, 1970), STUDY AREAS The study was undertaken at three locations near Armidale, NSW, Australia, in summer (January and February) 1983. Each site com- prised 1.5km of stream bank. They were struc- turally different but all supported large populations of water skinks, The heterogeneity of cach habitat was determined by quantifying the emergent ground cover in 10 m®* sections every 100m along the stream bank. The percent- age of each section covered with small racks (surface area< 1 5em’), medium-sized rocks (15- 100em*), large rocks (>100cm"), vegetation and fallen timber, and bare ground, was recorded. Rocks 15-100cm’ seemed to be optimal basking sites. Stream speed was measured at each section by timing the rate of passage of a floating ball down the fastest flowing 10m. The characteristics of the three sites are presented in Table 1. Boorolong Creek, 26km southwest of Ar- midale, was a wide, slowly flowing stream usually divided into very large pools by granite outcrops or fallen trees; the banks were densely lined with reeds, grasses, blackberries (Rubus sp.) and Casuarina. The Gara River, 15km north of Armidale, was a narrow, rapid stream with many small pools located at the periphery. The Stream contained many large rock outcrops but small rocks were comparatively scarce. The banks were densely lined with reeds, grasses and bottlebrushes (Callistemon sp.). Blue Hole was located 15km downstream from the Gara River study site and 15km northwest of Armidale. The banks were thickly covered with granite boulders of a variety of sizes with a correspond- ing reduction in streamside vegetalion. Although MEMOIRS OF THE QUEENSLAND MUSEUM TABLE 1. Charactetisties of the three streams Used as Study sites expressed as x% + SE. (n), Open Bank Small rocks Medium sized rocks Large rocks Vegetation Flow rate (ms) midstream flow is rapid, the rocks allow the formation of numerous small, quiet pools, The vegetation consisted primarily of grasses, black- berries and casuarinas. All sites were located in dry sclerophyll forest or open woodlands. In summary, Bluc Hole was much rockier, more open and Iess vegetated than the other twa sites, and Gara River had a higher rate of stream flow than the other two. MATERIAL AND METHODS ECOLOGICAL FACTORS Observations were undertaken only on fine sunny days with li(tle or no wind and air tempera- ture of 25-30°C. Light falls of rain occurred intermittently during the study period but were not sufficient to greatly alter stream morphology or rate of flow. The stimulus used to induce escape was a person approaching on foot al a moderate speed from the landward side. The same person was used throughout, For each escape event the fol- lowing data were recorded: (1) initial microhabitats and body orientation of the lizard with respect to the water and to bank cover, (2) execution of changes in direction from the initial orientation, but prior to fleeing, (3) microhabitat selected for escape and (4) escape tactic employed. Data from adults and juveniles were recorded separately; juveniles were defined as water skinks shorter than 85 mm snout to vent (Veron, 1969), The categories of initial microhubitat were: (1) water, (2) on a rock surrounded by water, (3) on a rock on the bank and (4) on the bank. A lizard’s initial orientation was tallied as (I) facing the water, (2) facing the bank or (3) paral- lel with both. Alteration in orientation prior to fleeing was listed as nearest to 0°, 45°, 90", 135°, or 180° from original orientation. The covers selected for protection were 33.92 6.5 (15) 19.3= 4.4 (15) (0,0+ 3,7 (15) 6.9+ 2,2 (15) 30.3 6.4 (15) 0.10 0,05(15) classed as (1) reeds, (2) rocks, (3) the bank or (4) open water, Water skinks utilised terrestrial (running), or aquatic (swimming and/or diving) escape tac- tics. Swimming was always undertaken on the surface with the head above water, while diving was defined as a descent to the bottom of a pond and remaining motionless there in leaf litter or tock crevices, usually for several minutes. Oc- casionally lizards did nol move when ap- proached and were casily caught. Another escape response, active surface swimming fol- lowed by a dive, was only occasionally observed and represents the onl y overlapping of categories of escape tactics. Values for the breadth of each habitat niche were calculated using Simpson's (1949) diver- sily index (DS), Microhabitat niche overlap and escape behaviour overlap were calculated using Pianka’s (1973) index, Statistical analyses employed the chi-squared goodness of fit and the G test for independent samples (Sokal and Rohlf, 1969; Snedecor and Cochrane, 1978). Since observations were made at different times of day, a lest was made to see whether there was a temporal effect. Morning and afternoon values were not significantly different for microhabitats prior to escape (G = 0.86; P>0,05; niche overlap between morning and afternoon = 0.999), escape tactic (G = 4.26; P>0.05), original oricntation (G=1.24; P> 0.05) and degree change in orientation (G = 4,10; P>0.05). Consequently, data from both periods of the day were pooled for subsequent analyses. Datla on microhabitat prior to escape were also pooled for juvenile and adult lizards as at all sites there were no significant differences between the age classes (G = 1.94, 5.40, and 6.80 for Blue Hole, Gara River and Boorolong Creek respec- tively: P>0.05 in all cases; niche overlap be- tween juveniles and adults in the three respective areas were 0,988, 0.922 and 0.938). There were LIZARD ESCAPE BEHAVIOUR too few juveniles for testing the other results for Gara River. However, there were no significant differences between adulis and juveniles in es- cape tactics (G= 5.12 and 5.64 for Bluc Hole and Boorolong Creek respectively, P>0.05 in both cases) original orientation (G = 1.96 and 0.94; P>0,05) and degree change in orientation (G = 2.28, 4.34; P>0.05), and data were pooled for these as well, PHYSIOLOGICAL FACTORS Eulamprus quoyit were tested to determine voluntary diving time, swimming speed, stamina, running speed and running stamina ac- cording to methods already described (Daniels, L984b, 1985), Lizards were collected by hand from Tea Tree Creck and Boorolong Creck be- tween February 1982 and December 1983, Adulls and juveniles were both maintained as previously described (Daniels, 1984b, 1985). In the experiments, unless otherwise slated, air temperatures of 30-32" and water temperatures of 19-20°C were used. These temperatures ap- proximate air and water temperatures during summer at creeks inhabited by the species (Pidgeon, 1978) and straddle its mean ereli body temperature of 29,6°C (Spellerberg, 1972a,c), Diving time determined by timing 30 dives for each of 21 lizards al a water temperature (mean +§.D,) of 19.1° £ 1.7°C and an air temperature of 31.6° + 1.1°C. Swimming stamina was ex- amined by maximally exercising 16 lizards until exhaustion (trial 1). The time taken before the lizard stopped swimming was recorded. A second swimming trial (trial 2) was conducted on the same lizard 20 minutes later ta determine recovery capacity. Two trials per lizard were conducted every day for 10 days. Running speeds were determined by stimulating 17 lizards to run the length of a 161 x 40 x Siem glass tank with a sand substrate at Ta = 31.h° + L.L°C. Ten trials were conducted for each lizard With 20 minutes rest between trials. Juvenile £. quoyii were timed while running a distance of 68cm, Running stamina was also determined at an ambient temperature of 18.1°+0,8°C in atank (24,5 4 38.5 x 37.5em) with a dirt substrate. Nine lizards were chased, one at a time, around the periphery of ihe tank until they refused to move after five consecutive taps on the base of the tail, A low ambient temperature was chosen to match that of the water temperature during the swim- ming slamina experiments, so as lo make direct comparisons between the two types of stamina 377 possible. A second running endurance trial Was conducted for each lizard 20 minules aller the first; lwo such trials were candocted each day for 10 days. All times were recorded using a Lauris Stop- watch accurate to 0.02 seconds, From the repeti~ tions of all experiments mean values were calculated for each individual; the maximury value was also analysed. All results were ex- pressed as means + siandard errors (S.E.) except air and water temperatures which were ex- pressed as means + standard deviations (S.D.}. Statistical analysis employed Student t-test and paired t-tests (Sokal and Rohlf, 1969). Speeds were expressed in metres per second (ms''), and stamina and diving limes were expressed in seconds (s). In addition to these laboratory studies, voluntary diving, swimming and run- ning times were determined in the held. Water skinks were chased into the water at Blue Hole and Boorolong Creek and timed till they swam tocoveror, if they dived, until re-emergence. Air and water temperatures were recorded at the site of entry into the water. Running times were recorded at Blue Hole and Boorolong Creek for lizards chased from their basking sites towards rocks or teeds. Air temperatures were recorded at the basking sites, RESULTS MICRONABITAT Undisturbed water skinks were usually atsun- lit sites, particularly on rocks near the waler (Table 2). The second highest usage was of the stream bank, Only uccasionally were lizards in the water, floating or resting on algal mats. Thus, there seems to be a clear preference for emergent streamside sites rather than for the water itself, There were three major terrestrial microhabitats available, rocks, open banks and vegetation (Table 1). When the observed fre- quencies of their use by lizards were tested by Chi-square analysis against the values.cxpected on the basis of their relative cover, the observed values departed significantly from the expected ones (X*, Blue Hole 54,1, Gara River 298.0, Boorolong Creek 86.4; P<<0.005 in all cases), indicating the lizards were not randomly as- sociated with substrate type. No lizards were found in the vegetation at any locality despite it having mean cover values in the three areas uf 7. 5-55.0% bolrocks and open habitats had rather high frequencies of usage. Thus, the lizards seemed to stlect non-vegetated microhabitats. + MEMOIRS OF THE QUEENSLAND MUSEUM TABLE 2, Microhabitat preferences (frequency of use) of Eulamprus quoyii prior to escape. In Waler On rock in water On rock on bank On bank Niche breadth ‘To test whether rock or open banks were favoured, a second Chi-square analysis was per- formed testing observed frequencies of use of ihese microhabitats relative to their proportional representation in non-vegetated stream sides, At Blue Hole where rocky habitat was abundantand bare bank relatively rare (Table 1) the lizards inhabited banks more often than expected by ifs relative cover (X?= 25.4; P<<().005). Atthe Gara River where open bank was well represented and rocks less common, the hzards favoured rocks (X? = 96.0, P<P>0,025). It may be that the lizards tend to preferentially use whichever is the rarer of these two microhabitats. Niche breadth was greatest at Boorolang Creek, the locality with the most even coverage of hubitat types, and least at Blue Hole where habitat diversity was lowest, racks ac- counting for over 78% of total microhabitat (Tables 1, 2). Overlap in microhabitat niche was high among sites (Table 2). Gara River and Boorolong Creek, the two sites most similar in habitat charac- teristics (Table 1), did not differ significantly in the frequency of use of different microhabitats by the lizards (Table 3), The greater dissimilarity in habitat characteristics of Boorolong Creek was reflected in a significant difference in fre- quency of microhabitat use by lizards there in comparison to the other two places (Table 3), ESCAPE BRNAVIOUR When approached, lizards usually escaped sig- miicantly more often by running rather than by MiCROHABITAT BLUE HOLE GARA RIVER 0.048 0.136 0.552 0.264 i24 204 188.86 (1) 2.19 (1) <(0,05 >0.05 swimming or diving. Rocks were the preferred cover sought regardless of the escape route or mode of locomotion employed (Table 4), The diversity of escape responses was lowest at Blue Hole, which had the greatest uniformity of microhabitat and the greatest proportion of rocks, the preferred escape cover (Tables 1, 4), Although the overlap in escape tactics was high umong areas, cach arca differed significant- ly lrom every other in the frequency with which different tactics were employed, and with one exception (Gara River vs Boorolong Creek), in the frequency with which different modes of locomotion were used (Table 3). The most notable differences were that in the rockiest area (Blue Hole) lizards ran to rocks more often for escape than at the other localities, and employed aquatic escape less frequently (Table 4). In the terrestrial situation water skinks at Blue Hole employed terrestrial escape tactics significantly more often than did those from Boorolong Creek (Gy) = 11.26; P<0,05); those that originally faced pardlle! to or toward the banks escaped Icss frequently to water at Blue Hole than did those with similar initial orientation at Boorolong Creek (Gj = 4,96; P<0.05). It would seem that when rocks are abundant they are preferentially used for escape, bul when they are less abundant, escape by diving or swimming increases in fre- quency. Rate of stream flow also may be a factor, Blue Hole, where aquatic escape was lowest, had the swiltest current (Table 1). Unmolested water skinks most frequently faced the water, and once disturbed, they tended to move away in the direction they were original- ly facing, usually to the nearest rock (Table 5). Thus, either the lizards tended to face toward predetermined escape routes or perhaps merely LIZARD ESCAPE BEHAVIOUR 379 TABLE 3, Comparison among localities of the microhabitat selection and escape tactics of Eulamprus quoayii. GARA RIVER VS BooROLONG CREFK MICROHABITAT Niche overlap 0.884 G (df) 20.18 (3) P <0.05 ESCAPE TACTICS 0.933 17.76 (9) <(0).05 Overlap G (df) P Escape TACTIC SUBTOTALS G (df 3.16 (3) P >0.05 fled in whatever direction they happened to facing. Approach was from the landward side and that may have influenced direction of escape. The 90° and 180° turns occurred mast often when the lizard was initially facing the direction from which the person approached. However, a landward approach would be expected to dive the lizards into the Water, rather than along ter- restrial escape routes. Such did not occur and direction of approach did not seem to be an overriding factor. Unfortunately, data on original orientation and degree change of direction are available only for two sites (Bluc Hole and Boorolong Creck). These two areas did not differ in regard to the direction undisturbed lizards faced (Table 5). By contrast, those from Boorolong Creck tended to alter their original orientation in order to escape, significantly more often than did those trom Blue Hole (Table 5). This is probably because the abundance of rocks at Blue Hole provided escape avenues in almost all directions, but at Boorolong Creek where rocks were less than half as abundant, a lizard would more often have to change direction in order to head for rocky cover. Comparison of the antecedent behaviour of lizards escaping aquatically with those escaping terrestrially, revealed a number of important dif- ferences, For example, at all study sites lizards that were already in the water or on rocks sur- rounded by water used an aquatic avenue of escape more often than a landward escape; those initially located on land escaped significantly more often to the land than to water (Table 6). Thus, the microhabitat occupied at the time of BOOROLONG CREEK vs BLUE HOLE BLUE HOLE vs GARA RIVER 0.830 51.92 (3) 0.10; Blue Hole r= -0.4, P>0,10). The data (Table 6) again confirmed the greater ten- dency toward terrestrial escapes at Blue Hole compared to Boorolong Creek. In neither area, however, are lizards that escuped in the direction they initially faced more (or less) prone to escape toward a particular medium than are those which showed greater amounts of turning after distur- bance. LOCOMOTOR PFRFORMANCE Adult and juvenile Eulamprus quoyii both ex- hibited similar diving behaviours. Lizards often submerged vertically in a crevice in the diving tank, und remained motionless with eyes closed throughout the dive. Both size groups exhibited a mean voluntary diving time in the laboratory MEMOIRS OF THE QUEENSLAND MUSEUM TABLE 4. Escape tactics of Eulamprus quoyii at three sites. Swim : lo Reeds : to Rocks ‘to Bank : then Dive >in Open Water rin Reeds : in Rocks > to Rocks : to Reeds Remain Motionless Dive Run 137.74(9) <0,05 2.95 X* (df) P Escape Diversity ESCAPE TACTIC SUBTOTALS Swim Dive Run Remain Motionless of between S and 6 minutes (Table 7), and the difference between the two proups was not sig- nificant (tuo, = 0.65, P>0.05). However, the adults did have significantly longer maximum dives (tiv) = 1.86, P<0.05). The longest dive for an adult E. quoyii was 35.4 minutes and the longest dive by a juvenile was 15.1 minutes. There was.a significant trend in both groups for mean voluntary dive time to decrease as the num- ber of completed dives increased (Daniels, 1984a,b, 1985). In the field, water skinks dived for only short periods. Mean voluntary diving time was 2.5 + 2,00 minutes (n = 12) with the longest dive being 12.5 minutes (water tempera- ture 24.1°+ 0.51°C; air temperature 28.2° + 087°C). The skinks usually swam in an anguilliform manner (Batholomew et al., 1976). The legs were held laterally against the side of the body and thrust was developed by lateral undulations of the body and tail. As the lizards became ex- hausted, tail undulations became weaker and propulsion was maintained by body movements and, prior to complete exhaustion, with flipper- like actions of the forelegs. The size groups exhibited significantly different mean speeds (ts) = 85.37, P<0.05) and maximum speeds (t15) = 122.77, P<0.05) (Table 11). The fastest adult S. quoyii swam at 1.34ms"' and the fastest 165.50(9) <0.05 5.14 157.82(9) <0.05 4.97 juvenile at 0.4ms". In the first swimming trial, lizards continued swimming for more than two minutes. Juveniles had slightly less stamina (mean swimming stamina: twa) = 1.10, P>0.05). The longest endurance for an adult in trial 1 was 181s and for a juvenile 163s, If adult water skinks could sustain a mean swimming speed of 0.64ms' for two minutes, then they could swim up to 80m before becom- ing exhausted. However, 124 adults and juveniles inhabiting New England creeks and chased into water of 21.0 + 0.86°C (air tempera- ture 23.5 + 1.09°C) swam for 2.19 + 0.15s, with the maximum swimming period being 79.5s. There was no significant difference in these swimming times between adults and juveniles(adults: 3.69 + 0.94s, n= 88; juveniles: 1.65 + 0.19s, n = 36; Student t-test: ty22) = 1.37, P>0.05). Lizards rarely swam more than 2-3m before emerging onto rocks or the bank. Twenty minutes after being exercised maxi- mally, lizards could only sustain short bursts of swimming (Table 7). Mean swimming stamina was significantly less in trial 2 than in trial 1 for both groups (adults: tiasy= 20.73, P<0.05; juveniles: tim) = 13.39, P<0.05). The greatest stamina for an adult in trial 2, was 121s and for a juvenile 100s. Adults swam for significantly longer than juveniles in trial 2 (mean stamina: LIZARD ESCAPE BEHAVIOUR 381 TABLE 5S. Original orientation and degree change in direction by Eulamprus quoyii from Blue Hole and Boorolong Creek. ORIGINAL ORIENTATION Facing Water Parallel to Water/Bank Facing Bank DEGREE CHANGE IN DIRECTION tua) = 2.28, P<0.05; maximum stamina: ti4) = 2.78, P<0.05). The intraspecific differences observed for swimming stamina were not observed for run- ning stamina (Table 7). Juveniles could sustain activity for as long as adults in trial 1 (mean Tunning stamina: tq) = 0.45, P> 0.05; maximum stamina : t(7) = 0.21, P>0.05). Maximum running endurance for an adult in trial 1 was 142s and for a juvenile 144s. Adults and juveniles also pos- sessed similar recovery capacities (Table 7) (mean running stamina in trial 2: t7) = 1.05, P>0.05; maximum stamina t7) = 0.42, P> 0.05). During trial 2 the longest endurance of an adult was 103s and that for a juvenile was 108s. As in the swimming stamina trials, the mean exercise times in trial 2 were significantly shorter than those of trial 1 (adults, t(43) = 12.47, P<0.05; juveniles, ts) = 13.00, P<0.05). No significant differences existed between the mean running and swimming stamina of adults (trial 1: tasi) = 1.65; P>0.05; trial 2: tas) = 1.66; P>0.05). Juveniles swam longer than they ran in trial 1 (tos) = 2.56, P<0.05). BLUE HOLE 38.15(2) <0.05 91.802(4) <0.05 BOOROLONG CREEK 25.44(2) <0.05 2.54(2) >0.05 73.654(4) <0.05 Adults ran faster than hatchlings (Table 7), (mean speed: tis) = 27.89, P<0.05; maximum speed t(1s) = 34.78, P<0.05). The fastest speed for an adult was 1.52ms" and that for a juvenile was 0.76ms'!. When approached in the field, lizards utilised only short bursts of running. The mean time it took lizards to reach cover, once dis- turbed, was 1.90 + 0.80s (n = 56) (air temperature = 28.2° + 0.87°C. DISCUSSION DIVERSITY OF ESCAPE TACTICS Water skinks are potential prey for over 100 different species of vertebrate predators (Daniels and Heatwole, 1984), yet they exhibited a rela- tively low index of escape diversity. The values of 2.95-5.14 (Table 4) rank in the bottom third of the 14 species of lizards so far studied (Table 8). The small diversity value reflects the high proportion of individuals that retreated to a single type of cover (rocks). It may be that such a tactic is effective against a wide range of the common predators of water skinks. Another pos- MEMOIRS OF THE QUEENSLAND MUSEUM al nm ‘(\UBOTTIUSIS JOU) GO'O<_ BLOIPUT SYStIAISe ON “(JUBSTFIUSIS ATYSIY) TOO >d RIIPU! SyStaise OMT ‘(URDITIUSIS) CO'O>JO q B OARS SIOqUINU [RN}OR dy] UO Jsaj JeNbS-1yD & SayVOIpU! SLIOT9}eI UONPIUANO [PUI PUR TEIIGeYOIOIA 1OJ Sanjea Nj 1B YSayse uy USI oStl V6 o$P ol) NOILVLNAIYO NI AONVHD 3qeaDdd yurg suloRy DOYS YEA JaT[RIE JAR AA SUTRA] NOILVINAINO TYNISINO yurg uQ yurg uo yooy uQ JarAA UL YOY UO JayeM UT IVLISVHOUIA TVINLSAMHA OLLVNOY TVIMLSANAAL OILWANOY TVIMLSAMHA OV AOY MAAN ONOTONOOY YaATY VaVH ATOH AVIG “UONIAIIP UI aSuryo aa1dap pur voNR}UavO JRUIsIUO ‘uOTDaTaS JeIGeYyoroU oO} AON snaduvjny Jo sadvosa [eNsaiay pur sNeNbe Jo aseuaciag’9 ATAVL 383 LIZARD ESCAPE BEHAVIOUR VL*8'98 OCFSLI YP COFO SCL 9ZFT9OL =F «66 TFBL9 OTFEL 90'0* PS'0 TO'OF6E0 Opps L0F8'0 LO1F0'0L yecr06S 9 SOFE PST SecrL sll 9 LIFL69 DOFEL cO'0F9€°0 TO'0*670 Or re L080 8'6SF6E8 9 OTL cle L 9TFL69 DOFEL (ww) ydua} (8) awit, u JUdA-INOUS —- SSE STUNAANL SOOFLET 10°0*96'0 0*6't6 SL+9 PL L0'0*S8'0 10'0*'9'0 Ocal FTSLOT ELIF 09e ow OT OL val u QEFO'SOL BCFB'ST 61F0' COL SFP ET S1+0'COL L1*6'Se 6TF0TOL S1FPET STF COL S1F9 PC (ww) ysua| (3) jUdA-jNOUS ssp sLINdy SO0FTS8I SOFT 81 T1L+9'Te C86! Vc*8'ol L'0*T61 winipayw OIFO'TE (IFO LE PI+9 le PFO le PIF Te PiFole UIF9'Lle qty (9.) FUNLVYTd WALL wn “XP Ss uRo|A] cf [PHL Ss “XP Ss ural T [PUL VNIAVLS ONINNAY su “XR Su UPd aagdS ONINNOY 8 “XPIA s uPaA Z IPL s “XBIAL s uray | PULL VNINVLS ONINWIMS sui “XP su una aaadS ONINWIAS I- s “XR S ura ONIAIG S.LINA ASIOYAXY ‘ndonb snadunjng ayuaant pur ynpe yo sanioedes asiosaxa ayy “2 ATE VL sibility is that this species adjusts its escape tactic to a particular kind of potential predator, e.g. using different tactics for snakes than Tor birds, or for humans, By using only one stimulus (humans) we may have observed only the subset of escape tactics given to large land animals. Analysis of predator-specific escape responses might be a fruitful line of research, We would like to caution against over-inter- pretation of such diversity indices. The index is influenced by the number of categories of escape response. Had we used each of our rock sizes us a separate category (Table I) rather than “rocks” as w single one, the diversily index would have been altered. We sugyvest thal some of the varia- lion in escape diversity may merely ceflect dif- ferences in investigator's classifications of escape tactics, The utilisation of water by primarily tereestrial lizards for predator avoidance is not uncommon, being practised by over 50 species fram 10 families (Daniels, 1984a), Most of (hese lizards swim across open Water to avoid terrestrial predators. A number submerge und res| on [he bottom until a predator departs, Water skinks ulilise Short aerobically sustained dives (Daniels, 1984a). The predator may still be in the vicinity when the lizard emerges, but because the skink is not exhausted it can submerge again if threatened, Moreover, because anacrobiosis has not been extensively employed, the skinks are still capable of rapid swimming to avoid predators, Thus, short dives keep all escape op- tions open. In addition, the use of swimming, diving and/or running in sequence has the ad- vantage of providing flexibility if one method of escape is suddenly unavailable. For example, if disturbance of rock cover removes running or olher terrestrial tactics a8 4 Viable escape method, swimming or diving still remain. The second major tactical advantage concerns locomotor capacity. Water skinks can dive for up to 35 minutes and swim or run continuously for aver 2 minutes, However, in natural situations, these lizards cither do not, or do not have to, perform at these maximal capacities, By exhibit ing dives of 2-5 minutes and sprints of 2 seconds in the field, water skinks are utilising only a small proportion of their maximal escape capability, Such behaviour enables the lizards.to avoid # potential predator in a given encounter and still be able to resume normal daily activities (or fee from vther predatars) almost immedi- ately, These lizards have flexibility in escape tactics and the behavioural and locomotor MEMOIRS OF THE QUEENSLAND MUSEUM cupacity to maximise the effectiveness of any of the escape tactics selected, INTRASPECIFIC: DIFFERENCES Ontogenetic changes. in diving performance have been observed in several ectotherms and Butler and Jones (1982) suggested that dive duration should be proportional to (body mass)" for ectotherms. The values obtained on water skinks in the present study are consistent with that view. Ontogenetic increases in swim- ming speed have been observed for Amblyr- hvnchus cristaius (Bartholomew et al,, 1976) and some anurans (Taigen and Pough, 1979; Miller, 1983), Running speed increases. on- togenetically for several African lizards (Huey, 1982; Huey and Hertz, 1982). The swimming stamina and recovery capacily of juvenile water skinks was much lower than that of adults, al- though the running stamina of the two size groups were similar. In view of the lower capabilities of juveniles for aquatic life it is surprising that their escape responses did not differ significantly from those of adults. However, use of water for escape is low in all age groups, and even the adults have nol developed unusual physiological adaptations for diving (Daniels, Oakes and Heatwole, 1987). EFFECT OF HABITAT ON ESCAPE TACTICS F. guoyil is a territorial, solitary heliotherm, Sunlit rocks, especially (hose on a soit substrate suitable for burrowing, are the preferred microhabitat for basking and other activities. (Spellerberg, 1972a,b,c,d, 1974; Dane and Heat- wole, 1977), Because of their close proximity to cocks and their use of them for other uctivitics, it is not surprising that water skinks also use racks as a primary shelter during escape. Schall and Pianka (1980) observed thal the escape diversity of Cnemidophorus was 40 cor- telated with environmental heterogencity (plant volume diversity or percent vegetative cover) but rather with presumed predation pressure (high incidence of tail breaks). Aside from the doubtfulness of incidence of broker tails as an indicator of predation pressure (¢.g. intraspecific fighting also may injure tails), such an explana- tion cannot be used to explain the differences in escape diversily be(ween Blug Hole and Boorolong Creck lizards in the present study. Samples Irom both sites bad similar frequencies of animals with regenerated tails; Blue Hole, original tails 54%, regenerated tails 46% (N = 15); Boorolong Creek, original tails 55%, 385 LIZARD ESCAPE BEHAVIOUR ‘roded ay] ul pajuasaid eyep wos) payepNoyeo sonjea = (epg61) sjaiueg (€g6T) ejoquiig (€g61) aMioquig (6L61) ZOUDN 3 oIsyeL (6L61) ZOUNN 2% oIsyeL (1861) Us0y] 29 Suamneg (1861) us0y], 7 suamneg Apmis SIWL. (0861) ByUeld %F TTEYIS (OR61) PXUEId F NPYS (O861) BURL, WF TTEYOS (0861) BXURld % [TEY9S (O86T) BAUEI, 7 [TPYOS (O861) BURL 7 TTEY9S (€g61) sdwieys SSDUSIOJOY ssajuonoy] SuULUIeWIY jo Aduanba4 « ‘OAIP = Q “WIMS = S “UNI = y (997) 0} 991) WO) duin[= f¢ -ssayUONOW Urea = anenby /eIsaa nanansay snyousisdyd jes10qiy sisuaiqoy? "7 jeaoqiy sisuadva snjAja0posayT jerisaliay, SNIDISMUUA] "'T [ewsaay, snosnf{ snwavjorT yenysayaL sa[eulay [BLySo1l9 J, So] RW 4 DADAIAIA DU49ID'T JPlsaq9L, onenby ndonb snadunjng [PlSa119|, ssi) *D [e1ysaqs3 |, snjojassat "9D yeiisai3 SnyvUudoul “Dd [elaysaua yp, supjns “2 JPINSO19 |, sinduvsxa “9 JBlysauay | (jjesaao) snsoydopiwauy sajtuaant jelsaay, 4 SNauav SOUY sasseg | aourproay |Aisisarq yeyigey saisadg jo | 10jeparg 105 oor Joquiny | sanbruysoy | aderosq ‘sprezty Aq Aitjiqowuit Jo sorjoRid ayy pue sonar) adeasa yo AysI2AIq *8 TIGV.L regenerated tails 45% (N = 22). These frequen- cies are not significantly different (Ga) = 0.04, P>0.05) between the two areas. Contrary to the prediction of Schall and Pianka (1980), mode of escape was related to the physi- cal characteristics of the habitat. Blue Hole was the rockiest site and a greater proportion of water skinks there escaped to rocks than at any other locality. There was a significant increase in the utilisation of aquatic escape tactics from the rocky to the vegetated sites and from the fast- flowing to the slow-flowing sites. It appears that if rock cover is available, then lizards prefer to run to it rather than utilise aquatic escape be- haviours. Running to rocks may represent the energetically least expensive form of escape be- cause the sprint distance is frequently short and rocks represent a secure form of protection. Moreover, diving into cold water results in decrease in body temperature which may retard locomotor capacity (Hertz et al., 1982; Daniels, Heatwole and Oakes, 1987). Fast stream flow may pose the threat of sweeping the lizard away and probably requires greater energy expendi- ture during swimming. In conclusion, water skinks alter their escape tactics depending on their immediate location and orientation and the physical characteristics of the habitat, and seem to select the energcetical- ly least demanding option. As a result they use only a small part of their capacity in a given escape attempt, leaving sufficient reserves for repeated attempts, or for other activities. ACKNOWLEDGEMENTS We are grateful to Janet Taylor, John Murray, Guy Jenkins, Malumo Simbotwe, Maria McCoy and Stephen Phillips for assistance, to Victor Bofinger and Stuart Cairns for statistical advice, to A.F. Bennett, T. Nicholas, R.V. Baudinctte, K. Christian and R. Shine for critical comment, and to H. Jones, J. Dodwell, Viola Watt and Sandra Higgins for typing. Part of this project was supported by a Commonwealth Postgraduate Research Award. REFERENCES BARTHOLOMEW, G.A. BENNETT, A.F. AND DAWSON, W.R. 1976. Swimming, diving and lactate production of the marine iguana. Amblyr- hynchus cristatus. Copeia 1976; 709-19. BAUWENS, D. AND THOEN, C, 1981. Escape tac- MEMOIRS OF THE QUEENSLAND MUSEUM tics and vulnerability to predation associated with reproduction in the lizard Lacerta vivipara. J, Anim. Ecol. 50: 733-43. BUTLER, P.J. AND JONES, D.R, 1982. The com- parative physiology of diving in vertebrates. Adv. Comp. Physiol, Biochem. 8: 179-364. DANIELS, C.B. 1984a. The adaptations to a riparian habitat by the water skink Sphenomorphus quoyit, (Unpublished Ph.D. Thesis, University of New England). 1984b. The effect of infection by a parasitic worm on swimming and diving in the water skink Sphenomorphus quoyit. J. Herpetol. 19: 160-2. 1985. The effect of tail autotomy on the exercise capacity of the water skink Sphenomorphus quoyii, Copeia 1985: 1074-77. DANIELS, C.B. AND HEATWOLE, H. 1984. Predators of the water skink Sphenomorphus quoyii. Herpetofauna 16: 6-15. DANIELS, C.B., HEATWOLE, H. AND OAKES, N. 1987. Heating and cooling rates in air and during diving of the Australian water skink, Sphenomorphus quoyii. Comp. Biochem. Physiol. 87A: 487-92. DANIELS, C.B., OAKES, N. AND HEATWOLE, H. 1987. Physiological diving adaptations of the Australian water skink Sphenomorphus quoyii. Comp. Biochem. Physiol. 88A: 187-99. DONE, B.S. AND HEATWOLE, H. 1977. Social behaviour of some Australian skinks. Copeia 1977: 420-30. HERTZ, P.E., HUEY, B.B. AND NEVO, E. 1982, Fight versus flight: Body temperature influences defensive responses of lizards, Anim. Behav. 30; 676-79. HUEY, R.B. 1982. Phylogenetic and ontogenetic determinants of sprint performance in some diurnal Kalahari lizards. Koedoe 25: 43-8. HUEY, R.B. AND HERTZ, P.E. 1982. Effects of body size and slope on sprint speed of a lizard (Stellio (A gama) stellio). J. Exp. Biol.97: 401-9. JAKSIC, F.M. AND NUNEZ, N. 1979. Escaping behaviour and morphological correlates in two Liolaemus species in central Chile. Oecologica 42: 119-22, MILLER, K. 1983. The role of lactate production in the metabolic support of locomotion by clawed frogs Xenopus laevis. Physiol. Zool. 56: 580-4. PIANKA, E.R, 1973, The structure of lizard com- munities. Ann. Rev. Ecol. Syst. 4: 53-74. PIDGEON, R.W.J. 1978. Energy flow in a small stream community, An evaluation of the effects of different riparian vegetation. (Unpublished Ph.D. Thesis, University of New England). SCHALL, J.J. AND PIANKA, E.R. 1980. Evolution LIZARD ESCAPE BEHAVIOUR of escape behaviour diversity. Amer, Nat. 115: 551-66. SIMBOTWE, M.P. 1983. On the spacing patterns and diversity of escape tactics in diurnal geckos (Lygodactylus) in Kafue flats, Zambia. Amph.- Rept. 4: 35-41. SIMPSON, E.H. 1949, Measurement of diversity. Na- ture 163: 688. SNEDECOR, G.W. AND COCHRANE, W.G. 1978. ‘Statistical Methods.’ 6th edit. (Iowa State Univ, Press: Ames). SOKAL, B.B, AND ROHLF, F.J. 1969. ‘Introduction to biostatistics.’ (W.H. Freeman and Co.: San Francisco). SPELLERBERG, I.F. 1972a. Temperature tolerances of south-east Australian reptiles examined in relation to reptile thermo- regulatory behaviour and distribution. Oecologica 9; 23-46, 1972b. Thermal ecology of allopatric lizards (Sphenomorphus) in south-east Australia. I. The environment and lizard critical temperatures. Oecologica 9: 371-83. 1972c, Thermal ecology of allopatric lizards (Sphenomorphus) in south-east Australia. II. Physiological aspects of thermoregulation. Oecologica 9: 385-98. 1972d. Thermal ecology of allopatric lizards (Sphenomorphus) in south-east Australia. III. Behavioural aspects. Oecologica 11: 1-16. 1974. Influence of photoperiod and light intensity on lizard voluntary temperatures. Br. J. Her- petol. 5: 412-20, STAMPS, J. 1983, Territoriality and defense of predator refuges in juvenile lizards. Anim. Behay. 31: 857-70. TAIGEN, T.L. AND POUGH, F.H. 1979, On- togenetic changes in activity metabolism of postmetamorphic toads (Bufo americanus), Am, Zool. 19: 492. VERON, J.E.N. 1969. The reproductive cycle of the water skink Sphenomorphus quoyii. J. Herpetol, 3: 55-63. VERON, J.E.N. AND HEATWOLE, H. 1970, Temperature relations of the water skink Sphenomorphus quoyii. J. Herpetol. 4: 141-53, AN INEXPENSIVE FORCE PLATFORM FOR USE WITH SMALL ANIMALS: DESIGN AND APPLICATION TIM HAMLEY Hamley,T. 1990 09 20: An inexpensive force platform for use with small animals: design and application, Memoirs of the Queensland Museum 29(2): 389-395. Brisbane, [SSN 0079-8835 A force platform was designed and manufactured to meet the following criteria: 1, be inexpensive and able to be constructed fromm readily available components, 2, be suitable for use with animals of less than 1kg and able to to provide an indication of the direction and timing of the forces produced by a lizard during locomotion. 3. be small enough ta record forces from a pair of ipsilateral feet only, but large enough to allow a reasonable chance for a running lizard to place its feet on the platform, A force analysis of Bearded Dragons (Amphiholurus barbatus) and Water Dragons (Physignathus lesueurii) indicated that unlike the more erect mammals, the legs of lizards apply no forward force to the ground during the limb cycle. Instead of the accelera- tion/deceleration cycle thal occurs in the limbs of erect mammals, the lizards apparently apply a ‘rotational® force that simply alters the angular momentum of the limb. These findings are discussed. (1) Lizards, locomotion, force platform. Tim Hamley, 609 Fairfield Road, Yeronga, Queensland 4104, Australia; 15 July 1989. The locomotor performance of animals can be investigated in a number of different ways: trackway analysis, for example, can provide es- timates of speed, stride length, animal size ete. even for animals that have long been extinct (Thulborn, 1982; Thulborn and Wade, 1984); anatomical studies can elucidate certain locomotor constraints (e.g. Russell and Rew- castle, 1979); and kinematic analysis has been used to provide comparative data useful in un- derstanding locomotor abnormalities (Parker and Bronks, 1980). However, none of these methods can provide more than an estimate of the way in which forces are transmitted to the ground during locomotion; to measure these ground reaction forces, a force platform is re- quired. Unfortunately, force platforms are not always readily available, are often complicated, sometimes not completely suitable and usually extremely expensive. In this paper I describe a force platform that was used as part of a larger study of the locomotion of two species of agamid lizard. The platform is cheap, simple to construct and suitable for use with small animals. THE FORCE PLATFORM DESIGN CRITERIA To produce an inexpensive force platform that could be constructed from readily available com- ponents, was suitable for animals less than Ikg, and would provide an indication of the direction and timing of the typical forces produced during the locomotion of the lizards used in this study. The platform was to be small enough to record forces from a puir of ipsilateral feet only, but big enough to allow a reasonable chance for a run- ning lizard to place its feet on the platform. TRANSDUCER ELEMENTS The most expensive components of a force platform are the transducer elements used to convert variations in the applicd forces into sig- nals that can be recorded and analysed. In the system described here, inexpensive, commer- cially-available crystal microphone elements were modified (Fig. ] and see below) to act as force tranducers. The output from these crystal elements varies with the rate of change of the force (i.e. the first derivative of the force, Fig. 2) thereby making an estimation of the actual value of applied forces an integrating process (either electrical or mathematical). However, as. the primary purpose of the force platform was to provide an indication of the direction and timing of the forces applied to the ground by a moving lizard, actual values of the forces were not re- quired and integration of the crystal output was considered unnecessary. CONSTRUCTION (Fig. 3) The surface plate for the force platform was microphone case ) MEMOIRS OF THE QUEENSLAND MUSEUM acoustic diaphragm aluminium bridge FIG. |. Exploded view of the crystal microphone element used to provide the force transducers. A: shows the position of attachment of the glass bead. provided by a piece of aluminium plate 10cm square and 3mm thick. An aluminium cube 25mm to a side was cemented to the centre of the plate and a transducer clement was attached to each of the five remaining surfaces of the cube. The transducer clement, when removed from the | i i] | iJ ! | i ! U 1 \ on ,; constant ! off ' microphone case and acoustic diaphragm, con- sisted of a peizo- electric crystal wafer with two small rubber ‘feet’ mounted on diagonally op- posite corners of one side. On the other side an aluminium bridge spanned the two remaining corners. Before fitting the transducer elements to FIG. 2, Force diagrams: Transducer output, the first derivative of the force (A), compared with the actual force (B) its integral. A FORCE PLATFORM FOR LIZARDS 391 FIG. 3. Exploded view of force platform. 1: Lateral view of plate and base; 2: Ventral view of plate (note the arrangement of transducer elements); 3: Dorsal view of base. the surface plate, a 3mm diameter glass bead was cemented at the apex of the bridge of each ele- ment. The surface plate was recessed into and supported by the body of the force platform which was also made from aluminium. Initially the surface plate was separated from the body of the force platform by a thin layer of low density foam but this arrangement was found to be over- ly sensitive and a more rigid, silicone jointing compound (silastic) was later used. The silastic effectively damped the plate and the amount used was varied, in conjunction with the degree of signal amplification, to suit the size of the lizard that was running over the platform during any particular set of trials. Five fine-thread brass screws were mounted in the body of the platform in such a way that they could be screwed up until just touching the glass bead of the transducer element. To reduce friction, and hence ‘crosstalk’ between the transducers, the ends of the screws were machined and polished and coated with teflon grease. Coarse grade sandpaper was glued to the top of the force plate which was recessed into the floor on one side of the runway. Electrical leads from each of the transducers were connected to a Grass 79D four- pen chart-recorder and the force platform was calibrated in situ before and after a set of trials for cach lizard. As there were five transducers, but only four recorder pens, the output from lateral and medial transducers was duplexed to a single chart-recorder pen so that a laterally 392 MEMOIRS OF THE QUEENSLAND MUSEUM VERTICAL FIG. 4, Calibration of the force platform. A,B,C,: Horizontal transducer outputs; D: Vertical transducer outputs. 1-9 directions of application of force to the plate. directed force resulted in an initial pen displace- ment in one direction and a medially directed force in the other. CALIBRATION Fig.4 is an example of the calibration record obtained from the force platform before and after each set of trials for a particular lizard. As shown, repetitive sequences of force were applied at 45° angles through 360° of a horizontal plane and vertically downwards. Force increments in each series were 50g and the forces applied ranged from 50g to 250g. As nearly as could be determined: 1. each of the transducers reacted instantaneously to the ap- plication a force; 2. each transducer reacted if a component of the applied force acted on it, and; 3. ‘crosstalk’ from transducers that had no component of the force acting on them was minimal. PERFORMANCE Hegland (1981, p.333) has listed eight at- tributes of an ideal force platform. It should * (1) be able to resolve the vertical, forward and lateral components of the force; (2) have low “crosstalk” between the measured components of the force, (3) have sufficient sensitivity and resolution for the subject of interest; (4) have a linear response; (5) have a response independant of where on the plate surface the force is exerted; (6) have a high natural frequency of oscillation; (7) have sufficient safety margin to protect both the plate and subject from damage due to failure; and (8) be simple and inexpensive.’ The force platform described here meets most of these requirements but is perhaps questionable in three of them: (a) Linearity of response (no.4) - this was not critically assessed but is, to a large extent, dependent on the transducer elements. Improvement in this area would require better, hence more expensive, crystal elements, which is not necessary under the stated design criteria. (b) Response independent of position on the plate (no.5) - the response of this plate was found to vary slightly the farthur a foot was placed from FIG. 5. Derivatives of the components of [he ground reaction forces: (As applied by the lizard). A: Lateral force component; B: Forward component; C; Backward component; D: Vertical (downward) component, f and h are the points at which fore and hind feet respectively were placed on the force platform, A FORCE PLATFORM FOR LIZARDS QUADRUPEDAL BIPEDAL 0.71m/s 1.70m/s 2.31m/s | Mra A . UVa QUADRUPEDAL 0.58m/s 1.04m/s ene re pen ee A. barbatus 393 the centre of the plate. This factor was controlled during the study by only using records where the foot or feet had been placed centrally on the plate. (c) Frequency of oscillation (na.@) - the frequency of oscillation is a function of the joint- ing compound used to damp the plate and can be controlled only marginally. However, in terms of the design criteria of the platform, this is relatively unimportant. In general, the force plat- form deseribed here was found to fulfill the design criteria adequately and be capable of providing records of the direction and timing of the components of the ground reaction forces produced by a running lizard. METHODS Two Bearded Dragons (Amphibolurus bar- bars: snout-vent lengths of 180mm and 239mm and Weights of 148,5p and 403,5g respectively), and two Water Dragons (Physignathus lesweurtt: snout-vent lengths of 172mm and 212mm and weights of 180,9g and 337,5g respeclively) were used in the trials. Each lizard was encouraged to tun in cither direction along the runway as often aS Was Necessary fa produce acceptable force records for right and left ipsilateral feet, Trials were extended to obtain force records for the hind feer of Water Dragons moving bipedally, RESULTS Fig.5 shows typical records of the forces ex- erted by both species of lizard dunng quad- rupedal locomotion and by Water Dragons during bipedal locamotion, Force records for left and right feel were essentially the same and only records from right feet are presented here for comparison. All records exhibit three distinct components of the force; a lateral component, a hackWard component and a vertical component. None of the records gave any indication that there was a forward component to the force exerted by a lizard during locomotion. Two major peaks are evident in cach of the foree components for all animals during quadrupedal locomotion and these correspond to the rates of force application by a front foot followed by a hind foot. At lower specds, minor peaks can be seen within the major peaks for each foot, but as speed increases the minor peaks in the profiles become less obvious. All three components of the force for each foot were initiated at the same time and, at lower speeds, have approximately the same rate and duration of application. At MEMMIRS OF THE QUEENSLAND MIISFLIM higher speeds, however, the rate of application of the backward force, by a front foot, was less than that for the vertical and lateral components of the same foot. As is to be expected, the dura- tion of the power stroke at higher speeds was considerably less than it was for lower speeds, Farce records for bipedal locomotion are similar to records for hind feet during fast quadrupedal locomotion - cach component of the force con- sisting of only one major peak with a relatively smooth profile, The initiation and duration of all three force components was the same but the maximum rate of application of the backward component was consistently less than that for the other {wo components. The duration of the power stroke during bipedal locomotion varied only slightly with speed and was similar to the duration of the power stroke during fast quad- rupedal locomotion, DISCUSSION Because the force platform had to be preloaded lo different degrees during cach trial, the force records presented in Figure 5 can not be used to provide estimates of the absolute magnitudes of the forces applied to the plate by a running lizard. However, the components of the force can be compared within each trial to provide an indica- tion of their relative importance. For con- venience, forces are discussed here in terms of “as applied by the lizard’; for example, a back- ward force means the foree, or force component, applied by the lizard in a caudal direction result- ing in the animal moving forwards. Perhaps the most surprising result of the force analysis in this investigation is the lack of any indication of a forward component to the horizontal force exerted on the ground. The al- ternation of forward and backward forces during a locomotor cycle is responsible for the fluctua- tion in kinetic energy and consequently for a large part of the energy cost of locomotion. Yet Alexander (1977) showed mathematically that for a range of bipedal and quadrupedal animals il is energetically more efficient to incorporate both backward and forward components of force into the locomotor cycle. lt would seem from Alexander’s (1977) formulation that the locomo- tion of lizards is energetically inefficient and yet Bakker’s (1972) investigations give evidence to the contrary, Although a mathematical analysis of the forces applied during lizard locomotion is beyond the scope of this investigation there may be 4 simple solution to this apparent conflict, A FORCE PLATFORM FOR LIZARDS Alexander’s model of applied forces relies (as he points out) on the height at the hip being greater than half the step length (step length = distance the animal moves while a foot is on the ground), which it apparently is for most animals (Alexander, 1977). Hip height in a lizard, how- ever is less than 30% of total step length. In fact, the way lizards apply force to the ground may even be the reason for their slightly greater locomotor efficiency (Bakker, 1972) than mammals of a similar size. Although Jenkins (1971) has shown that significant femoral abduction occurs in many species of small mammals, the movement of their limbs is still essentially forwards and backwards: energy must be expended to accelerate and decelerate the limb in both directions. By comparison, the movement of the limb of a lizard during locomo- tion is essentially rotatory: energy need be ap- plied only to change its angular momentum. Although this may be an oversimplification, work by Fedak et al. (1982) has shown that the energetic cost of the changes in the potential energy of the limbs of some bipeds and quad- tupeds during locomotion was not as high as expected because the essentially parasagittal movement of the limb in these animals also included a slight rotatory component. Further investigation in this area should prove instruc- tive. ACKNOWLEDGEMENTS I am indebted to Dr Neil Gribble for his con- tribution to the design and operation of the force platform, and to Dr Tony Thulborn for his en- couragement and assistance through all stages of the project from which this work was taken. 395 LITERATURE CITED ALEXANDER, R.MCN. 1977. Mechanics and scal- ing of terrestrial locomotion. pp. 93-110. Jn Ped- ley, T.J. (ed.), ‘Scale effects in animal locomotion’. (Academic Press: London). BAKKER, R.T. 1972. Locomotor energetics of lizards and mammals. The Physiologist 15:76. FEDAK, M.A., HEGLUND, N.C. AND TAYLOR, C.R. 1982. Energetics and mechanics of ter- restrial locomotion. Journal of Experimental Biology 79:23-40. HEGLUND, N.C. 1981. A simple design for a force plate.to measure ground reaction forces. Journal of Experimental Biology 93:333- 338. JENKINS, F.A. 1971. Limb posture and locomotion in the Virginia Opossum Didelphis marsupialis and in other non-cursorial mammals. Journal of Zoology, London 165: 303-315. PARKER, A.W. AND BRONKS, R. 1980. Gait of children with Down Syndrome. Archives of Physical Medicine and Rehabilitation 61:345- 351 RUSSELL, A.P. AND REWCASTLE, S.C. 1979. Digital reduction in Sitana (Reptilia: Agamidae) and the dual roles of the fifth metatarsal in lizards. Canadian Journal of Zoology 57:1129- 1135. THULBORN, R.A. 1982. Speeds and gaits of dinosaurs. Palaeogeography, Palaeoclimatol- ogy, Palaeoecology 38:227-256. THULBORN, R.A. AND WADE, M. 1984. Dinosaur trackways in the Winton formation (Mid- Cretaceous) of Queensland. Memoirs of the Queensland Museum 21:413-517. 396 TROPIDONOTUS MAIRIT VS BUFO MARINUS: — Bufo marinus was released in sugar-growing districts of eastern Queensland in 1935-6 and now occurs widely in that State and in northeast New South Wales. It is a highly toxic species (Meyer and Linde, 1971, p.522), The bulk of the venom is contained in the parotid glands. Venom is also secreted by smaller glands that cover the whole animal and toxins have been identified in other parts of the body, e.g. blood and ovaries (Meyer and Linde, 1971). The eggs also contain toxins (Licht, 1967) and, although there are no data on toxicity of the larvae, it does not seem unreasonable to assume they are also toxic. Several native vertebrates can utilise B. marinus as a food source by eating eggs, larvae, newly metamorphosed young (e.g. Jungle Perch, Kuhlia rupestris;Snapping Tortoise, El- seya latisternum; Green Tree Snake, Dendrelaphis punct- ulatus; Common Keelback, Tropidonotus mairii) and selected body organs (e.g. Crow, Corvus sp.; Swamp Hen, Porphyrio porphyrio; White Ibis, Thresikiornis mollucca; Water Rat, Hydromys chrysogaster); or by ‘washing’ adults prior to ingestion (captive Estuarine Crocodiles, Crocodylus porosus have been observed washing B. marinus by shaking them vigorously in water for several minutes prior to success- ful ingestion) (Covacevich and Archer, 1975;Hamley and Georges, 1985; G. Ingram, pers. obs.). Australian frog-eating snakes are known to be particularly susceptible to the toxin of B. marinus with one exception. This species the Common Keelback Snake (Tropidonotus mairii) has been regarded as the most successful and only regular native predator of B. marinus. Itis known to consume large numbers of eggs, larvae, and newly metamorphosed young (Lyon, 1973; Covacevich and Archer, 1975; C. Tanner pers. comm.). 7. mairii, a colubrid snake (subfamily Natricinae), is closely related to other natricine species oc- curring in North and South America, Asia and Europe (Mal- nate, 1960) where Bufo spp. also occur naturally. Its apparent high tolerance of Bufo marinus venom was attributed to the long evolutionary association of natricines and bufonids in these areas. In December, 1976 a dead specimen of Tropidonotus mairii (total length 85cm) with a young adult Bufo marinus (head width 2.5 cm), one third ingested from the vent (rather than the head as is usual with snakes), was found in the dry MEMOIRS OF THE QUEENSLAND MUSEUM bed of Richter’s Creek, 10km north of Cairns, NEQ (Queensland Museum registration no. J 28417 ). There were no marks on the snake to suggest death from an encounter with a possible predator, and the toad is no larger than other frogs or small mammals commonly consumed by snakes of comparable size. Death from toad toxin is the only apparent explanation. The discovery of this single known unsuccessful encounter between T. mairii and B. marinus is not conclusive evidence that larger B. marinus are invariably toxic to T. mairii, but this species is apparently more susceptible to B. marinus toxins than was supposed, particularly because in the case reported here, the snake had begun to ingest the toad from the rear, thus avoiding toxin concentrations in the parotid glands immediately behind the head. Literature Cited Covacevich, J and Archer, M. 1975. The distribution of the Cane Toad, Bufo marinus, in Australia and its effects on indigenous vertebrates. Memoirs of the Queensland Museum 17(2): 305-10, pl.41. Hamley, T. and Georges, A. 1985. The Australian snapping tortoise Elseya latisternum: a sucessful predator on the introduced cane toad. Australian Zoologist 21: 607- 610. Licht, L,E, 1967. Death following possible ingestion of toad eggs. Toxicon 5: 141-2. Lyon, B. 1973. Observations on the Common Keelback Snake, Natrix mairii, in Brisbane, south-eastern Queensland. Herpetofauna 6(1): 2-5. Malnate, E.V. 1960. Systematic division and evolution of the colubrid snake genus Natrix, with comments on the subfamily Natricinae. Proc. Acad. Nat. Sci. Philad. 112(3); 41-71. Meyer, K. and Linde, H.1971. Collections of toad venoms and chemistry of the toad venom steroids. pp. In Bucherl, W. and Buckley, E.E. (eds), ‘Venomous Animals and their Venoms’. Vol. 2. (Academic Press: New York). G.J. Ingram and J. Covacevich, Queensland Museum, PO Box 300, South Brisbane, Queensland 4101, Australia; 16 August, 1990. THE GENERIC CLASSIFICATION OF THE AUSTRALIAN TERRESTRIAL ELAPID SNAKES M. N. HUTCHINSON Hutchinson, M. N. 1990 09 20: The generic classification of the Australian terrestrial elapid snakes. Memoirs of the Queensland Museum 29(2): 397-405. Brisbane. ISSN 0079-8835. The generic arrangement for the Australian elapid snakes has been unstable because, in part, of a lack of phylogenetic data by which monophyletic taxa could be recognised. Recently there have been advances in our understanding of Australian elapid phylogeny. These are summarised and a revised classification is proposed. This is based, as far as the data will allow, on monophyletic units. Evidence for monophyly is drawn from karyotypic, electrophoretic, immunological and internal and external anatomical data. [ Serpentes, Elapidae, taxonomy, phylogeny, Australia. M. N. Hutchinson, South Australian Museum, North Terrace, Adelaide, South Australia 5000, Australia ; 1 June 1990. The Australian terrestrial elapid genera have had an unstable recent taxonomic history. Mengden (1983) thoroughly reviewed the his- tory of Australian elapid snake nomenclature, pointing out the conflicting views of the workers who have tackled this problem, and also noting areas where lack of data inhibited resolution of taxonomic problems. Cogger (1985) also reviewed elapid taxonomy, concluding that its history of largely intuitive analysis of mor- phological variation was responsible both for the prolonged instability of elapid systematics and the present disagreements over generic boun- daries. He anticipated that an objective, biologi- cally well-based taxonomy would only be achieved following a clear understanding of phylogenetic relationships. For much of the twentieth century snake workers (including Kinghorn [1929; 1956] in his influential guides) followed Boulenger’s (1896) arrangement, with relatively few genera diag- nosed by features of anal and subcaudal scala- tion, head shield modification, number of maxillary teeth and general appearance. Worrell (1961; 1963) expressed his conviction that the fauna was more diverse by proposing several new genera, although his views were ignored until McDowell’s (1967; 1969a; 1970) studies supported some of Worrell’s suggestions. McDowell’s comparative anatomical data led him to identify what he called ‘natural groups’, implying monophyly. Instead, some of his own analysis indicated that he formed some groups based on their /ack of the derived character state for a feature, so that some, but not all, of his groups are grades, not clades. It is not surprising therefore, that his different data sets did not always coincide, resulting in a partially incon- clusive revision of elapid taxonomy. Cogger (1975 et seq.) adopted a highly subdivided generic arrangement where most diagnosable groups were accorded generic status. Storr (Storr, 1985; Storr et al., 1986), however, has resisted this generic subdivision and has clustered together groups of species which have several external morphological features in com- mon. Typological thinking has thus led to the defini- tion of diagnosable units (e.g. McDowell’s ‘natural groups’) whose monophyly is assumed but untested. Clearly, as long as genera are defined in this way, classifications will continue to be accepted - or not - on the basis of authority or ‘gut feeling’, making discussion of the merits or biological validity of competing classifica- tions very difficult. Recently, data on elapid phylogeny became available in the series of articles forming part of the symposium volume edited by Grigg et al. (1985). These articles presented phylogenetic hypotheses based on karyology and allozyme electrophoresis (Mengden, 1985a; 1985b), im- munological comparison of serum proteins (Schwaner, et al. 1985) and soft anatomy (Wal- lach, 1985). None of these studies was complete, in that, for each, certain taxa were unavailable or their relationships were not clearly indicated, and the individual authors were not in a position to benefit from the others’ insights. Neverthe- less, the different data sets corroborate one another on several points and, more importantly, there are no obvious discordances among the 1 conclusions arrived at by the different authors. Thus, while a fully resolved, highly corroborated phylogeny for the Australian elapids has not yet been achieved, sufficient data are now available io sel up a taxonomic scheme in which the tn- cluded genera can be defined so as to be monophyletic as well as morphologically cohesive. For the remainder of this paper | set out the genera which I propose should be recognised, with annotations concerning the evidence for monophyly and the reasons, where appropriate, for the points at which this generic arrangement differs from those accepted by Storr or Cogger. One of the problem areas discussed by Mengden (1983), namely whether *Elapidae’ is the ap- propriate family name for the Australian protcroglyphs, will not be discussed here, Al- though biochemical (Maoetal., 1983; Schwaner et al., 1985) and morphological! (McDowell, 1967) evidence suggests that the Australian region proteroglyphs (including sea snakes) are monophyletic, suprageneric taxonomy will not be finalised until relationships to exotic proteroglyphs.,, and other colubroids, are better known. Through this article, ‘elapid’ is used as a convenient adjcctive, rather than as a position statement on higher taxonomy. In arriving at a generic scheme [ have used the following guidelines: !. Genera must be truly monophyletic (holophyletic). Paraphyletic groups have been avoided by making genera either more inclusive or by complete splitting of terminal taxa. Monophyly has been based on the data in Grigg et al. (1985) and on McDowell’s data on the derived states of adaptive complexes in venom gland musculature, palatal morphology and hemipenial structure. 2. Apart from the restrictions imposed by (1), genera are composed of species with strong phencetic similarities and ecologies. 3. Where a choice has been possible, genera have heen inclusive (‘lumped’) rather than sub- divided (‘split’) in order to show where species groups have clear sister-groups. 4. Aside from (1)-(3), at least one generic decision (concerning Nolechis, Austrelaps, Trapidechis) has been taken (pro tem) to main- lain usage of medically significant generic names. The generic scheme adopted here is sum- marised in Table 1 and compared to those of Cogger (1986) and Storr (1985; Storr et al., MEMOIRS OF THE QUEENSLAND MUSEUM 1986). For generic synonymies see Cogger etal. (1983 Acanthophis Daudin, 1803 REMARKS A chromosomally conservative but mor- phologically highly derived genus, biochemical- ly well-separated from its nearest relatives, the other viviparous specics with entire anal and subcaudal scales. DIAGNOSIS See Cogger (1986). INCLUDED SPECIES antarcticus (Shaw and Nodder, 1802); praelongus Ramsay, 1877; pyrrhus Boulenger, 1896. Austrelaps Worrell, 1963 REMARKS See remarks for Notechis. DIAGNOSIS See Cogger (1986). INCLUDED SPECIES: superbus (Giinther, 1858), As presently defined, this is a species-complex (Rawlinson and Hutchinson, in prep.) Cacophis Giinther, 1863 REMARKS A genus of small cryptozoic snakes associated by most workers with the species here included in Furina, Its species retain the primitive Glyphodon type of venom gland musculature and (apart from the autapomorphic squamulosus) the primitive karyotype. Features which argue for their monophyly with respect to Furina are the hemipenis, which is of the derived single type (fide Wallach, 1985) and the charac- teristic (probably derived) colour pattern of a nuchal pale blotch extending forward over the lores while Furina species are possibly monophyletic with respect to Cacophis based on theit uniformly dark eyes (iris paler than pupil in Cacaphis). Thus | tentatively support separate recognition of Cacephis pending more thorough phylogenetic study. GENERIC CLASSIFICATION OF AUSTRALIAN ELAPID SNAKES 399 TABLE 1. Correspondence between generic classification proposed and the generic schemes of Cogger (1985) and Storr (Storr, 1985; Storr et al., 1986). Cogger Acanthophis Austrelaps Drysdalia Echiopsis Elapognathus Notechis Demansia Cacophis Furina Hemiaspis Hoplocephalus Denisonia ———_ ,,, } —_ Suta ee." Suta Unechis —————_ P*" Oxyuranus = Oxyuranus Parademansia Pseudechis Pseudonaja Tropidechis Neelaps ee Simoselaps Simoselaps Vermicella Pseudechis Pseudonaja Tropidechis Vermicella DIAGNOSIS See Coggert (1986), INCLUDED SPECIES harriettae Krefft, 1869; krefftii Giinther, 1863; squamulosus (Duméril, Bibron and Duméril, 1854). Demansia Giinther, 1858 REMARKS A chromosomally unique genus whose mem- bers have a derived morphology (convergent on Holarctic racers) for highly active diurnal forag- ing. Biochemical evidence indicates wide diver- gence from its nearest relatives (Pseudechis and Pseudonaja). DIAGNOSIS See Cogger (1986). INCLUDED SPECIES calodera Storr, 1978; olivacea (Gray, 1842); papuensis (Macleay, 1877); psammophis (Schlegel, 1837); reticulata (Gray, 1842); rufes- cens Storr, 1978; simplex Storr, 1978; torquata Acanthaphis Austrelaps Drysdalia —_____ ar Echiopsis part Elapognathus «——**" Notechis Demansia Cacophis «ee Furina Glyphodon Hemiaspis Hoplocephalus Denisania ps see ere 2 part Riineplaeephe beg. nt Rhinoplocephalus Cryptophis Present Work Acanthophis part oe Notechis Demansia Cacophis Furina Hemiaspis Hoplocephalus part. Denisonia "= Rhinoplocephalus Cryptophis Oxyuranus Pseudechis Pseudonaja Tropidechis FF _ pat——____Vermicella (Gunther, 1862); vestigiata (de Vis, 1884) (from Storr et al., 1986; Ingram, 1990; and pers. obs.). Denisonia Krefft, 1869 REMARKS It is clear from all of the studies in Grigg et al. (1985) that this genus, even in the restricted sense of Cogger (1986), is polyphyletic. The type (maculata) and devisi are sister species, but more closely related to Drysdalia than to the other species retained in Denisonia by Cogger (fasciata and punctata) or the species placed by Storr in his expanded Denisonia. The pronounced difference in morphology and ecol- ogy between the two species retained here in Denisonia and their nearest relatives, Drysdalia (nocturnal, broad head-and-body species with glossy scales, elliptical pupils, versus diurnal, gracile species with matt scales and round pupils) argues for separate generic status for these two groups. AMENDED DIAGNOSIS As in Cogger (1985) with the following addi- tions: Pupil vertically elliptic, iris of eye pale. d00 Distinguished from some superticially similar species of Suta by venom gland musculature of the Oxyuranus type (versus the Pseudechis type), retention of the deeply forked hemipenis (simple in Suta), diploid number of 34 with pair 5 sex chromosomes (versus 30 with pair 4 sex chromosomes) and upper labials strongly barred with white and dark brown, INCLUDED SPECIES devisi Waite and Longman, 1920; maculata (Steindachner, 1867). Drysdalia Worrell, 1961 REMARKS The distinctive pair 5 (rather than pair 4) sex chromosomes (shared with Denisonia 5.5.) separate this morphologically cohesive group of small diurnal skink predators from Nofechis and its relatives, D. coranata, which lacks the chromosomal synapomorphy, is nevertheless close to the other three species based on analomi- cal features (Wallach, 1985). On biochemical evidence (Schwaner et al., 1985; Mengden, 1985a) these snakes are less closely related ta Notechis than are several morphologically cdiver- gent genera, notably Hoplocephalus and Trapidechis, DIAGNOSIS See Cogger (1986). INCLUDED SPECIES coronata (Schlegel, 1837): ceronoides (Giinther, 1858); mastersii (Krettt, 1866); rhodogaster (Jan, 1873). Echiopsis Fitzinger, 1843 REMARKS Undoubtedly a close relative of Notechis on the basis of strong internal anatomical (Wallach, 1985) and biochemical similarities (Schwaner et al., 1985; Mengden, 1985b), as well as the phenetic similarities noted by Storr (1982). However, its derived Acanthaphis-like habitus, including the subdivided temporal scalation noted by Mengden (1985a) and vertically cllip- tical pupil, and the absence of the derived Notechis karyotype, support separate generic status for at least curta. Mengden (1985a) also reported that curta showed venom properties with Acanthophis, adding to the list of features suggesting a possible sister-group relationship MEMOIRS OF THE QUEENSLAND MUSEUM between these two taxa, rather than between eurta and Notechis scutatus, Brachyaspis atriceps Storr, 1980, has not been studied and may, as Storr suggests, be closer to his Denisonia (Suita in my sense) than to curta. DIAGNOSIS See Cogger (1986). INCLUDED SPECIES curta (Schlegel, 1837), Elapognathus Boulenger, 1896 REMARKS The general relationships of this monotypic genus clearly lic with the large group of viviparous species having entire anal and sub- caudal scales. Storr (1982) partly expressed this in synonymising Elapognathus with Notechis. However, the precise sister species of FE. minor is not identified by the available data. It retains the primitive 2n = 36 karyotype and is biochemi- cally rather divergent from its relatives. Wallach’s analysis fails to consistently identify a sister taxon. In ‘gestalt’, E. minor is most similar to juvenile copperheads (esp. superbus, s.s.) and Storr placed it in his Notechis on the basis of shared similarities with scutatus, super- bus and Drysdalia. Storr dismissed the single generic character (no post-fang maxillary teeth) by making a general statement about the cautious use of dental characters in snakes. Nevertheless, his taxonomic characters for Notechis (s.1.) define a paraphyletic taxon (Tropidechis, Hoplocephalus, Denisonia s.s.. and possibly even Acanthophis, should all be included) so that his data have defined only a grade of organisa- tion (the primitive morphology for this group of genera?) rather than a strictly monophyletic taxon. My conclusion is that Elapognathus is, like Echiopsis, morphologically distinct (fang only on the maxilla, a derived character state) and lacks the apomorphic chromosomal feature of either Notechis or Drysdalia. Its single species, &. minor, is, on the basis of biochemical data, a very distinct species with no obvious sister taxon, and | favour its recognition. DIAGNOSIS See Cogger (1986). INCLUDED SPECIES minor (Giinther, 1863) GENERIC CLASSIFICATION OF AUSTRALIAN ELAPID SNAKES Furina Duménil, 1853 REMARKS Furina and Glyphodon ate similar small snukes characterised by divided anal and sub- caudal scales, a cryptozoic way of life and, in mast species. white to red nuchal patches. This pair has been least studied chromosomally and juchemically, so that there is rather little well- constructed phylogenetic data available. The genera are separated by features (presence/ab- sence of a divided nasal) of unknown phylogenctic significance. In Wallach’s analysis, the species included in the two genera tend to fall out as cach-other’s closest relatives, but do not form sub-groups matching the current generic boundaries; indeed, with every alteration in algorithm, the branching order changes. Mengden, on the basis of unspecified data, groups the pair as a monophyletic cluster of carly divergence, with Cacophis the sister group of Furina (diadema only) plus Glyphodon (includ- ing F. ornata). The boundary between Furina and Glyphodon seems tenuous, especially the intermediacy of F. ernata with tespect to #. diadema (gencrotype) on one hand and G, tristis (generotype) on the other. This lineage needs more study, pending which feel there is insufficient data of phylogenetic significance by which the two genera can be justified. Uniting them under the oldest available name does, with seeming reliability, give rise to a monophyletic unit, which is moreover, relatively cohesive in ecology. DIAGNOSIS A genus of glossy-scaled (15-21 rows at mid- body), snakes with a divided anal and divided subcaudals. Often (not F. barnardi or F. dunmal- li) with a pale (white to red) nuchal blotch. Five or more teeth on each maxilla behind the fang. Head somewhat to markedly wider than neck and lacking a canthus rostralis. INCLUDED SPECIES barnardi (Kinghorn, 1929); diadema (Schlegel, 1837): dunmalli (Worrell, 1955); or- nata (Gray, 1842); tristis (Giinther, 1858). Hemiaspis Fitzinger, 1860 REMARKS A karyotypically unique pair of species, show- ing the unusual combination of divided anal, entire subcaudals and viviparity. Electrophoretic (? and chromosomal) data of Mengden sug- atl gested a very close relationship between the two species, as did Wallach’s morphological data. The sister group of these two is not well estab- lished, but if seems a well-defined taxon. DIAGNOSIS See Cogger (1986). INCLUDED SPECIES damelii (Ginther, 1876); signata (Jan, 1859). Hoplocephalus Wagler, 1830 REMARKS Another chromosomally unique and mor- phologically well-defined genus, whose mem- bers possess the arboreal adaptation of angular ventrals and have markedly broad heads distinct from the narrow neck, Very closely related, on immunological (Schwaner et al., 1985) and mor- phological (Wallach, 1985) data to Notechis.and Tropidechis. The phylogenetic position of this genus gives one of the strongest indications that Storr’s concept of Notechis is paraphyletic, im- plying that his generic diagnosis is based at least in part on symplesiomorphies. DIAGNOSIS See Cogger (1986). INCLUDED SPECIES bitorquatus. (Jan, 1859); bungeroides (Schlegel, 1837), stephensii Krefft, 1869. Notechis Boulenger, 1896 REMARKS Al present a controversial genus containing either two or a single species (Cogger, 1986; Schwaner pers. comm.), or a cluster of species which are surface-dwelling, viviparous, have en- tire anal and subcaudal scales and are otherwise morphologically conservative (Storr, 1982). Im- munological, chromosomal and morphological studies all indicate that Storr’s concept is paraphyletic, * *And nomenclaturally invalid. Storr (1982) dis- missed Echiopsis Fitzinger as 2 nomen oblitum al- though declaration of a name as ‘forgotten’ could no longer be made atter 1 January 1973, Fitzinger’s (1843) names are widely regarded as available and are in wide use (including Echiopsis, see Cogger et al., 1983). Thus the correct name for Stosr's genus should have been Echiopsis Fitzinger, 1843, nat Notechis Boulenger, 1896. up However, there is a strong indication that scutatus is very closely related to the Ausirelaps superbus complex, the two sharing (with Tropidechis) a uniquely derived karyotype and being very similar biochemically, anatomically and ecologically. Accordingly, | would favour the elimination of Austrelaps and the transferral of the superbus complex to Notechis. However, the precise relationships of scuifatus, the super- bus complex and Tropidechis carinatus are not yet established. The three taxa differ in minor features of proportions and scalation and are, based on the chromosomal synapomorphy, each- other's closest relatives. Amalgamation of the three would be a simple answer, except for the nomenclatural problem of the synonymisalion of Notechis. under the older Tropidechis. Because of the widespread use of the junior name, phylogenetic data Would need to be compelling before such a destabilising revision of the exist- ing taxonomy would be justified - in fact it could well lead to appeals to the ICZN to suppress Tropidechis in favour of Notechis, Pending detailed phylogenetic knowledge, Tropidechis and Notechis, and therefore Ausirelaps should remain separale genera, although the close relationship between them should be borne in mind. Tropidechis is derived with respect to Notecltis (s.s.) in its kecled scalation and in- creased midbody scale count, while Austrelaps differs in its derived Pseuelechis type (rather than Oxyuranus type) of venom gland musculature. DIAGNOSIS See Cogger (1986). INCLUDED SPECIES scutatus (Peters, 1861). Oxyuranus Kinghorn, 1923 REMARKS Covacevich et al. (1981) set out a range of characteristics which argued for sister-species relationship and congeneric status of Pseudechis seuiellams Petets, 1868 and Diemenia microlepidota McCoy. 1879. Cogger preferred to continue recognition of a monotypic Parademansia for microlepidota, but the addi- tional data from the 1985 symposium reinforce the close relationship of these two species and further argue for their inclusion in asingle genus, DIAGNOSIS See Covacevich et al. (1981). MEMOIRS OF THE QUEENSLAND MUSEUM INCLUDED SPECIES microlepidotus (McCoy, 1879); seutellatus (Peters, 1868). Pseudechis Wagler, 1830 REMARKS A morphologically cohesive group retaining a primitive karyotype, but closely-related on the basis of immunological data (Schwaner et al., 1985) and monophyletic based on morphology (Wallach, 1985) and allozyme comparisons (Mengden et al., 1986). DIAGNOSIS See Cogger (1986). INCLUDED SPECIES australis (Gray, 1842); butleri Smith, 1982; colleui Boulenger, 1902; gutratus de Vis, 1905; papuanus Peters and Doria, 1878; porphyriacus (Shaw, 1794). Pseudonaja Giinther, 1858 REMARKS Another well defined and monophyletic genus, although its alpha taxonomy is presently very unsatisfactory. Wallach’s (1985) contention that modesta was not allied to the other species in this genus was refuted on several ground by Mengden (1985b). DIAGNOSIS See Cogger (1986). INCLUDED SPECIES affinis Giinther, 1872; guttata (Parker, 1926): ingrami (Boulenger, 1908); modesta (Gunther, 1872); auchalis Giinther, 1858: textilis (Dumeril, Bibron and Duméril, 1854). Rhineplocephalus Miiller, 1885 REMARKS Another genus treated discordantly by Cogger and Storr. Long regarded as a monotypic genus {on the strength of the fused internasals and nasals) Storr greatly expanded the genus to in- clude the other small, pale-bellied and dark-eyed semi- fossarial/nocturnal species placed by Cog- ger in Unechis, Mengden's and Wallach’s studies partially support Storr, in that they indi- cate that some other species are closely related GENERIC CLASSIFICATION OF AUSTRALIAN ELAPID SNAKES to bicolor, the type of Rhinoplocephalus, these being the two species of Cryptophis plus the type species of Unechis, U, boschmai (formerly U. carpentariae) and possibly L/, nigrastriatus. However, the other small black-headed snakes (the gouldii complex) show a closer relationship to Sura and ‘Denisonia’ punctata and ‘D.’ fas- ciata than they do to bicolor, Thus | favour expanding Rhinoplocephalas to include the four close relatives mentioned above (including the iypes of Cryptophis and Unechis), but transfer- ting the remaining species of Storr’s Rhineplocephalus to Suta (see belaw). REVISED DIAGNOSIS A group of small to moderale-sized species lacking. contrasting dark head colouring (apart from R. nigrosiriatus), with glossy midbody scales in 15 rows, anal. and subcaudals entire, cye small with black itis ,indistinguishable from pupil. Head. slightly to moderately depressed, no canthus, Distinguished externally from some su- permite similar species of Suta by deeper, blunter head, absence of contrasting colour pat- tern (except black-headed R. algrestriatus) and/or longer tails (subcaudal counts exceed 40 in most species [not beschmaij, versus 40 or fewer in most Su/a). Further distinguished from other genera by the unique karyotypes (not present in bicolor), 2n = 36 (20 M, 16 M) or 40). INCLUDED SPECIES bicalar Miller, 1885; boschmai (Brongersma and Knaap-Van Meeuwen, 1961); mgrescens (Giinther, 1862); nigrostriats (Krefft, 1864); pallidiceps (Gimher, 1858). Simoselaps Jan, 1859 REMARKS Storr and Cogger both noted that the small, mostly cross-banded fossorial snakes of arid Australia fall into several distinct subgroups, based on body, head and head-shicld proportions which reflect ecological specialisation (Shine, 1984). Cogger separated some of these as dis- linet genera, but Storr united all in Vermice/la, while identifying subgeneri¢ groups having similar morphologies, Karyolypic data show that at least two of Cogger’s genera (Neelaps and Vermicella s,s.) retain the primitive karyotype, while the types of two other genera or sub-genera (Simoselaps and Brackyurophuis) have derived karvotypes. Of the latter pair, Mengden (1985a) derived the karyotype of (Brachyurophis) semi ani fasciatus from that of (Simoselaps) bertholdi, implying a phylogenetic relationship hetween these taxa. Wallach's (1985) analysis shows all of the burrowing group consistently forming @ monophyletic lineage, but the branching order within the group is not unequivocal. Only Cacophis warro de Vis, 1884, fails ta fall out with the other burrowers, but Mengden’s report of its showing the uniquely derived karyotype of bertholdi would argue for retention in the same pom as this species at least. No authors seem to ave taken account of McDowell's (1969a) report of the distinctive biting apparatus present in all but annudata (and presumably multifas- ciata). This functional complex argues strongly for monophyly of at least all of the species except the type of Vermicella, Furina annulata Gray, 1841 isa remarkably primitive species, retaining the plesiomorphic state of the karyotype, venom gland musculature, hemipenis and palatine bone, As the true bandy- bandys (Vermicella s.s.) show none of the synapomorphies which unite some or all of the remaining species, there are no strong grounds, as MeDawell (1969a) pointed out, for placing annulata with the other bur- rowers grouped together here as an expanded Simaselaps. Although morphological subgroups certainly exist within Simoselaps, relationships among them are abscure, and | prefer to recog- nise the probable monophyly of ths group rather than itemising variation of uncertain phylogenetic significance. DIAGNOSIS A group of small (less than 0.6 m total length), glossy scaled semi-fossorial snakes with anal divided, short tails with 35 or fewer paired sub- eaudals and showing variation in snout shape and body proportions analogous to (hose seen in Ramphotyphlops, Rostral always projecting but varying in prolile from bulbous (e.g. bimaculatus) to wedge-shaped (e.g, fusciolatys) (0 upturned and angular (¢.g. sempfasciatus), Nu canthus rostralis. Body short and dumpy to elon- gate, but ventrals fewer than 230, Dark parietal and nuchal blotches always present, body usual- ly yellow, orange or reddish, generally with darker reticulated or cross-banded pattern. INCLUDED SPECIES anomala (Sternfeld, 1919): approximans (Glauert, 1954); australis (Kreffl, 1864): ber- theldi (Tan, 1859); bimaculata (Dumeéril, Bibron and Dumeril, 1854); caloneia (Duméril, Bibron and Duméril, 1854); fascrolatus (Ginther, 404 1872); incincta Storr, 1968; littoralis Storr, 1968; minima (Worrell, 1960); roperi (Kin- ghorn, 1931); semifasciatus (Ginther, 1863); warro (de Vis, 1884). Suta Worrell, 1961 REMARKS Preceding discussion on Denisonia and Rhinoplocephalus has alluded to the fact that several species share a close relationship with the type species of Suta (Hoplocephalus sutus Peters, 1863). The most compelling evidence is the uniquely shared 2n = 30 karyomorph, present in Suta, ‘Denisonia’ fasciata and ‘D.’ punctata, and the ‘Unechis’ gouldii species group. All are morphologically similar in being, like Rhinoplocephalus, small, glossy scaled cryp- tozoic species with entire anal and subcaudal scales. All species tested by Schwaner et al. (1985) also proved to be close to the Notechis group, and all of these species fall out as each other’s closest relatives in Wallach’s analysis. All have the derived Pseudechis type of venom gland musculature (McDowell, 1970). Storr et al. (1986) separated the three species with pale iris colour (suta, ordensis and fasciata) from the remainder, and place them with superficially similar species in his Denisonia. However, the pale iris is evidently a retained primitive feature, and eye colour is known to be intraspecifically variable in ordensis (Storr et al., 1986, p. 75). REVISED DIAGNOSIS A group of small to moderate-sized snakes with a consistent colour pattern of dark head markings (often a black to brown cap) and lighter brown body (cross-banded in fasciata), midbody scales very glossy, in 15-21 rows, anal and sub- caudal scales entire. Head slightly to markedly depressed; no canthus rostralis. Eye uniformly black in most species, but paler than pupil, which contracts to a vertical ellipse, in suta, fasciata and most ordensis. Further distinguished by uni- que 2n = 30 karyomorph. (See also Rhinoplocephalus.) INCLUDED SPECIES dwyeri (Worrell, 1956); fasciata (Rosén, 1905); flagellum (McCoy, 1878); gouldii (Gray, 1841); monachus (Storr, 1964), nigriceps (Giinther, 1863); ordensis (Storr, 1984); punctata (Boulenger, 1896); spectabilis (Krefft, 1869); suta (Peters, 1863). Probably MEMOIRS OF THE QUEENSLAND MUSEUM Brachyaspis atriceps Storr, 1980, belongs here as well. Tropidechis Giinther, 1863 REMARKS See remarks for Notechis. DIAGNOSIS See Cogger (1986). INCLUDED SPECIES carinatus (Krefft, 1863). Vermicella Giinther, 1858 REMARKS See remarks for Simoselaps. Aside from the unique colour pattern of black and white bands, the two species of Vermicella share a very to extremely attenuate body. Ventral counts in V. annulata range as high as 243 (Storr et al., 1986) (250 or fewer, generally less then 200, in other Australian elapid taxa) while in V. multifasciata counts range up to 290 (Cogger, 1986). Western populations of this genus have been described as a subspecies, snelli Storr, 1968, which Cogger (1986) placed with annulata but Storr et al. (1986) placed with multifasciata. The very high ventral count of snelli (to 318 ; Storr et al., 1986) indicates a closer relationship to multifasciata, although it lacks the latter’s derived fusion of internasal and prefrontal shields. DIAGNOSIS See Cogger (1986). INCLUDED SPECIES annulata (Gray, 1841); multifasciata (Long- man, 1915). ACKNOWLEDGEMENTS I am grateful to the several people who com- mented on an earlier draft of the manuscript: K. Aplin, H.G. Cogger, J. Covacevich, A.J. Coventry, G.J. Ingram and R. Shine. Their com- ments markedly improved it, although the final interpretations are my responsibility. I thank D. Lowery for typing the manuscript. LITERATURE CITED BOULENGER, G.A. 1896. ‘Catalogue of snakes in GENERIC CLASSIFICATION OF AUSTRALIAN ELAPID SNAKES the British Museum (Natural History)’. Vol 3. xiv+727pp. (Taylor and Francis: London). COGGER. H.G, 1975 (revised and expanded edition, 1986). ‘Reptiles and amphibians of Australia’. (A.H. and A.W. Reed: Sydney.) 1985. Australian proteroglyphous snakes - An histori- cal overview. pp. 143-154. /n Grigg et al, (1985). COGGER, H.G., CAMERON, E.E. AND COGGER, H.M. 1983. Amphibia and Reptilia. Zoological Catalogue of Australia 1; 1-313. COVACEVICH, J., MCDOWELL, S.B., TANNER, Cc. AND MENGDEN, G.A, 1981. The relation- ships of the taipan, Oxyuranus scutellatus. and the small-scaled snake, Oxyuranus microlepidotus (Serpentes: Elapidae). pp. 160- 168. Jn Banks. C.B. and Martin, A.A. (eds), ‘Proceedings of the Melbourne Herpetological Symposium, 1980’, (Zoological Board of Vic- toria: Melbourne.) FITZINGER, L.J. 1843. ‘Systema Reptilium’. (Braumiller und Seidel; Vienna.) vit 106pp. GRIGG, G.C., SHINE, R. AND EHMANN, H. 1985. ‘The Biology of Australasian frogs and reptiles’. (Surrey Beatty and Sons: Sydney.) xvi+S27pp. INGRAM, G.J. 1990. The works of Charles Walter de Vis, alias ‘Devis’, alias ‘Thickthorn’. Memoirs of the Queensland Museum 28: 1-34. KINGHORN, JR, 1929 (second edition, 1956). ‘The snakes of Australia’, (Angus and Robertson: Sydney.) MAO, S., CHEN, B., YIN, F. AND GUO, Y. 1983. Immunotaxonomic relationships of sea snakes and terrestrial elapids. Comparative Biochemistry and Physiology 74A; 869-872. MCDOWELL, S.B. 1967. Aspidomorphus, a genus of New Guinea snakes of the family Elapidae. with notes on related genera. Journal of Zoology, London 151; 497-543. 19693. Toxicocalamus, a New Guinea genus of snakes of the family Elapidae. Journal of Zool- ogy, London 159: 443-511. 1969b. Notes on the Australian sea-snake Ephalophis greyi M. Smith (Serpentes: Elapidae: Hydrophiinae) and the origin and clas- sification of sea-snakes. Zoological Journal of the Linnaean Society 48: 333-349, 405 1970. On the status and relationships of the Solomon Island elapid snakes. Journal of Zool- ogy, London 161: 145-190. MENGDEN, G.A, 1983, The taxonomy of Australian elapid snakes: a review. Records of the Australian Museum 45; 195-222. 1985a. Australian elapid phylogeny: a summary of the chromosomal and electrophoretic data. pp. 185-192, In Grigg et al. (1985). 1985b. A chromosomal and electrophoretic analysis of the genus Pseudonaja. pp.193-208. In Grigg et al. (1985). MENGDEN, G.A., SHINE, R. AND MORITZ, C, 1986. Phylogenetic relationships within the Australasian venomous snakes of the genus Pseudechis. Herpetologica 42: 215-229. SCHWANER, T.D.. BAVERSTOCK, P.R., DES- SAUER, H.C. AND MENGDEN, G.A. 1985. Immunological evidence for the phylogenetic relationships of Australian elapid snakes. pp. 177-184. In Grigg et al. (1985). SHINE, R. 1984. Ecology of small fossorial Australian snakes of the genera Neelaps and Sumoselaps (Serpentes: Elapidae). pp. 173-183. In Seigel, R.A., Hunt, L-E., Knight, J.L., Malaret, L. and Zuschlag, N.L. (eds), ‘Ver- tebrate ecology and systematics - A tribute to Henry §. Fitch’. (University of Kansas: Lawrence.) STORR, G.M. 1982. The genus Notechis in Western Australia, Records of the Western Australian Museum 9: 119-123. 1985. Phylogenetic relationships of Australian elapid snakes: external morphology with an em- phasis on species in Western Australia. pp,221- 222. Jn Grigg et al. (1985). STORR, G.M., SMITH, L.A. AND JOHNSTONE, R.E. 1986, ‘Snakes of Western Australia’- (Western Australian Museum: Perth.) WORRELL, E, 1961, Herpetological name changes. Western Australian Naturalist 8; 18-27, 1963. A new elapine generic name. Australian Rep- tile Park Records 1: 2-7, WALLACH, V. 1985. A cladistic analysis of the terrestrial Australian Elapidac. pp.223-253, Jnr Grigg el al. (1985). 406 HERPETOLOGISTS AND SNAKE-BITE:— Snake-bite is an occupational hazard for professional and amateur her- petologists alike. A series of clinical case studies of reptile keepers billen by snakes has revealed several ‘atrisk’ themes. This note documents these in the hope that the [requency and morbidity of bites might be reduced, and convalescence hastened. Snake-bite involving experts can happen at any time, in- cluding while attempting to catch snakes. The majority of clinical cases, however, occur during cleaning of viyaria or during other husbandry activities. Many cases occur al nigh. Unlike other syndromes of human snake-bite (the “Big Game Hunter’ scenario, for example. where an amateur, often intoxicated, tries to catch a snake for an audience), the viclim is often working alone. The victim, despite experience in handling reptiles, sometimes is unsure whether a strike has actually occurred, A snake can strike a hand in a cage and recoil with lightning speed. As well, the skin lesions caused by the majority of Australian elapid snakes are trivial. Often no lesions or blood specks are visible in the first few minules after strikes. Many reptile fanciers who are bitten are envenomed by dangerous Australian species. A significant proportion of the very severe clinical envenomations seen in practice resulls from the bites of Oxvuranus, Pseudonaja and Tropidechis. A special polential medical problem that involves her- petologists is the scenario of envenomation by exotic snakes. I is essential that specific antivenom to African, Asian and American species is held at the national reference antivenom facility (Commonwealth Serum Laboratories, Parkville, Melbourne) if one is keeping such imported species. Another medical problem for herpetologists is the multiple MEMOIRS OF THE QUEENSLAND MUSEUM (serial) bite syndrome. Snake venom is highly allergenic and many reptile fanciers become sensitised to the venom. In- sofar as the initial (transient) collapse of some victims is thought to be due to hypotension, this problem may be of particular relevance to the serial victim. Secondly, the risk of anaphylaxis rises with repeated (serial) injections of horse serum (antivenom) and the risk of such reactions probably rises above ten percent in those who have had several lifesav- ing infusions following previous bites. One problem en- countered in practice is the fear of such reactions in an expericnced herpetologist - sweating, tachycardia and faint- ness can be signs of true envenomation or incipient allergic reactions, or can be those simply of apprehension itself. The skin lesions of elapid snakes are pleomorphic - typical two fang punctures (often with oozing blood specks) occur in 6U percent of cases only. Single punctures, multiple fang marks, and combined fang and teeth (both maxillary and pulatine) lesions are common. If multiple strikes have oe- curred - anotuncommon scenario for both herpetologists and toddler children (both tend lo hang on to the snake) - the potential for severe envenomation is greally increased. I is essential that all herpetologists should be trained in first aid, and have a compressive bandage (preferably an Esmarch bandage) and splint handy when they are working, With proper first aid treatment, elapid snakebite does not necessarily mean severe envenomation and severe envenomation does nol mean morbidity or death. Join Pearn, Director of Training, St John Ambulance Australia, e/- Department of Child Health, Royal Children's Hospital, Brisbane, Queensland 4029, Australia; 17 August, 1990, FOUR NEW SPECIES OF STRIPED SKINKS FROM QUEENSLAND G.J. INGRAM AND G,V. CZECHURA Ingram, G.J. and Czechura. G.V. 1990 09 20: Four new species of striped skinks [rom Queensland. Memoirs of the Queensland Museum 29(2); 407-410. Brisbane. ISSN 0079- S835. Citenotus nullum sp. noy., C. hypatia sp. noy, and C, lerrareginae sp. nov. are skinks of rocky substrates from northeast Queensalnd. C. aphrodite sp. nov. is found in the arid country of the southwest of the State.) Crenotus, skinks, new species, Queensland, Glen J. Ingram and Gregory V. Czechura, Queensland Museum, PO Box 300, South Brishane, Queensland 410), Australia; 17 August, /990, Clenalus is a vety speciose genus of skinks. Wilson and Knowles (1988) listed 81 species and illustrated several additional new species. In this paper, we describe four new species from Queensland. Crenotus was described in 1964 by Dr Glen Stor of Perth, Western Australia, and most of the species of the genus were sub- sequently described by him (see Wilson and Knowles, loc. cit., for a list of papers). Sadly. Glen Storr recently passed away. Australians owe him u debt for his thirty years of ceaseless work towards elucidating the species of reptiles of this continent. He will be sorely missed by his colleagues, Our paper is dedicated to his memary- In the following, pattern nomenclature is of Wilson and Knowles (1988), Abbreviations are: SV, distance from snout to vent in mm; HW, head width at widest part as % SV; HL, length of hindlimb as % SV; TL, tail length as % SV; QM, Queensland Museum; AM, Australian Museum. Ctenotus nullum sp, nov. Ctenoius sp. (4). Wilson and Knowles, 1988, p,278. MATERIAL EXAMINED HoLoTyPe: OM J32424, sandstone escarpment, 2km W of Melvor River Crossing, Cape York Peninsula, (15°07°S. 145"04'E), Queensland, collected by G.J. Ingram on 15 July, 1976. PARATYPES: nr Isabella Falls (QM J41023-5, 142768- 9); Finch Bay, 1.7 km SE of Cooktown (AM R71031); ESE side of Mt Simon (AM R71033); Black Moun- tain, S of Cooktown (QM 24647); Shiptons Flat (QM J42736); Spit Rock Gallery, S of Laura (QM J37999- 38001); Quinkan, S of Laura (QM J24705). DIAGNOSIS A moderately large (maximum SV 79) Crenotus with 4 pattern of stripes and an upper lateral row of pale squarish blotches which are usually confluent with the midlateral stripe; a black vertebral stripe begins on the nuchals and always extends to at Ieast the fore back; a brow that does not conceal the supraciliaries; supralabials usually 8; subdigital lamellae wide- ly calloused, 25-28 on fourth toe; four supraoculars: 4-6 ear lobules: and 26-28 mid- bady scales. For an illustration of the species see Wilson and Knowles (1988, p,278, photo no. 424). DISTRIBUTION Sandstone areas around Laura and those to the near north of Cooktown; also at the base of the boulder mountains of Trevethan Range; and Shiptons Flat sauth of Cooktown. DESCRIPTION SV: 45-79 (N = 9, mean 62,2). HW; 13-17 (N = 9, mean 14.8). HL: 46-55 (N = 9, mean 51.2). TL: 245-257 (N = 4, mean 251.8). Snout sloping, slightly rounded in profile. Nasals separated. Nasal groove absent. Rostral and frontonasal in narrow contact. Prefrontals large, separated or conlacling and forming a short suture along the midline, Frontal long and narrow, contacting the prefrontals (and the fron- tonasal in 66% of specimens), the first three supraoculars, and the frontoparietals. 2-4 en- larged nuchals on either side (N = 18, mean 2.7). Four supraoculars, second much the larger. Supraciliaries 9-11 (N = 18, mean 9.5), first largest. Supralabials &, uncommonly 9 (N = 18, mean 8.1); sixth under the eye and enters the orbit. Ear aperture large, 4-6 (N = 8, mean 4.8) pointed lobules on anterior border, Midbody scale rows 26-28 (N = 9, mean 27.8). Number of scales from chin to vent 60-66 (N = 9, mean 408 62.4), Toes long, compressed; subdigital lamel- lae widely calloused, 25-28 (N = 9, mean 26,4) under fourth toe. Upperparts olive to reddish brown with a nar- row black vertebral stripe edged with white paravertebral stripes continuing for varying dis- tances down back or tail, There are 2 to 4 white- lined black dashes on the hind edge of the parictals. A white dorsolateral stripe begins at the last supraocular and continues down back and along tail. This is edged broadly (sometimes narrowly) by a black laterodorsal stripe at least as far as the base of the tail. Upper lateral zone black with squarish white, or fawn with white centres, spots that are usually confluent with the midlateral stripe (in life, the upper lateral spots ure usually red). The white midlateral siripe begins behind the nostril and continues back through ear, along the body and tail. This stripe is lined darkly below, with white blotches againsta greyish background on the lower lateral surface. Sometimes there is a suggestion of a white ventrolateral stripe. Limbs with three black stripes. Ventrally cream to white, ETYMOLOGY “Nullum' is a patronym from the Ko-ko-ya-o language of castern Cape York Peninsula, Queensland. REMARKS On specimens from the coast near Cooktown, the black vertebral stripe extends from the nuchals to the hind back or base of the tail. On specimens from the sandstones near Laura, this stripe only extends to the fore-back, The range and habitat of C. null is similar to that of C. quinkan, except the latter has only been collected on sandstone (Ingram, 1979), Ctenotus hypatia sp. nov, Ctenotus sp, (3). Wilson and Knowles, 1988, p.278. MATERIAL EXAMINED HOLOTYPE: OM J42092, granite gorge, 13km W of Mareeba, (17°00"S, 145°17°E}, NEO, by D, Knowles on 17 October, 1983. DIAGNOSIS A medium sized (SV 54) Crenotus with a pat- iern of stripes and an upper lateral zone of lan or white blotches on chocolate-brown; supralabials 7; subdigital lamellae keeled, 20 on fourth toe; four supraoculars; 3-4 ear lobules; and 28 mid- MEMOIRS OF THE QUEENSLAND MUSEUM body scales. For an illustration of the species see Wilson and Knowles (1988, p,278, photo no. 423), DISTRIBUTION Known only from the holotype, which was tuken in 4 granite gorge. DESCRIPTION SV: 54. HW: 12. HL: 44. TL: 215. Snout sloping, slightly pointed in profile. Nasals separated. Nasal groove absent, Rostral and frontonasal in narrow contact, Prefrontals large, separated. Frontal long and narrow, con- tacting the prefrontals, the frontonasal, the first three supraoculars, and the frontoparietals. 4-5 enlarged nuchals. Four supraoculars, second much the larger. Supraciliaries 9, first largest. Supralabials 7, fifth under the eye and enters the orbit. Ear aperture large, 3-4 pointed lobules on anterior border. Midbody seale rows 28. Number of scales from chin to vent 75. Toes long, com- pressed; subdigital lamellae keeled, 20 under fourth toe. Dorsal ground colour orange-brown. Black vertebral stripe begins at nuchals and continues onto tail for about 15mm, White paravertebral stripes begin at nuchals and fade out at base of tail. Black laterodorsal and white dorsolateral Stripes begin above and behind the eye; Jaterodorsal fades out at fore back while dor- solateral breaks up at mid back but continues as white dashes onto the base of the tail. Upper lateral zone chocolate-brown with tan or white blotches that give the effect of barring but breaks up into speckling on the tail. White midlateral stripe begins in front of ear and ends at hindlimb. Lower lateral zone grey with white blotching. Side of head blotched with chocolate-brown stripes. Ventrally white. ETYMOLOGY Named for Hypatia of Alexandria who oc- cupied the chair of Platonic philosophy. She was murdered in 415AD by a Christian mob. Ctenotus terrareginae sp. nov. Clenotus sp. (6), Wilson and Knowles, 1988, p.278, MATERIAL EXAMINED HOLOTYPE: QM J41996, Warrawilla Ck, Hinchinbrook ts, (18°22’S 146°15'B), NEQ, collected by D. Milton on 14 duly, 1983. NEW QUEENSLAND SKINKS DIAGNOSIS A large (SY 91) Ctenotus with a pattern of stripes and an upper lateral zone of white blotch- ing on black; supralabials 7; subdigital lamellae widely calloused, 28 on fourth toe; four supraoculars; 3 car lobules; and 28 midbody scales, For an illustration of the species see Wilson and Knowles (1985, p.278, photo no. 426). DISTRIBUTION Known from Hinchinbrook Island and about 60km south of the island by the turnoff to Paluma on the Bruce Highway, on the coast. (GYC, pers. obs.}. DESCRIPTION SV: 91. HW: 11. HL: 44. TL; 227. Snout sloping, slightly rounded in profile. Nasals separated, Nasal groove present, Rostral and frontonasal in narrow contact. Divided prefrontals large, separated. Frontal long and narrow, contacting the prefrontals, the fron- tonasal, the first three supraoculars, and the fron- toparietals. 3-4 enlarged nuchals. Four supraoculars, second much the larger. Supraciliaries 10, first largest. Supralabials 7; fifth under the eye and enters the orbit. Ear aperture large, 3 pointed lobules on anterior bor- der. Midbady scale rows 28. Number of scales from chin to vent 74. Toes long, compressed; subdigital lamellae widely calloused, 28 under fourth toe. Dorsal ground colour dark brown. Short black vertebral stripe begins at nuchals and terminates on upper back. Head scales marbled with black. Black lateradorsal and white dorsolateral stripes begin above and in front of eye and continue along nearly the full length of the tail, Black upper lateral stripe begins in front of cye and continues to beyond base of tail where it be- comes divided into two black stripes by a brown mid-stripe. White midlateral stripe begins be- hind nasals and continues back Io and along length of tail, Lower lateral zone black with white blotching from head to midlateral area (white blotches sometimes coalesce with mid- Jateral stripe to give an effect of black barring) where blotches merge to form a black lower lateral stripe that continucs along the tail. Limbs reddish brown with black and white stripes, Ventrally white, REMARKS The holotype was found under rock exfoliation aq amongst shrubs along a creek, ETYMOLOGY Named for Queensland. Ctenotus aphrodite sp. nov. MaTERIAL EXAMINED HOLOTYPE: OM J41814. Oorida area, Diamantina Lakes, (23°46'S 14.1°08"E), SWQ, collected by R. Atherton and G. Porter on 12 March, 1983. DIAGNOSIS A moderately large (SV 72) Ctenotus with a pattern of stripes (five black and four white dorsal stripes between the dorsolateral stripes} and an upper lateral zone of white blotches on black; supralabials 8; subdigital lamellae nar- rowly calloused, 34 on fourth toe; four supraoculars; 3-4 ear lobules: and 31 midbody scales, DISTRIBUTION Known only from the type locality. DESCRIPTION SV: 72. HW: 12. HL: 53. TL: 229. Snout sloping, pointed in profile, Nasals separated. Nasal groove abseni. Rostral and frontonasal in narrow contact. Prefrontals large, widely separated. Frontal broad, contacting the prefrontals, the frontonasal, the first three supraoculars, and the frontoparjetals, Two en- larged nuchals on cither side. Four supraoculars, second much the larger, Supraciliaries 8, first largest. Supralabials 8; sixth under the eye and enters the orbit. Ear aperture large, 3-4 pointed lobules on anterior border. Midbody scale rows 31. Number of scales from chin to vent 76. Toes long, compressed; subdigital lamellae narrowly calloused, 34 under fourth toe. Dorsal ground colour coppery brown. Top of head marbled with black. There are five black and four white dorsal stripes between the dor- solateral stripes. Black vertebral and white paravertebral stripes fade out above hindlegs, Other dorsal stripes break up just beyond mid back. In alcohol, the pale dorsal stripes can ap- pear to be coloured bluish-white or copper depending upon the angle of the light, Upper lateral zone black with distinct white blotches but, from hindlegs to along tail, the blotches are absent. White, wavy midlateral stripe sometimes broken with black, Lower lateral zone grey to 410 MEMOIRS OF THE QUEENSLAND MUSEUM black with white blotching; continues as a grey lower lateral stripe outlining below the white midlateral stripe along the tail. Limbs light brown with black stripes. Ventrally pinkish white. ETYMOLOGY Named for Aphrodite, the Greek Goddess of love. ACKNOWLEDGEMENTS We thank Stephen Wilson, Queensland Museum, for his kind help. LITERATURE CITED INGRAM, G.J. 1979. Two new species of skinks, Genus Ctenotus (Reptilia, Lacertilia, Scincidae), from Cape York Peninsula, Queensland, Australia. Journal of Herpetology 13(3): 279- 282. STORR, G.M. 1964. Ctenotus, a new generic name for a group of Australian skinks. Western Australian Naturalist 9: 84-85. WILSON, S.K. AND KNOWLES, D.G. 1988. ‘Australia’s reptiles. A photographic reference to the terrestrial reptiles of Australia,’ (William Collins: Sydney). ELEMENTS IN THE PROCESS OF RECOVERY BY CROCODYLUS POROSUS (REPTILIA : CROCODILIDAE) IN THE EAST ALLIGATOR RIVER AND ASSOCIATED WETLANDS ROBERT W.G. JENKINS AND MALCOLM A. FORBES Jenkins, R.W.G. and Forbes, M.A. 1990 09 20: Elements in the process of recovery by Crocodylus porosus (Reptilia : Crocodilidae) in the East Alligator River and associated wetlands. Memoirs of the Queensland Museum 29(2): 411-420. Brisbane. ISSN 0079-8835. This paper reports the results of spotlight surveys from 1977 to 1987 of Crocodylus porosus populations in the tidal East Alligator River and its associated freshwater wetlands. Comparative data on the size and structure of the tidal and freshwater ‘subpopulations’ are analysed and recovery assessed since protection of the species in 1971. The population in the tidal river has increased significantly at an annual rate of 0.06. Hatchling production in the tidal river has increased significantly at an annual rate of 0.14. In contrast to the absence of any significant long term increase in the numbers of non-hatchling crocodiles in the mid and downstream sections of the tidal river, non-hatc- hling crocodiles in the upstream section (>55km) have increased significantly at an annual rate of 0.14. This increase in the number of crocodiles in the upstream section is largely accounted for by animals >1.2m in length which have increased significantly at an annual rate of 0.11. The data reveal major differences in the population structure between the tidal river and freshwater wetlands. Recruitment into the population is essentially confined to the tidal subpopulation, and is concentrated in the midsection of the river. The absence of suitable nesting habitat severely limits successful nesting in freshwater. The tidal and freshwater subpopulations do not appear to be mutually exclusive. Increases observed in the freshwater subpopulation although not statistically significant, are surmised to have been derived from the tidal subpopulation. Regular seasonal movement of non- breeding crocodiles >1.2m in length between the tidal and freshwater habitat occurs during the wet season and following dry season. This movement is localised in the upstream section of the river and is thought to be the principal mechanism by which animals enter freshwater habitat. Crocodylus porosus populations in the East Alligator River System have responded positively to protection and the amelioration of habitat degradation that has resulted from the active control of feral Asiatic water buffalo Bubalus bubalis in Kakadu National Park by the Australian National Parks and Wildlife Service. (J Crocodiles, Crocodylus porosus, East Alligator River, survey. Robert W.G. Jenkins and Malcolm A. Forbes, Australian National Parks and Wildlife Service, PO Box 636, Canberra City, Australian Capital City 2601, Australia; 20 August, 1988. Crocodylus porosus populations inhabiting the tidal rivers of northern Australia have been described by Messel et al (1979, 1981) as a result of comprehensive spotlight surveys. In that work the authors indicated that saltwater crocodile populations, despite a number of years of protec- tion, were still in a depleted state. They sug- gested, however, that the extensive freshwater swamps associated with the tidal rivers that com- prise the Alligator Rivers Region may act as important recruitment centres or rearing stock- yards for sub-adult crocodiles and hence the rate of recovery in these rivers could be expected to be more rapid relative to most other tidal rivers in northern Australia. STUDY AREA The East Alligator River drains the escarpment country of western Arnhem Land and flows ina generally northerly direction through extensive sub-coastal floodplains into Van Diemens Gulf. The river is tidal for a distance of 84.5km upstream from its mouth (Fig. 1). The influence Ald of all tides in the extreme upstream sections of the East Alligator River except those associated with full ard new moon phases is largely im- peded by the presence of a concrete causeway at Cahills Crossing (84.5km). Fringing vegetation, salinity and temperature profiles for the East Alligator River are described by Messel ct al. (1979). Magela Creek traverses subcoastal floodplains and enters the East Alligator River 49.7km upstream from the mouth, ll is characterised asa series of discrete fresh waterbodies (billabongs) of varying size and depth in the late dry season (October - November). During the wet season when rainfall and run-off substantially increase water levels,individual billabongs become con- nected and inundate adjacent low-lying countryside forming extensive areas of fresh water with emergent vegetation. Mean annual rainfall in the study area is 1556mm, 82% of which occurs between Decem- ber and March (Bureau of Meteorology - Jabiru Recording Station), The result of this rainfall pattern is a distinctive wet summer and dry winter. SURVEYS TIBAL RIVER Surveys Were conducted from 1980-1985 under varying seasonal conditions. However, greajest effort was concentrated in the early (April-May) and late (October-November) dry seasons to coincide with hatchling recruitment and minimum discharge of freshwater from the calechment respectively, No survey was under- taken when the volume of fresh water discharge was sufficient to breach the banks of the tiver. Crocodiles inhabiting the tidal section were counted at night from a boat using a 12 volt’ LO0 wall sealed beam spotlight. The survey area ex- tended [rom the confluence with Cooper Creek 13km from the mouth) to Cahills Crossing {34 sho from the mouth), Cooper Creek was not surveyed, however, twa tidal crecks intersecting the west bank of the river (Creek A iat 32.9km and Magela Creck at 49.7km) were included in ihe surveys (see Fig. 1). Survey procedure described by Messel et al. (1981) was adopted with the following modification. The east and west banks of the East Alligator River between Cooper Creek and the 30km mark were surveyed on consecutive nights. Because of the extreme width of the river below 50km and limitations in personnel and equipment both banks could not MEMOIRS OF THE QUEENSLAND MUSEUM be surveyed simultaneously. Most ©. porosus are Sighted at the water's edge; we assumed that movement between banks was minimal and, if it does occur, movement east would be the same as movement west. Surveys were carried out in the East Alligator River from Cooper Creek in an upstream direc- lion on a half rising tide during periods coincid- ing with high tides in excess of 6 metres (ic. full or new moon phases). The extreme upstream section (74km—84.5km) of the river is not navigable on tides <5.5 metres because of sand and rock bars. Each crocodile located in the spotlight beam was approached and its total length estimated in one foot categories. Animals greater than 10f in length were assigned to the one category. ( > 1 0ft). For analyses, size class data were grouped into the following categories; hatchling, 2-3 ft (0.6- 0.9m), 3-4 ft (0,9-1,2m), 4-6 ft (1.2-1,8m) and > ft (>1.8m). Animals that could not be ap- proached were scored as ‘eyes only’, Animalsin the ‘eyes only’ category were allocated to the >1.8m size class if they were sighted in midstream, or if the spacing between the eyes indicated a large animal. This category repre- sented on average approximately 23 per cent of observations. The ‘eyes only’ component of the population that could not be ascribed with cer- tainty to the >1.8 m category was treated in the following manner, The size frequency composi- tion of crocodiles for cach 5 km length of the tiver was determined and the remaining ‘eyes only’ component for cach section was allocated proportionately among size classes >0.9m. Ex- perience has shown that crocodiles <0.9m are easily approached and sized. If these crocodiles submerge on being approached, they surface nearby almost immediately and their size can be estimated, The location of each animal sighted was plotips onto a calibrated river map compiled by Messel et al.(1982), FRESHWATER WETLANDS On the basis of broad vegetation type the area was stratified into the following three zones: (a) floodplain; (b) woodland; and (c) Melaleuca cor- ridor separating the first two zones. Sample bil- labongs (see Fig. 1) were selected randomly from a pool of accessible billabongs in cach stratum, This paper deals only with those situated in strata (a) and (c). Each billabong was surveyed during October or November prior to the onset of the monsoonal wet season when the areca of surface water was at its minimum. RECOVERY BY CROCODILES IN EAST ALLIGATOR RIVER 413 Pt, Farewell Creek A Cahills Crossing 84-Skm ® \Mudginberri Station FIG. 1. The study area within Kakadu National Park showing the location of survey billabongs on Magela Creek in relation to the East Alligator River. 414 MEMOIRS OF THE QUEENSLAND MUSEUM 8 . y=5.29+0.06x r?=0.56 p<0.05 log, (number of crocodiles) 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 52 g ©) re = 48 3 y=3.04+0.14x Fs) 44 r?=0.66 5 p<0.05 S 40 8 e 36 =} t& ne 3.2 1=)) 2 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 55 r 58 DB 51 = y=4.03+0.14x 8 49 r?=0.81 » “7 e re) p<0.001 al - a 47 ry no oO ie} _o 14 _ BR 45 in r= y=3.84+0.11x 2 2 43 ed r°=0.78 — p<0.001 38 4d 3.9 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 FIG, 2. Regression line relationships between late dry season numbers of crocodiles (natural logarithms) sighted in the tidal East Alligator River and year. (a) Total number of crocodiles in the river. (b) Total number of hatchling crocodiles in the river. (c) Total number of crocodiles in the river (solid datum points and line) and number of animals > 1.2m in length (hollow datum points and dashed line) upstream of 55km. Regression formulae apply x=1,2,3 etc. where 1=1977,2=1978 etc. Lx eee a2) y bo KOO > Pale) Z2 C5 Vou 2ul Ss St6cy ° asd Pi > r able ear regre sis of the transformed late dry season data in the East Alligator River, 1977-83. (N=351 where N represents the sum of maximum number recorded annually RECOVERY BY CROCODILES IN EAST ALLIGATOR RIVER of the billabong on foo and size er i counted with a spotli ro Distribution of hatchling C. porasus 3 IDAL RIVER FIG ( against year on the late dry s sighted during of hatchling cro of C, porosus spotlight surveys in the tidal East Alligator River The numbers ab hey The rate of increase in the number of hatchling crocodiles jn the river does not reflect a uniform distribution of this age class in the river, Maxi- mum numbers of hatchling C. porosus recorded each year from [977 to 1979 (Messel etal., 1979, 1980) and 1980 to 1983 have been pooled for 5 km segments and presented in Fig, 3, Survey data for 1985 and 1987 were gathered ina manner that did not facilitate this form of analysis, Regular annual recruitment during this period has been generally restricted to the mid-section of the river (30km-55km) and Creek A. In comparing the 1977 late dry season data (Messel et al,, 1979) with those far November 1983 (Fig. 4), the distribution of crocodiles in the river has changed significantly (X°=58.423, p<0.001), This difference has resulled from a highly significant increase in the abundance of >1,2m long crocodiles in the upstecam section of the river above 55km(t=9.771, p1,2m long (Fig. 2c) which have been increasing an- nually In this section of the river at 11 percent (r'=0.78,p<0.001). Although the slopes of regression lines for crocodiles >1.2m in length elsewhere in the river for the period 1977-1987 are positive, the relationships are not significant. FRESHWATER WETLANDS The number and sizes of C_ poresws recorded in freshwater billabongs which characterise Magela Creck in the late dry season are sum- marised in Table 2, There is some difficulty in interpreting crocodiles recorded as ‘eyes only’ in freshwater habitat, Phe ‘eyes only” category in freshwater habitats constitutes a greater com- ponent of the sightable population than in tidal rivers. It is not possible to approach animals in shallow water, very often located amongst fallen limber and/or grasses, ta determine a size. We assumed that the frequency distribution of animals able ta be sized isa reflection of the enlire population and apporlioned the “eyes unly’ component among the size classes recorded. The size structure of the freshwater population Uiffers from that inhabiting the tidal East Al- ligator River in that there is an almost total MEMOIRS OF THE QUEENSLAND MUSEUM absence of hatchling and yearling (0.6-0,9m) crocodiles (Table 2). These data suggest that recruitment within the freshwater populations of C. porosus is absent or at best minimal. Nosiatistically significant increase in numbers of crocodiles in comparable billabongs was detected for the period 1980 to 1985 (Table 3). Whilst there has been an increase in the number of C. porasus insome of the sampled billabongs, there has been little change in others. Corn- parisons with surveys of a small series of Magela Creek billabongs undertaken in 1977 (Messel, pers, comm,) indicate an obvious increase in abundance of erocodiles between 1977 and 1980. However, the increase is not statistically significant. This may in part be explained by the inability 10 sight crocodiles in some billabongs in the latler stages of the study and the small sample size, Surveys of Leichhardt Billabong could not be conducted in 1984 and 1985 be- cause the entire surface of this waterbody was covered with the introduced aquatic plant Sal- vinia molesta. This plant was also present on Jabiluka and Nankeen Billabongs during the 1985 surveys and severely hindered progress on the water and the ability to locate crocodiles in previously accessible arcas. DISCUSSION Messel et al. (1981) modelled the dynamics of C. porasus in tidal rivers of northern Australia based on (#) the suitability of a river for crocodile breeding being determined by its salinity charac- teristics and (b) movement of non-reproductive animals from breeding rivers into non-breeding wreas, In considering the East Alligator River system, Messel et al. (1981) recognised the potential importance of the freshwater swamps but were unable to quantify it. Jenkins and Forbes (1985) found that for the Fast Alligator River the distribution as well as the size class structure of C, porasus inhabiting the tidal river in the late dry season (October - November) differed significantly from that during the carly dry season (April - May). They also found that generally C. poresus abundance in the river Was greatest at the end of the dry season, Analysis of the total numbers of crocodiles sighted in the river during the late dry season for the period 1977-1987 indicates an annual rate of increase of 0.06 (Fig. 2a). This rate is marginally lower than the annual rate of increase of 0.07 derived by Bayliss (1987) and Webbetal. (1989) RECOVERY BY CROCODILES IN EAST ALLIGATOR RIVER 40- OCTOBER 1977 N=235 30-7 20-4 SY . \N & ” 104 \N 3 NS \X 9 S Q EN p 3 : =~ 2) c.) So 2 oO i} g E 105 z iy] oO 2 205 30- NOVEMBER 1983 N=380 40-7 Total crocodiles Crocodiles >1.2m 50- Ll 1 1 1 1 iz j|___t ! 1 417 eseneyene & tetesetece eee x ® o 2 Oo a o Da a = Ae 1 1 t 1 1 _t i°) 10 20 30 40 50 60 70 84.5 East Alligator River (kilometres upstream from the mouth) FIG. 4.Distribution of C. porosus sighted during late dry season spotlight surveys in the East Alligator River system in 1977 (above line) and 1983 (below line). 1977 data from Messel et al. (1979). for the period 1977- 1985. This difference in the rates of increase may be due to the two additional years recovery and the omission from this study of (a) Cooper Creek - a major tributary of the East Alligator River, and (b) results of surveys undertaken during the early dry season to avoid incorporating in the regression analysis large numbers of hatchling crocodiles that enter the population during the period but are subject to high mortalities between the early and late dry season (Jenkins and Forbes, 1985). Not- withstanding, it is to be expected that the rate of 418 MEMOIRS OF THE QUEENSLAND MUSEUM TABLE 1. Numbers of C. porosus in various size classes sighted in the tidal East Alligator River system during spotlight surveys 2-3ft 3-4ft Hatchlings (0.6-0.9m) — (0.9-1.2m) June 1980 26 October 1980 21 July 1981 23 November 1981 17 April 1982 89 November 1982*| 29 June 1983 52 November 1983 78 May 1984 34 November 1984 65 November 1985 83 October 1987** 4-6ft >6ft Eyes (1.2-1.8m) (>1.8m) only N 39 30 45 198 51 41 43 213 56 46 34 201 58 54 36 204 20 28 32 207 49 79 44 223 19 81 39 260 60 142 30 380 19 45 21 160 57 74 68 320 64 110 48 367 28 159 74 428 *partial survey (13-30km not surveyed). ** ex—Conservation Commission of the Northern Territory increase in a recovering population will tend towards zero in the long term. It is therefore of interest that the rate of increase in the East Al- ligator River is lower than the mean annual rate of increase of 0.08 for all crocodiles calculated by Webb et al. (1989) for 26 tidal rivers in the Northern Territory. Hatchling production has been variable in the East Alligator River. When initial mortalities have been discounted through regression analysis of late dry season sightings, the annual rate of increase for the period 1977-1987 has been significant at a rate of 0.14 (Fig. 2b). Nest- ing activity has generally been restricted to the mid section of the East Alligator River (30km- 55km) including Creek A and Magela Creek although in later years there has been increased nesting activity in the upstream sections above 55km. Recruitment to the population is essentially confined to the tidal East Alligator River. Hatc- hling production in freshwater is minimal or absent because of the limited availability of suitable nesting habitat (Wells, 1980). In the absence of such habitat nesting effort in fresh- water is generally unsuccessful as nests are sub- ject to inundation with concomitant high egg mortality. The size structure and distribution of the population inhabiting the river have changed significantly since 1977. This study demonstrates a highly significant increase in upstream numbers of crocodiles at an annual rate of 0.14 (Fig. 2c) has occurred in the period 1977-1987. When the size structure of this seg- ment of the population is examined, animals >1.2m in length largely account for the overall increase in the upstream section increasing at an annual rate of 0.11 (Fig. 2c). The marked seasonal difference in the num- bers of crocodiles inhabiting the upstream sec- tion of the river results from a significant reduction in the abundance of >1.2m long crocodiles in the early dry season. Jenkins and Forbes (1985) attributed this difference to wet scason movement of animals out of the river into adjacent freshwater habitat being facilitated by the flooding characteristics of the East Alligator River. These animals can be considered to be non-breeding as limited nesting activity occurs in this section of the tidal river. The observed increase in the freshwater population can thus be explained as having been derived from wet season movement from the breached upstream sections of the river. In the absence of any sig- nificant long-term increase in abundance of non- hatchling crocodiles elsewhere in the river and the virtual absence of hatchling crocodiles in freshwater habitat, the upstream section of the river may be considered the major dispersal cor- ridor to freshwater habitat for animals derived from downstream breeding and nursery areas. The annual concentration and movement of crocodiles >1.2m in length from the upstream section of the tidal river into adjacent freshwater wetlands may also explain the lower annual rate of increase in overall numbers of crocodiles in the East Alligator River relative to the mean rate RECOVERY BY CROCODILES IN EAST ALLIGATOR RIVER No. of billabongs Hatchlings 2-3ft (0.6-0,9m) October 1980 4 Navember 1981 6 November 1982 6 November 1983 8 November 1984 7 419 3-4ft 4-6Ft >6fl Eyes N (0.9-1.2m) (1,2-1.8m) >1.8m only 3 22 10 22 Ss7 4 19 19 26 68 2 17 25 31 77 1 26 64 22 113 10 31 63 35 147 TABLE 2. Numbers of C. porosus in various size classes sighted in freshwater billabongs of Magela Creek during spotlight surveys calculated by Webb et al. (1989) for 26 tidal rivers in the Northern Territory. Movement of C. porosus from the East Alligator River to adjacent freshwater swamps was also suggested by Mes- sel et al. (1980) to explain differences observed in the population in 1977 to 1979. Webb et al. (1983) also recorded marked differences in the number of C. porosus sighted during wet season surveys in the Adelaide River compared with surveys undertaken in the dry season. This be- haviour is consistent with movement docu- mented for other crocodilians in response to rainfall and flooding in Uganda and Northern Rhodesia (Cott, 1961), Louisiana (Chabreck, 1965) and Venezuela (Gorzula, 1978). The restriction of regular recruitment to the mid-section of the East Alligator River suggests that recovery by the population in the tidal sys- tem and associated freshwater wetlands since protection has been derived principally from this area. The most important agent responsible for the destruction of C. porosus nesting habitat is the Asian water buffalo Bubalus bubalis (Letts et al., 1979, Fogarty, 1982). Excessive numbers of this large feral herbivore have been a major constraint to the rate of recovery of C. porosus by limiting available nesting habitat. The ap- pearance of creches of hatchlings in the upstream section of the tidal East Alligator River in later years has accompanied the re-establishment of suitable riverside nesting vegetation following the removal of large numbers of water buffalo by the Australian National Parks and Wildlife Ser- vice under the Plan of Management for Kakadu National Park. Similarly, it has been demonstrated elsewhere in Kakadu National Park that floating grass mats become re- estab- lished in the freshwater wetlands following the removal of significant numbers of water buffalo (Jenkins, unpubl. data). The increasing stability of these vegetation platforms with time may result in successful breeding by saltwater crocodiles in freshwater habitat. The results of this study confirm the ability of C. porosus to exhibit a rapid response to habitat and harvest protection similar to that reported for other crocodilians, viz. Crocodylus niloticus (Blomberg, 1976, Graham, 1976) and Alligator mississippiensis (Campbell, 1978). The continu- ing recovery of the East Alligator River popula- tion of C. porasus together with those inhabiting the other rivers and associated wetlands of Kakadu National Park that are managed under a regime of national park legislative protection TABLE 3. Numbers of C. porosus recorded in freshwater billabongs of Magela Creekin 1977 (Messel pers. comm.) and 1980-1983 n.s.= not surveyed 1977 1980 1981 Buffalo ns 2 J Island 8 11 14 Hidden ns ns 4 JaJa ns 15 10 Jabiluka I ns 9 Nankeen § 28 30 Magela Point ns ns ns Leichhardt 5 ns ns *100% surface coverage of Salvinia sp. 1982 1983 1984 1985 l ] 4 2 14 11 18 11 4 1 7 7 14 1s 23 18 11 Q 17 Ore 23 20 18 14** ns 43 56 30 ns 10 ns* ns* **Substantial coverage (>30%) of surface with Salvinia sp. provide a sound basis for continued commercial ranching outside of the conservation reserve nel- Work, It also demonstrates the need for manage- ment policies for the C, porosus resource to be cognizant of and tesponsive to the increasing potential for interaction between visitors to Kakadu National Park and C. porosus. ACKNOWLEDGEMENTS The authors wish to thank those staff of \he Australian National Parks and Wildlife Service who provided assistance to this study. Special thanks are extended to Jan Morris and Alex Carter who devoted much their time to assist in obtaining field data sometimes under adverse conditions. We are grateful to Professor Messel, University of Sydney, and the Conservation Commission of the Northern Territory for frecly making data available to be incorporated into this study. Charlie Manolis provided valuable comments on earlier drafts of the manuscript. Figures were prepared by Peter Tieman. LITERATURE CITED BAYLISS, P, 1987, Survey methods and monitoring within crocodile management programmes. pp.157-175. in Webb G.J.W., Manolis $.C. and Whitehead PJ, (eds), ‘Wildlife management: Crocodiles and alligators.” (Surrey Beatty and Sons: Sydney). 552pp. BLOMBERG, G.E,D. 1976, Feeding and nesting ecology and habitat preference of Okavango crocodiles, pp. 131-139. /m ‘Proceedings of the Symposium on the Okavango Delta and ils fu- ture utilisation”, (Gaborone: Botswana). 350pp, CAMPBELL, H.W. 1978, American Alligator, AL ligator mississippiensis (Daudin), (Mimeo- gtaphed report prepared by U.S. National Fish and Wildlife Laboratory, Gainesville, Florida). 14 pp. CHABRECK, R.H. 1965. The movement of alliyalors in Louisiana, Proc, Southeast, Assoc, Game Fish Comm, 19: 102-110. COTT, A.B. 1961. Scientific results of an inquiry into the ecology and economic status of the Nile crocodile Crocodylus niloticus in Uganda and Northern Rhodesia. Trans, Zool, Soc, Lond. 29(4); 211-337, FOGARTY, P. 1982. ‘A preliminary survey of en- vironmental damage associated with activity of feral buffalo,’ (Conservation Commission of the Northern Territory: Darwin), GORZULA, S.J, 1978. An ecological study af MEMOIRS OF THE QUEENSLAND MUSEUM Caiman crocodilus crocodilus inhabiting savan- na lagoons in the Venezuelan Guayana. Oecologica Berl. 35: 21-34. GRAHAM, A, 1976, A management plan for Okavan- go crocodile. pp.223-234. In ‘Proceedings of the Symposium on the Okavango Delta and its furure utilisation’. (Gaborone: Botswana). 350pp. JENKINS, R.W.G. AND FORBES, M.A, 1985. Seasonal variation in abundance and distribution of Creeodylus porosus in the tidal East Alligator River, Northern Australia. pp. 63-69. In Grigg, G., Shine R, and Ehmann H. (eds), ‘Biology of Australasian frogs and reptiles”. (Surrey Beatty & Sons: Sydney). 552pp. LETTS, G.A., BASSINGTHWAITE, A. AND DE VOS, W.E.L, 1979. ‘Feral animals in the North- em Temtory.” (Report of the Board of Inquiry. Dept. of Primary Production: Darwin). 234 pp, MESSEL, H., WELLS, A.G, AND GREEN W.J. 1979.‘ The Alligator Region River systems. Sur- veys of tidal river systems in the Northern Ter- ritory of Australia and their crocodile populations. Monograph 4.’ (Pergamon Press: Sydney). 70pp. MESSEL, H., VORLICEK, G.C., WELLS, A.G. AND GREEN, W.J. 1980. ‘Tidal waterways of Van Diemen Gulf, Surveys of tidal river systems in the Northern Territory of Australia and their crocodile populations. Monograph 14.' (Per- gamon Press: Sydney). 104 pp. 1981, ‘The Blyth-Cadell Rivers System Study and the status of Crocodylus poresus in tidal water- ways of Northern Australia. Surveys of tidal river systems in the Northern Territory of Australia and their crocodile populations. Monograph 1.’ (Pergamon Press; Sydney), 463 PP: 1982, ‘Work maps of tidal waterways in northern Australia, Surveys of tidal river systems in northern Australia. Monograph 15,’ (Pergamon Press; Sydney). 463 pp. WEBB, G.J.W., MANOLIS, S.C, AND SACK, G.C, 1983, Crocodylus johnstoni and Cracodylus porosus co-existing in atidal river. Aust. Wildl. Res. 10: 639-650. WEBB, G.J,W., BAYLISS, P. AND MANOLIS, S.C. 1989. Population research on crocodiles in the Northern Territory, 1984-1986, pp, 22-59, In “Proceedings of the 8th Working Meeting of the IUCN/SSC Crocodile Specialist Group. Oct. 1986’, (Quito: Ecuador). 204pp. WELLS, A.G. 1980. Kakadu, A sanctuary for crocodiles? Tigerpaper 7: 2-4. GROWTH AND CALCIUM METABOLISM OF EMBRYOS OF THE LONG-NECKED TORTOISE, CHELODINA LONGICOLLIS (SHAW). JEFFREY D. MILLER AND MENNA E. JONES Miller, J and Jones, M. 1990 09 20: Growth and calcium metabolism of embryos of the long-necked tortoise, Chelodina longicollis (Shaw). Memoirs of the Queensland Museum 29(2): 421-436. Brisbane. ISSN 0070-8835. Growth (change in mass) and calcium metabolism of embryonic turtles have received little attention. The present study extends this small data set to include the long-necked tortoise, Chelodina longicollis, Shaw (Testudinata: Chelidae). Eggs were 3/4 buried in vermiculite and incubated under controlled moisture (-200kPa) and temperature (30°C) conditions. Embryonic growth was described by the regression of log dry mass on log day of incubation. Total calcium in the egg did not change but was redistributed. The demand for calcium during embryogenesis exceeded the amount available from the yolk and albumen; the additional calcium required for osteogenesis was supplied by the shell. L] Growth, calcium metabolism, tortoise embryos. Jeffrey D, Miller and Menna E. Jones, Department of Zoology, University of New England, Armidale, NSW, 2351; Present Addresses: J.D.M., Queensland National Parks and Wildlife Service, Pallarenda, Townsville, Queensland 4810, Australia; M.E.J., Department of Zoology, University of Tasmania, Hobart, Tasmania 7000, Australia; 17 August, 1988. The few studies that have considered growth (change in mass) of embryonic turtles have been restricted to cryptodires (Ackerman, 1980, 1981, Cheloniidae: Caretta, Chelonia; Miller, 1982, Cheloniidae: Caretta, Chelonia, Eretmochelys, Natator; Morris et al., 1983, Chelydridae, Chelydra serpentina; G. Packard et al., 1983, Emydidae, Chrysemys picta). Growth of these embryos follows an exponential or logistic form. There are no published data on the growth of embryonic pleurodiran turtles. Four studies have considered the pattern of calcium metabolism within incubating eggs of oviparous reptiles (Packard et al., 1984a, Chelydridae: Chelydra serpentina; Packard et al., 1984b, Colubridae: Coluber constrictor, Packard et al., 1985, Agamidae: Amphibolurus barbatus; Packard and Packard, 1986, Emydidae, Chrysemys picta). Other studies con- cerning the utilization of calcium in reptilian eggs only identified that the shell was the primary source (Simkiss, 1962, 1967; Bustard et al., 1969: Dermochelyidae, Cheloniidae). In eggs of the American snapping turtle, Chelydra serpentina, about 56% of the calcium for embryogenesis originates in the shell (Packard et al., 1984a); embryos of sea turtles obtain be- tween 60 and 80% of the required calcium from the shell (Simkiss, 1962, 1967; Bustard et al., 1969). The present study was designed to extend the knowledge of growth and calcium metabolism in reptilian embryos to include the long-necked tortoise, Chelodina longicollis (Shaw, 1794) (Testudinata: Chelidae). The specific objectives were (a) to describe the pattern of embryonic growth, (b) to determine the amount of calcium in each of the compartments of the egg (yolk, albumen, shell) at oviposition and (c) to describe the pattern of calcium utilization and the relative contribution from these sources during develop- ment of the embryo. Preovulatory follicles were also analyzed for calcium content. METHODS Gravid tortoises were collected by netting dams between 13 November and 7 December, 1983 at Herbert Park (30°27’S, 151°50’E), ap- proximately 25 km northeast of Armidale, NSW. Tortoises were housed in individual aquaria at ambient temperature (18-25°C). No tortoise was retained longer than 10 days. The straight carapace length (SCL) of each tortoise was measured to the nearest 0.01cm with calipers from the anterior edge of the nuchal (cervical) scute to the posterior edge of the postcentrals (12th marginals). Each tortoise was weighed to the nearest 5 grams using a Pesola spring balance. Tortoises were palpated in the inguinal area to determine if eggs were present. Oviposition was 422 22 MEMOIRS OF THE QUEENSLAND MUSEUM induced by intracoelomic injection of oxytocin at a dosage of 1 iu/100g of total body mass (Ewert and Legler, 1978). If the first injection did not produce eggs within one hour, a second injection was given. If no eggs were expelled, a third injection was given. No more than three injections were given to any tortoise; all eggs were laid in water. Within 2 minutes of egg laying, eggs were removed from the water, wiped dry, and num- bered using permanent ink. Eggs were weighed to the nearest 0.01 gram ona Bosch p115 top pan balance and measured (length and width) to the nearest 0.01em with calipers. Eggs were incubated in 2.5 litre plastic con- tainers sealed with tight-fitting lids. Eggs were 3/4 buried in vermiculite moistened to 15 % water by weight (approximately -200kpa). The moisture level was maintained by adding 3-8m] distilled water every 7-10 days during incuba- tion, The temperature was maintained at 307/. 0.7°C throughout incubation. The position of cach container was regularly shifted within the incubator to minimize the potential effects of temperature gradients (Bull et al., 1982). If most eggs in a clutch did not exhibit a white area on the uppermost part of the shell (an indica- tion of viability) within 7 days of oviposition, the clutch was not used for the calcium metabolism experiments. One cgg from each clutch was collected at predetermined times during incubation (0, 20, 30, 35, 40, 45, 50, 55, 60 days, and hatching). Each egg was re-weighed and re-measured at the time of sampling then opened and the contents were separated into embryo, yolk, membranes and fluids, and shell. The fresh mass of cach component was obtained to the nearest mil- ligram. The embryos were staged according to Yntema (1968). To determine hatching success, it was assumed that normal embryos obtained from eggs sampled during incubation would have developed into normal hatchlings and it was assumed that abnormal embryos would not have produced hatchlings. Follicles (n=66) of different measured diameters (+0.01cm) were removed from the ovaries of four decapitated tortoises and weighed (+ 0.01g). Each egg component and follicle was dried to constant mass at 60°C. Embryos were ground in a mill into fine particles; yolks and follicles were homogenized by hand using a mortar and pestle. Subsamples of 300mg or entire samples were digested in boiling, concentrated nitric acid aided by 30 % hydrogen peroxide (both reagent grade). Samples were brought to 100m1 volume with distilled water. Further dilutions of diges- tate were made using 0.02 % strontium chloride (1ml sample + 9ml dilutent) dispensed through an auto-dilution system (Hook and Tucker In- struments, New Addington, England). Five samples of water from vermiculite that had been soaked for 20 days in distilled water were also analyzed for calcium. The calcium content of each sample was deter- mined using a Pye Unicam SP 190 single-beam Atomic Absorption Spectrophotometer (using an air/acetylene flame) coordinated with an SP 450 automatic sample changer following stand- ard procedures. The data were analyzed using one-way analysis of variance with the initial egg mass as the potential covariate (programme BMDP2V, Dixon et al., 1981). Further examination of the data for calcium in yolks and embryos was done by comparing the regressions for the variables among the clutches. This approach was taken because only one egg was collected from each clutch at the sampling times which precluded use of analysis of co-variance with groupings by clutch and day of sampling. Data for the regres- sion analysis were truncated to eliminate time zero because preliminary analysis demonstrated no significant differences occurred between data at time zero and day 20. Truncation simplified the curvilinear nature of the remaining data and allowed better comparison on a linear basis. Be- cause of this, however, these regression lines do not fully describe the entire data set. Sample sizes vary among the treatments; P = 0.05 was uscd to establish significance. Least Significant Difference (LSD) values were calculated using Statistix (Analytical Software, St Paul, MN, USA) computer program; calculation of other statistical tests followed Zar (1974), RESULTS A total of 64 female tortoises were captured and injected with oxytocin; only 16 (25%) produced eggs (Table 1), usually after a third injection. The general effectiveness of oxytocin on the tortoises was low. The time interval be- tween injection and oviposition was variable. The shortest period was 1.5 hours and the longest was nearly 15 hours. All tortoises required at least two injections before oviposition was in- duced. Most tortoises were not obviously dis- tressed by handling or injection; all animals GROWTH AND CALCIUM METABOLISM OF EMBRYONIC TURTLES 423 TABLE 1. Summary on tortoises, clutches, incubation period, and hatching success. SCL (cm) FEMALE NUMBER NUMBER OF EGGS WEIGHT (g) USED IN CALCIUM DETERMINATION 2a MEAN STD DEV MINIMUM MAXIMUM NUMBER 19.34 17.84 Me OwWwshouw 78 19.81 99 1 102 17.72 103 20.18 19.25 19.023 0.932 MINIMUM 17.72 MAXIMUM 20.18 NUMBER remained watchful and active during their cap- tivity. Oviposition was usually preceded by a slight increase in activity. In total, 175 eggs were induced from 16 tor- toises giving a mean clutch size of 10.93 eggs (sd=3.47, range=4-17) (Table 1). Non-viable eggs were found in 7 clutches containing 73 eggs. SIZE AND WEIGHT OF FEMALE TORTOISES The mean straight carapace length (SCL) of female tortoises captured during the study was 19.lem (sd=2.229, Range=14.1-26.5, n=59). There was no significant difference between the mean SCL of tortoises that laid eggs (19.1cm, sd=0.867, range=17.72-20.18, n=11) and the mean SCL of those that did not (19.1cem, INCUBATION PERIOD (days) HATCHING SUCCESS (%) =) ooRooOoO nN sd=2.411, range=14.12-26.5, n=48) (t=0.0831, df=57). In the group of tortoises that yielded eggs, there was no significant difference be- tween the SCL measurements of those tortoises whose eggs were used in the calcium experiment and the others (t=0.254, df=10). The mean mass of all female tortoises was 855.4g (sd=229.7, range= 364-1440, n=62). There was no significant difference between the mean mass of the tortoises which yielded eggs (909g, sd=130.2g, range=720-1080, n=14) and the mean mass of those that did not (824.5g (sd=266.1, range=364-1440, n=48) (t=1.162, df=61). There was no significant difference be- tween the masses of those females whose eggs were used in the calcium experiment and the others (t=0.252, df=12). Further comparisons between the tortoises uscd in the calcium experiment and the others that oviposited revealed no significant differen- ces in the number of eggs (t=1,344, df=14), total clutch weight (t= 1.5889, df=14), or in the per- centage of the female weight represented by the total clutch weight (t=1.346, df=12). The relationship between the straight carapace length and mass of all female tortoises was described by the equation: SCL = 0.00835 female mass + 11.983, R-=0.86 where SCL, is in cm and weight is in grams. The relationship between SCL and log of the mass for the tortoises which yielded eggs was described by the cqua- lion: SCL (cm) = 21.597-0.3572 log female mass (9), R°=0.26. SIZE, AND WEIGHT OF EGGS Based on the [02 viable eggs which were used in the study of calcium metabolism, the mean mass of freshly oviposited eggs grouped by clutch was 6.0g (sd=1,04, range=4.85-7.83, n=9 clutches), The mean length of these eggs was 3.09¢m (sd=0.1824, range=2,81-3.38, n=9 elutches); the mean width was 1.82cm (sd=0.16, range=1,51-2,04, n=9 clutches). The relationships between ege length, width and mass were described best by the equations: Mean Egp Width = -0.3162! (Mean Ege Length) + 2.7797, R*=0,074, P=0.05: Mean Egg Width = 0.14245 (Mean Ege Fresh Mass) + 0.9742, R-=().725, P<0),01 Mean Egg Length = 0.03018 (Mean Egg Fresh Mass) + 2.5811, R°=9.037, P>U.05. There was a tendency for the width of eggs to increase as mass increased, but the mean epg length did not increase proportionally. Female mass was not cogrelated with the num- ber of eggs produced (R~= 0.099, P>0.05) but was slightly correlated with mean egg mass (R-= ().247, P<0,.05) (Fig. 1). The poor correlation between egg mass and female mass is partially explained by the ineffectiveness of the oxytocin on the tortoises; il is passible (hat some females were induced to oviposit incomplete clutches, The correlation between female SCL and total clutch mass was also poor (R°=0,269) and so was that between female mass and total clutch mass (R°=(1.512). The mass of clutches that contained viable eggs represented an average of 7.4 % {range= 3_16-9,02, n=9) of the female mass. There was a slight decrease in the mass of all MEMOIRS OF THE QUEENS AND MUSEUM eggs during incubation (egg mass = -U.042 + 6.863 days, n=10)2). The slopes of the regres- sions of the change in egg mass grouped by clutch were not significantly different (F 0.05 [8,62}]=1.123); however the elevations were (F 0.05 [8,70]=56.64). Analysis of variance among the initial masses of the eggs from different clutches showed that significant differences oc- curred between clutches (Table 2). Among the 102 eggs used in the calcium study, 95 eggs (93.1 %) produced normal embryos or hatchlings (given the assumptions above); only 6.9 % of {hese eggs failed to develop normally. The mean duration of incubation at 30°/.0.7°C was 70,41 days (sd=5.85, range=63-82, n=12 cggs from 7 clutches) when averaged from the clutch averages. When considered inde- pendently of clutch, the mean was 69.2 days (sd=5,25, range 63-82, n=12). WET AND DRY MASS The relation between wet mass and dry mass of follicles followed the power curve {log dry mass = 0,9608 + 0.6693 log wel mass, R*=0.993). The diameter and wet mass of fol- licles were related according to the equation: log wet mass = 2.896 + 0.4514 log diameter, R-=0.871, There was no significant difference between the mean mass of the 15 largest follicles and the mean mass of yolks of ovipositional eggs (wet mass: t=1,13, df=22, P>0,05). However, there was a significant difference between the dry masses (t=3.71, df=22, P<0.05), The mean dry mass of the follicles was heavier than that of the ovipasitional egg yolks, this may: have resulted from differences between clutches. At oviposition, a turtle egg is comprised of shell, albumen, yolk and embryo, However, the embryo is at the early gastrula stage (Yntema, 1968; Cunningham, 1922; Lynn and von Brand, 1945; Miller, 1985) and contributes little to the (oral mass, Because of the difficulty of removing, the blastodise from the yolk, its actual contribu- tion Was ignored until itcould be retrieved (about Stage 10, Yntema, 1968). The proportions of the egg components changed little during the first third of incubation but thereafter the amount of yolk and albumen decreased and the amount of embryo increased (Table 3). The mass of the fresh egg shell decreased significantly during incubation (F 0.0S5[10,S7] =3.885) but the dry mass of the shells did not decrease significantly (F 0,05[6,28] = 0,186). Although there was a significant decrease in GROWTH AND CALCIUM METABOLISM OF EMBRYONIC TURTLES 425 9 P R 8 E K 2 T fe = 7 Oo uJ J N 3 : © U © 6 I WW | tf Qs L D F 5 { } M 4 600 700 800 900 1000 oo 1200 FEMALE WEIGHT (9) FIG. 1. The relation between female body weight and mean egg weight grouped by clutch. Letters identify clutches: D=15, E=102, F=16, H=82, I=89, J=41, K=122, L=118, M=6b, P=78, Q=27, R=11, S=6a, T=61, U=103. the wet mass of the albumen beginning at about one-third of the incubation period (F 0.05[8,29] = 6.265), there was not a significant decrease in the dry mass of the solid material of the albumen during incubation (F 0.05[8,29] = 2.201). The wet mass of the yolk decreased sig- nificantly during incubation (F 0.05[10,55] = 22.866). This was mirrored by the decrease in dry mass of the yolks (F 0.05[10,55] = 13.277) (Fig. 3). The combined mass of the fresh yolk and al- bumen decreased significantly during incuba- tion (F 0.05[9,42] = 15.464). The combined mass of dry yolk and albumen exhibited a significant decrease (F 0.05[9,42] = 26.416). The mass of water in the albumen, yolk and combined yolk and albumen decreased during incubation in concert with the decrease in the total mass of each. However, the percentage of water in the albumen and yolk remained relative- ly constant (Table 4). Water comprised ap- proximately 95.3% of the mass of fresh albumen throughout incubation. Water contributed ap- proximately 69.3 % of the total mass of the yolk at the beginning of incubation but only 56.3 % at hatching. There appeared to be little change in the proportion of water in the yolk during the first 426 MEMOIRS OF THE QUEENSLAND MUSEUM TABLE 2. Results of analysis of variance between clutches for ovipositional weight, length, and width of eggs. ea 2 SQUARES OVIPOSITIONAL EGG WEIGHT 103.561 7.787 REGRESSION RESIDUAL OVIPOSITIONAL EGG LENGTH REGRESSION RESIDUAL OVIPOSITIONAL EGG WIDTH REGRESSION RESIDUAL one-half of incubation but during the last half, the proportion of water decreased. Even though there was a nett decrease in the solid material in the yolk, the percentage of solids increased be- cause, proportionally more water than solids was removed. The fresh mass of the embryos increased sig- nificantly (F 0.05[9,61 ] = 24.157) during incuba- tion. In embryos, as the amount (mass) of water increased, the percentage of water decreased (Table 4). The dry mass of embryos increased DEGREES FREEDOM MEAN SQUARE 12.945 154.597 | 0.001 0.984 0.389 53.943 | 0.001 0.007 0.262 15.821 | 0.001 0.017 significantly during incubation (F 0.05[9,60] 20.588). STAGE OF DEVELOPMENT The stage of development reached by the embryos at any given time during incubation at 30°C was described by the equation: log stage = 1.448 + 0.4325 log days, R°=0.651. The stages of development increased rapidly during the first half of incubation and began to plateau shortly thereafter (Fig 2). This is predictable because more stages are defined to occur during this TABLE 3. Percent composition of fresh egg components at selected times during incubation. ND=not rr == uy OF nie vars WEIGHT OF ALBUMEN MEAN STD DEV NUMBER WEIGHT OF YOLK MEAN STD DEV NUMBER WEIGHT OF EMBRYO MEAN STD DEV NUMBER GROWTH AND CALCIUM METABOLISM OF EMBRYONIC TURTLES 427 26 @2 e @25 e 24 e e er = @23 22 @22 e@2 20 @20 = @19 18 eu = O16 @16 al uw >14 eu a 6 - HATCHLING 12 eV - YOLK SAC REDUCED, BODY PIGMENTATION COMPLETE O CARAPACE VERY DARK, BODY MARKINGS LIGHT TARAPACE PIGMENTED 10 ¢ = SKIN PIGME NTAT ION PALL Ww LOWER EYE LID CROSSES LOWER MARGIN OF LINE MARGINAL LAMINAL DEP INED 4 8 LOWLR £ YI LID INDICATED = - CARAPACE MARGIN COMPLETE w 6 - ALLANTOIS GREATER THAN CROWN RUMP LENGTH = Li ALLANTOIS PRESENT MIDDLE GASTRULATION 4 2 0 10 20 30 40 50 60 70 DAYS OF INCUBATION FIG. 2. Stages of development of Chelodina longicollis based on the table of normal development for Chelydra serpentina (Yntema, 1968). The box provides some of the characteristics for determining the stage. period when differentiation is occurring rapidly (see Yntema, 1968). Development of the allan- tois began before the mid-point of incubation. As the membrane grew into position to participate fully in metabolic activity, embryonic growth accelerated. CALCIUM IN FOLLICLES AND EGGS The calcium available from the incubation medium (vermiculite + distilled water) averaged (0.0 19mg (range=0.025-0.015, n=5). The pattern of increase in calcium in follicles of increasing dry mass followed the formula:log Ca = 1,3483+ 1.2086 log dry follicle mass (R°=0.792). The amount of calcium contained in the 15 largest follicles did not differ significantly from that in ovipositional yalks (t = 0.268, P > 0.05). Twelve samples of shell were analyzed to pro- vide an estimate of total calcium available for translocation into the developing embryo. Shells were collected from early in incubation because later shells tended to fragment. The mean of the samples was 184.9mg Ca (sd=41.054, range=135.6-248.7), Eighteen samples of albumen were analyzed to determine calcium content. The mean was ().6252mg Ca (sd=0.4422. range=0.119-1.637). The samples were accepted only as an indication and were not subjected to further analysis be- cause of the potential of contamination by other materials (e.g. blood, extra-embryonic fluids and granules of shell) during preparation. The amount of calcium in the yolk did not change significantly during the first 30 days of incubation, then it declined rapidly for about 20 days after which it leveled off at about 2% of the mean starting value (Fig. 4; Table 5). Analysis of variance with initial egg mass as the potential covariate showed there was significant decrease in the amount of calcium in the yolk (F 0,05[9,69] = 135,40). Comparison of the slopes of regression lines of the amount of calcium in the yolk against time for each clutch were not 428 1 1:0 a 4 EMBRYO w YOLK DRY WEIGHT IN GRAMS B 0 10 20 30 MEMOIRS OF THE QUEENSLAND MUSEUM 40 50 60 0 DAYS OF INCUBATION FIG. 3. The change in dry mass of embryos and yolks during incubation of Chelodina longicollis eggs. Vertical lines indicate one standard deviation; numbers are sample sizes. Means that are separated by at least one LSD bar are significantly different at the 0.05 level. significantly different (F 0.05[8,50] = 1.594) but the elevations were (F 0.05[8,58] = 3.058). The concentration of calcium in the yolk remained relatively stable for the first half of incubation but decreased thereafter to hatching (Fig. 5). Analysis of variance with initial egg mass as the potential covariate revealed that the concentration of calcium in the yolk decreased significantly during incubation (Table 6). Dif- ferent clutches displayed significantly different slopes indicating that the concentration of cal- cium during incubation was significantly dif- ferent (Table 6). The amount of calcium in the embryos did not increase significantly until after 30 days of development. The rate of change was slow at first then increased rapidly until just before hatching. The amount of calcium in the embryo of ovipositional eggs was assumed to be negli- gible; this was confirmed by analysis of slightly older embryos (Table 5). Analysis of variance with initial egg mass as the potential covariate demonstrated that significant differences oc- GROWTH AND CALCIUM METABOLISM OF EMBRYONIC TURTLES 429 TABLE 4, Percentage of water and solids in egg components al selected times during incubation. ND= not determined. PERCENT WATER ALBUMEN MEAN Stp DEV NUMBER YOLK MEAN Stp DEV NUMBER EMBRYO MEAN Stp Dev NUMBER PERCENT SOLIDS ALBUMEN MEAN Stp DEV NUMBER YOLK MEAN STD DEV NUMBER EMBRYO MEAN STp DEV NUMBER curred in the amount of calcium in the embryos during incubation (Table 6). The concentration of calcium in the embryos increased steadily from about 20-30 days until hatching (Fig. 5), The increased variance following 60 days of incubation resulted from either of two sources; (1) contamination by extra-embryonic fluids, blood, or granules of shell or (2) non-dis- crimination between late stage embryos and hatc- hlings. Analysis of variance with initial egg mass as the potential covariate substantiated that the concentration of calcium increased significantly during incubation (Table 6). The total amount of calcium in the egg, exclud- ing the shell, increased significantly during in- cubation (Table 6). The pattern of change in the total calcium followed that of the yolk for slight- ly more than the first half of incubation and followed the increase in the embryo thereafter (Fig. 5). The demand for calcium by the embryo for osteogenesis exceeded the amount available from the yolk and albumen combined. Because only a low concentration of calcium was avail- able from the incubation medium and because eggs showed a net loss of water during incuba- DAY OF INCUBATION tion, it is assumed that the extra requirement was supplied by the shell. DISCUSSION References to studies of the ecology and general biology of Chelodina longicollis are provided by Cogger et al. (1983); Parmenter (1976) studied the ecology of this tortoise in the Armidale region and provided a comparative review of the older Jiterature. Parmenter (1976, 1985) found that larger females tended to lay more eggs than smaller ones, The average number of eggs reported was 13.9 (sd=4,29, range 6-23, n=74). The average clutch size in the present study was 10.93 (range 4-17) eggs. Both of these values are consistent with other reports on the number of eggs per clutch (Harrington, 1933: n=up to 20; Goode, 1967: n=10-15; Krefft. 1865; n=15-20: Mc- Cooey: 1887, n=15-36; Lucus and Le Souef,. 1909: n=7-21; Vestjens, 1969: n=13-24). Atan incubation temperature of 30°C, incuba- tion requires 60-69 days (Parmenter, 1976, 1985), 73-78 days (Goode and Russell, 1968). 53-76.5 days (Legler, 1985) and 63-82 (this 430 MEMOIRS OF THE QUEENSLAND MUSEUM 17 20 4 EMBRYO @ YOLK @ TOTAL ve LSD CALCIUM CONTENT (mg) fe} fe) 10 20 30 40 50 60 70 DAYS OF INCUBATION FIG. 4. The content of calcium in embryos, yolks and total egg (excluding the egg shell) during incubation of Chelodina longicollis eggs. Vertical lines indicate one standard deviation; numbers are sample sizes. Means that are separated by at least one LSD bar are significantly different at the 0.05 level. GROWTH AND CALCIUM METABOLISM OF EMBRYONIC TURTLES 431 26 24 22 20 18 A EMBRYO gw YOLK 16 14 12 LSD 10 © CALCIUM CONCENTRATION (mg/g dry wt) 0 10 20 30 40 50 60 70 DAYS OF INCUBATION FIG. 5. The concentration of calcium in embryos and yolks during incubation of Chelodina longicollis eggs. Vertical lines indicate one standard deviation; numbers are sample sizes. Means that are separated by at least one LSD bar are significantly different at the 0.05 level. study). Because the duration of incubation varies Parmenter (1976) reported a very strong cor- inversely with temperature, eggs incubated relation between the log of female mass and under natural conditions require longer to hatch _ straight carapace length (SCL) (R°=0.97). In the (Goode, 1967: 130-137d; Goode and Russell, present study, the linear regression had the best 1968: 131-145d; Vestjens, 1969: 118-150d; Par- correlation between female mass and SCL. Par- menter, 1976: 105-1234). menter (1976) also found a significant, positive MEMOIRS OF THE QUEENSLAND MUSEUM TABLE 5. Amounts of calcium (mg) in egg components at selected times during incubation, ND=noat determined. ALBUMEN MEAN STD DEV NUMBER YOLK MEAN STp Dev NUMBER EMBRYO MEAN STD DEV NUMBER SHELI. MEAN STD DEV NUMBER ALBUMEN GROUPS NOT SIGNIFICANTLY DIFFERENT Anova' F 0.05 [4,9]=0.5819 EMBRYO GROUPS SIGNIFICANTLY DIFFERENT ANOVA: FO.001[3,24]=60.284 correlation between the size of the female and the total number of eggs in the clutch. This was not supported in the present study but may simp- ly reflect differences in sample sizes. Although there was a reasonable: correlation (R° =0,77) between egg diameter and the SCL of the female. the correlation between egg Jength and SCL was poor (R°=0.18) (Parmenter, 1976), The peneralization that egg length does not increase substantially with SCL but that egg diameter does was supported by the present results. The correlation between mean egg width and mean egg mass is associated with an increase in follicle diameter in larger females. Although not demonstrated in the present study, the increase in follicle diameter affects the diameter and in- crease in mass of the eggs in marine turtles (Miller, 1982). Because the eggs are oval and the yolks are round, there is little albumen between the vitelline membrane of the yolk and the inner portion of the shell membrane. The albumen is situated primarily towards the ends of the egg. Embryonic growth in Chelodina longicollis as indicated by the change in dry mass of embryos and yolks is similar to that reported for Chelydra serpenlina (Morris et al., 1983) and follows the general pattern of embryonic growth in marine turtles (Ackerman, 1981: Miller, 1982), DAY OF INCUBATION 20 YOLK GROUPS SIGNIFICANTLY DIFFERENT ANOVA: FO.O01 [4,36] =28.311 SHELL GROUPS NOT SIGNIFICANTLY DIFFERENT ANova: F 0.005{2,7]=0.3705 The distribution of calcium in fresh eggs of Chelodina longicollis is similar to that reported for other species of oviparous reptiles (see Pack- ard and Packard, 1984). At oviposition the al- bumen contained only a small quantity of calcium, The yolk contained more, but less than the egg shell (Table 5). At the end of incubation, calcium reserves in the albumen had not been significantly reduced. Those of the yolk were reduced and calcium in the embryo had in- creased significantly. The yolk contributed about 30% of the total calcium required by the developing embryo; the remaining 70% of the embryonic requirement was derived from the shell. This compares favorably with the contribution (% Ca) made by (he egg shells of sea turtles (Chelontidae and Dermochelyidae) of between 60 and 80 % (Sim- kiss, 1962, 1967; Bustard et al., 1969; Miller and Jones, unpub data). The contribution made by the egg shell of Chelodina longicollis is about 15 % higher than occurs in Chelydra serpentina (56%, M. Packard et al., 1984b) and is much higher than the contribution made by the poorly calcified egg shell of the snake Coluber constric- tor (21 %, M. Packard et al., 1984a) and the lizard Amphibolurus barbatus (40 % M. Packard etal,, 1985), GROWTH AND CALCIUM METABOLISM OF EMBRYONIC TURTLES 433 TABLE 6, Analysis of variance tables for calcium content and concentration in embryos, yolks, and the total egg (excluding shell) with initial egg weight as the potential covariate based on log (value + 1) NS=Not Significant. SOURCE SUM OF DEGREES MEAN F SQUARES FREEDOM SQUARE YOLK Dry WEIGHT (2) Day of incubation Initial egg weight Error EMBRYO DRY WEIGHT (g) Day of incubation Initial egy weight Error YOLK CA CONCENTRATION (mg/g) Day of incubation Initial egg weight Error Empryo CA CONCENTRATION (mg/g) Day of incubation Initial egg weight Error YOLK CA (mg) Day of incubation Initial egg weight Error EMBRYO CA (mg) Day of incubation Initial egg weight Error 112.512 0.625 TOTAL CA (mg) IN EGG EXCLUDING SHELL Day of incubation Initial egg weight Error The total calcium in the entire egg (shell, al- bumen, yolk, embryo) apparently does not change but is redistributed. The total egg con- tains enough calcium for embryogenesis without obtaining any from the environment. Hawever, the combined yolk and albumen cannot supply the requirements for embryogenesis without augmentation from the shell. The pattern of embryonic growth and incorporation of calcium indicates that the two sources (yolk, shell) are utilized more or less sequentially. More calcium is derived from the shell later in incubation as reserves in the yolk decline. This is consistent with the pattern of development of the allantois and indicates that extraction from the egg shell is by the chorioaliontic membrane. Further, the two sources of calcium are not utilized equally by the embryo. By the end of incubation the calcium in the yolk was nearly exhausted; whereas the shell contained sufficient calcium to supply all the embryonic demand (based on samples taken early in incubation, Table 5), The pattern of calcium utilization by embryonic Chelodina longicollis is similar to that reported for Chelydra serpentina (Packard etal., 1984a) and Chrysemys picta (Packard and Packard, 1986). The amount of calcium in the yolk is al first relatively stable suggesting little use of calcium by the embryo. However, the amount of yolk declines sharply after about 50 “% of incubation, Similarly, (here is vo detectable change in calcium during the carly stages of embryonic differentiation. However, (he amount of calcium increases rapidly after the embryo begins the growth phase of development, The only noticeable variation in the two patterns occurs lale m incubation of Chelodina longicol- Irs when the rates of calcium uptake in the embryo and loss by the yolk is slow. This may be the case in reptilian eggs when the incubation period is variable. The incubation ranged be- tween 63 and 82 days under constant conditions. This range is equal to one-third of the fastest developmental time and one-quarter of the slowest, At present the variation cannot be evaluated because few clutches were incubated, und the possibility of a subile influence by minor temperature gradients cannot be discounted. There may also have been a minor influence resulting from forced oviposition, as has been shown in lizards (Jones, 1983), Rupid mobilization of calcium into the embryo in the litter half of incubation coincides with the osteogenic phase of development, The primary use of calcium by the embryo is in the building of hones (Simkiss, 1967), Embryonic calcium levels increase in concer| with the decrease in yolk caleium. The yolk supplies the calcium necessary for the initiation of the embryonic growth phase. The somatic development of the embryo requires calcium (albeit small quantities) al a time when the allan- tois has not developed sufficiently to give ready access to the reserves.in the shell. As the vitclline (nembrane extends around the yolk, the arcu between the sinus terminalis and the embryo beconres vascularised. This occurs close enough to the inner part of the shell membrane to allow the necessary respiratory exchange; only a thin aver af albumen and the chorion lic between them. This degree of apposition may allow some translocation of calcium. Certainly, as the arca vasculosa of the vitelling increases, before the allantois extrudes between it and the chorion, some calcium may be acquired from the shell, However, at this time the demand is small and the vascularised surface of the vitelline is active- ly interacting with the yolk for the general nutri- tion of the embryo and apparently selectively removes calcium (See Packard ct al., 1984a) Packurd and Packard, 1984). The development of the allantois prior to (or simultancously with) the increase in demand for oxygen and calcium ensures support for the growth phase. The role of respiratory exchange in the trans- MEMOIRS OF THE OUFENSLAND MUSEUM location of calcium in reptilian eggs has not been demonstrated but gas exchange plays an impor- lant role in translocation during avian develop- ment (Crooks and Simkiss, 1974), Packard and Packard (1984) provide critical speculation about the function of the chorioallantois in trans- location of calcium but no experimental data are available. During incubation, embryos of domestic fowl (Gallus domesticus) store calcium derived from the shell in the yolk fo such an extent that by hatching the yolk contains more calcium (50-75 %e) and ut a higher concentration than at oviposi- tion (Johnston and Comar, 1955; Romanoff, 1967; Simkiss, 1967; Crooks and Simkiss. 1974), In contrast, the amount of calcium in yolks of Chelydra serpentina (Packard et al, 1984a), Chrvsenrys picra (Packard and Packatd, 1986) and Chelodina longicollis decrease during incubation, The decrease in concentration of cal- ciim in the yolk indicates that these embryos selectively remove calcium from the yolk, Sim- kiss (1967) reported a decline in the amount of calcium in yolks of Curetta caretia during in- cubation, although the temporal changes in the quantity and concentration were not determined: he alsa teported that the *ealcium in the ceg contents. increases rapidly in the latter part of incubation and is five times greater at hatching than in the fresh egg” (Simkiss, 1967, 9.229). Barhier, Simkiss (1962) demonstrated a fourfold increase of calcium within the ege of Dermoche- lvs coriacea. Cleurly these oviparous reptiles follow a similar pattern of utilization of calcium that is quite different from the pattern followed by birds. Although all the caletum for embryogenesis is available from the yolk and shell, the ultimate source of calcium ts (he female who seeretes both of these structures. Very litthe work has focused on this aspect of the overall role of calcium in reproduction. Data derived from a number of different reptiles (see review by Sim- kiss, 1967) provides only a partial picture. Reproduction by an oviparous reptile can be subdivided into two phases. The first phase in- eludes the preparation of the follicles prior to ovulation, This may. require only a month or so in some lizards or as long as two years ar more (e.g. Vipera berus). The second phase is the deposition of the shell around the ovulated ova. The process of vitellogenesis provides the yolk proteins containing calcium aver a period of time thalas typically longer than the period required for deposition of the shell. The former does not GROWTH AND CALCIUM METABOLISM OF EMBRYONIC TURTLES put as much stress on the calcium budget of the female as does the latter. The proximal source of calcium for vitellogenesis and shell deposition is bone (Dessauer and Fox, 1959) but the ultimate source is the diet of the female. ACKNOWLEDGEMENTS Tortoises were collected under permit A213 issued by the New South Wales National Parks and Wildlife Service. The research was funded by a grant (82/n/1) from the Australian Nuclear Science and Technology Organisation (formerly the AAEC). Lynda Bridges assisted with the husbandry of eggs and with analysis of samples. The Department of Botany, UNE, graciously allowed use of their Atomic Absorption Spectrophotometer. R. Hobbs gave advice on the statistical analysis. G. Rawlins aided computer analysis. G.C. and M.J. Packard freely discussed many ideas oncalcium metabolism in embryonic reptiles. C. Limpus, H. Heatwole and R. Harden critically read the manuscript; two anonymous reviewers made useful suggestions. LITERATURE CITED ACKERMAN, R.A. 1980. Physiological and ecologi- cal aspects of gas exchange by sea turtle eggs. Amer, Zool. 20: 575-583. 1981. Growth and gas exchange of embryonic sea turtles (Chelonia, Caretta). Copeia 1981 (4): 757-765. BULL, J.J.. VOGT, R.C. AND BULMER, M.G. 1982. Heritability of sex ratio in turtles with environmental sex determination. Evolution 36: 333-341, BUSTARD, H.R., SIMKISS, K. AND JENKINS, N.K. 1969. Some analyses of artificially in- cubated eggs and hatchlings of green and logger- head sea turtles. J. Zool. Lond. 158: 311-315. COGGER, H.G., CAMERON, E.E. AND COGGER, H.M. 1983. ‘Amphibia and Reptilia. Vol 1. Zoological catalogue of Australia.’ (Australian Government Printing Service: Canberra). CROOKS, R.J. AND SIMKISS, K. 1974. Respiratory acidosis and eggshell resorption by the chick embryo. J. Exp. Biol. 61: 197-202. CUNNINGHAM, B. 1922 . Some phases in the development of Chrysemys cinerea. J.Elisha Mitchel Soc.38:5 1-73. DESSAUR, H.C. AND FOX, W. 1959. Changes in ovarian follicle composition with plasma levels 435 of snakes during estrus. Am. J. Physiol. 197: 360-366. DIXON, W.J., BROWN, M.B., ENGLMAN, L., FANE, J.W., HILL, M.A., JENNRICH, R.I. AND TOPOREK, J.D. 1981. ‘BMDP Statistical Software.’ (Univ. Calif. Press: Davis). EWERT, M.A. AND LEGLER, J.M. 1978. Hormonal induction of oviposition in turtles. Her- petologica 34: 314-318. GOODE, J. 1967. ‘Freshwater Tortoises of Australia and New Guinea.’ (Lansdowne Press: Mel- bourne). GOODE, J. AND RUSSELL, J. 1968. Incubation of eggs of three species of chelid tortoises, and notes on their embryological development. Aust. J. Zool. 16: 749-761. HARRINGTON, K.H. 1933. Breeding habits of Chelodina longicollis. S.A. Natur. 15: 25-27. JOHNSTON, P.M. AND COMAR, C.L, 1955. Dis- tribution and contribution of calcium from the albumen, yolk and shell to the developing chick embryo. Am. J. Physiol. 183: 365-370. JONES, M.E. 1983. Some aspects of the developmen- tal physiology of eggs of the bearded dragon, Amphibolurus barbatus (Cuvier). (Unpublished Honours thesis, Univ. New England, Armidale, N.S.W). KREFFT, G. 1865. Vertebrata of the lower Murray and Darling. Trans. Phil. Soc. N.S.W: 24-25. LEGLER, J.M. 1985 Australian Chelid turtles: reproductive patterns in wide-ranging taxa. pp.53-61.J/n Grigg, G., Shine, R. and Ehmann, H. (eds), ‘Biology of Australasian frogs and reptiles.” (Surrey Beatty & Sons: Royal Soc. N.S.W : Chipping Norton). _ LUCUS, W.G. AND LE SOUEF, W.H. 1909. Chelonia. /n ‘The animals of Australia. Mam- mals, reptiles and amphibians.’ (Whitcombe and Tombs; Melbourne). LYNN, W.G. AND VON BRAUN, T. 1945. Studies on the oxygen consumption and water metaboloism of turtle embryos. Biol. Bull. 88:112-125. MCCOOEY, H.J. 1887. Notes on the method adopted by the female of the common freshwater tortoise, Chelodina longicollis, in the excavation of the burrows in which her eggs are to be deposited. Proc. Linn. Soc. N.S.W 12: 107-108. MILLER, J.D. 1982. Embryology of marine turtles. (Unpublished Ph.D. thesis, Univ. New England, Armidale, N.S.W.). MILLER,J.D. 1985. Embryology of marine turtles. pp 269-328. Jn Gans,C., Billet, F. and Maderson, P.(eds), ‘Biology of the Reptilia Vol.14. 436 Development A’. (John Wiley and Sons: New- York). MORRIS, K.A., PACKARD, G.C. BOARDMAN, T.J., PAUSKTIS, G.L. AND PACKARD, M.J, 1983. Effect of the hydric environment on growth of embryonic snapping turtles (Chelydra serpentina). Herpetologica 39: 272-285. PACKARD, M.J. AND PACKARD, G.C. 1984. Comparative aspects of calcium metabolism in embryonic reptiles and birds. pp. 155-179. In Semour, R, (ed.), ‘Respiration and metabolism of embryonic vertebrates’. (D.W. Junk: The Hague). 1986. Effect of water balance on growth and cal- cium metabolism of embryonic painted turtles, Chrysemys picta. Physiol. Zool. 59(4): 398-405. PACKARD, M.J., PACKARD, G.C., BOARDMAN, TJ., MORRIS, K.A, AND SHUMAN, R.D. 1983. Influence of water exchanges by flexible- shelled eggs of painted turtles Chrysemys picta on metabolism and growth of embryos. Physiol. Zool. 56: 217-230. PACKARD, M.J., PACKARD, G.C. AND GUTZKE, W.H.N. 1984b. Calcium metabolism in the embryos of the oviparous snake Coluber con- strictor, J, exp. Biol. 110: 99-112. PACKARD, M.J., PACKARD, G.C., MILLER, J.D., JONES, M.E. AND GUTZKE, W.H. 1985. Cal- cium mobilization, water balance and growth jn embryos of the agamid lizard Amphibolurus bar- batus, J. exp. Zool, 235; 349-357. PACKARD, M.J., SHORT, T.M., PACKARD, G.C. MEMOIRS OF THE QUEENSLAND MUSEUM AND GORELL, T.A. 1984a. Sources of calcium for embryonic development in eggs of the snap- ping turtle Chelydra serpentina. J. exp. Zool. 230; 81- 87. PARMENTER, C.J. 1976. The natural history of the Australian freshwater turtle Chelodina longicol- lis Shaw (Testudinata, Chelidae). (Unpublished Ph.D. thesis, Univ. New England, N.S.W.), 1985. Reproduction and survivorship of Chelodina longicollis (Testudinata:Chelidae). pp. 53-61. Jn Grigg, G., Shine, R. and Ehmann, H. (eds), “Biology of Australasian frogs and reptiles’. (Surrey Beatty & Sons: Royal Soc. N.S.W.: Chipping Norton). ROMANOFF, A. 1967. ‘Biochemistry of the avian embryo.’ (Jahn Wiley: New York). SHAW, G. 1794. The zoology of New Holland. Il- lustrated by J. Sowerby. (London). SIMKISS, K. 1962. The sources of calcium for the ossification of the embryo of the giant leathery turtle. Comp. Biochem. Physiol. 7: 71-99. 1967. ‘Calcium in reproductive physiology.’ (Rein- hold: New York). VESTIENS, W.J.M. 1969. Nesting, egg-laying and hatching of the snake-necked tortoise at Canber- ta, A.C.T. Aust. Zoologist 15: 141-149. YNTEMA, C.L. 1968. A series of stages in the embryonic development of Chelydra serpen- tina. J, Morph, 125: 219-251, ZAR, J.R. 1974. ‘Biostatistical analysis’. (Prentice- Hall: Englewood Cliffs, N.J.). NEW CRANIAL ELEMENTS OF A GIANT VARANID FROM QUEENSLAND R, E. MOLNAR Molnar, R.E. 1990 09 20: New cranial elements of a giant varanid from Queensland. Memoirs of the Queensland Museum 29(2): 437-444. Brisbane. ISSN 0079-8835, Two massive varanid frontals and matching parietal from the eastern Darling Downs (Queensland) Pleistocene derive from a large varanid, probably Megalania prisca . The frontal is characterised by a Sagittal crest and low ornamentation on the dorsal surface. The parictal has relatively longer lateral and supratemporal processes than in modern varanids, and a relatively smaller area roofing the braincase. Confluent contacts on the [frontal for the prefrontal and postfrontal-postorbital and the encroachment of the supratemporal fossa onto the dorsal surface of the parietal suggest that M. prisca was a more derived varanid than any now existing in Australia. The frontal appears quite thick and the endocranial cavity small; these both are probably allometric effects. () Queensland, Australia, Pleistocene, Varanidae, Megalania, sagittal crest R. £. Malnar, Queensland Museum, PO Box 300, South Brisbane, Queensland 4101, Australia; 19 December, 1988, The giant varanid, Megalania prisca (Qwen, 1859), is among the most distinctive Australian fossil tetrapods, as well as the largest known terrestrial lepidosaur. It is known from the remains of one skeleton, or possibly two, (Rich, 1985) from the eastern Darling Downs of Queensland, and isolated remains from there and other localities in the eastern half of Australia (Lydekker, 1888; Hecht, 1975). Fossils of M, prisca are known only from the Pleistocene. Smaller vertebrae attributed to Megalania sp. are known from the Pliocene of Chinchilla, western Darling Downs, Queensland (Hecht, 1975), Recently discovered or recognised material sheds new light, and raises new ques- tions, tegarding this animal. The material described here suggests that the skull of M. pris- ca Was unusual in its construction. Specimen numbers prefixed with ‘J’ or ‘F’ are held in the Queensland Museum, that prefixed with “Y' in the lan Sobbe collection and that prefixed by ‘BMNH’ in the British Museum (Natural History). DESCRIPTION In about 1984, Mr Ian Sobbe recovered an unusual bone (F16783) from the Pleistocene deposits at Pearson’s Locality, King Creek, east- erm Darling Downs, Queensland. In August of 1985 a second, worn specimen (V0033), was recovered, also by Mr Sabbe, from the ‘Sutton Bed’, King Creek west of Clifton. Both elements are left frontals, approximately equal in size (Table 1), During preparation of this paper, Mr Sobbe donated a large lacertilian right parietal (F16792), collected from King Creek about ten years ago. In form the frontals are basically like those of Varanus salyadorii (Figs 1,2). In dorsal view the element resembles a reversed L, the stem repre- senting the body of the frontal and the lower bar, the lateral process that contacts the fused postfrontal-postorbital distally and the parietal posteriorly. The nasal contact is like that of Varanus varius, with the dorsal surface of the frontal projecting anteriorly along the midline. This would give the frontonasal contact a V- shape, with the apex anteriorly directed. A shal- low horizontal flange dorsally limits the prefrontal contact: there is no such flange in either V. salvadorii or Y. varius, The lateral process of the frontal is anteroposteriorly nar- Midline length Maximum length Maximum width Minimum width at orbit Maximum thickness TABLE 1. Giant varanid frontals (mm). 438 MEMOIRS OF THE QUEENSLAND MUSEUM FIG. 1. Left frontals of a large varanid, probably Megalania prisca . Unworn frontal (F16783) in dorsal (A) and lateral (C) views. Worn frontal (V0033) in dorsal (B) and lateral (D) views. Scale bar lcm. GIANT VARANID CRANIAL KILEMENTS rawer, relutive lo ils length, than in the modern species examined (in addition to those noted, Voranus eouldii and Varanus mertenst were seen) The contact surface for the prefrontal meets thal forthe postfrontal-postorbital, with no indication of a free orbital margin on the frontal, as (here is in modern varanid skulls, F16783 is shorter than the frontal of V, selvaderti, more similar in its proportions to that of Vi varius. However the orbital emargination is placed rela- tively further back, Ventrally, the frontals are similar in form to those of the modern species. The subolfactory processes are well developed and extend to the midline. hence would be in contact medially, A small foramen penctrates the processes along the midline contact, The Pleistocene frontals differ from those of moder varanids in three particulars. A distinct sugittal crest is present (Fig. 1C), which ter- minates posteriorly in front of the parictal con- tact. This indicates that the crest was limited to the frontal, Lateral to the crest the dorsal face of the frontal is ornamented with low, rounded, mostly parallel ridges (Fig, 1A), Low parallel ridges are often found on vertebrae of M. prisea, abutting articular surfaces, but | know of no other varanid with such ornament or, for that matter, any other reptile, However, a low dorsal ridge at the frontal symphysis is apparentin the skulls of several varanids, including Varants indicus (HILO 7 and T1018), Varanus spencer’ (J42U22 and J47915), Veranus tristis (JSO724) and V. varius (J15361, 116156 and J47065), (is ab- sent in V, gouldér (J16135) and ¥. ddertensi (146280). Finally, and most obviously, the King Creck frontal is massive. At the parietal contact the depth of the fronpil is 1/4 its midline length; in V. Salvadori this tatio is less than 1/10, The parietal (/'16792) is worn, although less then V0033. However the anierior suture pattern is lost. [t is u creseentic element (Fig. 2). Ap- parently both parictals were fused medially, as In modern varanids, but this specimen is broken along the midline. Anteriorly the lateral process prajeets perpendicular to the longitudinal axis, and posteriorly a longer supratemporal process projects posterolaterally at an angle of 35 degrees to the longitudinal axis. In proportions the parietal js basically similar to those of V. salvadorii and V. varius, but differs in hitving a proportionately shorter body. In conjunction with this (he supratemporal process of the parie- tal, Which distally contacts the paroccipital process, is relutively longer, Al the untenor ice mination of the mediolaterally compressed 459 Suprautemporal process a praminent horizontal pit penetrates lhe body of the parietal. Such a pit is also present in wl least VL vertus, Although in form basically similar to those of modern varanids, this parictal differs in seven oints. A large parictal foramen is present, }om in diameter, but set less (han Jem back from the frontal contact. Thus it is more anterior than in the modern varanids seen, The supratemporal process is horizontal and nat declined posteriorly as in living varanids, The dorsal margin of this process is distinctly elevated from the dorsal face of the body. This, together with the extension of the supratemporal fenesira over the top of the parictal to the midline, suggests powertul development of the jaw adductors. Correlated with these differences, the flat dorsal face of the parietal extends posteriorly from the frontal con- tact only to the parictal foramen, unlike the modern yaranids available where this surface extends trom frontal margin to occipital face. The mediolaterally compressed supratemporal process bears a distinct medial shelf along its entire length but less prominent distally. Such a shelf was not seen on any of the modern varanid material available. Ventrally the area of the parictal roofing the endocranial cavity is strong- ly reduced compared lo the condition in V, varivs (J47065) and V, salvadorit (J14498), The linear dimensions of the endocranial roof are twice those of J47065 (¥, varies ), but the lengths of the lateral and supratemporal processes are three to four times those of that specimen. This reduc- tion of amount of the parretal forming the en- docranial roof is reflected in the extension of the parietal laterally beyond the lateral walls of the braincasc, The broken face of the parietal shows a depth of 2.5em of which the top 0,5em is compact bone und most of the remainder is spongy bone, A thin (0.20m) layer af compact bone farms the ventral surface. SCALING The large size of the King, Creek varanicd cranial material leads to questions of ils scaling. This is relevant to the following taxonumie dis- cussion and interesting in its own right, ‘l'we issues will be raised: whether the apparent thick vess of the frontals results only from their large size and the relative size of the endocranial cavity, Could (he appearance of thickness of the fron- iuls and parictal fram King Creek simply he tre result of sealing’? McMuhaon's elastic scaling 440 MEMOIRS OF THE QUEENSLAND MUSEUM FIG, 2. Left frontals and right parietal of a large varanid, probably Megalania prisca . Worn frontal (V0033) SIANT VARANID CRANIAL ELEMENTS (McMahon, 1973; McMahon and Bonner, 1983), for which there is some evidence when applied to the anatomical analogues of columns (Hamley, 1990; McMahon, 1975), recognises that transverse linear dimensions scale as the 3/2 power of longitudinal lincar dimensions. Mc- Mahon assumes that the orthogonal transverse dimensions will be equal, that is D° = L° where D js the transverse dimension and L the length. However if the two orthogonal transverse dimensions Were nol equal, as is here the case, then it would follow from the derivation that D1! x D2 = L’, In this case if D1 is the width of the frontal and D2 its thickness, we wish ta find the value of D2 expected from knowing L and D1, if the large frontal Were lo have the same resis- tance to bending as the frontals of smaller modern varanids (here V. salvadorii and V. varius ), This analysis treats the frontals asa plate principally resisting bending stresses imposed in biting, and transmitting the forces then im- pressed to the parietal and occipital regions of the skull. It also assumes that the frontals can be regarded as simple plates with resistance to bending proportional only to the cross-sectional area. [t ignores any possible role in stiffening the frontals of the subolfactory pracesses, which in varanids make the posterior part of the frontals into a flattened tube. It also ignores the role of the sagittal erest of the King Creek frontals. However these effects will be ignored here for wo teasons, first they are technically difficult to treal, and second both considerations would act {o increase the resistance to bending of the fron- tals. Thus consideration of both factors would lend to decrease the estimate of thickness for scaled up frontals. [ wish to determine il the King Creek frontals are thicker than expected from arguments of scaling and hence wish to err (if at all) on the side of estimating too thick rather than too thin. Working with the dimensions of the two avail- able skulls, 314498 (V. salvadorii ) and J47065 (V. varius ), it appears. that the thickness of the King Creck frontals is such as would be predicted from elastic scaling. Scaling up the skull of V. salvadorii would give a frontal aboul 20 mm thick, which is close to the thickness of the F16783 (19.8 mm), while scaling up that of VW. varius would predict frontals even thicker, about 45 mm thick. In view of the approxima- tins used in making these calculations, this is viewed as reasonable agreement (i.e, within one order of magnitude), providing no evidence that the frontals from King Creek are unusually thick, 44] For purposes of an order of magnitude calcula- tion the endocranial cast of a varanid may be approximated by a six-faced irregular but bilaterally symmetric polyhedron thai ap- proximates the endocranial cavity. The ventral surface of the parietals forms the upper face of the polyhedron. This polyhedron was defined from examination of the figures 1), 17 and 18 of Starck (1979) and of a skull of ¥. varius (J1656) that retains some of the soft connective lissuc walling the endocranial cavity. The figures of Starck (1979) indicate that in V, sa/wator at least virtually all of the brain is included within this volume, although not filling it, The similarity in form of the parietals of V, varius to that from King Creck, suggests that this polyhedron may be used to approximate the endocranial cavity of that form as well. Because the same polyhedron is used in both instances, if the ratio of the areas of the corresponding face of each of the two polyhedra is known, the ratio of the volumes can be calculated. The endocranial surface of the parietals of F16792 is about 5 times greater than thal of ¥, varius (47065). Using the relationship that volume is proportional to the 3/2 power of area, this gives a ratia of volumes of about 11 to 1, This result gives no indication that the ef- docranial cavity of the King Creek varanid was relatively smaller (han in the ntudern V. varius in spite of the fact that relatively less of the ventral face of the parietal roofs the endocranial cavity in the fossil form than in the living one. In modern vatanids the brain is substantially smaller than the endocranial cavity aiid $0 does net closely conform 16 the endocranial surfaces (Starck, 1979, figs 17 and 18). Thus no inferen- ces regarding relative brain size will be essayed here. TAXONOMIC IDENTIFICATION Varanoid frontals are characterised by the structure of the subolfactory processes (Pregill et al, 1986), which are well developed and come in contact medially. Thus the King Creek fron- fals are varanoid, Pregill e1 al. (1986) cite a mediolaterally compressed supratemporal process of the parictal as characteristic of varantds, hence this parietal derives from a Varanid, The parietal matches in the size the two fron- tals, Suggesting thal boll elements derive from the same species. Unfortunately the anterior su- ture pattern on the parietal has been worn, so a direct Comparison of their forms ts Hor possible, However, some similarity is evident. The paric- tal conluct face of the frontal is stepped. its medial centimetre situated slightly forward of the lateral portion, The anterior face of the parie- tal shows a corresponding step, with its medial centiinetre set slightly fonward. The dorsoventral thickness of the lateral process (2.1enm) matches that of the frontal (2.0cm), so that the two ele- ments could have derived from the same in- dividual, The parietal is also consistent in size with the occipital segment of Owen (1880), ul- though thal comes from Gowrie, not King Creek. The similur thickness and form of the frontal- purictal contact indicates that the frontals and parietal probably derive from the same species, Further evidence for common derivation could be given if the parictal had a similar pattern of ornament. Unfortunately there is no indication on the dorsal face of the parietal of the unique sculpture or the sagittal crest seen on the frontal, Because the dorsal face of the parictal has been worn und the sculpture of the frontal is very subdued at its posterior margin, sculpture may have been present and lost from wear. A sagittal crest, however, should have been sulficiently murked to have survived this degree of wear, Were any crest present on the parietal, Presumably this material pertains to Megalania prisca. The holotype of Mf. prisca vonsists of wo and half dorsal vertebrae (BMNH 32908a, 32908b and 32908c: Lydekker, 1888), and so reference to this species must depend on companson with associated material. No fron- tals OF parietals of M, prisce were previously known (Rich, 1985, figure on p. 154). However both King Creek frontals were found in associa- tion with material of M. prisca, vertebrae and tecth at Pearson's locality and yvertebrac and a ibis at Sutton'’s bed, Bul much other tetrapod material has also been found at these localities, $0 no firm conclusion may be drawn from this. However M. prisca is the only large varanid known from Pleistocene Australia, and since these skul] roof clements derive from a large varamd, relerence to M. prisea is reasonable. Further conclusions may be drawn regarding the evuluyjonary position of the beust from which these elements derived, A close approach of the prefrontal to the pastfrontal above the orhit is a derived feature (Pregill et al., 1986). Thus con- fluent contact surfaces for the prefrontal and postfrontal-pastorbital is a derived feuture. So these frontals representa more denved eandition than any surviving Australiun varanids cx- MEMOIRS OF THE QUEENSLAND MUSEUM amined. "The panetal appears less derived, in that it retains the parietal foramen (Pregill et al., 1986), and a large one al thal, However [ would suggest. by analogy with the evolution of the cranial roof in large theropod dinosaurs (Walker, 1964), that reduction of the flat dorsal surface of the parictal by encroachment of the supratem- poral fenestrae is also a derived feature in varanuids. This suggests that Megalania reprte- sents a more derived varanid than now exists in Australia. DISCUSSION WITH SPECULATIONS The frontals and parietal from King Creck appear obviously thicker than the maxillae and denlary attributed to Megalania prisca - Either the skull roaf was considerably thicker than the trophic apparatus, or the roof elements derive from an individual larger than those from Which the jaws are known, or there Was variation, such as Sexual dimorphism, in thickness of the. skull elements, A dentary, F6562, from an animal presumably approximately equal in size to that from which the cranial roof elements derive, is al the base of the teeth (where it is thickest) only 60% as thick us the frontal F16783, A maxilla (F12370), also apparently from an animal of this size, is equally thin compared to the frontal. This is not the case in the skulls of living varanids, where the frontal and dentary are approximately equally thick, The only other ammotes known to me with the skull roof significantly more massive than the trophic apparatus are the herbivorous pachycephalosaurian dinosaurs (Maryanska and Osmolska, 1974), ‘These are quile different in cranial form. The tooth form of M. prisca tim- plics that it was most likely either a predator or scavenger. In neither case is the braincase ex- pected to be more robust than the trophic ap- paralus: such construction is unknown among living predators and scavengers, The frontal and parictal appear to be ap- proximately of the size expected to match the known maxillac and dentary, to judge from com- parison with living varanids. Unless its cranial proportions were very different from modern Varanids they would nol dernve from an in- dividual 3% larger than those from which the jaws come, So the possibility that they derive from individuals of different sizes seems remote. Possibly one sex. presumably the male, had a more robust skull, orat least skull roof, than the other, There is at present no Way of testing this GIANT VARANID CRANIAL ELEMENTS possibility. Sexual dimorphism is unknown in living varanids, but the environmental cir- cumstances of Megalania were doubtless dif- ferent and sexual dimorphism is known in some mammalian top carnivores, e.g. lions. The frontal crest suggests habits different from those of living varanids. It may have been a weapon, or display (species recognition) struc- ture. M. prisca would presumably have been a top carnivore of the Australian Pleistocene (cf. Rich, 1985) and thus, at least in some respects, analogous to the large theropod dinosaurs of the Mesozoic. Large theropods bore cranial orna- ment, usually horns or crests (Molnar, 1977; Kurzanov, 1976; Welles, 1984; Bonaparte, 1985), thus it is not unreasonable to suggest that M. prisca too might have had cranial ornament. The frontal crest may have been used in head to head shoving contests, as among the marine iguana Amblyrhynchus cristatus (Carpenter, 1978). Living varanids are not known to engage in such contests (Stamps, 1977; Carpenter, 1978), but the circumstances of the life of M. prisca, as a large terrestrial top carnivore were unlike those of modern varanids. A different speculative significance of the crest has also been suggested. It is well known that aquatic lizards (including some varanids) usually show lateral compression of the tail and sometimes the trunk, Furthermore many show some development of a dorsal ridge or crest along the back and tail, as in Hydrosaurus am- boinensis (although rarely so prominent). In some species of Basiliscus these crests are com- plemented by a crest on the skull roof. Possibly the sagittal crest of the King Creek frontal indi- cates aquatic or amphibious habits. A cranial crest is found in some arboreal lizards, such as Corythophanes. We secm safe in presuming, however, that the giant King Creek varanid was not arboreal. If the King Creek varanid was amphibious or aquatic, one might expect that crocodiles would have been rare in its habitat, Indeed, crocodile remains are rare (Pearson’s locality) or absent (Sutton’s bed) from the localities and levels at which the frontals were found (Sobbe, pers. comm., 1988; also cf. Bartholomat, 1976). This suggests that competition for the niche of a large aquatic predator would have been weak or ab- sent. It also suggests that predation on a large aquatic lizard would have been weak or absent. CONCLUSIONS Two frontals and a parietal from King Creek, eastern Darling Downs, Queensland, indicate the presence of a giant varanid. This form, presumably Megalania , was more derived than living varanids in two features: the contact of the articular surfaces for prefrontal and postfrontal- postorbital and the encroachment of the supratemporal fenestra over the top of the parie- tals. Both the appearance of unusually thick fron- tals and of a relatively small endocranial cavity scem to result from scaling effects. ACKNOWLEDGEMENTS The persistence and acute eye of Mr Ian Sobbe in finding the cranial material and his kindness in donating it made this paper possible. Two of the speculations discussed in above - sexual dimorphism and aquatic habitus - were sug- gested by Tony Thulborn and Greg Czechura, respectively. They are not, of course, responsible for my treatment of their suggestions. Valuable assistance was also given by Drs M. Borsuk- Bialynicka (Warsaw), R. Estes and G. K. Pregill (both in San Diego). LITERATURE CITED BARTHOLOMAI, A. 1976. Notes on the fos- siliferous Pleistocene fluviatile deposits of the eastern Darling Downs. Bull. Miner. Resourc. Geol. Geophys. Aust. 166: 153- 4, BONAPARTE, J. F. 1985. A horned Cretaceous car- nosaur from Patagonia. Nat. Geog. Research 1: 149-51. CARPENTER, C. C. 1978. Ritualistic social be- haviors in lizards, pp. 253-67. Jn Greenberg, N. and MacLean, P. D. (eds), ‘Behavior and Neurology of Lizards.’ (National Institute of Mental Health: Rockville). HAMLEY, T. 1990. Functions of the tail in bipedalism of lizards and dinosaurs. Mem. Qd Mus, 28: 153-8. HECHT, M. K. 1975. The morphology and relation- ships of the largest known terrestrial lizard, Megalania prisca Owen, from the Pleistocene of Australia, Proc, R. Soc. Vict. 87: 239-49. KURZANOV, S. M. 1976. Novie pozdnemelovoi kar- nozavr iz Nogon-Tsava, Mongoliia. Sovmest, Sovet.-Mongol, Paleont.Eksped., Trudy 3: 93- 104. LYDEKKER, R. 1888. ‘Catalogue of the Fossil Rep- tilia and Amphibia in the British Museum 444 (Natural History), Cromwell Road, S.W. Part 1.’ (British Museum: London). 309pp. MARYANSKA, T. AND OSMOLSKA, H. 1974. Pachycephalosauria, a new suborder of ornithis- chian dinosaurs. Palaeont. pol. 30: 45-102. MCMAHON, T. A. 1973. Size and shape in biology. Science 179: 1201-4. 1975, Allometry and biomechanics: limb bones in adult ungulates. Amer. Nat. 109: 547-63. MCMAHON, T. A. AND BONNER, J. T. 1983. ‘On Size and Life.’ (Scientific American Library: New York). 255pp. MOLNAR, R. E. 1977. Analogies in the evolution of combat and display structures in ornithopods and ungulates. Evol. Theory 3: 165-90. OWEN, R. 1859. Description of some remains of a gigantic land-lizard (Megalania prisca *, Owen) from Australia. Phil. Trans. R. Soc. London 149: 43-8. 1880. Description of some remains of the gigantic land-lizard (Megalania prisca ), from Australia.- Part II. Phil. Trans. R. Soc. London 171: 1037- 50. PREGILL, G. K., GAUTHIER, J. A. AND GREENE, H. W. 1986. The evolution of helodermatid MEMOIRS OF THE QUEENSLAND MUSEUM squamates, with description of a new taxon and an overview of Varanoidea, Trans. San Diego Soc. Nat. Hist. 21: 167-202. RICH, T.H. 1985. Megalania prisca the giant goanna. pp. 152-5. In Rich, P. V. and van Tets, G. F. (eds), ‘Kadimakara’. (Pioneer Design: Mel- bourne). STAMPS, J. A. 1977. Social behavior and spacing patterns in lizards. pp. 265-334. Jn Gans, C. and Tinkle, D. W. (eds), ‘Biology of the Reptilia, 7, Ecology and Behaviour A’. (Academic Press: London). STARCK, D. 1979. Cranio-cerebral relations in recent reptiles. pp. 1-38. Jn Gans, C., Northcutt, R. G. and Ulinski, P. (eds), ‘Biology of the Reptilia. Vol. 9, Neurology A.’ (Academic Press: Lon- don). WALKER, A. D. 1964. Triassic reptiles from the Elgin area: Ornithosuchus and the origin of car- nosaurs, Phil. Trans. R. Soc, London, B 248: 53-134. WELLES, S. P. 1984. Dilophosaurus wetherilli (Dinosauria, Theropoda) osteology and com- parisons. Palaeontographica A 185: 85-180. PUTATIVE LOWER CRETACEOUS AUSTRALIAN LIZARD JAW LIKELY A FISH R,E, MOLNAR AND G,V, CZECHURA Molnar, R.E. and Czechura, G.V. 1990 09 20; Putative Lower Cretaceous Australian lizard jaw likely a fish. Memoirs of the Queensland Museum 29(2): 445-447. Brisbane. ISSN 0078-8835. A jaw fragment from Lower Cretaceous beds of Victoria has been identified as that of a Jacertilian. If correct this would be the oldest known lacertilian cranial material from Australia. The acrodont, transversely broadened teeth of uniform size lack cusps and are well separated from one another in the tooth row. We have not been able to match these features in any. Mesozoic replile. Thus we conclude that the jaw fragment probably derives froma large teleost fish. Early Cretaceous, Victoria, Australia, Koonwarra, teeth, lizard. R.E. Malnar, and G.V. Czeckhura, Queensland Museum, PO Box 300, South Brisbane, Queensland 4101, Australia; 20 February, 1989, A fragmentary jaw from the Lower Cretaceous of Victoria has been labelled as lizard, The jaw was collected by a Field Museum of Natural History field party, probably in 1964. According to a letter fram W, Turnbull (Chicago) to J. Warren (Melbourne), the locality is uncertain. Collecting at that time was carried out at the a | yo : FIG, 1. The jaw fragment (PR1425) from the early Cretaceous of Victoria labelled as lizard. The leeth are visible toward the lop: the bevelled appearance may be seen of the second and third from the right: the first and fourth are clearly broken. Scale in mm. Koonwarra pond deposits and at coastal deposits at Cape Paterson, Comparison of the matrix in which the fragment was found with samples from both Koonwarra and Cape Paterson was inconclusive. An impressive fauna of fishes (Waldman, 1971), mostly teleosts, but with some dipnoans and invertebrates (Jell and Duncan, 1986), has been recovered from Koonwarra. These rocks are assigned to the Strzelecki Gr., af Valagianian to Aptian age (approximately 113- 138 million years ago: Dettman, 1986). The Cape Paterson deposits also are the Strzelecki Gr., and have been studied extensively by T. and P. Rich. They have yielded a fauna of fish, turtles, dinosaurs (Rich and Rich, 1989) and a single humerus probably from a lizard (Molnar, 1980). If the identification of the jaw is correct, it is the oldest known Australian lizard material iden- tifiable to a level below suborder. Molnar’s (1980) likely lizard humerus - from the Strzelecki Gr. at Eagles Nest, on the Gippsland coast - lacked the articular ends and thus ts nat identifiable more precisely than as ‘lacertilian’. The oldest Australian lacertilian material iden- tifiable to familial level or below is of Miocene age (Estes, 1984; Molnar, 1985, and references cited therein; Covacevich et al., 1990). Since lizards were present in the Jurassic, potentially a very substantial portion of lacertilian history in Australia is completely unknown. Thus study of the Victorian fragment is potentially very sig- nificant for understanding the evolution of Australian lepidosaurs. We here describe this specimen and show that it seems, after all, not to be a lizard but is most likely a teleost fish. (The reference of Molnar (1985) to lepidosaur 44h FIG, 2. Occlusal view of the teeth of PR1425: the gaps between successive leeth can be scen al [he arrow. Seale in mm material fram the Lower Cretaceous. Toolebuc Fm. of Queensland, is also incorrect: this ver- tebra appears to derive from a diminutive archosaur). The specimen is catalogued as Field Museum of Natural History (Chicago) PR1425: it consists ofa fragment of jaw 16mm long, 5mm in maxi- mum thickness and 7mm in maximum depth (Fig. 1). The entire edge opposite the dentigerous margin is broken. The fragment is now em- bedded in transparent resin to reinforce the fragile bone. Six teeth are preserved, three ap- parently complete, and three broken apically, with spaces forfour more present. Each crown is separated from its neighbours by a distinct gap, apparently of uniform width along the series, The blunt teeth are acrodont, and triangular in anterior aspect. They are anteroposteriorly com- pressed (Fig. 2). In Jateral view the basal half of the crown has almost parallel margins, but in the MEMOIRS OF THE QUEENSLAND MUSEUM apical half one margin becomes inclined so as to intersect the other (Fig, 1), giving the crown a bevelled orchisel-like appearance. The teeth are set al a very Slight inclination to the dentigerous margin and are uniform in size. None show cusps, striae, denticles or other such structures, There are no resorption pits or other indication af tooth replacement, The bone of the jaw js not sculptured, and lacks foramina. One face of the jaw fragment is slightly convex, while the other (the embedded face) is concave, so that ihe bone thins away from the dentigerous edge, and ap- parently broadens abruptly to form that margin. It is unfortunately impossible to be certain of this because of the resin, Because of the fragmentary nature of the specimen, it is also impossible to determine if this fragment derives from the upper or lower jaw. The acrodent tooth emplacement contraindi- cates reference ta such Cretaceous reptiles as have thecodont (archosaurs) or pleurodont tecth. Most modern Iepidosaurs have pleurodont tecth: acrodont dentitions are found among agamids and chamaeleonids (Edmund, 1969), Acrodonty is also found in amphisbaenians (Gans, 1960) - which may be eliminated because of their quite different tooth form - and sphenodontians, Tran- versely broadened teeth are found in some sphenodontians (Throckmorton et al., 1981; Fraser, 1986) and trilophosaurs (Gregory, 1945; Robinson, 1957), Trilophosaurs have tricuspate, wedge-shaped, offen dilated crowns, distinct from those of the Victorian fragment. Mostof the few sphenodontians with transversely widened teeth have teeth that noticeably increase in size posteriorly (Throckmorton et al., 1981, fig. 5; Fraser, 1986, fig. 5). The most similar dentition is that of the sphenodontian Eilenodon robustus (Rasmussen and Callison, 1981) from the Upper Jurassic Morrison Fm, of Colorado. Its teeth are transversely broadened pyramids, showing con- siderable wear. They are placed in the jaw very close to one another, without the distinct separa- tion shown in the Victorian fragment. Sphenodontians seem either to lack transversely broadened teeth. or where such teeth are present, to lack teeth that are distinctly separated. Thus neither trilophosaurs. nor sphenodontians seem sufficiently similar to the Victorian fragment for reference. Comparison was carried out with a variety of Australian modern and Miocene (Riversleigh) Jacertilian material. Overseas material, both modern and fossil, was compared using the literature. Lacertilian teeth, when compressed, LIZARD IS LIKELY A FISH are almost always longitudinally rather than transversely compressed: one of the few excep- tions is Polyglyphanodon sternbergi (Gilmore 1942). This longitudinal compression is espe- cially obvious for acrodont lacertilians, such as agamids and chamaeleontids. The lacertilian (and sphenodont) teeth examined by us uniform- ly taper toward the tip and do not show bevelled form of the Victorian crowns (including those of Polyglyphanodon). The only reported bevelled lacertilian crowns are those of the teiid Macro- cephalosaurus ferrugenous (Gilmore, 1943). Only a single crown, however, was considered, with some doubt, to be unworn and this exhibited low cusps. We have been unable to find any convincing match between the crowns of the Victorian frag- ment and the teeth of known Mesozoic or later reptiles. Thus we conclude that this jaw fragment is not demonstrably reptilian: it is clearly not teferable to any of the (few) known Lower Cretaceous Australian reptiles (mostly ar- chosaurs). Presumably it derives from one of the many teleost taxa known from this deposit. ACKNOWLEDGEMENTS Drs James Warren, William Turnbull and Ms Betty Thompson kindly assisted this work. LITERATURE CITED COVACEVICH, J., COUPER, P., MOLNAR, R., YOUNG, B. AND WITTEN, G. 1990. Fossil dragons from Riversleigh: new data on the his- tory of the family Agamidae in Australia. Memoirs of the Queensland Museum 29(2): 339-360. DETTMANN, M.E. 1986. Early Cretaceous palynoflora of subsurface strata correlative with the Koonwarra Fossil Bed, Victoria. Association of Australasian Palaeontologists, Memoir 3: 79- 110. EDMUND, A.G. 1969. Dentition. pp. 117-200. In Gans, C., Bellairs, A. d’A. and Parsons, T.S. (eds), ‘Biology of the Reptilia’. Vol. 1, Morphol- ogy A. (Academic Press: London). ESTES, R. 1984. Fish, amphibians and reptiles from the Etadunna Formation, Miocene of South Australia. Australian Zoologist 21; 335-343. FRASER, N. 1986. New Triassic sphenodontids from south-west England and a review of their clas- 447 sification. Palaeontology, 29; 165-86. GANS, C. 1960. Studies on amphisbaenids (Amphis- baenia, Reptilia). I. A taxonomic revision of the Trogonophinae and a functional interpretation of the amphisbaenid adaptive pattern. Bulletin of the American Museum of Natural History 105: 1-142. GILMORE, C.W. 1942. Osteology of Polyglyphanodon, an Upper Cretaceous lizard from Utah. Proceedings of the United States National Museum 92: 229-65. 1943. Fossil lizards of Mongolia. Bulletin of the American Museum of Natural History 81:361- 84. GREGORY, J.T, 1945. Osteology and relationships of Trilophosaurus. University of Texas Publi- cations 4401: 273-359. JELL, P.A. AND DUNCAN, P.M. 1986. Inver- tebrates, mainly insects, from the freshwater, Lower Cretaceous, Koonwarra Fossil Bed (Kor- rumburra Group), South Gippsland, Victoria. Association of Australasian Palaeontologists, Memoir 3: 111-205. MOLNAR, R.E. 1980. Australian late Mesozoic ter- restrial tetrapods: some implications. Mémoires de la Société géologique de France (ns)139: 131-43. 1985. The history of lepidosaurs in Australia. pp. 155-8. In Grigg, G., Shine, R. and Ehmann, H. (eds), ‘Biology of Australian Frogs and Reptiles’. (Royal Zoological Society of New South Wales: Sydney). RASMUSSEN, T.E. AND CALLISON, G. 1981. A new herbivorous sphenodontid (Rhynch- ocephalia: Reptilia) from the Jurassic of Colarado. Journal of Paleontology 55: 1109-16. RICH, T.H.V. AND RICH, P.V. 1989. Polar dinosaurs and biotas of the Early Cretaceous of southeastern Australia, National Geographic Research 5: 15-33. ROBINSON, P. 1957. An unusual sauropsid denti- tion. Linnean Society’s Journal - Zoology 43: 283-93, THROCKMORTON, G.S., HOPSON, J.A. AND PARKS, P. 1981. A redescription of Toxolophosaurus cloudi Olson, a Lower Cretaceous herbivorous sphenodontid reptile. Journal of Paleontology 55: 586-97. WALDMAN, M. 1971. Fish from the freshwater Cretaceous of Victoria, Australia, with com- ments on the palaeoenvironment. Special Paper Palaeontology 9: 1-124. MEMOIRS OF THE (QUEENSLAND MUSEUM BRISBANE © Queensland Museum PO Box 3300, South Brisbane 4101, Australia Phone 06 7 3840 7555 Fax 06 7 3846 1226 Email qmlib@qm.qld.gov.au Website www.qm.qld.gov.au National Library of Australia card number ISSN 0079-8835 NOTE Papers published in this volume and in all previous volumes of the Afemoirs of the Queensland Museum maybe teproduced for scientific research, individual study or other educational purposes. Properly acknowledged quotations may be made but queries regarding the republication of any papers should be addressed to the Editor in Chief. Copies of the journal can be purchased from the Queensland Museum Shop. A Guide to Authors is displayed at the Queensland Museum web site A Queensland Government Project Typeset at the Queensland Museum 448 SPIDER PREDATORS OF REPTILES AND AMPHIBIA: — Predation of vertebrates by spiders in not uncommon, In Australia, Whisting Spiders (Selcnonpus plumipes, Theraphosidae) have been reported drageing 4 young chicken 50 feet fram its enclosure and then attemping to drag it into a small burrow 1,25 inches in diameter (Chisholm, 1919). Predation of reptiles is less known. Two cases reported from Australia are that of the Whistling Spiders (Selenacos- mia and Selenotypus spp.) feeding on the frog Helioporus centralis (Main and Main, 1956) and the Funnel-web spider (Arrax Jormidabilis) feeding on Hyla (=Litoria) caerulea (McKeown, 1952). Those predators, however, are large power- ful trapdoor spiders that would seize prey on the ground. Predation of vertebrates by web building spiders is also known. Best known are the very strong webs of Golden Orb-Weavers (Nephila) that snare and kill small birds but evidently do. not consume them. McKeown (1952) shows 4 mouse caught and hoisted in the web of 4 Redback Spider (Latrodectus hasseltii). He also reported cases of a Funnel-web spider (A frax robustus) taking a chicken, Water spiders (Dolomedes) taking fish. and web- building spiders laking small native birds, bats, and reptiles, A skink, Lygosoma, had also been taken by a Redback. One account discussed a black snake that had been tied head to tail and killed by an adult Redback spider. The spider's young were evidently feeding on the snake: Twa cases of Redback Spider predation on reptiles have been noted in qurcallections. A female Redback spider had buill.a web in the fold of a blanket left hanging on a clothes line to dry, presumably for some days. On removing the blanket, two Wall Skinks (Cryptobelpharus virgatus) were found dead and partially cansumed in the web. (My nomenclalure for reptiles and amphibians. follows Cogper, 1983).In the second case (Fig. 1), the web of the female Redback had been built close to the ground and it had snared a Verreaux’s Skink (Anomalopus verreauxii). Apparently, the very sticky lower vertical lines of the Redback had trapped the skink and lifted its head high off the ground. The spider then moved repeatedly to the underside of the skink la inflict its bile (arrow shaws blood stained scar), The spider did not consume the lizard but merely immobilised it. Raven and Gallon (1987) suggested that the Redback is an intro- duced spider and hence may rank along with Bufo marinus in ils effects on native vertebrates, A third instance, observed personally, was thatofa Wolf Spider and a young frog (Litoria lesyeyri). | had caught. identified and placed the frag back on the rocks of a creek hed. Immediately, a large female Lycosa lapidosa jumped onto the frog, impaling it on its fangs, The live spider and ils MEMOIRS OF THE QUEENSLAND MUSEUM CMa ‘90 FIG,1, Redback and snared Verreaux's Skink. prey were taken in a vial back to camp where the vial was opened, Only a grey liquid mound remained of the frog, no hard tissue could be felt with a wooden probe. Little more than five minutes had passed since the spider had seized the frog. Literature Cited Butler, W-H_and Majin, B.Y. 1959. Predation on vertebrates by mygalomorph spiders. Western Australian Naturalist 7: 52. Chisholm, J.R. 1999. Spider and chicken, Emu 18; 307, pl, XLVI. Cogger. H.R. 1983, ‘Reptiles and amphibians of Australia.’ 4th edit, (Reed Books; Frenchs Forest). Main, B.Y., and Main, A.R. 1956. Spider predator on a vertebrate. Western Australian Naturalist 5: 139. McKeown, K,C. 1952. ‘Australian spiders. Their lives and habits.’ 2nd edit. (Angus & Robertson: Sydney). Raven, RJ. and Galton, J.A. 1987. The Redback Spider. pp. 306-11, In Covacevich, J., Davie, P., and Pearn, J. (eds), “Toxic plants and animals. A guide for Australia” (Queensland Museum: Brisbane), Robert J, Raven, Queensland Museum, P.O. Box 300, Sourh Brisbane, Queensland. 4101, Avstralia,; 20 August, 1990. TREATMENT OF CLOACAL PROLAPSE IN THE ESTUARINE CROCODILE LYALL NAYLOR Naylor, L. 1990 09 20: Treatment of cloacal prolapse in the Estuarine Crocodile, Memoirs of the Queensland Museum 29(2): 449-451. Brisbane. ISSN 0078-8835. The prolapsed and inflamed genitalia of two captive male specimens of Crocodylus porasus were successfully repaired following sedation of the specimen; cleansing and replacement of the genitalia and surrounding tissues; and suturing of the vent to prevent repetition of the prolapse. ] Treatment, cloacal prolapse, Cracodylus porosus. Lyall Naylor, C/- Wild World, Palm Cove, Queensland 4879, Australia; 18 August, 1988. The Estuarine Crocodile, Crocodylus porosus farms in Australia. One potentially serious prob- is a species which is kept frequently in, and lem with captive crocodiles results from ter- generally adapts well to, captivity. Many _ ritorial conflicts between males, even between specimens are held in zoological collections and individuals who have shared ponds harmonious- FIG. 1. A. Prolapsed and inflamed genitalia of C. porosus. B. Abraded tissue. C. Cleaning method. D. Replacement of genital tissue. 450 TABLE I. Doses of Flaxedil and Valium required to sedate Crocodylus porosus specimens. Drug Dose (mls) ly for long periods. Here, the circumstances lead- ing to cloacal prolapse with subsequent irritation of the genital organs and successful treatment of this problem are reported for two specimens. When two males fight, displaced sexual arousal can expose their genital organs. These can be severely injured, either by abrasive con- tact with concrete (most battles take place in concrete-lined ponds) or by an opponent (Fig. 1A,B). The prolapse can be aggravated by sand and debris. Captive crocodiles bask in favoured MEMOIRS OF THE QUEENSLAND MUSEUM sites that become denuded of grass cover and sand quickly enters the genital area and exacer- bates the problem. Abrasion of the hemipenis, testicles, and surrounding soft tissue results in a serious discomfort. Crocodiles suffering from the combination of prolapse and inflamed, abraded genitalia walk with their hind legs fully extended to help hold the injured tissue above the ground to avoid further discomfort. Even at rest, they favour elevating their hindquarters. REPAIR PROCEDURE The crocodiles were sedated prior to surgery with Valium and Flaxedil-Gallamine (Table 1). These were administered via a syringe attached to asuitably modified 2m aluminium pole. Seda- tives were injected via the neck, hind legs or butt of the tail depending on the animal’s position and disposition. Once the specimens were sedated, their jaws were bound and their eyes covered. FIG. 2. A-C. Suturing of cloacal opening. Note plastic tubing to prevent tearing. D. Completed repair. TREATMENT OF CLOACAL PROLAPSE IN CROCODYLUS POROSUS Then the animals were positioned dorsum down on a hessian mattress. With the specimens ‘safe’ and in position for surgery, Betadine surgical scrub solution (7.5% uv povidone-iodine) was applied liberally to the whole area affected. This was then rinsed with tap water (Fig. 1C). (With hindsight it is now felt that sterile saline — 0.9% sodium chloride solu- tion — would be a more appropriate rinse). To ensure removal of all foreign bodies and cleanse the wounds this procedure was repeated several times. The genitalia and exposed tissues were replaced in the cloaca (Fig. 1D) after the applica- tion of 50gms of socatyl-sulphonamide paste (active constituent formosulphathiazole) to facilitate placement and reduce the risk of infec- tion. The vent of crocodiles has an elliptical shape and this, sutured to reduce its size dramatically, 45] assisted prevention of repeated prolapse until healing was complete. To prevent tearing of sutures small sections of plastic tubing were affixed to entry and exit points (Fig. 2). Finally, after suturing, the whole area was washed with Betadine. The animals were isolated during the recovery period and were not fed until the sutures were removed some six weeks after the repair. Both crocodiles have made uneventful recoveries, though the ability of these animals to breed is doubtful and, as yet, untested. ACKNOWLEDGEMENTS Dr D.A. Johnson, Veterinary Surgeon to Wild World via Cairns, supervised this procedure. I am also indebted to K. Day, M. O’Brien, C. Shaw and S. Wallis for advice and assistance. MEMOIRS OF THE (QUEENSLAND MUSEUM BRISBANE © Queensland Museum PO Box 3300, South Brisbane 4101, Australia Phone 06 7 3840 7555 Fax 06 7 3846 1226 Email qmlib@qm.qld.gov.au Website www.qm.qld.gov.au National Library of Australia card number ISSN 0079-8835 NOTE Papers published in this volume and in all previous volumes of the Afemoirs of the Queensland Museum maybe teproduced for scientific research, individual study or other educational purposes. Properly acknowledged quotations may be made but queries regarding the republication of any papers should be addressed to the Editor in Chief. Copies of the journal can be purchased from the Queensland Museum Shop. A Guide to Authors is displayed at the Queensland Museum web site A Queensland Government Project Typeset at the Queensland Museum DIGIT 1 IN PAD-BEARING GEKKONINE GECKOS:ALTERNATE DESIGNS AND THE POTENTIAL CONSTRAINTS OF PHALANGEAL NUMBER ANTHONY P. RUSSELL AND AARON M. BAUER Russell, A,P. and Bauer, A.M. 1990 09 20: Digit [ in pad-bearing gekkonine geckos: allernate designs and the potential constraints of phalangeal number. Memoirs of the Queensland Museum 29(2); 453-472. Brisbane. ISSN 0079-8835. Sprawling locomotion is typical of lizards and dictales the kinematics of locomotion. While lateral undulation of the body is still an important component in the production of locomotor thrust, the kinematics of the limb joints coupled with the marked asymmetry of foot structure resultin most of the thrust being directed posteriorly and little of it being oriented laterally. The asymmetry of pedal design, however, leaves digit f with only two phalanges. With the independent acquisition, in many lineages, of subdigital adhesive pads in gek- konine geckos there is a potential ‘problem’ jn incorporating a pad into digit I and enabling it io operate effectively, All other digits have three or mare phalanges, but the first lacks the fundamental prerequisites to permit hyperextension of the digit and the deployment ot subdigital setae. Such inherent limitations have resulted in a variety of solutions of this problem and there is preal variance in the structure of digit 1 of pad-bearing gekkonines. Some lineages have reduced the first digit and abandoned it as an effective locomotor device, while others have modified itin a variety of ways in order to permit the functioning of an adhesive apparatus. This paper documents the alternate designs that have evolved in the various lineages of gekkonine geckos and relates them to perceived phylogenetic and functional design constraints. Comparison with diplodactyline geckos and anoline iguanids further exemplify the fundamenta} constraints involved. [| Gekkonidae, Gekkoninae, first digit, adhesive pads, functional morphology, evolutionary constraint. Anthony P. Russell, Department of Biological Sciences, The University of Calgary, 2500 University Drive N.W., Calgary, Alberta, Canada, T2N 1N4; Aaron M. Bauer, Biology Department, Villanova University, Villanova, Pennsylvania 19085, U.S.A.; 16 August, 1988 The pedal asymmetry that is associated with the three-dimensional kinematics of sprawling locomotion in lizards (Rewcastle, 1983) places constraints upon the morphological potential of foot structure. In both the forelimb (Renous and Gasc, 1977) and hindlimb (Rewcastle, 1983) the complex, three-dimensional movements of the components of the limb skeleton during the limb cycle dictate that a line of contact perpendicular to the body long axis must be made by the foot in question for posteriorly- directed thrust to be developed. In the pes this line of contact is delineated by the metatarsophalangeal line, which connects the distal ends of the first three or four metatarsals (Fig. 1). The asymmetry of the pes is emphasised by the unequal lengths of the digits, with the first being primitively the shortest and the fourth the longest (Fig. 1). This asymmetry in digital length is important in the maintenance of contact with the sub- stratum as the locomotor thrust is delivered (Brinkman, 1980, 1981), Ultimately the foot rolls onto its mesial border as the proximal end of the fifth metatarsal is raised. This is deter- mined by the geometry of the knee and mesotar- sal joints (Rewcastle, 1980, 1981, 1983) and results in complex and controlled movements and deformations in the pes (Rewcastle, 1983). Similar deformations occur in the manus (Landsmeer, 1981, 1984). In terms of digital structure, asymmetry is reflected not only in the relative lengths. of the digits, but alsa in phalangeal formulae. The primitive manual phalangeal formula for lizards is 2-3-4-5-3, and that for the pes is 2-3-4-5-4 (Romer, 1956), The ventral surfaces of the feet and digits are subjected to tractive and torsiona) forces as the powerstroke of the limb is delivered (Landsmeer, 1981, 1954). This is correlated with the presence of friction pads (Figs. 2, 3) and enlarged plates in these regions. Friction pads may be defined as enlarged scales that are 454 FIG. |. Dorsal view of the left pes of /guana iguana. Note the asymmetry of the foot, the sub-parallel disposition of the first four digits and the offset nature of the fifth. The dashed line is the metatarsophalan- geal line and represents the line of contact of the foot towards the end of the power stroke. This line con- nects the distal ends of the first three metatarsals and is normal to the body long axis. The foot rolls of the substrate such that the fifth metatarsal is raised,the fourth distal tarsal pivots about the astragalocal- caneum and the mesial border of the first digit is pressed into contact with the substratum. Abbrevia- lions: ac, astragalocalcaneum; dt iii, dt iv, third and fourth distal tarsals; mt/ph line, metatarsophalangeal line; I - V, digits one to five. MEMOIRS OF THE QUEENSLAND MUSEUM present in areas of potentially intensive loading that are thickened and robust. They have some- times, in pad-bearing forms, been included in a count of total lamellae, but they differ from true scansors (Russell, 1975) in that they are not controlled by the lateral digital tendon system (Russell, 1986) and lack an internal hydrostatic (Russell, 1981) or similar support system. Such friction pads contribute to the minimisation of rotational slippage (Padian and Olsen, 1984) and are arranged in association with the fundamental asymmetry of the foot (Schaeffer, 1941; Snyder, 1952; Robinson, 1975). This asymmetry is of long standing in tetrapod reptiles and is found in a variety of fossil forms, including thecodonts such as Euparkeria (Ewer, 1965), Chas- matosaurus (Cruickshank, 1972) and many pelycosaurs (Romer, 1956). Charig (1972) noted that the development of a symmetrical foot, as seen in crocodilians, dinosaurs and mammals, is correlated with erect limb posture and two- dimensional limb kinematics. Thus, asymmetry of the manus and pes is a primitive lacertilian characteristic and is inherently associated with the normal sprawling locomotor mode of lizards. While the basic characteristics of sprawling locomotion and pedal asymmetry are typical of lizards in general, there have been significant departures from this pattern. In chamaeleons, for example, semi-erect posture is correlated with a zygodactylous grasping foot (Gasc, 1963; Peter- son, 1984). Limb reduction coupled with body clongation is associated with simplification and ultimate loss of the feet and ultimately the ap- pendages (Essex, 1927; Lande, 1978; Raynaud, 1985; Greer, 1987). Perhaps one of the most remarkable examples of pedal redesign is seen in those lizards bearing a subdigital adhesive apparatus. Variations on this theme are evident in a wide varicty of gekkonid lizards (Solano G., 1964; Russell, 1976), as well as in the anoline iguanid radiation (Peterson, 1983a, b), and in an incipient form in some scincids (Smith, 1935; Williams and Peterson, 1982). Modifications as- sociated with the operation of these adhesive systems (Dellit, 1934; Russell, 1975) place fur- ther constraints on digital and pedal design. The adhesive pads are placed into contact with and removed from the substrate by way of digital hyperextension (Russell, 1975). Here the limita- tion of the number of phalanges primitively present in the first digit of both manus and pes renders problematic the incorporation of a fully- developed subdigital pad into digit I. Thus, in pad-bearing geckos there is a wide array of varia- DIGIT | OF GEKKONINE GECKOS FIG. 2. Ventral surface of the left pes of Varanus prasinus indicating the enlarged friction plates on the ventril surfaces of the digits (d), the mesial burder of the hallux (h), und the heel region (he). Drawn from a photograph in Greene (1986), tion in the structure of the first digit and even a tendency to jts reduction in some lineages. Such trends are not seen elsewhere among lizards and reflect the compromise of either modifying the first digit into an effective adhesive device or abandoning it as such a structure. We here survey the subfamily Gekkoninac and document the variation in form of the first digit. We attempt to correlate certain design con- straints and modifications with the construction of an adhesive apparatus in general. Numerous adaptive radiations are evident within the widespread and diverse Gekkoninae (Russell, 1976) and cach of these serves as a test of ideas put forward for other radiations. Additionally. the situation found in the subfamily Diplodac- tylinae acts as another independent test. as does that found in the iguanid anoline radiation. MATERIALS AND METHODS The structure of the first digit of gekkonine geckos was surveyed by examination of museum specimens. The chief collections employed were those of the British Museum (Natural History) (BMNH), the California Academy of Sciences (CAS), the Transvaal Museum, Pretoria (TM) and the University of Calgary Museum of Zool- ogy (UCMZ). Additionally specimens in the col- lection of the senior author were also used (APR). Russell’s (1972) groupings were employed as the units of comparison, with com- parison being carried out both within and be- tween groups. Specimens were examined by one or more of acombination of approaches - a survey of overall external morphology; a survey of skele- tal preparations; a survey of cleared-and-stained preparations; radiography; and histology. RESULTS For completeness and to establish certain parameters of baseline information some non- gekkonids and primitively padless gekkonines were examined. This permitted the plesiomor- phic state of the first digit of lacertilians to be established. Following this, a group by group survey of pad-bearing gekkonines was made in order to document the structure of the pollex and hallux. Finally, for the sake of comparison, the 456 structure of the first digit in the iguanid Anolis is examined. While not wanting to presuppose the discussion, the sequence of presentation of the basic data is given in the same order as it is considered in the discussion. In such a broad survey some order must be imposed in order to ensure a logical discussion of the results. PADLESS NON-GEKKONIDS. A variety of non-gekkonid taxa, in addition to Iguana iguana (Fig. 1), was examined to estab- lish the basic parameters of the form of digit I. In none of these was the pollex or hallux found to be reduced in size. For the purpose of discus- sion, the first digit of Lacerta dugesii will serve as an example (Fig. 3). In both the manus and pes digit I is considerably shorter than the second, bears a strongly-developed claw and has contact points with the substrate at the claw tip and the base of the digit. There are two phalanges present but there is no inflection between them. The most extensively- developed area of friction plates occurs at the base of the digit, beneath the metacarpophalangeal and metatarsophalangeal joints (Fig. 3). The scales beneath the free part of the digit are plate-like and transversely widened, but are not the main load/friction-bear- ing areas of the digit during normal terrestrial locomotion. The remaining digits of Lacerta dugesii also lack marked inflections and their subdigital scale architecture is similar to that of digit I. In this type of foot design the first four digits are sub- parallel, with the fifth markedly diver- gent in the pes (Fig. 1). The metatarsophalangeal line connects the distal ends of the first three metatarsals. MEMOIRS OF THE QUEENSLAND MUSEUM NAKED-TOED GEKKONINES. Many genera of gekkonine geckos primitively lack subdigital pads (Russell, 1976). Of these, two examples have been chosen as extremes. One of these bears digits without inflections, the other with. The genus Teratoscincus may be taken as an example of the situation in primitively padless gekkonines in which purely terrestrial locomo- tion is practised and in which the digits are not inflected. [Indeed, they are modified for sand- walking and burrowing - Luke, 1986]. As with Lacerta dugesii (above) the first digit is consid- erably shorter than the second, bears a strongly- developed although somewhat less curved claw, and has normal contact points with the substrate at the claw tip and base of the digit (Fig. 4). In Teratoscincus there are no friction plates at the base of the digits or enlarged plates beneath the free phalanx, but such situations do exist in other primitively padless gekkonines, such as Bunopus and Agamura (Fig. 5). Contact with the substrate may be made with the entire ventral surface of the digits and foot (Teratoscincus) or with more restricted areas (Agamura), this being determined by locomotor substrate preference and by the condition of the skin of the plantar surface of the foot (‘puffy’ in Teratoscincus and taut in Agamura). The first four digits are sub-parallel (Figs. 4, 5) and the metatarsophalangeal line is similar to that for Lacerta dugesii and Iguana iguana (Fig. 1). In contrast to the straight, uninflected digits described above, those of many species of the (probably paraphyletic) genus Cyrtodactylus are strongly inflected (Fig. 6). Some species of Cyr- FIG. 3. Mesial view of the first and second digits, right pes of Lacerta dugesii. The enlarged friction plates (fp) at the base of the first digit are evident, representing the area of contact as the foot rolls off the substratum during normal terrestrial locomotion. The contact points of the first digit during normal resting posture are marked with asterisks (*). Scale bar = 2mm. DIGIT I OF GEKKONINE GECKOS 457 FIG. 4. Outline of Teratoscincus scincus illustrating the asymmetry of the pes and the essentially flat digits. Redrawn from Lanza (1972). FIG. 5. Outline of Agamura persica illustrating the asymmetry of the pes and the essentially flat digits. This species bears friction plates on the ventral surface of the digits and soles of the feet. Redrawn from Lanza (1972). de ‘A % rodactylus (sensu lato) even bear ineypient pads (Russell, 1976, 1977). In general, enlarged fric- tion pads occur al the melacarpophalan- geal/metatarsophalangeal joints and beneath the inflections (Fig. 6). Transversely widened plates are evident between these points. Beyond the inflection the distal phalanges are held above the substratum and are clud ventrally in smaller scales. All digits are strongly clawed. Digit | bears no enlarged plates bencaith its proximal phalanx, but enlarged friction plates occur al the base of the digit (Fig. 6), while the digit proper is clad ventrally in smaller scales. The enlarged plates at the base of digit I are similar in position and form to thase seen in Lacerta dugesti (Fig. 3). PAD-BEARING GEKKONINES Within the radiations of pad-bearing gek- konines there have been lineages that have developed pads from the base of the digits that have spread distally, and Jineages that have developed pads distally that have spread proximally (Russell, 1976), Asa result of this, notall groups can be considered ina single linear sequence or as part of a ere morphotypic series. We have chosen examples that appear to exemplify evolutionary trends in pad structure and in the form of digit 1, but the sequences we have chosen should be treated as being illustra- live of morphological rather than phylogenetic trends. We begin by considering forms that are believed to have developed pads from the basal portions of the digits, and then consider as- semblages where pads appear lo have begun distally, Both of these have had different influen- ces on the first digit, but there are also parallels between the two types. Genera with basally-derived pads Working on the premise established by Russell (1976, 1977) thal the pedal structure seen in the genus CyHtodaecrylus could be, morphologically, a precursor of the Hemidactylus patiern, we begin here with a consideration of the Hemudac- rvlus group, Here the digits hear basal pads (Fig. 7) in which the scansors are borne beneath the modified digital inflection (Russell, 1977, Fig, 3) similar to that seen in Cyriodacy lus (Fig. 6). The scansors of Hemidactylus grade into friction plates at the bases of the digits (Pig, 7}. The distal phalanges are free of the pad and there is na evidence that they have ever been included within its confines. As with other pad-bearing geckos the disposition of the digits is more sym- MEMOIRS OF THE QUEENSLAND MUSEUM metrical than in padless forms (Migs, 6, 7), this symmetry being brought about largely by a grealer divergence of digit 1V from digit ILI. The first digit still possesses only two phalan- ges and hence the pad borne on this digit is relatively small (Figs. 7, 8). The ability to hyper- extend (Russell, 1975) the pad is rather limited and occurs al the junction of the first metacar- pal/metatarsal and the first phalanx (Fig. 9). In all other digits hyperextension occurs between two phalanges rather than between the basal phalanx and a metapodial. The size of the pad in digit Lis thus restricted by the dimensions of this digi} and the absence of a free basal portion (Pigs, 7, 8). Although there is a great deal of variation within the genus Hemidactylus, the situation described above prevails in all species. Similar arfangements are encountered in the satellite genera Cosymbotus (=Platyurus) and Terato- lepis, while Dravidogecko (Fig. 10) exhibits what may essentially be taken to be a stage morphologically intermediate between Hemidactylus and the digitally more derived species of Cyriodactylus (Russell, 1976). Here again, however, digit | exhibits only incipient pad development. The most marked departure from the basic ‘Hemidacitvlus’ condition is seen in the monotypic genus Briba. Here both the pollex and hallux are small (Vanzolini, 1968a, b), and both are clawless (Amaral, 1937, Fig. 2). The distal- most phalanx (ungual phalanx) in each 1s some- what clongated and supports the distal extremity of the pad (Fig, 11), rather than being extended beyond it (Figs, 7.8). Thus, both phalanges in digit | support the pad and are involved in hyper- extension of the pad. While the first digit is relatively small in size itis more fully committed to involvement in the operation of the pad. Genera with distally-derived pads. The complex of generic groupings to be con- sidered here represents the result of multiple evolutionary pathways (Bock, 1959) rather than a single morphotypic sequence. The basic as- sumption made is that adhesive pads in all of these assemblages arose distally and thalin some lincages they subsequently spread towards the base of each digit. Data derived from internal morphology of the digits is consistent with this interpretation (Russell, 1974). In this context (he structurally simplest case within the Gekkoninae is exemplified by Phy/- ladactylus (sensu lato) and us putative allies. The DIGIT I OF GEKKONINE GECKOS basic morphological pattern is apparent in such forms as Phyllodactylus porphyreus (Fig. 12), although there is much variation within this genus and its allies (Asaccus, Ebenavia, Paroedura and Urocotyledon). Typically in Phyllodactylus the adhesive plates are present as paired, Ieaf-like structures at the distal extremity of each digit (Fig. 12), with the claw being disposed between them. The proximal end of each plate is located adjacent to the joint between the penultimate and ungual phalanx. Proximal to this the ventral surface of each digit is clad in expanded, plate-like scales that merge gradually into the scales of the palm/sole (Fig. 12). There are no enlarged friction plates in the palm/sole area. Although radiation of the digits is some- what pronounced, digit IV is not markedly diver- gent from digit III (Fig. 12). Digit I again contains two phalanges and is relatively short (Fig. 12). The distal, leaf-like pads (Fig. 13) are strongly developed, however, and may be raised from the substratum by hyper- 459 FIG. 6. Ventral aspect of the right pes of Cyrtodac- tylus peguensis. Note the digital inflections, the friction plates at the bases of the digits and those beneath the first metatarsal (far right). Specimen UCMZ (R) 1981.3. FIG. 7. Ventral view of the right pes of Hemidactylus brookii. Note the free distal portions of the digits, the basal pads and the relative symmetry of the foot. Uncatalogued specimen. FIG. 8. Ventral view of the scansors of digit I, right pes of Hemidactylus brookii. Note the relative lack of expansion of the pad and its relatively small size. The scansors grade into the friction plates at the base of the digit. Uncatalogued specimen, extension of the distalmost digital joint and the metacarpophalangeal/metatarsophalangeal joint. The genera Ptyodactylus and Uroplatus were included within the Phyllodactylus group by Russell (1972), based upon certain morphologi- cal similarities of digital structure. From a func- tional viewpoint the replacement of a terminal single pair of leaves by a subdivided, radiating, fan-like array of scansors has had little impact on the disposition of other digital components. The first digit remains strongly developed, the scan- sor area is still confined to the extreme distal end of the digit, and the basal components of the digits are still devoid of any scansor-like struc- tures (see Smith, 1935, Fig. 24 for Ptyodactylus and Duméri]l and Bibron, 1836, plate 31 for Uroplatus). In all members of the putative Phyllodactylus group (Russell, 1972) the digits are flat and lack inflections. Thus, while the adhesive pads are situated distally the majority of the ventral sur- 460 MEMOIRS OF THE QUEENSLAND MUSEUM FIG. 9. Dorsal view of the skeletons and pad outlines of the first and second digits, left pes of Cosymbotus platyurus. In the second digit the greatest expansion of the pad occurs beneath the penultimate phalanx and this can be hyperextended on the first phalanx. In the first digit hyperextension is limited to the metatar- sophalangeal joint. Drawn from a cleared-and- stained specimen - CAS 18565. Scale bar = 2mm. Abbrevia- tions: mt, metatarsal; pen.p, penultimate phalanx; ppe, paraphalangeal element; I, II, digits one and two. face of each digit is free to contact the locomotor substratum. The gekkonine genera Afroedura and Calodactylodes represent a continuation of a morphotypic series begun by Phyllodactylus (Russell and Bauer, 1989). Here the scansors are more elaborate. The terminal pair of leaf-like scansors has been augmented by more proximal plates. While these plates are not as fully demar- cated as the distalmost pair, they are quite dis- tinct from the more proximal scales and bear distinct fields of setae (Fig. 14). Digits I] - V exhibit this proximal encroachment of the ad- hesive apparatus. The digits are without inflec- tion and digit IV is considerably divergent from digit III. As has been pointed out elsewhere (Russell, 1979), the proximal sets of scansors are controlled by a different musculotendinous com- plex than the distalmost ones. Digit I has much less scope for the proximal elaboration of scansors (Fig. 15). The terminal pair of scansors is still well- developed, but the more proximal ones are much less prominent than those of the other digits (Figs. 14, 15). The more proximal scansors of digit I are present beneath the raised component of the first phalanx, but there is little in the way of a free portion of the digit for the incorporation of a more elaborate pad. The claw is still prominent on digit I (Fig. 15) and in all digits the more proximal scansors are located proximal to the ungual joint, beneath the raised portion of the penultimate phalanx. Based upon the scenario outlined by Russell (1979) which suggests that distally originating scansors migrate proximally in their extent, the next example in the proposed morphotypic series can be taken from Aristelliger. Here the second to fifth digits have broadly expanded series of scansors beneath the proximal part of the penul- timate phalanx and the distal part of the next most proximal one (Fig. 16). More proximal still there is a free basal portion to each digit, and the penultimate phalanx and claw are carried free beyond the distal end of the pad (Fig. 16). In digit I, however, the more proximal part of the pad is only incipiently developed (Fig. 17), while a single, asymmetrically disposed leaf-like plate remains present distally. Digit I remains clawed but has no free distal portion and remains much less distinctly pad-bearing than the other digits DIGIT |! OF GEKKONINE GECKOS Tp FIG. 10. Ventral aspect of the tight pes of Dravidagecko anamallensis (BM(NH) 82.5.22.79). Note the undivided nature of the scansors and the basal friction plates/scansors on digit J. Scale bar = 2mm, (Fig. 16). Terminal, leaf-like pads are no longer evident on digits II - V. Beyond this situation a number of groups have modified the first digit in a variety of ways. In the Gekko group the nominal genus bears a strongly padded first digit, the pad being ex- panded in the same manner as that of the other digits (Fig. 18). Examination of foot structure reveals, however, that digits [1] - V are strongly clawed while digit I is clawless. Internally digit I can be seen to possess an elongate distalmost phalanx, and it is beneath this that the pad is positioned (Fig. 19). Thus, the distalmost phalanx may be hyperextended on the first phalanx of this digit, and the internal control mechanism is the same as that for the other digits (Russell, 1975), Within the genus GeAyra a similar situation obtains in some cases, with divided scansors being borne on the first digit (Figs. 20, 21). Some species of Gehyra exhibit a somewhat inter- mediate stage in which the claw is still evident but small and needle-like (Fig. 22), The distal- most phalanx is somewhat elongate and the pad is partially expanded, Other genera in this as- semblage that exhibit a clawless or minutely clawed first digit with an expanded pad are Lepidodactylus, Luperosaurus, Pseudogekko and Prychozeon. Some genera belonging to the Gekko group (sensu Russell, 1972) do not, however, exhibit elongation of the distalmost phalanx and loss of the functional claw in digit 1. In both Hemiphyl- lodactylus and Perochirus the first digit remains small and unexpanded. In Hemiphyllodactylus the first digit of both the manus and pes is minute but clawed, and pads are not evident. In Perachirus the pollex is rudimentary and claw- less (Fig, 23), while the hallux is rudimentary bul bears a claw (Fig. 24). The digits of Perochirus, except for the first, are widely expanded (Figs. 23, 24), The first digit has essentially been sup- pressed by the second and does not form a func- tional pad. The Ailuronyx group includes the nominal genus plus Phelsuma, Lygadacrylus, Micros- calabotes and Millotisaurus. Only in Ailuronyx is the first digit expanded and pad-bearing, There are relatively few scansors but the digit is Tat St Saes CS aS o <2 -, — ss c= rt —— “a= e<: et: FIG. 11. Ventral view of the pes of Briba brasiliana. Redrawn from Amaral (1937). The first digit lacks a claw and [he free portion of the pad is enhanced by incorporation of the ungual phalanx into it. 462 MEMOIRS OF THE QUEENSLAND MUSEUM FIG. 12. Ventral view of the left pes of Phyllodactylus porphyreus. CAS 167593. Note the terminal, leaf-like scansors on each digit and the broad plates borne more proximally. FIG. 13. Ventral aspect of digit I, left pes of Phyllodactylus porphyreus. CAS 167593. FIG. 14. Ventral view of the left pes of Afroedura hawaquensis. CAS 167638. Proximal encroachment of scansors on digits II - V is evident. FIG, 15, Ventral aspect of the left hallux of Afroeaura hawaquensis. CAS 167638. The proximal setose plates remain poorly differentiated, FIG. 16. The lefi pes of Aristelliger praesignis in ventral view. The adhesive pads on digits II - V are well-developed, but digit I lacks this differentiation. Uncatalogued specimen. FIG. 17. Ventral view of the left hallux of Aristelliger praesignis. The single, terminal leaf-like scansor is evident distally (demarcated by an asterisk - *). The claw (arrow) lies to the medial side of the terminal plate. DIGIT 1 OF GEKKONINE GECKOS FIG. 18. Ventral view of the left pes of Gekko gecko. Note the broadly expanded pad on digit I and the absence of a claw from this digit. Abbreviations: I - V, digits one to five. clawed, although the penultimate phalanx is not raised. The pad on digit I of the manus is rela- lively narrow, while thal on the pes is slightly wider but still considerably narrower than the pads on the other digits, The entire padded por- tion of digit I is free. In all other genera of this putative assemblage the first digit is greatly reduced in size. In Phel- suma it is minutely clawed, and only friction plates are present ventrally (Figs. 25, 26). The other digits bear widely expanded pads and the claws are reduced (Fig. 25). Functionally and proportionally digit I is similar to the pollex of Perochirus (Fig. 23). Similarly, in Lygodactylus and Micrascalabotes (Fig. 27) the first digit is markedly reduced in size, although here it remains prominently clawed. In Millotisaurus the pollex is essentially absent (Pasteur, 1964, Plate 1) and the distalmost (ungual) phalanx is absent. The hallux is small, clawed and of much the same form as that of Microscalabotes (Fig, 27). In contrast to the cases of reduction in size of the first digit or expansion of the pad by virtue of elongation of the distalmost phalanx, as out- lined above, the Pachydactylus group is charac- terised by a rather prominent digit I. Here there is a relatively long free basal portion capped, in those species not showing reduction of the ad- hesive apparatus, with an expanded pad. In all genera included within the Pachydactylus group 463 (Pachydactylus, Rhopiropus, Colopus, Chondrodactylus, Kaokagecko, Palmatogecko, Tarentola and Geckonia) there is a hyperphalan- gy of the first digil, with the phalangeal formula being 3-3-4-5-3/3-3-4-5-4. Thus, in combination with clawlessness of the first digit (or the posses- sion of needle-like elongate claws) the distal twa phalanges of the pollex and hallux may be hy- perextended while the stoutly developed proximalmost phalanx remains as a stable base for the digit, with friction plates beneath (Fig. 28). Friction plates are also present at the bases of the other digits (Fig. 28). When not climbing, members of the Pachydactylus group walk on the bases of their digits with the distal ends held in a permanently hyperextended position. Even in situations Where the pads are reduced in size (Figs. 29.30, 31) the first digit remains relatively long, with the remnant of the pad being displaced distally. In the genus Homopholis (Fig. 32) the claw of digit [ is needle-like and elongate and the pad is relatively strongly dilated. The condition in the manus and pes is not identical, however. The pollex (Fig. 32) exhibits hyperphalangy similar to the situation outlined for the Pachydactylus Broup (above) while the hallux has the normal compliment of two phalanges, but with the distal one being elongated to support the pad. The latter condition is similar to that found in Gekko (see above). Thus in Homopholis two independent Vie, FIG. 19, Dorsal view of the skeleton of the right pes of Gekko gecko, Note the elongaied terminal phalanx on digit . Compare this figure to Fig. 1. Abbrevia- tions: | - V, digits one to five. 464 solutions to the operation of a functional pad on the first digit have occurred, one in the manus and one in the pes. The genus Geckolepis was included within the Homopholis group by Rus- sell (1972). Members of this taxon bear a needle- like claw on both the pollex and hallux, MEMOIRS OF THE QUEENSLAND MUSEUM FIG. 20. Ventral view of the right pes of Gehyra oceanica (CAS 159706). Fig. 21. Ventral view of the left hallux of Gehyra punctata (APR 48). FIG. 22. Ventral aspect of digit I, left pes of Gehyra interstitialis (CAS 89686). Note the minute claw visible distally (arrow). FIG. 23. Ventral view of the right manus of Perochirus ateles (CAS 159768). The pollex lacks a claw. _ FIG. 24. Ventral view of the left pes of Perochirus ateles (CAS 159768). The hallux bears a claw. associated with a distally elaborated pad (Figs. 33, 34). In this genus both the manus and pes exhibit elongation of the terminal phalanx of digit I, similar to that found in the hallux of Homopholis. The final group to be considered is the DIGIT I OF GEKKONINE GECKOS 465 FIG. 25. The left pes of Phelsuma sundbergi (CAS 167553) in ventral view. Note the small size of digit I and the friction plates at its base. FIG, 26. The hallux of Phelsuma sundbergi (CAS 167553) in ventral view. A minute claw is still evident. FIG. 27. The first and second digits of the left pes of Microscalabotes bivittis (CAS 126289) in ventral view. The hallux is extremely small. FIG. 28. Ventral view of the right pes of Rhoptropus bradfieldi (CAS167673). The pad on digit I is prominent. FIG. 29, Ventral aspect of the right pes of Pachydactylus maculatus (CAS 167613). FIG. 30. The hallux of Geckonia chazaliae (CAS 134579) in ventral view. Although the pad is reduced this digit is still relatively long. 466 MEMOIRS OF THE QUEENSLAND MUSEUM FIG. 31. Ventral aspect of the right pes of Pachydactylus kochi (CAS 126170). The pads are greatly reduced but all digits remain long. FIG. 32. Ventral view of the right manus of Homopholis wahlbergi (TM 57464). FIG. 33. The right pes of Geckolepis maculatus (CAS 156886) in ventral view. The hallux is to the extreme right. FIG. 34. Ventral aspect of the right hallux of Geckolepis maculatus (CAS 156886). Thecadactylus group (Thecadactylus, Bogertia and Phyllopezus) of South America. In Phyl- lopezus digit I has a prominent claw and a free terminal phalanx, but only unexpanded friction plates at the base of the digit. In appearance the pollex and hallux are similar to those of Aristel- liger (Fig. 16), but without the asymmetrically placed terminal leaf-like scansor. The remaining digits bear broadly expanded pads. In the closely related Bogertia the first digit is minute and the claw small (Fig. 35), while the remainder of the digits are broadly expanded. In Thecadactylus the claw on digit I, as on all other digits, is drawn up into a sulcus and scansors are present in two distinct rows. Digit I is strongly developed, broadly expanded and the penultimate phalanx is long and permits hyperextension at the metacarpophalangeal/metatarsophalangeal joint. The incorporation of a sulcus for the claw of digit I has permitted distal extension of the pad on this digit. THE GENUS ANOLIS —AN INDEPENDENT TEST CASE Other than gekkonid lizards, the group that exhibits the greatest development of a subdigital adhesive apparatus is the iguanid Anolis and its immediate allies (Peterson, 1983a,b). Here pads appear to have been developed beneath digital inflections, with the scansors being derived from friction plates beneath the digital inflections (see Peterson, 1983a, fig. 5 for a comparative illustra- tion of this between the primitive anoline DIGIT | OF GEKKONINE GECKOS FIG, 35. Dorsal view of digits 1 and II of Bogertia lutzae. Redrawn from Vanzolini (1968b), Chamaelinorops and the pedally derived Anolis occultus). An examination of digit [ in this as- semblage will be instructive in placing the obser- vations on gekkonines into perspective. In Anolis (Fig. 36) digit I lacks a pad but bears enlarged friction plates basally. The remainder of the digit is held out of contact with the sub- strate and the claw contacts the locomotor sur- face distally. All other digits bear similar friction plates proximally but have adhesive pads located more distally at the region of the junction of the penultimate and antepenultimate phalanges. Thus, as in many gekkonine geckos, the first 467 digit lacks a pad and is not functionally a part of the adhesive apparatus. DISCUSSION The results presented above indicate a great deal of variability in the structure of the first digit of gekkonine geckos. Such variability may be correlated with functional constraints. With the inception of the hyperextension mechanism of pad control (Russell, 1975) the mode of opera- tion of the feet during locomotion changed. No longer is the majority of locomotor thrust developed as the foot rolls onto its mesial border and makes contact at the metatarsophalangeal line. The adhesive apparatus has brought along with it a re-establishment of pedal symmetry in many groups of gekkonines and has dictated that thrust is generated in a different way (Russell, 1975). The metatarsophalangeal/metacar- pophalangeal area is no longer the primary site FIG, 36, Ventral view of the left pes of Anolis garmani (APR 257) showing the digital proportions and the pads on digits I] - V, Digit I bears only friction plates al its base, similar to those al the bases of the other digits (illustrated for digits Il and III). Scale bar = Smm. 468 of application of thrust to the substratum ii pad- bearing gekkonines and the digits play a greater role in providing fixed areas of attachment over which the body moves. Protection of the ad- hesive apparatus also dictates the way in which locomotion can be brought about (Russell, 1975). The fundamental asymmetry of the lizard foot resulted in only two phalanges being incor- porated into the first digit. This condition is retained in pad-bearing gekkonines and thus places restrictions on the incorporation of an adhesive pad into this digit. This has resulted in cither the omission of an adhesive system from this digit, or modifications of it such that an adhesive apparatus can be incorporuled. The perceived possibilities of digit [ evolution in gekkonines are outlined in Fig. 37, Beginning from the primitively padless condition digits may remain (al or may develop inflections to assist in climbing. In either of these circumsian- ces pads may develop, In the case of the former they initially develop distally, while in inflected digits they develop basally. The Hemidactylus group appears to have developed pads in a basal position from a morphotypic ancestral state such ag that exemplified in the genus Cyrtodactylus. As digit | bearsonly two phalanges, no inflection is present within the digit. Hence, any enlarged scansor-like plates that develop do so beneath the distal portion of the penultimate phalanx. Constraints of digital form dictate that the pad remains relatively rudimentary. The allernate solution js to elongate the pad by losing the claw and clongating the distalmost phalanx, This provides more support fo the pad and permits it to extend distally, This condition is found in Briba, although the pollex and hallux are not greatly claborated in this genus. Thesimplest situation found in distally derived pads is that exemplified by the genus Phyllodac- rvfus. Here cach digit bears a single pair of ter- minal, teaf-like scansors that are associated with the distalmost digital joint. The plates on each pair of digits are equally developed and the digits romain essentially Mat, One form of elaboration of this pattern is ta subdivide the distal plates buy to not extend the adhesive apparatus any further proximally along the digit, Such a situation is scen in the genera Ptyodactylus and Uraplatus und the adhesive apparatus on digit [ is as well- developed as it is on the other digits. The other means of elaborating the adhesive apparatus that is seen in the Gekkoninae is to permit it to encroach onto the proximal regions MEMOIRS OF THE QUEENSLAND MUSEUM of the digit. This is more feasible for digils Il- V than it is for digit J. Thus in genera such as Afreedura, where this occurs, the pads become further elaborated on all but the pollex and hallux where they remain essentially terminal, This trend is continued in genera such as Aristefliger where digits If- V develop elaborate, multiscan- sorial pads. In this genus, however, pad develop- ment on digit | ts suppressed and proximal elaboration does not occur, The internal digital structure in the Gekko group is more complex than that in Aristelliger, but if the claw is retained on digit I as a free and independently controllable structure, then elaboration of a pad on digit lis suppressed. Such 4 Siluation is seen in genera such as Hemiphy!- lodacrylus and Perachirus. Some members of the Gekko group have developed an enlarged pad on digit I, but this has come at the cost of a freely controllable claw on this digit, Thus, in Gekko the distalmost phalanx of digit | has become elongate and supports the elaborated and en- larged pad. In essence these two choices have been adopted by members of the Ailuronyx group also. In Adluronyx the first digit is moderately well padded and retains its claw al- though this is not held free of the pad. In Phel- suuma and Lygodactylus and its relatives the pad on the first digit is suppressed. An alternative solution to the maintenance of a well-developed pad on digit | is found in the Pachydactylus group, Here, instead of the distal- must phalanx becoming elongated, an additional phalanx has been incorporated resulting in hy- perphalangy in digit |. The genus Homephalis has employed both solutions, being hyper- phalangic in the pollex and having an elongate distalmost phalanx in the hallux. Yet another solution is seen in (he genus Thecadactylus. Here the claw is retracted into a sulcus, in digit I as well as the other digits, allowing the pad to become elaborated distally without hindrance from the claw. The situation seen in Anolis illustrates that the problem of elaborating 2 pad on the first digit exists outside of the family Gekkonidae, Here the constraints of only two phalanges have resulted in this digit lacking the development of a subdigital adhesive pad. Thus, although theoretically all digits may be thought of is possessing approximately the same potential in terms of evolutionary modification, limitations occur that are imposed by functional demands, The same problem has manifested itself inessen- tially the same way in the Gekkonidac and the 469 DIGIT I OF GEKKONINE GECKOS -AuadoyAyd e se pojasdsajur ag Jou pynoys awayos ay) pure sodAjoyd sejnoyed jo sojdurexa se pasn ase payeaipur e1auad ayy ‘soyseS ouruoyyas ul | 1131p Jo OIN[OAS ay) UI Spuad Saeed jo SAE paar aa UO! puoD ssaj ped A] aA! }IWlig payoajjul $716ig yej4 uieway s3/5ig pat ASS s}6iq {|e uo SMIALOVGOLYAD sped a4!|-jea] jeulWsay [SMIALOVGOTTAHd | spe jeseg we pedojeraq sseq T speq jo quawyoeosouy |eWIXOld paplaipgns spey |eullWsa) parebuo) 3 xueledd jenBuy) = Ayequauiipny Ped VunGIOusV SMIALOVGOAIA yso7] MEID poureyjoy Me|D value SNTIALOVGINWAH juawyoeosug JBWIXOlg JBYYNY Y¥4d91 7135 1S1NV passaiddng ped ee ee pauleyey Med J jo ABuejeydiedAy VWAS1SHd SNIALOVGAHOVd aej4 JON ng peuleyey me MSF Hey payesogely JON ped XANOUYNTIV MB|D SUIE}BY T Vi ld¥aD08 pejyebuo| 3 Ajyeysiq SpueyxyQ Ped xuepjeyd jsow|e3s!q snajns ul Me|D OWA SMALOVGV OSHL 470 Iguanidae. Within the Gekkonidae similar trends are seen in the diplodactyline geckos, a radiation parallel to that of the Gekkoninae but restricted to the Australasian region (Kluge, 1967a,b; Bauer, 1986). In Diplodactylus digits morpho- logically very similar to those of Phyllodactylus (see above) are present and a single pair of terminal leaf-like plates is present on all digits. Modifications of this pattern are also evident in the satellite genera Crenadactylus and Rhynch- oedura. In the closely related Oedura proximal elaboration of the pads is seen in digits Il - V in a fashion very similar to that seen in Afroedura (sce above). In digit I, however, proximal en- croachment of scansors is much less marked, again paralleling the situation seen inAfroedura. In the tribe Carphodactylini the simplest pads are seen in the genus Naultinus. Here digits II - V bear pads consisting of multiple scansors beyond which is a free, clawed distal portion of the digit. Digit I retains its claw and lacks a definitive pad, there being only friction pads at the base of this digit. Distally there is an asymmetrical, leaf-like pad similar to that seen in Aristelliger (see above). A similar situation obtains in Hoplodac- tylus although the pads on digits II - V are somewhat broader. Digit I bears an asymmetri- cal, leaf-like pad very much like that of Naul- tinus. The genera Rhacodactylus, Bavayia and Eurydactylodes bear more elaborate pads on digits II -V but digit I remains much less ex- panded and again retains the terminal, leaf-like plate (Bauer, 1986). In all of these genera the claw is retained on the pollex and hallux, and the incipient pad at the base of digit I is not ex- panded. Thus again, although those digits with a free basal portion have elaborated substantial pads, the inherent limitations in the design of digit I have suppressed such an expression here. The genus Pseudothecadactylus, now subsumed as part of Rhacodactylus (Bauer, 1986) has divided scansors on digits II - V somewhat rem- iniscent of those of Thecadactylus, with the claw at least partially recessed into a sulcus between the rows of scansorial plates. Digit I lacks a claw and the possession of a sulcus has permitted distal extension of the pad, again similar to the situation seen in Thecadactylus. Thus Pseudo- thecadactylus has a pad on digit I that is much more extensively elaborated than that in any other carphodactylines and is constructed on similar lines to the first digit of the gekkonine Thecadactylus. The discovery of constraints and limitations in morphology (Zweers, 1979) gives us some in- MEMOIRS OF THE QUEENSLAND MUSEUM sight into why things are the way they are (Seilacher, 1970). The repetition of pattern in different evolutionary lineages serves to recipro- cally illuminate the concepts being postulated. In the above example the constraints placed upon the first digit of lizards by virtue of its inherent design have had a strong influence in the elaboration of the adhesive apparatus in gek- konid lizards, This in turn has had, presumably, some effect on the locomotor mechanics of pad- bearing geckos, with different selective factors being influential in the particular outcome in a given set of ecological circumstances. In ques- tions such as this, however, our background in- formation is still woefully inadequate to attempt to assess what these influences might be. In almost all groups of lizards trends towards limb reduction are evident (Gasc and Renous, 1976), and in these cases digital reduction is a progressive and essentially symmetrical event (Lande, 1978; Raynaud, 1985). In some lacer- tilians, however, there are patterns of digital reduction that occur independently of those of entire limb reduction and that are associated with quite different locomotor modifications. In the agamid genus Sitana, for example, the fifth digit of the pes is lost (Russell and Rewcastle, 1979) while the other digits remain seemingly unaf- fected. A similar reduction is present in the teiid Teius (Presch, 1970). The standard models that have been proposed to account for digital reduc- tion (Raynaud, 1977; Lande, 1978) have as- sumed an equal functional importance for all digits. In both Sitana (Russell and Rewcastle, 1979) and in a variety of gekkonine geckos this cannot be assumed to be the case, however. Indeed, the various trends towards simplification and reduction of the pollex and hallux in gek- konine geckos, as outlined above, can be directly associated with proposed functional constraints that place a different set of selective pressures on digit I than they do on the other digits. Thus in gekkonines there are trends to both elaboration and reduction of the first digit in association with the acquisition of a setal ad- hesive apparatus. This in itself is instructive as it indicates to us that there are a variety of ways to achieve a functional adhesive system. In the case of the lack of incorporation of digit I into this system one can invoke Underwood’s (1976) dis- tinction between simplification (reduction in the complexity of structure without loss of full anatomical and histological differentiation) and degeneration (reduction with loss of precise dif- ferentiation). The distinction is not always sharp, DIGIT | OF GEKKONINE GECKOS but Underwood (1976) stated further that simplification is involved with modification of function without a break in continuity of fune- tion, while degencration is associated with reduction or loss of original function, In this context the reduction of digit | in a variety of gekkonines 1s probably best categorised as simp- lifieation, with sufficient of its structure remain- ing to permit a variety of already extant anatomical relationships to persist. In no case is the entire complement of components of digit [ last, even if its external manifestation becomes negligible. ACKNOWLEDGEMENTS We thank E.N. Arnold (British Museum (Natural History)), Robert Drewes (California Academy of Sciences) and Wulf Haacke (Transvaal Museum) for access to and Joans of extensive series of specimens examined in this study. The Natural Sciences and Engineering Research Council of Canada provided funds by way of grant No, A9745 to A.P.R., and that the University of Calgary funded a postdoctoral fel- lowship for A.M.B, that enabled completion of this work and travel to Australia to present the results. Assistance in the laboratory was provided by C.C. Chinnappa, Darcy Rac and Ed Grimley. We thank Herb Rosenberg and W.B. Williams for discussion of ideas incorporated herein, Susan Stauffer typed the manuscript in her usual efficient manner. LITERATURE CITED AMARAL, A, 1937. New genera and species of lacet- tilians from Brazil. Compt. Rend, XI} Congr. Int. Zool. Lisbon {1935]: 1701-1707. BAUER, A.M. 1986. Systematics, biogeography and evolutionary morphology of the Carphodac- tylini (Reptilia; Gekkonidae), (Unpublished Ph.D. thesis, University of California, Berkeley). BOCK, W.J. 1959. Preadaptation and multiple evolu- (ionary pathways. Evolution 13: 194-211. BRINKMAN, D. 1980. Structural correlates of tarsal and metatarsal funcuioning in /euara (Lacertilia; TIguanidae) and other lizards, Can, J. Zool, 58: 277-289. 1981. The hind limb step cycle of Tpuana and primitive reptiles, J, Zool, Lond. 181! 91-103. CHARIG, A.J. 1972. The evolution of the archosaur pelvis und hindlimb: an explanation in function- al lerms. pp. 121-155. Ja Joysey, K.A. and 41 Kemp, T.S. (eds), ‘Studies in vertebrate evolution’ (Oliver and Boyd; Edinburgh), CRUICKSHANK, A.R.1. 1972, The proterosuchian thecudonts. pp. 89- 119. a Joysey, K.A. and Kemp, T,S. (eds), ‘Studies in vertebrate evolution’. (Oliverand Boyd: Edinburgh). DELLIT. W.D. 1934, Zur Anatomie und Physiologie der Geckozehe. Jena Z. Naturwiss. 68; 613-656. DUMERIT., A.M.C. AND BIBRON,G. 1836. ‘Erpétologie Générale ou Histoire Naturelle Complete des Reptiles.” Vol. 3, (Roret: Paris}. ESSEX, R. (927. Studies in reptilian degeneration. Proc. Zool. Soc. Lond, 1927: 879-945, EWER, R,.F. 1965, The anatomy of the thecodont teplile Euparkeria capensis Broom. Phil. Trans. Roy. Soc. Lond, Ser, B, 248: 375-435, GASC, J-P. 1963. Adaptation a la marche arboricole chez le cameleon, Arch. Anat. Hist. Embryol. Normal, Exp. 46: 81-115. GASC, J-P, AND RENOUS,S. 1976, Les caracteres morphologiques des formes apodes chez les rep- tiles et leur evolution. Bull Soe, Zool France (OL: 47-59, GREENE, H.W. 1986. Diet and arbyrealily in the emerald monitor, Varanus prasinus, with com ments anthe study of adaptation, Pieldiana Zool, ns, 31) 1-12, GREER, A.E, 1987, Limb reduction in the lizard genus Lerisfa, |, Varialion in the number of phalanges and presacral veriebrae. J. Herpetol. 21 267-276, KLUGE, A.G. 1967i. Higher laxonomie categories ot vekkonid lizards and their evolution, Bull, Amer. Mus, Natl, Hist. 135: 1-60. }967b, Systematics, phylogeny, and zoogeography of the lizard genus Diplodactylus Gray (Gek- konidae). Aust. J. Zool. 15; 1007-1108, LANDE, R. 1978. Evolutionary mechanisms ul limb loss in tetrapods. Evolution 32: 73-92. LANDSMEER, J.M.F. 1981. Digital morphology in Varanus and /guana. J. Morphol, 168; 289-295. 1984, Morphology of the anterior limb in relation to sprawling gait in Varanus, Symp. Zool. Soc. Lond, 52) 27-45. LANZA, B. 1972. ‘I Vertebrati Inferjori del] Bwrasia,” (Istitura Geografico Mililare; Firenze), LUKE, C. 1986, Convergent evolution of lizard toc fringes. Biol. J. Linn. Soc. 27; 1-16. PADIAN, K. AND OLSEN, P.B. 1984, Footprints of the Komodo monitor and the trackways of fossil tepules, Copeia L984: 662-671, PASTEUR, G 1964, Recherches sur l'évolution des lygodactyles, lézurds Afro-Malgaches actucls, Trav. Inst. Sci, Cherilien Ser. Zool, 29: 1-132. PETERSON, J.A. 19838, Vhe evolution af the 472 subdigital pad in Anolis. 1. Comparisons among the anoline genera, pp. 245- 283. In Rhodin, A.G.J. and Miyata, K. (eds), ‘Advances in her- petology and evolutionary biology’. (Museum of Comparative Zoology: Cambridge, Mas- sachusetts). 1983b. The evolution of the subdigital pad in Anolis. 2. Comparisons among the iguanid genera related to the anolines and a view from outside the radiation. J. Herp. 17: 371-397. 1984, The locomotion of Chamaeleo (Reptilia: Sauria) with particular reference to the forelimb. J. Zool. Lond. 202: 1-42. PRESCH, W. 1970. The evolution of macroteiid lizards: an osteological interpretation. (Un- published Ph.D, thesis, University of Southern California). RAYNAUD, A. 1977. Somites and early mor- phogenesis of reptile limbs. pp. 373-385. In Ede, D.A., Hinchliffe, J.R. and Balls, M. (eds), ‘Ver- tebrate limb and somite morphogenesis’. (Cambridge University Press: Cambridge). 1985. Development of limbs and embryonic limb reduction. pp. 59- 148. Jn Gans, C. and Billett, F, (eds), ‘Biology of the Reptilia’. Vol. 15. (John Wiley and Sons: New York). RENOUS, S. AND GASC, J-P. 1977. Etude de la locomotion chez un vertébré tétrapode. Ann, Sci. Nat, Zool. Paris, (12) 19: 137- 186. REWCASTLE, S.C. 1980. Form and function in lacertilian knee and mesotarsal joints; a con- tribution to the analysis of sprawling locomo- tion. J, Zool, Lond. 191: 147-170. 1981. Stance and gait in tetrapods: an evolutionary scenario. Symp. Zool. Soc. Lond. 48: 239-267. 1983. Fundamental adaptations in the lacertilian hind limb: a partial analysis of the sprawling limb posture and gait. Copeia 1983: 467-487. ROBINSON, P.L. 1975. The functions of the fifth metatarsal in lepidosaurian reptiles, Colloq. Int. C.N.R.S. 218: 461-483. ROMER, A.S. 1956. ‘Osteology of the reptiles.’ (University of Chicago Press: Chicago). RUSSELL, A.P, 1972, The foot of the gekkonid lizards: a study in comparative and functional anatomy. (Unpublished Ph.D. thesis, University of London). 1975. A contribution to the functional analysis of the foot of the tokay, Gekko gecko (Reptilia: Gekkonidae). J. Zool. Lond. 176: 437-476. 1976. Some comments concerning interrelation- ships amongst gekkonine geckos. pp. 217-244. In Bellairs, A. d’A. and Cox, C.B. (eds). ‘Mor- phology and biology of reptiles’. (Academic Press: London). MEMOIRS OF THE QUEENSLAND MUSEUM 1977. The phalangeal formula of Hemidactylus Oken, 1817 (Reptilia: Gekkonidae): acorrection and a functional explanation. Zbl, Vet. Med. C. Anat, Hist. Embryol. 6; 332-338. 1979, Parallelism and integrated design in the foot structure of gekkonine and diplodactyline geckos. Copeia 1979: 1-21. 1981, Descriptive and functional anatomy of the digital vascular system of the tokay, Gekko gecko. J. Morphol. 169: 293- 323. 1986. The morphological basis of weight-bearing in the scansors of the tokay gecko (Reptilia: Sauria). Can. J. Zool, 64: 948- 955, RUSSELL, A.P. AND BAUER, A.M. 1989. The morphology of the digits of the golden gecko, Calodactylodes aureus (Reptilia: Gekkonidae) and its implications for the occupation of tupicolous habitats. Amph.-Rept. 10: 125-140. RUSSELL, A.P. AND REWCASTLE, S.C. 1979, Digital reduction in Sitana (Reptilia: Agamidae) and the dual roles of the fifth metatarsal in lizards. Can. J. Zool, 57: 1129-1135. SCHAEFFER, B. 1941. The morphological and func- tional evolution of the tarsus in amphibians and reptiles. Bull. Amer. Mus. Nat. Hist. 78: 395- 472. SEILACHER, A. 1970. Arbeits konzept zur Konstruktionsmorphologie. Lethaia 3: 393-396. SMITH, M.A. 1935. Reptilia and Amphibia. Vol. 2. Sauria. /n Sewell, R.B.S. (ed.), ‘The fauna of British India, including Ceylon and Burma’. (Taylor and Francis: London), SNYDER, R.C. 1952. Quadrupedal and bipedal locomotion of lizards. Copeia 1952: 64-70, SOLANO G., H. 1964. Adaptive radiation in the Family Gekkonidae. Pub. Oc. Mus, Cienc. Naturales, Caracas Zool. 8: 12 pages (un- paginated). UNDERWOOD, G. 1976. Simplification and degeneration in the course of evolution of squamate reptiles. Colloq. Int. C.N.R.S. 266: 341-352. VANZOLINI, P.E. 1968a, Lagartos brasileiros da familia Gekkonidae (Sauria). Arg. Zool. S. Paulo 17; 1-84. 1968b. Geography of the South American Gek- konidae (Sauria). Arg. Zool. S. Paulo 17: 85- 112. WILLIAMS, E.E. AND PETERSON, J.A. 1982. Con- vergent and alternative designs in the digital adhesive pads of scincid lizards. Science 215: 1509-1511. ZWEERS, G.A. 1979. Explanation of structure by optimization and systemization. Neth. J. Zool. 29: 418-440. OFEDURA AND AFROEDURA (REPTILIA: GEKKONIDAE) REVISITED: SIMILARITIES OF DIGITAL DESIGN, AND CONSTRAINTS ON THE DEVELOPMENT OF MULTISCANSORIAL SUBDIGITAL PADS? ANTHONY P. RUSSELL AND AARON M. BAUER Russell, A.P. and Bauer, A.M. 1990 09 20: Oedura and Afroedura (Reptilia: Gekkonidae) revisited: similarities of digital design, and constraints on the development of multiscan- sorial subdigital pads? Memoirs of the Queensland Museum 29(2); 473-486, Brisbane. ISSN 0079-8835. The gekkonid genera Oedura (Diplodactylinae) and Afroedura (Gekkoninae) possess digits that ate very similar in external morphology. These are characterised by the possession of a large, lerminal pair of leaf-like scansors and .a series of scansor-like plates that gradually grade into the scales of the digital bases. Such genera appear to have developed an elaborate subdigital adhesive system by encroachment of the scansorial system proximally. The genera Diplodactylus and Phiyllodactylus provide potential morphotypic precursors of the digital form seen in Qedura and Afroedura respectively, Proximal encroachment of the adhesive system involves changes in external morphology, the internal muscular and tendon systems and the integument. In both Afroedura and OQedura the perceived elaboration of the adhesive system from an external perspective is nol tracked exactly by internal changes. Not all of the plates that become enlarged and hypertrophied are converted into true scansors — structures that possess some form of internal hydraulic support system in association with musculotendinous control systems and a seta-bearing integument. Only those plates that occur beneath the arcuate penultimate phalanx become elaborated into true scansors. The genera Oedura and Afroedura may both represent independent trends towards the eluboration of multiscansorial pads. They can be employed to represent a stage in a morphotypic series towards this end, but lack the fealures found in multiscansorial systems, such as subdivision of both the scansors and associated sinus system beneath the penul- timate phalanx, overlap of scansors, and the development of a free margin on the scansors, The development of such features may operate as constraints on the evolutionary elabora- tion of multiscansorial pads. CL] Gekkonidae, Oedura, Afroedura, digits, scansors, function- al morphology, evolutionary constraint. Anthony P. Russell, Department .of Biological Sciences, The University of Calgary, 2500 University Drive N.W., Calgary, Alberta, Canada T2N 1N4; Aaron M. Bauer, Biology Department, Villanova University, Villanova, Pennsylvania 19085, U.S.A.; 16 August, 1988. The similarity of digital form in the diplodac- tyline genus Oedura and the gekkonine genus Afroedura was at one time thought to be of sufficient significance to unite these taxa in a single genus (Boulenger, 1888), notwithstand- ing their great geographic separation, Not until much later (Loveridge, 1944) was this con- generic status seriously questioned, and here cer- tain external digital features were employed to advocate separation. Further credence was given to the separation of these two taxa as the sys- tematics of gekkonid lizards became better known (Underwood, 1954) and different sets of characters became employed. This led to a fur- ther evaluation of their status (Cogger, 1964) and the examination of a wide array of anatomical systems. Thus, Cogger (1964) ably demonstrated that similarities between Oedura and Afroedura were due ta convergence and brought into focus questions about why such overt similarities should be so, Despite the sub- sequent unequivocal systematic separation of gekkonine and diplodactyline geckos (Kluge, 1967a, b), however, the basis for such similarity has remained largely unstudied. A brief attempt at addressing this question was made by Russell (1979), but the consequences of particular digital design in these two genera were only partially pursued, and the potential constraints on further elaboration of this. particular design remained 47¢ Jargely undddressed. It was proposed by Russell (1979) that the digital pattern seen in Oedura has been developed from au condition similar lo that seen in the ostensibly closely related genus Diplodac- ivlus and its satellite genera Crenadacsylus and Rhyachoedura, (see King, 1987, however, foran alternative but less traditional view thal Oedura is 4 curphodactyline), The basis of this proposed evolutionary trend was that the simple, single pair of terminal, leaf-like scansors of Diplodae- ‘vive and its satellite genera had become modified into a more elaborate system by proximal recruitment of additional adhesive plutes. Some species of Diplodicrylus, such as D, strophures and DB. ciliaris possess what ap- pear to be more proximally located incipient scunsors on digits two to tive, and it was proposed (Russell, 1979) (hat such structures Were the morphological precursors of the more yroxtmal enlarged plates in Oedura. Only the Wer genus within the Diplodaclylini passesses an adhesive system incorporating more than a single pair of scansors per digit, Within the Gekkoninae the genus Afroedura possesses digits that are externally similar in form to those of Oedare. Given what is known about the phylogeny of the Diplodactylinae (Bauer, 1986) and Gekkoninae (Kluge, 1967b, 1983) the digital patlerns of these two genera appear lo have been independently evolved, with that of Afroedura being derived from a Piryl lodacnlus-like ancestor (Russell, 1972; Russell and Bauer, 1989), The similarity is not merely superficial, however, but also involves certain aspects of the internal control mechanisms of the selta-bearing plates (Russell, 1979). Such similarity is worthy of further scrutiny, as the independent development of very similar sys- tems by distantly related taxa is indicative nal only of similar selective pressures but also of the consiftints (hat potentially govern the final form that systems will take. In this context we may initially assume-that the behavioural component of the organism - environment interaction (Bock and von Wahlert, 1965) has played a major role in influcncing the morphological parameters of the system, the major dictates being Ihe way in which subdigital setae can be employed as cffec- tive agents of adhesion (Russell, 1975), Morphological systems consist of integrated sets of components that must Operate together if the entire system is to funetion (Alexander, 1975; Zaeers, 1979). In the case of subdigital adhesive systems, a VYariely of functional and MEMOIRS OF THE VUUEENSLAND MUSEUM control criteria appear to be directly correlated with the evolution of such systenis in gekkonids (Russell 1975, 1976, 1981, 1986), The genera Oedura and Afroedura provide an instructive example of how canalisation (sensu Brundin, 1968) of the evolution of morphological features is involved in the elaboration of a system, Given a particular basic morphology and a particular ‘problem’ ta be ‘solved", there is only a limited amount of scope available within a given phylogenetic lineage, The recognition of such a great degree of digi- tal similarity between Qedura and Afroedura gives cause to pose questions about the function- il reasons for such convergence, [L also prompts investigation about the evolution of the system in each lineage and the potential that the in- creased complexity of the system has. Thus, we have here employed these two genera in order to analyse the basic features of the mechanical components (Gans, 1969) employed by each and lo attempt 10 make some predictions about the integration of components in the evolution of digital adhesive mechanisms in gekkonids. Taking the digits of Oedura and Afroedura as examples, we initially postulate that the adhesive systems evolved from conditions similar to those in the supposed outgroups, Diplodactylus and Phyllodactylus respectively, Here a single pair of leaf-like scansors, a means of hyperextension of the digits and the possession of 4 device for conforming the existing scansors to the sub- stratum is present. Assuming that more proximal scansors are evolutionarily governed by the same functional concerns, the following predic- tions can be made about their elaboration from ancestrally simpler structures (subdigital scales). {i) More seales will be added to the system in 4 Sequential manner, from distal lo proximal, and these will become modified into scansors. Thus, the distalmost of the newly acquired plates will be |he most elaborate and will grade into more proximal plates that are barely distinguishable from scales, and finally mto scales themselves. True scansors will be recognisable by a com- bination of discrete characteristics, (ii}The elaboration of additional scansors will be associated with the claboration of a muscular control system, Means of application of the scan- sors to the substratum and removal of the scan- sors therefrom will be associated with specific musculolendinous networks. There may not, however, be a direct and exact correlation of the recruitment Of true scansors, the elaboration of setal fields, and the differentiation of the mus- DIGITIAL PADS OF OEDURA AND AFROEDURA cular control systems and their tendinous net- works. Thus, true scansors and subdigital lamel- jae should be distinguishable from each other on anatomical and histological grounds. MATERIALS AND METHODS Gross morphological, internal anatomical and histological features of digits of the genera Oedura and Afroedura were examined and com- pared with each other and with similar features at Diplodactylus and Phyllodactylus, Dissection material and that for histological investigation was obtained from collections housed at the Australian Museum, British Muscum (Natural History), California Academy of Sciences and the Transvaal Museum. The chicf histological procedures employed were haematoxylin and eosin, Milligan’s trichrome and Mallory’s azan, protocols for which may be found in Humason (1979). Dissection and external examination was carried out on Qedura castelnani, O, coggert, O legueuril, O. marmorata, Q. ocellata,O. robusta, O7 tryout, Afroedura karroica, A, hawaqguensis, A. tivaria, A. pondolia, A. tembulica and A. inmisvaulica, as well as ona varicty of species of Diplodactylus and Phyllodactylus. Histological examination was conducted on Qedura mar- morata, O, manilis, O. tryoni, Afraedura africana, Diplodactylus streplhurus and benavia inunguis (a satellite genus of Phyl lodactylus), RESULTS Gross EXTERNAL. MORPHOLOGY The digits of both Phyllodactylus porplyreus (hig, 1) and Diplodactylus strophurus (Fig. 2) are similar in external form in that they are [ree, relatively fat throughout their length and bear a pair af expanded, leaf-like plates al the distal end, These plates are disposed symmetrically about the claw and their bases are coincident with the articulation between the ungual and penultimate phalanx. The seales on the ventral surface uf the digit are broadly expanded and extend back far proximally on the digit, bul show ne tendency to division or to becoming sctose in any macroscopically visible sense. In both Afreedura (Fig. 3) and Qedura (Fig. 4) the digits are also essentially flat and they each bear a pair of enlarged terminal, Icaf-like plates similar to those of Phylladactylus and Dipladac- tylus (Pigs. 1,2), More proximally, however, fur- ther scrics of enlarged plates are present (hit are both divided and setose (Figs. 3,4,5,6), These plates are borne proximal to the distalmost digi- tal joint and are, therefore placed beneath the penultimate and preceding phalanges. Their number varies [ram digil to digit in Oedura (Fig. 4), with the longer digits having the greater num- ber of elaborated plates, In Oedura marmorata, for example, there are three pairs of more proximal divided plates on digit two of the pes, four pairs on digits three and four, and three pairs on digit five. Digit one bears no additional divided plates (Fig, 4), Proximal to the divided plates a short scries of gradually diminishing undivided plates merges with the plantar scales. The extent of the sctose, more proximal divided plates is less marked in Afroedura in general (Fig. 3), with an additional one or two plates located on digits two to five (Onderstall, 1984). The basi¢ arrangement is very similar to that found in Oedura, however, including a lack of elaboration of further divided plates on the first digit. As in Oedura, he more proximal scales uf the digits gradually diminish in size (Fig. 3) and merge with the plantar scales (Onderstall, 1984: Fig, 6), In both Afroedura (Fig. 3) and Oedura (Fig..4) there is a tendency for the second pair (counting from distal to proximal) of additional divided plates to be mare broadly expanded and to ex- hibit greater faleromesial separation than the others. This is consistent in all digits and repre- sents a position at the base of the penultimate phalanx (see below), INTERNAL, GROSS ANATOMY OF THE DIGITS, In both Phyllodactylus (Fig, 7) and Diplodac- tylus (Pig. 8), the internal morphology of the digits is relatively simple. In both, the inter- mediate phalanges are short, depressed and cres centi¢ distally, this being associated with the process of hyperextension (Russell 1975, 1976), In Phylladactylus (Fig. 7) the dorsal interossei muscles do not traverse any of the phalanges Neshily and do not anastomose, bul instead inser! mainly at the level of the metapodial-phalangeal joint capsule of each digit. The short digital extensor muscles control the claw and the scan- sors, but their bellies do not exiend fleshily to cross any of the phalanges (Fiz, 7), At the distal end of the penultimate phalanx the tendon of each short digital extensor divides into three. with one branch continuing to insert mid- dorsal- ly an the claw and the other two diverging to insert distally an cach scansorial plate, On the flexor surface the plantar aponcurasis sends 476 MEMOIRS OF THE QUEENSLAND MUSEUM FIG. 1. Ventral view of the left pes of Phylladactylus porphyreus (California Academy of Sciences - CAS 167593). Note the terminal, leaf-like scansors and the more proximal transversely expanded plates. FIG. 2. Ventral aspect of the left pes of Diplodactylus strophurus. (Uncatalogued specimen). Note the leaf-like scansors and the more proximal expanded plates. The dashed lines and the numbers 11-15 represent the planes of the sections depicted in Figs. 11-15. FIG. 3. Ventral view of the left pes of Afroedura hawaquensis (CAS 167638). Note the distal, leaf-like scansors and the more proximal expanded plates. FIG. 4. Ventral view of the left pes of Oedura marmorata (CAS 75405). Note the terminal, leaf-like scansors and the more proximal expanded and divided plates. FIG. 5. Ventral aspect of digit IV, left pes of Oedura marmorata (CAS 75405), showing the relative sizes of the distal scansor pair and the more proximal plates. FIG. 6. Ventral aspect of digit V, left pes of Oedura rubusta (CAS 75671), showing the setal fields on the distalmost three pairs of plates. DIGITIAL PADS OF OEDURA AND AFROEDURA branches to each of the synovial metapodial- phalangeal joint capsules. Lateral digital tendons arise at these joint capsules and insert at the distal end of the antepenultimate phalanx, The lateral digital tendons thus have no contact with the scansors, The long flexor tendon extends the entire length of the digit mid-ventrally and divides distally in the manner of the short exten- sor tendon, to serve jhe claw mid- ventrally and the proximal borders of the scansors (Fig, 7), Interdigital tendinous webs are present but rela- lively weak, while Lransverse inlermetatarsal and intermetacarpal ligaments are strongly developed. The internal structure of the digits of Diplodactvlus (Fig. 8) is architecwurally very similar to that of Phyllodactylus described above. The short digital extensors give rise to tendons at the level of the metapodial-phalangeal joint capsules, and these traverse the phalanges (which are of the same basic form as those of Phyllodactylus) to insert on the claw and dislal ends of the scansors. Ventrally the long flexor tendon controls the claw and the scansor pair in the samme manner as that of Phyllodacryltus, and the more proximal ventral plates are nol can- nected with this system. The lateral digital ten- dons extend us far distally as the distal extremity of the antepenultimate phalanx, In Afroedura (Fig. 9) the internal anatomy of the digits is somewhal more complex. The dorsal interossei muscles extend fleshily to the penul- limate phalanx, and from here send a tendinous sheet to the distal extremity of the distalmosl scansor pair. The more proximal pair of scansors also receives a tendinous sheet trom the dorsal inlerosse) muscles, The claw receives its exten- sor control from the tendon of the short digital extensor, arising from the belly of this muscle at the base of the digit. Ventrally the branches of the plantar aponeurosis insert at the synovial metapodial- phalangeal joint capsules. These capsules are also linked by the interdigital tendinous webs and the transverse intermetacarpal and inter- metatarsal ligaments. The lateral digital tendons arise from the synovial joint capsules and insert on the proximal borders of the more proximal scansor pairs, The long flexor tendon is strongly developed and serves the claw and the distalmost scansor pair, as in Phyllodactylus. Thus, the dis- talmost and more proximal scansors are control- led by different components of the flexor system. The phalanges have the same basic form as those of Phyllodaervlus, but the penultimate phalanx i» ~~ | is slightly more arcuate. Comparing the internal anatomy of the digits of Oedura with those of Dipledactylus again reveals major differences (Figs. 8,10), Here the short digital extensors have anastomosed mid~ dorsally and extended fleshily as far as the proximal end of the penullimate phalanx (Rus- sell, 1979), The architecture of the modified short digital extensors (Fig. 10) is. similar to that of the dorsal interosset of Afroedura (Fig. 9}. Here, however, the mid-dorsal tendinous raphe gives rise lo individual tendons that insert distal- ly on cach of the scansor pairs, including the distalmast, A mid-dorsal tendon also continues distally to insert on the claw. The plantar aponcurosis and associated ligamentous strands are similar to thase of Afroedura. The long flexor tendon splits distally to serve the claw and the distalmost scansor pair, as in Diplodacryivs, while the lateral digital tendons serve the more proximal scansor paits ina rnanier similar te thal in Afroedura (Figs. 9,10). The phalanges of Oedura are similar to those of Diplaductylus, bu (he penullimate phalanx is more arcuate. HISTOLOGICAL DETAILS Gross dissection of the digits of all four genera in question teveals that blood sinuses are present in the digits. The extent of these and their tributaries is only evident, however, i sections of the digits are examined. In Phyllodactylus (Dellit, 1934:Fig. 13) and Dipladactylus (Fig. 11) a sinus is present bul is Testricted to the distalmost part of the digit and is associated with the penultimate and ungual phalanges, The sinus is a palred structure distal- ly, with the ungual phalanx intervening between ils lwo halves (Fig, 11), Immediately proximal to the distal scansor pair the sinus diminishes in size (Fig. 12) and finally disappears as a discrete structure in the hinge region between the scansor bases and the next most proximal plate (Fig. 15). The next most proximal plate is undivided and beats a smaller, bul none the less distinct, ex- panded sinus that is mostly concentrated over the central part of the plate (Fig. 14). This plate resides beneath the penultimate phalanx, The next more proximal plate is borne beneath the antepenultimate phalanx. lt ts single but shows some sign of incipient division centrally (Fig. 15). This plate bears selae but appears to be cushioned primarily by vacuolar adipose tissuc (Fig. 15). The difference in size of the sinuses of the distalmost and hext More proximal plates can be more fully appreciated in longitudinal section 478 MEMOIRS OF THE QUEENSLAND MUSEUM FIG. 7. Diagrammatic representation of the musculoskeletal system of digit IV of the pes of Diplodactylus. The block-like structures represent the metatarsal and the phalanges, those with chevron-shaped ends being the phalanges involved in hyperextension. The muscles are represented by black lines, with arrow heads indicating their insertions. The large, black arrow represents the relatively mild curvature of the penultimate phalanx. Blood sinuses are represented by stars, the primary one being solid and the secandary one open. The scansors are stippled. The hoop-like structure connecting the metatarsal (the large block to the far left) with the first phalanx (the block lo the immediate right of this) represents the metatarsophalangeal joint capsule. Abbrevia- tions: edb, extensor digitorum brevis; fdb, flexor digitorum brevis; fdl, flexor digitorum longus; id, interossei dorsales; pl/Idt, plantar aponeurosis/lateral digital tendon continuum. The small, vertical arrow represents the distal extent of the fleshy belly of the extensor digitorum brevis. FIG. 8. Diagrammatic representation of the musculoskeletal system of digit IV of the pes of Phylloductylus. The symbolism and abbreviations are as for those of Fig. 7. FIG. 9. Diagrammatic representation of the musculoskeletal system of digit IV of the pes of Afroedura. The symbolism and abbreviations are as for those of Fig. 7 except for the large, black arrow that represents a greater curvature of the penultimate phalanx. FIG. 10. Diagrammatic representation of the musculoskeletal system of digit TV of the pes of Oedura. The symbolism and abbreviations are as for those of Fig. 9. Note thal in Oedura the extensor digitorum brevis controls the hyperextension of the scansors, while in Afroedura this is achieved by the interossei dorsales. DIGITIAL PADS OF OEDURA AND AFROEDURA (Fig. 16), together with the absence of any ap- preciable sinus development in the most distal of the plates beneath the antepenultimate phalanx, The penultimate plate may be considered as an incipient scansor. The stratum compactum of the dermis of the distalmost plates is strongly developed at their bases (Fig. 17), This is associated with the long, flexor tendon, Jn the more proximal plates there is incipient development of a strong basal stratum compactum, especially in the most distal ofthe single plates, As one passes proximally the plates become less well-developed in this regard (Fig. 17), although they are cushioned by adipose-likc tissue. The more proximal manifes- tations of the stratum compactum may be as- sociated with the lateral digital lendons. In Oedura and Afroedura the extent of the setal fields ts greater than it is in Phylladactylus and Diplodactylus, and to some extent this ts reflected in the internal anatomy. In Qedura monilis the sinus associated with the distalmost pair of plates is massive and single, but sends branches into both sides of the terminal pad (Fig. [8). This. 1s associated with what appears to be vertically stacked columns of smooth muscle associated with the walls of the sinuses (Fig, 18). Further proximally the sinus diminishes in size and then appears again as an appreciable expan- sion in the next most proximal, and divided, plate. This pair is borne beneath the penullimate phalanx. Here, however, the sinus is somewhat smaller, more diffuse and more markedly paired (Fig. 19). A large amount of fibrous connective tissue is present centrally, above the division between [he two halves of the penultimate plate, As in Diplodactylus, only the distalmost.and the next mast proximal plites are housed ventral ta the penultimate phalanx. The next more proximal pair of plates is present beneath the antepenultimate phalanx (Fig. 20), and here the involvement of the vascular system is minimal, In longitudinal section the diminution in extent of involvement of the vascular system in the plates is evident (Fig. 21), as isthe arcuate nature of the penultimate phalanx and the basal development of the stratum compactum in the scunsors proper and the more proximal plates. More laterally the association of the columns of smooth musele cells with the sinus in the distal- most scansor pair, the extent of the sinus system in the next most proximal scansor pair and the sharp demmreahun between (hese and the moze proximal plates can be seen (Fig, 22). SPTAL DIFFEREN TIA FON, Comparing Diplodactylus with Oedura {anu Phyllodactylus with Afroedura) it can be secn that setac are associated with subdigital plates that exhibit a hypertrophied epidermis (Figs. 16,21). There is a sharp demarcation between these plates and more proximal, typical scales, Viewed with the light microscope there does not appear to be a significant diminution in size of the setae from distal to proximal, even though the internal structure of the plates that bear (hem differs considerably, ‘lhe more distal plates are associated with the blood sinus system while the more proximal ones have little or na such as- sociation and are instead filled with vacuolar, adipose-type tissue. Those plates exhibiting the seta-bearing (Schleich and Kastle, 1986: plate 6, fig. 3), hypertrophied epidermis occur beneath the hyperextensible phalanges (phalanges two (u five in digit four) (Figs. 16,21). A comparative survey of the potential differentiation of spines, spikes, prongs and setae on the various subdigital scales and plates, similar to that performed by Peterson (1983) for anoline iguanids, remains to be carried out. DISCUSSION The data outlined above make it possible to make some deductions about the development of a subdigital, adhesive apparatus and about the distinction between scansors and “lamellae”. The scenario outlined by Russell (1976) that lincages that initially develop distal scansors may expand the subdigital pads by extending the scansor scries proximally ts the premise upon which the deductions below are based. Comparisons of Phyllodacrylus and Afroedura from the Gekkoninae and Diplodactylus and Oedura from the Diplodactylinac provide nwo independent morphotypic series In Which trends can be compured. In both there is a tendency to elaborate the size of the proximal plates on digits two to five, to divide them and to increase the seli-bearing surface area, This may be poste lated to be a means of Increasing adhesive ef- ficiency, although empirical tests of this in standardised conditions, or rigorous com. parisons of the details and physical properties of preferred locomotor substrata are not available (Bauer and Good, 1986). In both putative lineages an adhesive system is prinvilively present, relying on the employment of a pair of distinet and enlarged distal leaves. These are morphologically sharply demarcated fram the 480 more proximal digital plates. The digits of both Phyllodactylus and Diplodactylus are capable of hyperextension (Russell, pers. obs.) and this ac- tivity is practised during normal locomotion. The musculotendinous systems that bring these movements about are relatively simple, but pos- sess all of the basic requisites deemed to be necessary for such activity (Russell 1975, 1976). The distal plates are true scansors (Russell 1981), as adjudged by their possession of setae, a mechanism enabling them to be hyperex- tended, and an internal hydraulic device that permits these plates to conform to irregularities MEMOIRS OF THE QUEENSLAND MUSEUM of the substratum and thus make optimal contact (Russell, 1981). In this case the hydraulic device is present as a vascular sinus system, but other mechanisms also exist (Russell, 1972, 1979). More proximally the next plate may be categorised as an incipient scansor. It bears setae, may be hyperextended, but possesses only a relatively small sinus system, making confor- mation with the substratum potentially less ef- fective. This more proximal plate is borne beneath the penultimate phalanx, which arches away from the substratum and permits the hous- ing of the central component of the sinus system DIGITIAL PADS OF QEPURA AND AFROEDURA beneath it (Russell, 1981), Proximal to this the plates do not possess the properties of true scan- sors in as much as the sinus system, orsome other hydrostatic device that can be pressurised, is not incorporated. ‘This appears to be duc largely to the morphology of the intermediate phalanges (phalanges two and three in digit four) that are associated with hyperextension, Typically these are depressed and wide (Russell, 1975) and remain firmly adpressed to the substratum (Rus- sell, 1976). Their morphology precludes the in- corporation of a centralised sinus or similar device beneath them. The presence of such phalanges in forms that develop distal scansors may later act as a constraint on the potential for further digital modification as they may preclude the development of true scansors beneath them (see below). Forms that have developed scansors. from the base of the digits distally have a dif- ferent constructional arrangement of the inter- mediate phalanges (Russell, 1977). Further elaboration of the contro! systems of the digits inAfroedura and Oedura, as compared to their putative morphotypic precursors, has not brought about a concomitant increase in the number of scansors, if the criteria for their recog- nition, as outlined above, are applied. Here the mechanisms of application of the subdigital 4a} plates to the substratum and their removal from it can apparently be more precisely controlled, as deduced from (he increased complexity of the musculoteninous features of the digits. Ex- amination of histological detail reveals, bow- ever, that although the sinus system remains elaborate in the distal pair of plates and hus became more cluborate in the next most proximal pair, it is not evident proximal to this. Thus, although externally the yet more proximal plates have become more prominent and mare scansor-like (more so in Gedura than Afroedura) in appearance, their internal differentiation is not so marked, Thus, tn the proximal cocroachmeut of subdigital pads from a distal beginning, more than external elaboration is required to convert these structures into fully-differentiated scat sors. It appears that one constraining parameter is the ability to incorporate a sinus sysicm into a d consisting of multiple scansors. Here, the incorporation of scansors beneath the penul- timate phalanx becomes eritical, This phenomenon was noted for the tokay (Gekko gecko) by Russell (1981: Fig. 7). Here, multiple scansors are present beneath the penultimate phalanx and these possess branches of an elaborate sinus system, Proximal to this the lamellae bear sctae but are morphologically FIG, (1, Cross section of digit IV, Jeft pes of Diplodactylus strophurus through the terminal, leaf-I|ke scansors al the base of the claw, See Figs, 2 und 16 for position of section, The sinus (s) is large and is positioned ventral to (he ungual. phalanx (up). Al this point the sinus is paired, one half being associated with éach leaf of the scansar (scan). The scale bar = 0.25mm. Mallory’s azan stain, FIG. 12. Cross séction of digit 1V, left pes of Diplodactylus strophurus through the base of the distal scanser pair, See Figs. 2 and 16 for position of the section, At this point the sinus (5) is paired but smaller and Lies beneath the distal cartilaginous epiphysis (ep) of the penultimate phalanx. The section is sotvewhat ublique, with the right leaf'and sinus being represented more proximally than the lef}, where the scansor base (scan) is Sul visible. The scale bar = 0.25mmm. Mallory’s azan stain. FIG, 13, Cross section of digit TV, left pes of Diplodactylus strophurus through the hinge region between the distal scansor pair and the next most proximal plate. See Figs. 2 and 16 for position of the section. The sinus (s) is much reduced at this point and is present only as a connecting channel between the scansors and the penultimate plate, The penultimale phalanx (pp) is depressed and transversely widened at this point. The scale har =0,25mm, Mallory’s azan stain, FIG. 14. Cross section of digit lV, left pes of Diplodactylus strophurus through the undivided penultimate plate. See Figs. 2 and 16 for position of the section. The sinus (s) is expanded again and is present beneath the penultimate phalanx (pp) and associated tendon of the flexor digitorum longus (fdl), The incipient scansar {incip, scan) bears no central thinned area. The scale bar = 0.25mm. Mallury’s azan stain. FIG, 15, Cross section of digit FV, left pes of Diplodactylus strophurus through the antepenultimate plate, beneath the anteperultimate phalanx (ap). See Figs. 2 and 16 for position of the section. The sinus is nol evident and the lamella (lam) is cushioned by vacuolar adipose tissue (at). The tunnel for the tendon of the flexor digitorum longus (fdl) is evident, The scale bur = 0.25mm, Mallory's azan stain. FIG. 16. Longitudinal section of digit TV, loft pes of Dipladactylus strophurus showing the relationships of the sinus (s), Scansors (scan), incipienl scansor (incip. scan), lamellae (lam), ungual (up), penultimate (pp) and jmepenultimate (ap) phalanges. The dashed lines und the numbers 11-15 represent the planes of the sections ilepicted in Figs. 11-15. The scale bar = (0.5mm, Mallory’s azan stain much less elaborate, bearing a closer tesemblance to true scales than do the scansors. Again there is probably a functional correlate of this differentiation, as the scansors are more likely to be able to maximise contact with the substratum (on at least some locomotor surfaces) due to their additional compliance. In the tokay (and many other geckos having multiple scan- sors) the adhesive properties are probably further enhanced by the overlapping nature of the scan- sors (Fig. 23) and the sequential effect of the reticular network of the sinus system of one scansor on the setae of the next most proximal MEMOIRS OF THE QUEENSLAND MUSEUM scansor (Russell, 1981). In Oedura and Afroedura the scansors do not overlap and do not have the free border typical of those of the tokay (Fig. 23), so again the adhesive efficiency is potentially limited. The digital conditions in Oedura and Afroedura appear to be indicative of a transition- al phase (in the sense of evolutionary morphol- ogy tather than phylogeny) between a single scansor system and a multiscansorial system. These taxa do, therefore, provide us with some insight as to how the latter may have evolved from a morphologically less complex condition. DIGITIAL PADS OF CEDURA AND AFROEDURA The transition as visualised in a morphotypic sequence is nol smooth, however, and the dif- ferences between the Gekko system (typifying multiscansorial pads) and the Oedura and Afroedura systems are quite marked. The dif- ferences seen may be indicative of a syndrome of characters that are synonymous with the evolution of a multiscansorial system, and func- tional constraints may operate either to govern the potential transition or to prevent it. The se- quence by which multiscansorial pads arose is not known, but exantination of the condition of the digits in Oedura and Afroedura provides one means of attempting to understand how this may have come about, Further anatomical and his- tological investigations of other genera will be helpful in attempting to assess the feasibility of the praposed scheme. A similar scheme erected for anoline iguanids (Peterson, 1983) suggests a similar sel of related morphological events, al- though her assessment was based entirely on external features. Given what can be potentially deduced about digital form and evolution in the Diplodactvlus ~ Oedura and Phyllodactylus - Afroedura mor- 453 photypic sequences, the following represents a relative chronology of changes in the adhesive system based upon the constraints as outlined, This chronology attempts to firstly account for the transition from a single pair of terminal, leaf-like scansors to the elaboration of more proximal pairs, as exemplified by Oedura and Afroedura. Secondly it allempts to explain how the latter conformation may be morphologically extended into a multiscansorial system (or alter- natively how it may be limited by the magnitude of the morphological paps that are evident). 1,The pattern begins with a pair of terminal, leaf-like scansors with extensive sctal fields and a well-developed blood sinus system. The digit is able to be hyperextended, The proximal part of the digit is flat and the transversely widened plates beneath arc endowed with setae as far proximally as the distal end of the first phalanx. The most distal widened plate is positioned ventral to the penultimate phalanx and possesses an incipient sinus. This is the condition ex- emplified by Phyllodactylus and Diplodactylus. This stage suggests that seta-bearing lamellie precede the development of true scansors both FIG. 17. Longitudinal section of digit [V, left pes of Diplodactylus strophurus showing the stratum campactum of the dermis of the scansors and penultimate plate (stnall arrows), this being associated with the long flexur tendon (large arrows). The scale bar = 0.5mm. Mallory’s azan stain. FIG, 18, Cross section of digit TV, lef pes of Ocdura monilix through the terminal, leaf-like scansor pair. See Fig, 21 far position of the section. The large central sinus (s) is evident, together with its associated smooth muscle columns (sm). The sinus resides beneath the ungual phalanx (up) and cushions the scansors (scan). The scale bar = 0.5mm, Mallory’s azan stain, FIG. 19. Cross section of digit IV, left pes of Qedura monilis through the penultimate scansor pair. See Fig. 21 for position of the section. The sinus (s) is parted and smaller than that of the distalmost scansor pair. The penultimate scansor palit (scan) resides beneath the penultimate phalanx (pp), Fibrous connective tissue (fet) fills the space above the cleft between the leaves of he scansor pair. The scale bar = 0.5mm. Mallory’s azan stain. FIG. 20, Cross section of digit I'V, left pes of Oedura monilis through the antepenultimate pair of plates. See Fig. 21 for position of the section. The lamellae (Jam) are cushioned only by adipose tissue (at) and reside beneath the antepenultimate phalanx (ap) and ils associated tendon of the Nexor digitorum longus (fdl). The scale bar = 0.5mm. Mallory's azan stain. FIG. 21. Longitudinal section of digit IV, left pes of Oedura monilis showing the relationships of the sinus (5), scansors (scan), lamellae (lam), ungual (up), penultimate (pp) and antepenultiniale (ap) phalanges. The stratum compacium of the dermis.of the scansors (smal! arrows) is evident, That of the distal scansor pair is associated with the tendon of the flexor digitorum longus (large arraw), while that of the penultimate scansor pair is associated with the lateral digital tendon (oul of the plane of the section, bul see Fig. 22), The dashed lings and the numbers 18-20 represent the planes af the sections depicted in Figs, 18-20, The scale bar = 0.5mm, Mallory’s azan stain. FIG. 22. Parasagittal section of digit [V, left pes of Oedara monilis, showing the columns of smooth muscle (sm) associated with the distal scansor pair and the extent of the sinus (s) of the penultimate seansor pair. The lateral digital tendon (dt) and its association with the stratum comipactum of the penultimate scansor pair is evident, Note the absence of a free distal margin on the scansors (compare with Fig. 23) and the sharp demarcation between the scansors and the more proximal lamellae, The scale bar = 0.5mm, Mallory's azan stain. 484 MEMOIRS OF THE QUEENSLAND MUSEUM FIG. 23. Scansor and lamellar differentiation in Gekko gecko. A, represents a longitudinal section through the proximal region of digit I'V, lefi pes. Here the subdigital plates bear short setae but lack a free distal margin and an involvement of the sinus system. B. represents a longitudinal section through the region of the antepenultimate phalanx. Here the free distal margin is slightly developed and the sinus system is partially involved, C, represents the region of fully differentiated scansors beneath the penultimate phalanx. Note the multiple scansors in this region, the extensive involvement of the sinus system, the extensive overlap of the scansors and the extensive free distal margin on each scansor. Abbreviations: ap, antepenultimate phalanx; fdm, free distal margin; lam, lamellae; pp, penultimate phalanx; s, sinus; scan, scansors. evolutionarily and positionally. This suggests that adhesive systems begin by employing ex- panded plates with setose surfaces (lamellae). Subsequently these may be modified to become rue scansors, and this may increase adhesive efficiency. 2.The adhesive system is elaborated by a proximal encroachment of scansor-like plates, enhancing the possibility of surface contact by elaborating the core of the plate by the develop- ment of vacuolar, adipose-like tissue. The distal scansors remain well-developed and the next most proximal plate gives rise to leaves that possess a greater elaboration of the sinus system. The penultimate phalanx is more arcuate and true scansors reside beneath this. Proximal to this the scansor-like plates are enlarged and prominent but lack branches of the sinus system. The proximal plates may thus increase the ad- hesive power of the digits, but they do not have the flexibility for conformity with the locomotor substratum that the true scansors do, with their sinus systems. The distal scansor pair is control- led in its flexor aspect in the same way as in DIGITIAL PADS OF OEDURA AND AFROEDURA Diplodactylus and Phyllodactylus, while the more proximal plates receive their flexar contro! from the lateral digital tendons. Only the mose distal of these is a true scansor, The extensor control of the plates has been taken over by the distal migration of muscle bellies into the digit proper, and these traverse all but the penultimate and distalmost phalanges, The scansors and mare proximal plates are controlled on the ex- tensor surface in essentially the same way, This is essentially the condition in Afroedura and Oedura, although in the former the plates have not progressed as far proximally as they have in the latter, 3.The trend so far outlined appears to be one \hat is leading towards the elaboration of multi- scansorial pads, but both Afroedura and Oedura lack certain features that are typical of a multi- scansorial system. In order to continue the trend io give rise lo such a pad the following changes would be predicted to take place (based on a comparative survey of scansor and pad design in general): The distal scansor pair would become reduced in size, giving the claw a greater degree of freedom fram the pad. In association with this the scansor pair beneath the penultimate phalanx would become subdivided as the penullimate phalanx became more arcuate (This is seen also in the gekkonine genus Calodactylodes). The sinus system would become subdivided, with the reticular networks of the scansors being confined to their bases. As the scansors subdivided there would be a trend to the elaboration of a free, distal, seta-bearing margin thal Would become overlain by the reticular network of the next most distal scansor. True scansors would be restricted to the area bencath the penultimate phalanx. Proximal to this, enlarged plates would bear ficlds of sctae but would not possess branches of the sinus system. These plates would lack the free margin and would remain essentially similar to the more proximal plates seen in Dipledac- tylus and Oedura, The constraints of the incor- poration of components of the venous sinus system and the restriction to elaboration of scan- sors to the region immediately below or adjacent io the penultimate phalanx dictate to a large degree how the system as seen in Qedura and Afroedura may be further elaborated. Cogger (1964) was intrigued with the phenetic similarity of, but apparent lack of phylogenctic affinity between, Oedura and Afroedura, He carefully documented the similarities and dit- ferences between these Iwo taxa end correctly surmised thatthe similarities were due to conver- 485 gence. Although Cogger (1964) Indicated thal similarities of fool structure were superficial, they are. in Fact, quite extensive. The arrange- ment of components and their integration indi- cate that the independent acquisition of this basic pallern has been governed by very similar selec- tive factors relating to the functional control of the scansor sysiem. Both genera appear to ex- hibit a morphologically intermediate condition between a relatively simple adhesive system and one that is considerably more complex, The potential limits and constraints on both genera, in terms of further elaboration of their subdigital adhesive apparatus, appear to be rather similar and represent an cxample of how functional demands can potentially canalise (Brundin, 1968) an evolving system. ACKNOWLEDGEMENTS We thank Darcy Rae for the preparation of the histological material used in this study, Specimens examined were made available from the collections of the Australian Museum (AMS), the British Museum (Natural History) {(BMNH), California Academy of Sciences (CAS), and the Transvaal Museum (TM). We thank Allen Greer, Nick Arnold, Robert Drewes and Wulf Haacke respectively for the loan of spirit preserved and skeletal material, and for giving permission to dissect specimens and prepare material for histological examination. Merle Marsden Jr provided general assistance in the laboratory, Financial support for the comple- tion of this work and to assist in travel to Bris- bane to present this contribution at the Australian Herpetological Conference was provided in part by a Natural Sciences and Engineering Research Council of Canada grant (No. A9745) ta A_P.R. and in part by 4 University of Calgary postdoc- toral fellowship to A.M,B. The manuscript was typed by Susan Stauffer. LITERATURE CITED ALEXANDER, R. MCN. 1975. Evolution of in- tegrated design, Amer, Zool, 15; 419-425, BAUER, A.M. L986. Systematics, biogeography and evolutionary morphology of the Carphodac- tylini (Reptilia: Gekkonidae). (Unpublished Ph.D. thesis, University of California, Berkeley). BAUER, AM. AND GOOD, D.A_ 1986. Sealing af scansorial surface atea in the genus Gekko, pp, 363-366, In Rocek, Z, (ed.), ‘Studies in 486 herpetology’. (Charles University: Prague). BOCK, W.J. AND VON WAHLERT, G. 1965, Adap- tation and the form- function complex. Evolu- tion 19; 269-299, BOULENGER, G.A. 1888. On new and little-known South African reptiles. Ann. Mag. Nat, Hist. (6)2: 136-141. BRUNDIN. L. 1968. Application of phylogenetic principles in systematics and evolutionary theory, pp, 473-495, /n Orvig, T. (ed.), ‘Current problems of lower vertebrale phylogeny, (Inter- serence; New York). COGGER, H.G. 1964. The comparative asteology and systematic status of the gekkonid genera Afroedura loveridge and Qedura Gray, Proe, Linn. Soc. N,S.W. 89; 364-372, DELLIT, W.-D. 1934. Zur Anatomie und Physiologie der Geckozehe, Jena Z, Naturwiss. 68: 613-656. GANS, C. 1969. Functional components versus mechanical units in descriptive morphology. J. Morphol. 128: 365-368, HUMASON, G.L. 1979. ‘Animal tissue techniques’. (W.H, Freeman and Co,: San Francisco). KING, M. 1987, Chromosomal evolution in the Diplodactylinae (Gekkonidae: Reptilia). 1. Evolutionary relationships and patterns of change. Aust, J. Zool. 35: 507-531. KLUGE, A,G, 19674. Systematics, phylogeny. and coogeography of the lizard genus Diplodactylus Gray (Gekkonidae). Aust. J. Zool, 15. L007- 1108, 1967b. Higher taxonomic categories of gekkonid lizards and their evolution, Bull, Amer. Mus, Nat. Hist. 135: 1-60. 1983. Cladistic relationships among gekkontd lizards, Copeia 1983: 405-475. LOVERIDGE, A, 1944. New geckos of the genera Afroedura, new genus, and Pachydacrv/us from Angola. Amer. Mus. Novitates [254: 1-4. ONDERSTALL., D. 1984. Descriptions of two new subspecies of Afroedura pondolia (Hewitt) and a discussion of species groups within the genus (Reptilia: Gekkonidae). Ann. Transvaal Mus. 33: 497-509, PETERSON, J.A, 1983. The evolution of the subdigi- MEMOIRS OF THE QUEENSLAND MUSEUM tal pad in Anolis. J. Comparisons among the anoline genera. pp. 245- 283. /n Rhodin, A.G.J., and Miyata, K. (eds), ‘Advances in herpetology and evolutionary biology’. (Harvard Universily Press: Cambridge, Massachusetts). RUSSELL, A.P. 1972, ‘The foot of gekkonid lizards: a study in comparative and functional anatomy’. (Unpublished Ph.D. thesis, Universily of Lon- don), 1975. A contribution to the functional analysis of the foot of the tokay, Gekko gecko (Reptilia: Gekkonidae). J. Zool. Lond. 176: 437-476. 1976. Some comments concerning inlerrelalion- ships amongst gekkonine geckos. pp. 217-244. /n Bellairs, A, d'A. and Cox, C_B. (eds), ‘Mor- phology and biology of reptiles’. (Academic Press; London). 1977. The phalangeal formula of Hemidacnrylus Oken. 1817 (Reptilia: Gekkonidae); a correction and a functional explanation. Zbl. Vet. Med. C. Anal. Hist. Embryol, 6; 332-338. 1979. Parallelism and integraled design in the fool structure of gekkonine and diplodactyline geckos. Copeia 1979; 1-21, 1981. Descriptive and functional anatomy of the digital vascular system of the tokay, Gekka gecko. J. Morphol. 169; 293-323. |}986, The morphological basis of weight-bearing in the scansors of the lokay gecko (Reptilia: Sauria), Can. J. Zool. 64: 948-955. RUSSELL, A.P. AND BAUER, A.M. 1989. The morphology of the digits of the golden gecko, Calodactylodes aureus (Reptilia: Gekkonidae) und its implications for the occupation of rupicolous habitats, Amph.-Rept, 10:125-140. SCHLEICH, H.H. AND KASTLE, W. 1986. Ultrastruktaren an Gecko- Zehen (Reptilia: Sauria: Gekkonidae). Amph.-Rept. 7: 141-146. UNDERWOOD, G. 1954, On the classification and evolution of geckos. Proc. Zool. Sac. Lond. 124: 469-492. ZWEERS, G.A. 1979, Explanation of structure by optimization and systemization. Neth. J. Zool. 29; 418-440. THE SCINCID LIZARD GENUS NANNOSCINCUS GUNTHER: A REVALUATION R.A. SADLIER Sadlier, R.A. 1990 09 20: The scincid lizard genus Nannoscincus Ginther: a revaluation. Memoirs of the Queensland Museum 29(2): 487-494. Brisbane. ISSN 0079-8835. A combination of skeletal, scalation, coloration and reproductive characteristics are used lo analyse the intrageneric relationships of the species included in Nannascincus (Sadlier, 1987), which is here shown to comprise 2 subgenera. Proposed is a monotypic subgenus Nannoseps n, subgen, for the Australian species N, maeecoyi (Lucas and Frost). The New Caledonian species N. muriei (Bavay), N. gracilis (Bavay), N. sleveni (Loveridge), N, rankini Sadlier and N. greeri Sadlier comprise the nominate subgenus, Within the subgenus Nannoscincus there appear to be 2 distinct species groups, the N. mariei species group (including N, mariei, N. grecriand tentatively N. rankini) and the N. gracilis species group (including N. gracilis and N. sleveni), () Scineidae, Nannoscincus, phylogeny, Australia, New Caledonia Ross A. Sadlier, The Australian Museum, P.O. Bax A285, Sydney South, NSW 2000, Australia; 9 July, 1990. Greer (1979) diagnosed 3 major lineages for the Australian scincid lizard fauna, the Egernia, Sphenomorphus and Eugongylus groups. These groups ure widespread within Australia and also include most scincid genera in the Pacific region. Outside Australia, the Egernia and Sphenomor- phus groups are distributed mainly over the In- donesian archipelago east to the Solomon Islands; members of the Sphenomorphus group in particular are prominent in closed forest habitats. By contrast, the distribution of the Eugongylus group outside of Australia is mainly over the Pacific islands to the cast of Australia (including New Guinea) and is poorly repre- sented in the Indonesian archipelago. Within the Eugongylus group (Greer, 1990) there is a distinc! subgroup of species that share a derived character state unique within lygzosomine skinks. This subgroup is diagnosed by having the atlantal arches of the first cervical verlebrue fused to the intercentrum. Within this subgroup a subset of species share a pattern of phalangeal reduction on the 4th digit of the manus nol observed in other Eugongylas group members. This subset of species comprises the genus Nannoscincus and includes: Anotis mariet Bavay, 1869; Lyeosoma gracilis Bavay, 1869; Saiphos maccoyi Lucas and Frost, 1894; Lygosama slevent Loveridge, 1941, Nannosein- cus rankint Sadlier, 1987; Nannoscincus greert Sadlicr, 1987. In addition to the pattern of phalangeal loss all members of this subgroup are small (maximum snout to vent length of SOmm in maccoyi the largest species) with elongate bodies and reduced Jimbs which fail to meel when adpressed to the body, They gencrally occur in closed forest or montane habitats, shel- tering beneath and within rotting logs or under stones, or within the fine, loose superficial sub- strate beneath these sheltering sites, SYSTEMATICS Greer (1974) in reviewing Letalapisma and associated species identified 2 groups (Groups and IIL of thal wark) which essentially comprise what is now regarded as the Eugongylus group (Greer, 1979). Greer( 1974) was however unclear as to whether Anagiis Bavay (a Group [1 member at that time comprising the Australian species A, maccoy! (Lucas and Frost), A. graciloides (Lénnberg and Anderson), and the New Caledonian species A, mariet Bavay.A. gracilis (Bavay), and A. sleveni (Loveridge)) was monophyletic or palyphyletic. Czechura (1981) noted that Anotis Bavay was preoccupied, and resurrected Nannoscincuy Ginther to replace it. A review of the New Caledonian scincids by Sadlier (1987) redefined Nannoscincus largely on the basis of the paticrn of phalange reduction in the 4th digit of the manus. Lygosoma graciloides with a pattern of phalange reduction in the Ist digit only was removed, and the genus then comprised the species NV. gracilis, N. mariei, N. sleveni, N, rankini, N. greeri and N. maccoyt. Note that Greer (1982) further defined Geomyer- sia When deseribing a second species in the 488 genus, G. coggeri; he listed as one of the diag- nostic features a phalangeal formula similar to Nannoscincus. Re-examination of Geomyersia shows it to have a primitive phalangeal formula and is for this reason not considered further here. Subsequent research on the species of Nannos- cincus, particularly osteology and soft morphol- ogy, has established the Australian species Saiphos maccoyi Lucas and Frost as a sister group warranting subgeneric recognition within a redefined Nannoscincus that also recognises the New Caledonian species as a monophyletic subgenus. METHODS AND MATERIALS Scalation and reproductive characteristics were assessed on whole alcoholic specimens. Phalange and presacral vertebrae condition were assessed from x-rays of selected samples. Ver- tebral and sternal characters were assessed from a combination of cleared and stained and whole alcoholic specimens, and skull characters from a combination of cleared and stained and skeletal preparations. Coloration characteristics were determined from my field observations. Polarities for characters, unless otherwise stated, are those used by Greer in determining relation- ships between lygosomine skinks, otherwise the primitive state is considered the widespread con- dition in the primitive Eugongylus group species. MEMOIRS OF THE QUEENSLAND MUSEUM Sources used in assigning character polarities are as follows: character state A, Greer (1974); char- acter state B, this work; character state C, Greer (1974); character state D, Greer (1983); charac- ter state E, this work; character state F, this work and Greer (1974); character state G, Lecuru (1968 :524, fig. 8a and 8b); character states H and I, Romer (1956) as cited in Greer (1983); character state J, Hoffstetter and Gasc as cited in Greer (1983); character state K, Romer (1956); character state L, Greer (1987). EVALUATION AND DISTRIBUTION OF CHARACTERS The following 13 characters were used in in- ferring relationships between species of Nannos- cincus. A. Prefrontal scales. Primitively the prefron- tals of lygosomine skinks are moderately large and either in contact medially or narrowly- moderately separated (A). From this primitive condition the prefrontals can be lost in 2 ways, either through fusion to the frontonasal (a1) or through diminution (a2). In N. maccoyi the prefrontals are absent. Loss of the prefrontal scales through fusion (a1) in N. maccoyi is indicated by a single ‘anomalous’ specimen from Bendigo, Victoria (Fig. 1a) in which the prefrontals are distinct and in broad contact, whereas in all other N. maccoyi ex- amined the prefrontals are absent (Fig. 1b) but retain an undulating shape to the frontonasal- FIG. 1. Dorsal views of the head scalation of: (a) aberrant N. maccoyi (note broadly contacting prefrontal scales); (b) lectotype of Saiphos maccoyi Lucas and Frost (NMV D1851); (c) N. mariei. THE SCINCID LIZARD GENUS NANNOSCINCUS frontal suture characteristic of species with well developed prefrontals (and also in the specimen of N. maccoyi from Bendigo mentioned above). The prefrontals in the remaining species of Nan- noscincus appear to have been reduced by diminution (Fig. 1c) as indicated by their small size and obvious separation. B. Contact between the Ist supraciliary and frontal scales, Contact between the prefrontal and Ist supraocular scales is considered the primitive (B) condition, Contact between the Ist supraciliary and frontal scale, thereby excluding contact betwcen the prefrontal and Ist supraocular is considered derived (b). Contact between the prefrontal and Ist supraocular occurs in N. maccoyi, and N. mariei, while N. gracilis, N. sleveni, N. greeri and N. rankini generally have the derived condition, C. Frontoparictal scales, In the primitive con- dition the frontoparietals are present as 2 distinct scales (C). Fusion of these scales along the mid- line to form a single scale 1s considered to be derived (c). The primitive condition occurs in N. maccoyi, N. mariei, N. gracilis, and N. sleveni, while the derived condition occurs. in N. greeri and N. rankini, D. Loreal scales. The presence of 2 distinct loreal scales between the nasal and preocular scales is considered to be the primitive condition for lygosomine skinks (D). In most primitive lygasomine skinks the anterior loreal is either square or slightly higher than wide, while the FIG, 2. Lateral view of the head of; (4) N, sleveni with the lower eyelid in the tuised position showing the ‘scaled’ lower eyelid typical of N, gracilis, N. sleveni and N, mariei; (b).N. greeri with the lower eyelid in the raised position showing the windowed condition typical of MN. maccoyi, N, greeri, and N. rankinl; (c) detail of the scalation of (a) above. 489 posterior loreal is either square or slightly wider than high, The primitive loreal condition in Nannoscin- eus occurs in N. gracilis and N. sleveni which have 2 reduced loreal scales (D), the anterior usually present as a semilunar scale positioned on the posterodorsal margin of the nasal and failing to contact the labials, and the posterior usually as high as the nasal but noticeably wider dorsally than basally, N. maccoyi, N. mariei, N. greeriandN., rankiniall have asingle loreal scale which is considered to be derived (d). E. Lower labial scales. Most generally primi- live Eugongylus group species have 6 lower labials which is considered the primitive condi- tion (E). Reduction in the number of lower labials, is considered to be derived (e). N. maccoyi, N. gracilis , and N. sleveni have the primitive condition of 6 lower labial scales, while N. mariei, N. greeri and N. rankini have 5 lower labials (c). F. Lower eyelid morphology. A scaled lower eyelid is considered primitive for lygosomine skinks, and derivations from this condition derived. The lower evelid of N. maccoyi, N. greeri (Fig. 2b) and N. rankini has a semi-transparent disc below a distinct palpebral rim. N. mariei, N, gracilis and N.sleveni (Fig, 2a and 2c) lack su- tures defining the palpebral rim and have the opaque central area of the lower eyelid divided by fine transverse sutures only (‘scaled’). It is however unclear whether these conditions 490 ——— Vine MEMOIRS OF THE QUEENSLAND MUSEUM i yh ff C4 y 13 (( ft Jew FIG. 3. The various types of mesosternal rib attachment found in the Nannoscincus subgroup: (a) N. maccoyi; (b) N. mariei, typical of N. mariei and N. greeri;(c)N. gracilis, typical of N. gracilis, N. sleveni, and N, rankini. (above) for Nannoscincus represent: 1) 2 inde- pendent patterns of evolution from an ancestor with the scaled condition, 2) the ‘scaled’ condition is secondarily derived from an ancestor having a semi-transparent disc in the lower eyelid. 3) the lower eyelid with a semi-transparent disc is derived from an ancestor with a ‘scaled’ lower eyelid. For this reason polarities have not here been assigned to either of the conditions. Of the above possibilities the most par- simonious would be the 2nd, with the apparent loss of the palpebral sutures defining the pal- pebral rim and reaquisition of fine, widely spaced, transverse sutures possibly the result of an extension of the palpebral rim down over the centre of the eye to its lower margin, from an ancestor with a semi-transparent disc. G. Mesosternal rib attachment. Contact of the 12th and 13th ribs with the mesosternum is con- sidered the primitive condition (G), Loss of the 13th rib attachment to the mesosternum (g) is considered to be the derived condition. The primitive condition exists in N, maccoyi (Fig. 3a) and in a slightly modified form in N. mariet and N. greeri (Fig. 3b), whereas N, rankini, N, gracilis and N. sleveni have the 12th rib only contacting the mesosternum (g) (Fig. 3c), and the 13th rib lying posterior and separate to the mesosternum. H. Phalangeal formula of the manus. The primitive phalangeal formula for the manus is 2.3.4.5.3. Loss of phalanges on the manus is a derived condition. The phalangeal formula for the manus of the ancestor of Nannoscincus is considered to be 2.3,4.4.3 (H), and is the condi- tion occurring in N. maccoyi, N. greeri, N. rankini and N, mariei. Loss of an additional phalange on the 3rd and 4th digits of the manus of N. gracilis (2.3.3.3,2) and N. sleveni (2.3.3.3.0) is considered derived (h) within the genus. Note the phalangeal formula for N. sleveni given previously by Sadlier (1987) was incorrect (read in reverse off x-ray plate) and is here corrected to a loss of the 5th (rather than Ist) digit - a condition unique within the Eugongylus group. I. Phalangeal formula of the pes. The primitive phalangeal formula for the pes is 2.3.4.5,4, (1), and reduction in phalange number derived (i). N. greeri, has the primitive phalangeal for- mula. The phalangeal formula for the remaining species of Nannoscincus is 2.3.4,4.3. Loss of a phalange on the 4th and Sth digits of the pes of these species is considered derived within the genus. J. Presacral vertebrae. In skinks the modal number of presacral vertebrae is 26, any devia- tions from this can be taken as progressive derivations. The species with the lawest number of vertebrae above 26 will be primitive for this condition, and those with a higher number of vertebrae derived. In Nannoscincus presacral vertebrae number is variable but falls roughly into 2 groups: N. mariei (31-32, mode 31), N. greeri (29) and N. rankini (29-30) with generally 31 or fewer presacral vertebrae which is considered primi- tive (J) for the genus; and N. maccoyi (34-37), N. THE SCINCID LIZARD GENUS NANNOSCINCUS N. maccoyi N. greerj N. rankini 491 N. mariei N. gracilis N. sleveni ABCDEGHIJKL FIG. 4. Phylogeny of the genus Nannoscincus subgroup. sleveni (31-34, mode 32) and N. gracilis (33-34) with generally 32 or more presacral vertebrae which is derived (j) within the genus. K. Atlantal vertebrae. In most generally primi- tive lygosomine skinks the atlas consists of 3 distinct elements, the 2 atlantal arches and the intercentrum. All species of Nannoscincus have the atlantal arches fused to the intercentrum. N. maccoyi has the atlantal arches distinct dorsally where they abut, this is considered the primitive condition (K) within the genus. The remaining species of Nannoscincus have undergone further fusion of the atlantal vertebrae, the atlantal arches being fused to one another dorsally, this is derived (k) for the genus. L. Oviduct. The presence of a pair of oviducts in females is considered the primitive condition (L) and occurs in N. maccoyi, N. gracilis (*bar 1 specimen see below), and N. sleveni. A single oviduct only occurs on the right side of the body in N. mariei and N. greeri, and is the derived condition (1). N. rankini is known only from 2 adult male specimens, the female reproductive trait for this species is therefore unknown. A single female N. gracilis examined had a single oviduct on the right side of the body containing a single enlarged yolked ovarian fol- licle. This individual is unusual in 2 other respects. Firstly it is the smallest reproductively active female N. gracilis examined, and second- ly it is from a geographically disjunct pointin the species range. At this stage it is unclear as to whether this specimen represents an aberrant individual of N. gracilis, or a sibling species to N. gracilis distinguished solely on the oviduct condition. If the latter it would represent another case of loss of the left oviduct. M. Ventral coloration. The polarity of features of coloration are uncertain, however they may in the future add to our knowledge of relationships. The venter of N. maccoyi in life has an orange flush predominantly in males and a yellow flush in females. No orange or yellow flush to the venter was observed in live (breeding and non- breeding) N. mariei, N. gracilis, N. sleveni or N. rankini | have observed. MEMOIRS OF THE QUEENSLAND MUSEUM 492 quasqe yuasqe FIG. 11. Dorsal view of fronto-nasal region of skulls of A. Tiliqua scincoides (AM R127901), B. T. gigas (AM R93222), C. T. nigrolutea (AM R127911), D. T. rugosa (AM R127916), E. T. occipitalis (AM R127925), F. T. multifasciata (AM R100984), G. T. adelaidensis (SAM R4307A), H. Egernia striata (WAM R25402), I. Cyclodomorphus gerrardii (AM R13084), J. C. casuarinae (AM R37706), K. C. branchialis (AM R127930), L. C. maximus (WAM R77042). Fr = frontal; ma = maxilla; na = nasal; pr = prefrontal. Scale bar = 1mm. generally present in Eumeces and Scincopus, but absent in Janetaescincus and Pamelaescincus. Although the evidence is not conclusive, the condition shown by Egernia, Corucia, Mabuya and Eumeces, several moderate to large lobules MEMOIRS OF THE QUEENSLAND MUSEUM along the rostral margin of the ear, is considered primitive, and the 0-2 small rounded lobules seen in Cyclodomorphus, Hemisphaeriodon and T. adelaidensis, derived. It is difficult to assess the condition of the rostral margin of the ear in Trachydosaurus, as the scales are thick and bony, and evenly grade into smaller bony scales deep within the external auditory meatus, but these thickened bony scales may be derived from the lobules of other Tiliqua species. THE TILIQUA LINEAGE The species variously assigned to Cyclodomorphus, Hemisphaeriodon, Tiliqua and Trachydosaurus share the derived condition in characters 1-13, and constitute a lineage, which may be defined as follows: Osteology: Prefrontal and postfrontal narrow- ly separated or in contact; jugal and squamosal in contact; lacrimal absent; medial palatine process of ectopterygoid strong, broadly con- tacting palatine; coronoid process of dentary laterally overlapping coronoid; single grossly enlarged tooth in maxilla (position 7 or 8) and dentary (position 10) in juveniles; presacral ver- tebrae 32-44; phalangeal formulae of manus and pes 2.3.4.4.3/2.3.4.4.3 or fewer. Scalation: Caudalmost supralabial divided into an upper and a lower scale; supraciliaries modally six or fewer. Coloration: tongue deeply pigmented, at least in juveniles, blue- black to bright blue; dorsal and lateral pattern on body and tail predominant- ly consists of narrow to broad bands or transver- sely aligned vermiculations or spots, at least in juveniles. THE HOLOPHYLY AND RELATIONSHIPS OF THE TILIQUA LINEAGE There seems little doubt that the Tiliqua lineage is holophyletic. Two characters seem particularly telling in this regard: the increase in number of presacral vertebrae and the pattern of phalangeal loss. Within the Egernia group, these characters readily separate the Tiliqua lineage from both Egernia and Corucia, with no evidence of intermediacy. The Egernia luctuosa species group is clearly not a member of the Tiliqua lineage on both characters, having the primitive number of presacral vertebrae and phalanges. No skinks currently outside of the Egernia GENERA TILIQUA AND CYCLODOMORPHUS group appear to be members of the Tiliqua lineage or likely close relatives. The cluster of genera closest (o the Egernia group, the Eugon- gylus group, rarcly show marked increases in number of presacral vertebrae or phalangeal loss, apart from the loss of the first finger in Carlia, Lygisaurus, Menetia, Ristella and Saproscincus tetradactyla (Greer, 1974, 1979a; pers. obs.), a derived state that does not occur within the Egernia group. The anly two excep- lions to this pattern are Graciliscincus, which has a similar number of presacral vertebrae to the Tiliqua lineage while still retaining the primitive phalangeal configuration, and Nannoscincus, in which there is a mosaic of taxa with elevated numbers of presacral vertebrae and phalangeal lass (Sadlier, 1987, pers, comm,), including the combination seen in the Tiliqua lineage. How- ever, it is apparent that this similarity between Nannoscincus and the Tiliqua lineage is due to convergence, as Nannoscincus is both monophyletic and clearly a member of the Eugongylus group rather than the Egernra group (Greer, 1974; Sadlier, 1987), and otherwise shows little resemblance to Tiliqua. Although Egernia has. been shown to be the genetically closest genus to the Tiligua lineage (Hutchinson, 1981), the nature of the relation- ship has nol previously been determined. Three types of relationship are possible: Egernia and the Tiliqua lineage are sister-groups; Egernia is primitive, possibly ancestral to the Tiliqua lineage, or the Tiliqua lineage is primitive, pos- sibly ancestral to Egernia, The latter hypothesis was favoured by Horton (1972). At first glance, the third hypothesis seems untenable, piven the above argument for the holophyly of the Tiliqua lineage. However, given the high frequency of parallel evolution and character reversal within the Scincidae, if the third alternative were the case, use of Fgernia as the primary outgroup would be inappropriate, potentially assigning er- roncous Character polarities. This is worrying, when iL is remembered that in almost all charac- ters used to define the Tiliqua lineage, either Egernia uniformly shows the ‘primitive’ condi- tion, or only a few Egernia species show the ‘derived’ condition. However, exclusion of the first outgroup does not reverse the inferred polarity of any character, and hence confirms the highly derived nature of the Triiqua lineage. In contrast, | have been unable to identify any synapomorphies with which to diagnose Egernia vis-a-vis the Tiliqua lineage. Previous diagnoses S0Y have also fatled lo demonstrate a sister-group relationship between the two groups. The moder concept of Egernia is derived Trom Boulenger (1887), who placed in one genus a range of species formerly spread over at least five genera. Boulenger’s diagnosis utilises only two derived characters compared to generally primitive lygosomine skinks: pterygoid teeth ‘few or absent’ and lack of supranasal scales, Although Hoffstetter (1949) also records pterygoid teeth in Egernia, | have been able to identify them only in one specimen of £, cunmin- ghami, Both characters are shared with Tiligqua, and the second also with Corucia. At best, the second character merely supports the monophy- ly of the Egernia group, and the firs? the munbphy)y of Egernia + Ttliqua, Mitchell (1950), Cogger (1975) and Storr (1978) have subsequently attempted to diagnose Egernia. However, none of these diagnoses offer any ad- ditional synapomorphies for Egernia. On present knowledge, therefore, the second hypothesis, that Egernia is primitive, possibly ancestral to the Tiligua lineage, and potentially a pataphyletic assemblage, seems to be (he most likely. Although there ate arguments for not recognising paraphyletic taxa (recently dis- cussed by Hutchinson and Maxson, 1987), the interrelationships of the recognisable lineages within Egernia remain obscure (Horton, 1972; Storr, 1975; Wells and Wellington, 1984, 1983; Shea. in prep.) and in the absence of firm evidence relating the Ti/iqua lineage to any one of these other lineages, | prefer to retain the Egernia assemblage as a generic unit distinct from the Tiliqua lineage, GENERA WITHIN THE TILIQUA LINEAGE On the basis of characters 14-19, I believe that (wo sisler-taxa can be recognised within the Tili- qua lineage. The first of these, comprising the species formerly placed in Tiliqua (8,s.) and Trachydosaurus and for which the name Tiligue is available, may be diagnosed as follows; Tiliqua Gray, 1825 Tiliqgva Gray, 1825; 201. Type species Lacerta sein- caides Shaw, 179), by subsequent designation (Cogger et al, 1983). Trachydesaurus Gray, 1825: 201. Type species, by monotypy, Trachydosaurus ragesus Gray, 1825. Trachysaurus Gray, 1827: $M), Unjustified emenda- tion pro, Trachydosaurus. 510 Cyclodus Wagler, 1828: pl. 6. Type species, by monotypy, Cyclodus flavigularis Wagler, 1828 [= T. gigas]. Brachydactylus Smith, 1834: 144. Type species, by monotypy, Brachydactylus typicus Smith, 1834 [= T, rugosa]. Tiligua Duméril, 1837: 16, Lapsus pro, Tiliqua. Keneaux Duméril, 1837: 16. Nomen nudum. Original- ly proposed without included species, ex Cocteau MS. Tachydosaurus Gray, 1838: 288. Lapsus pro. Trachydosaurus. DIAGNOSIS Moderate to very large skinks, with a complete subocular row of evenly enlarged scales separat- ing supralabials from lower eyelid, nuchals either a single variably expressed pair or absent, and a broad, winglike jugal. CONTENT Cyclodus adelaidensis Peters, 1864, Scincus gigas Boddaert, 1783, Tiliqua occipitalis multi- fasciata Sternfeld, 1919, Scincus nigroluteus Quoy and Gaimard, 1824, Cyclodus occipitalis Peters, 1864, Trachydosaurus rugosus Gray, 1825, Lacerta scincoides Shaw, 1790. See Boulenger (1887) and Cogger et al. (1983) for species synonymies. NOMENCLATURE Although Tiliqua and Trachydosaurus were both erected by Gray (1825), Mitchell (1950), acting as first reviser in the sense of Article 24(b) of the Code of Zoological Nomenclature, selected Tiliqua to have precedence over Trachydosaurus. The second taxon, comprising the species variably placed in Omolepida, Cyclodomorphus and Hemisphaeriodon, for which Cyclodomor- phus is the earliest available name, may be diag- nosed as; Cyclodomorphus Fitzinger, 1843. Cyclodomorphus Fitzinger, 1843: 23. Type species, by original designation, Cyclodus casuarinae Duméril and Bibron, 1839. Omolepida Gray, 1845: 71, 87. Type species, by monotypy, Cyclodus casuarinae Duméril and Bibron, 1839. Hemisphaeriodon Peters, 1867: 24. Type species, by monotypy, Hinulia gerrardii Gray, 1845. MEMOIRS OF THE QUEENSLAND MUSEUM Homolepida Liitken, 1863: 294. Lapsus pro. Omolepida. Omolepidota Frost and Lucas, 1894: 227. Lapsus pro. Omolepida. DIAGNOSIS Small to moderately large skinks lacking lateral rostral projections of frontal bone, or with them very reduced, leaving a A-shaped frontal margin; second and third supraoculars fused, leaving only three supraoculars, first two con- tacting the frontal; lobules along rostral margin of ear very reduced (both in size and number) or absent. CONTENT Hinulia branchialis Giinther, 1867, Cyclodus casuarinae Duméril and Bibron, 1839, Hinulia gerrardii Gray, 1845, Omolepida maxima Storr, 1976. See Cogger et al. (1983) for species synonymies. NOMENCLATURE Although Cyclodomorphus, a senior objective synonym of Omolepida, has been formally used only six times in the 145 years since its erection (Fitzinger, 1860; Wells and Wellington, 1984, 1985; Shea and Wells, 1985; Czechura, 1986; Shea, 1988), while Omolepida (or its emenda- tion Homolepida) has been frequently used as an available generic or subgeneric name over the same period, I do not believe that recognition of the priority of Cyclodomorphus over Omolepida disturbs stability or causes confusion (Articles 23(b) and 79(c) of the Code). Mitchell (1950), Hutchinson (1981) and Cogger (1983), while placing both names into the synonymy of Tili- qua, clearly recognised the priority of Cyclodomorphus. In the previous fifty years, Omolepida has been formally used only once in combination with the type species (Storr, 1976), although frequently used as the generic name for the C. branchialis complex and C. maximus in Western Australia. Use of Cyclodomorphus here recognises the rather different concept of the genus I have proposed, and clearly distinguishes this version from that to which the name Omolepida had formerly been applied. Romer (1956) and Cogger et al. (1983) list three additional names in the synonymy of Tili- qua and Trachydosaurus: Rachites, Homolep- ides and Silubolepis. All are apparently derived from an unpublished manuscript, Tabulae synopticae Scincoideorum, by J.-T. Cocteau, submitted to the Académie des Sciences in Paris, GENERA TILIQUA AND CYCLOBOMORFHUS and described by Duméril (1837). Al three names appear to be unavailable, Rachifes was published without any included species of description (Dumeéril, 1837; Dumeéril and Bibron, 1839: 523). There appears to be no jus- tification for associating Rachites with Tiliqua uther than the inclusion of both, along with Euprepis Wagler, 1830, Keneaux, Psammites, Heremites and Arne (the Jatter four similarly noming nuda) as subgenera of the vernacular Seléroblépharides by Dumeéril (1837), Keneaux Dumeéri), 1837 was subsequently associated with Tiliqua by the inclusion of wwe of Cocteau’s vernacular names, Kéneaux de |’Uranie and Kéncaux de Boddaert, in the synonymy of Cyelodus nisraluteus and C. hoddaertii (Dumeéri! and Bibron, 1839). Momolepides Agussiz, [$46 was based, again without included species, On Cocteau’s vernacular Omolepides, There is no indication provided by Dumeril (1537) as to (he status assigned to this name, otier than (hat it Was six divisions below a tribe anu, in turt, three divisions above Tiliqua. Con- sequently, there appears to be no basis for as- souraling Homolepides with the Tiliqua lineage, Siluholepis Dumeril and Bibron, 1839, a name assigned to Cocteau, appears only in the synonymy of Trachysaurus, and is not therefore available (Article 11(¢)). Ad alternative classification reflecting the sume relalionships as defined here would be to reeogoise Tiliqna and Cycladomarphius as sub- xenera within an expanded Tiliqua. This would emphasise the sister-group relationship between the two taxa. However, | prefer generic separa- lion for three reasons. Firstly the larger Tiliqua are frequently used as experimental subjects in comparative physiological and biochemical re- seurch, Genuric separation simplifies a numenclature frequently used by non- Wxonumists, Secondly, with the generic slatus.ol Kgernia still undetermined, generic status adds Iwo Well-defined monophyletic groups to an Everio group otherwise having Cornea as its anly other Jdetinable genus. Finally, the two Henerd ure also ecologically distinet. With the exception ol 7, adelaidensis, a small, probably extiner species of largely unknown habits (Eh- muon. (983), Tiliqua comprises large, mostly diurtally active species that forage widely in lurgely open habitats, while Cyelodomorphus species are mostly of small to moderate size and seeretive habits in generally ‘closed’ habitats and microhubitats, from closed loresi (C, gerrur- dil We Prredia Wussocks (C, braviciiialis), S11 PREVIOUS ARGUMENTS FOR THE SYNONYMY OF CYCLODOMORPHUS WITH TILIGUA Arguments for ihe synonymy of Cyclodomar- phus with Tiliqua are based on two lines of evidence: morphology (Duméril and Bibron, 1839; Duméril and Duméril, 1851; Strauch, 1866; Smith, 1937: Mitchell, 1950; Cogger, 1983) and immunology (Hutchinson, 1981), Hutchinsan (1981), using serum im- munoelectrophoresis with a single 7, rugosa an- liscrum, found little antigenic difference between 7. rugese and 7. seineoides, a greater divergence between 7. ruxasa and C, casuarinae, and C. gerrardii the most divergent. Hence, he concluded, ‘to separate T. rugase or T. casuarinae [from Tiliqua], and not T. gerrar- di, as has been suggested [by Storr, 1976], is quite inconsistent with the HEP results’ (Hutchin- son, 1Y51; 188), By comparison with Fgernia, which showed grealer intrageneric vanation to E. cunninghamt antiserum than occurred be- tween T. rugosa and C. gerrardit, yet was still (realed as a monophyletic unit, Cyelodomorphus was Tegarded a8 Synonymous with Tiliqua. However, as noted above, evidence for the monophyly of Egernia is wanting, and hence the comparison used by Hutchinson (1981) is ine valid, The classification proposed here salishies Hutchinson’s other major criticism by separating both ©. gerrardij and C. casuarinae (rom Tili- qua, Indeed, Hutchinson’s criticism of Storr's (1976) coneept of Omolepida is flawed. Al- though Storr did not specifically include gerrar- dii in Omolepida (perhaps duc tu lack of familiarity with the species), il possesses all of the diagnostic characters Stor’ propased for the gonus, and clearly should have been included. Of the morphological arguments for the synonymy ol Cyclodomerphus and Tiliqua, those of Duméri] and Dumeéril (1851) and Strauch (1866) are not explicit, but appear to he largely based on a combination of overall phenetic similarity and the synapomorphy of enlarged, molariform teeth, while one of the two chitacters employed by Smith (1937), complete separation of the parictals by the interparictal, is usymplesiomorphy (Greer, 1979a) and henee of no use in inferring, relationships. Most authors advacating synomymy on morphological grounds have recognised a basic division within Tiltiqua (si. Dumeéril and Bihron (1839) and Dumeril and Dumeéril (185)) separated C casuariiae fron) the two other Cydledus species then recognised in the first couplet of their keys, on the basis of lack of ear lobules, Strauch (1866) separated the subgenus Omolepida on the basis of lack of a postnarial groove. Smith (1937) and Mitchell (1950) separated casuarinae and the teanchialis complex from other Tiligua on the basis of a longer tail and incomplete subocular scale row. Using these criteria, C, gerrardii comes out with C. casua@rinae (Mitchell, 1950). The generic separation advocated here Goes not contradict any of these proposed taxonomies, apart From the level at which the distinction is made. Cogger (1983:8) introduced a more serious objectron to the recognition of Cyclodomorphus by slating ‘there is a continuum of character states linking the extreme expression of Tiliqua via Hemisphaeriodon with that of Omolepide (=Cycladomorp/ies)', 1 do not believe this to be the case. Hemisphaeriodon shows all of the synapomorphies used lo diagnose Cyclodomar- phus vis-a-vis Tiligua, most notably the supraocular pattern and the shape of the suture between frontal, nasals, maxillae and prefron- tals, and is plestomorphic vis-4-wis Tédiqua in all diagnostic characters. Within Cyelodomorphus, gerrardiié shares With casudringde one synapomorphy unique within the Tiliqua lineage, loss of the postnarial groove. and another synapomorphy rare in other taxa, ex- treme reduction of the single ear lobule. A derived behavioural pattern also Jinks the two species: tongue-flickering, wsed in both food location and defence (Shea, 1988, pers. obs.), in contrast to simple tongue protrusion in other species. Both species are primitive within. the Tiliqua lineage in possessing «a mode of eight premaxillary teeth (Greer, 1979a; Shea, pers, obs.), These characters in combination suggest io me that C. casuerinae and C. gerrardii are cach other's closest relatives, and thal any ap. parent phenesic similarity between C. gerraridii and Tiliqua is due to a position lor C. gerrardii close to the basal stock of the lineage. PREVIOUS ARGUMENTS FOR THE RECOGNITION OF TRACHYDOSAURUS Trachydasaurus rugasus possesses all of the diagnostic characters listed above lor Tiliqua, or further derivations from these, and is clearly a member of the Tiliqua (s.s.) radiation. Trachydosaurus has previously been differen- tiated From Tiliqua by only a few characters. Gray (1825), in describing Trachydosaurus, MEMOIRS OF THE QUEENSLAND MUSEUM used two characters: thick, bony scales on head and body, and a short, depressed tail. Wagler (1830) added to these a ditference in dentition: conical teeth in Trachydosaurus vs rounded, ob- tuse crowns in Cyclodus. These three characters were employed by all authors for over sixty years (Gray, 1827, 1831, 1838, 1845; Wiegmann. 1834; Duméril and Bibron, 1839; Duméri] and Duméril, 1851; McCoy, 1885), although Peters (1864) noted that the teeth of 7. adelatdensis had conical rather than rounded crowns, Boulenger (1887) recognised all three characters, and added a further two: the presence of an azygous occipi- ial scale and mostly divided subdigital lamellae- Mitchell (1950), in synonymising Traciydo- saurus with Teleqgua, considered only the dif- ference in tail morphology to be of potential value for generic separation, stating ‘the general scalation, dentition and osteolagy are identical with those typifying Tiliqua’ (Mitchell, 1950: 277), The tail shape he dismissed as a character by using as a parallel the placement of the similarly short-tailed depressa and stekesit in Egernia. However, as noted above, this argu- ment is invalid, as Egernia is plesiomorphic and may only be an assemblage. Copland (1953: xxi) wished to retain Trachydosaurus ‘if only on the grounds of its grass scalation', Mertens (1958) resurrected Trachydesaurus in describing the insular race T. r, komowi, but reserved his reasons for publication in a report on his 1957 Australian expedition, This appears not to have been pub- lished. Glavert (1960) used the blunt tail as a diagnosis for Trachydosaurus, while Worrell (1963) used both the tail and the rugose scala- lion. Cogger (1975) noted the short tail, rugose scalation, and mostly divided subdigital Jamel- lac, Cogger (1983: 8) justified his continued recognition of Trachydosaurus, stating ‘I believe... thal the available morphological, biological and aay et evidence suggests that the shingle-back/blue-tongue divergence was earlier than, rather than approximately con- temporaneous with, the radiation of the blue-ton- gued lizards in Australia’, apparently hypothesising a sister-group relationship wit Tiliqua (inclusive of Cyclodomorphus}. Haw- ever, no evidence was advanced in support of this hypothests. Insummary, previous arguments for the recog- nition of Trachydosaurus have rested on five morphological characters: a short, blunt tail, thickened, rugose scalation, divided subdigital lamellae, conical teeth and an azygous occipital scale. GENERA TILIQUA AND CYCLODOMORPHUS 313 a ____ e mul (n= 236) = occ (n= 267) ——_ ade (n=14) —_— ssc (n= 582) ——— sin (n= 173) _- RE gig (n=176) a nig (n=324) a rux (n=35) ——_ rko (n=64) ———_—__ rru (n= 188) = ras (n=309) 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 MIDBODY SCALES FIG. 12. Variation in number of midbady scales in Tiliqua species. Vertical bar is mean, solid rectangle is one standard deviation on each side of mean, horizontal line is range. Ade = T. adelaidensis, gig = T. gigas, mul =T. multifasciata, nig =T. nigrolutea, occ = T. occipitalis, ras = T. rugosa asper, tko = T. r. konowi, rru = T. r. rugosa, Tux = T. r. subsp. nov., sin = T. scincoides intermedia, ssc = T. s. scincoides. The latter two characters are of no use in diagnosing Trachydosaurus, as they also occur in Tiliqua species. Within Tiliqua, there is marked interspecific and ontogenetic variation in tooth shape (Shea, pers obs.). Only 7. gigas and 7. scincoides, the first two described species, have the rounded tooth crowns noted by Wagler (1830). The other species have more conical crowns, those of T. nigrolutea being more coni- cal than in Trachydosaurus. The presence of a median occipital is variable in Trachydosaurus, although it is present in most individuals. A median scale caudal to the inter- parietal is a derived character in skinks (Greer, 1968), and has been previously used as a major diagnostic character in one genus, Geomyersia. However, the median occipital of Trachy- dosaurus also occurs in T. adelaidensis (Fig. 9D), and is frequently present in T. nigrolutea, occurring in 42.1% (n = 321) of specimens ex- amined, Asymmetry in the scales bordering the caudal margin of the parietal/interparietal com- plex, a possible precursor to the differentiation of a median occipital, is common in other Tiliqua species. Similarly, although the grossly enlarged, thickened osteoderms characteristic of Trachydosaurus are unique within the Scin- cidae, T. nigrolutea also displays a trend in this direction. Enlargement of body scales can also be expressed as a reduction in number of scales. If number of midbody scales, paravertebral scales and ventral scales are compared (Figs. 12- 14), it can be seen that the values for T. rugosa overlap with 7. nigrolutea in two cases (midbody and ventral scales) while 7. nigrolutea also 514 a rox (n=33) = rko (n= 62) —— rru (n=186) ras (n= 299) MEMOIRS OF THE QUEENSLAND MUSEUM aE mul (n= 236) = oce (n= 271) a ade (n= 14) = ssc (n= 583) = sin (n= 172) $$ gig (n=176) oe nig (n= 323) 18 22 26 30 34 38 42 46 50 54 58 62 66 70 74 78 82 PARAVERTEBRAL SCALES FIG. 13. Variation in number of paravertebral scales in Tiliqua species. Conventions as in Fig 12. shows a trend towards 7. rugosa in number of paravertebral scales. The short, depressed, blunt-tipped tail of Trachydosaurus is also derived. However, there is geographic variation in tail length in Trachydosaurus, with the longest tails occurring in the south-west of Western Australia. Moreover, some Western Australian individuals have a distinctly conical tail tip (Fig. 15). 7. nigrolutea again shows some trend in the direc- tion of Trachydosaurus, having a short, thick tail which becomes depressed in emaciated in- dividuals, in contrast to the compressed tail seen in T. multifasciata and T. occipitalis. The division of subdigital lamellae seen in Trachydosaurus is uniquely derived within the Egernia group, with no trend in this direction, such as a median groove, seen in any other Tiliqua species. A number of other differences between T. rugosa and other Tiliqua (usually as represented by T. scincoides) have been noted in the course of more general comparative studies, though not previously utilised for formal taxonomic separa- tion (Arnold, 1984; Camp, 1923; Cope, 1892b; Greer, 1979a; Hoffstetter, 1949; Lécuru, 1968; Parker, 1868; Renous-Lécuru, 1973; Sieben- rock, 1892, 1895; Smith, 1976, 1982). I have re- examined all of these characters. In almost all cases, I find the purported differences to be less than diagnostic, either due to variation within T. rugosa, or Tiliqua species not previously ex- amined having the condition reported for T. rugosa. Only in the further reduction of phalan- geal formula (Siebenrock, 1895; Hoffstetter, 1949) is the difference clear-cut and consistent. In summary, T. rugosa differs markedly and consistently from other Tiliqua species in having some subdigital lamellae divided and in further teduction in phalangeal formula. In two other GENERA TILIQUA AND CYCLODOMORPHUS 515 — mul (n= 229) a occ (n= 260) = ade (n=13) SE EE ssc (n= 571) ——————_ sin (n= 172) a gig (n= 175) a ___ nig (n= 318) a rux (n= 34) —_ = rko (n=63) SE rru in= 190) a ras (n= 300) 48 52 56 60 64 68 72 76 80 84 88 92 96 100 104 108 VENTRAL SCALES FIG. 14. Variation in number of ventral scales in Tiliqua species. Conventions as in Fig 12. A B FIG. 15. Dorsal view of tails of A. Tiliqua rugosa asper (AM R123583), B. T. rugosa subsp. nov. (AM field series 15164), C. T. rugosa subsp. nov. (AM R102711), D. T. r. rugosa (AM R102594). 516 characters, tail shape and rugosity of body scala- tion, variation is largely non-overlapping with other Tiliqua species, although in both cases T. nigrolutea displays a trend in the direction of T. rugosa. In all of these characters, the state present in 7. rugosa is derived. However, to generically separate Trachydosaurus on these characters would leave Tiliqua an undiagnosable entity vis-a-vis Trachydosaurus, as there are as yet no identifiable synapomorphies to link the remaining Tiliqua species independent of T. rugosa. On the available data, 7. rugosa is mere- ly a highly derived member of the genus, phenetically most similar to 7. nigrolutea, and Tiliqua without T. rugosa is paraphyletic. Con- sequently, I retain Trachydosaurus in the synonymy of Tiliqua. ACKNOWLEDGEMENTS lam grateful to Dr A.E. Greer for the provision of facilities, frequent helpful advice and data on skink species not present in Australian collec- tions. J. Covacevich, A.J. Coventry, A. Edwards, R.H. Green, M. King, R.A. Sadlier, T.D. Schwaner, L.A. Smith, G.M. Storr and J. Wom- bey generously allowed me virtually unlimited access to material in their care, while P. Alberch, A. Allison, N. Ananjeva, A.P. Andrews, E.N. Arnold, R. Giinther, M. Hoogmoed, E. Kramer, J. Rosado, J. Sites, F. Tiedemann, H.K. Voris, G.R. Zug and R.G. Zweifel kindly loaned a number of valuable museum specimens over often long distances. N. Shea translated a num- ber of references. B. Jantulik prepared the final illustrations. B.R.H. Farrow, A.E. Greer, M. Hutchinson and R.A. Sadlier offered helpful criticisms of the manuscript while K. Jopson provided assistance and encouragement. LITERATURE CITED AGGASIZ, L. 1846. ‘Nomenclatoris Zoologici Index Universalis, continens nomina systematica clas- sium, ordinum, familiarum et generum animalium omnium, tam viventium quam fos- sillum, secundum ordinem alphabeticum unicum disposita, adjectis homonymiis plan- tarum, nec non variis adnotationibus et emendationibus’. (Jent et Gassmann: Soloduri). ARNOLD, E.N. 1980. Recently extinct reptile populations from Mauritius and Reunion, Indian Ocean. J. Zool., Lond. 191(1): 33-47. 1981. Estimating phylogenies at low taxonomic MEMOIRS OF THE QUEENSLAND MUSEUM levels. Z. f. zool. Systematik u. Evolutionsfor- schung 19(1): 1-35. 1984. Evolutionary aspects of tail shedding in lizards and their relatives. J. Nat. Hist. 18(1): 127-169. BODDAERT, P. 1783. Specimen Novae Methodi dis- tinguendi Serpentia. Nova Acta Physico-Medica Academiae Caesareae Leopoldino Carolinae Naturae Curiosorum 7: 12-16. BOULENGER, G.A, 1887. ‘Catalogue of the lizards in the British Museum (Natural History)’. Vol. Il. (Trustees of the British Museum: London). BRYGOO, E.R. 1981. Systématique des Lézards Scincidés de la région malgache IX. Nouvelles unités taxinomiques pour les Scelotes s.1. Bull. Mus. natn. Hist. nat., Paris A(4)3(4): 1193-1204. CAMP, C.L. 1923. Classification of the lizards, Bull. Am. Mus. nat. Hist. 48(11): 289-481. COGGER, H.G. 1975. ‘Reptiles and amphibians of Australia’. (A.H. & A.W. Reed: Sydney). 1983. Introduction. pp. 1-11. Jn Cogger, H.G., Cameron, E.E. and Cogger, H.M. ‘Zoological Catalogue of Australia. Vol. 1. Amphibia and Reptilia’. (Australian Government Printing Ser- vice: Canberra). COGGER, H.G., CAMERON, E.E. AND COGGER, H.M. 1983. ‘Zoological Catalogue of Australia, Vol. 1. Amphibia and Reptilia’. (Australian Government Printing Service: Canberra). COPE, E.D. 1892a. On degenerate types of scapular and pelvic arches in the Lacertilia. J. Morph. 7(2): 223-244. 1892b. The osteology of the Lacertilia. Proc. Am. Philos. Soc. 30(138):; 185-221. COPLAND, S.J. 1953. Recent Australian herpetol- ogy. Proc. Linn. Soc. N.S.W. 78(1-2): v-xxxvii. CZECHURA, G.V. 1986. Distant exiles: frogs and reptiles recorded from Kroombit Tops, southeastern Queensland. Qd Nat. 27(1- 4): 61- 67. DONNELLAN, S.C. 1985. The evolution of sex chromosomes in scincid lizards, (Unpublished Ph.D. thesis, Macquarie University). DUMERIL, [A.M.C.]. 1837. Rapport sur un ouvrage manuscrit de M. le docteur Cocteau, ayant pour titre: Tabulae synopticae Scincoideorum. C. r. hebd. Séanc. Acad. Sci., Paris 4(1): 14-17. DUMERIL, A.M.C. AND BIBRON, G. 1839. ‘Erpétologie Générale ou Histoire Naturelle Compléte des Reptiles’. Vol. 5. (Libraire Encyclopédique de Roret:Paris). DUMERIL, A.M.C. AND DUMERIL, M.A. 1851, ‘Catalogue méthodique de la collection des reptiles’. (Gide et Baudry: Paris). EHMANN, H. 1983. The natural history and conser- GENERA TILIQUA AND CYCLODOMORPHUS valion status of the Adelaide pigmy bluetongue lizard Tiliqua adelaidensis. Herpetofauna 14(1): 61-76. EL-TOUBI, M.R. 1938. The osteology of the lizard Scincus scincus (Linn), Bull, Fac, Sci. Egypt, Univ. (14): 5- 38. ESTES, R. 1983. ‘Handbuch der Palaoherpetologie Tei! 10A, Satria Terrestria. Amphisbaenia’. (Gustav Fischer: Stultgart). ESTES, R. AND WILLIAMS, E.E. 1984, On- togenetic variation in the molariform teeth of lizards. J. Vert. Paleont. 4(1); 96- 107, FITZINGER. L, 1843. ‘Systema reptilium, Fasciculus primus Amblyglossae’. (Briumiiller et Seidel; Vindobonae), 1860. Die Ausbeule der Gsterreichischen Natut- forscher an Saugethieren und Reptilien wahrend der Weltumsegelung sr. Majestit Fregatte Novara. Sber, Akad. Wiss, Wien 42(24): 383- 416. FROST, C. AND LUCAS, A.H.S. 1894. On a new skink lizard from Tasmania. Proc. Linn, Soc. N.S.W, (2)8(2): 227-228, FUHN. LE. 1969. The “polyphyletic” origin of the genus Ablepharas (Reptilia, Scincidae): a case of parallel evolution. Z. f. zool. Systematik u. Evolutionsforschung 7(1); 67-76. GLAUERT, L. 1960. Herpetological Miscellanea. XII. The family Scincidae in Western Australia. Part 1 - The genera Tiliqua, Trachysaurus and Egernia. West. Aust. Nal. 7(2): 67-77. GRAY,J.E. 1825. A synopsis of the genera of reptiles and amphibia, with a description of some new species, Ann, Philos, (2)10(3); 193-217. 1827, Replilia. pp. 424-434. /n King, P.P., ‘Narra- tive of a survey of the interiropical and western coasts of Australia. Performed between the years 1818 and 1822", Vol_2. (John Murray: London). 1831. Asynopsis of the species of the class Reptilia, pp. 483- 600. fn Griffith, E. and Pidgeon, E. ‘The Animal Kingdom arranged in conformity with its organization by the Baron Cuvier’, Vol. 9. (Whittaker, Treacher & Co.. London). 1838, Catalogue of the slender-tongued saurians, with descriptions of many new genera and species. (contd). Ann. Mag. nat. Hist. 2(10): 287-293. 1845, ‘Catalogue of the specimens of lizards in the collection of the Brilish Museum’. (Edward Newman: London). GREER, A.E. 1968. Geomyersia glabra, anew penus and species of scincid lizard from Bougainville, Solomon Islands, with comments on the relationships of some lygosomine genera. Breviora (302): 1-17, 317 1970a, A subfamilial classification of scineid lizards. Bull. Mus. Comp. Zool. 139(3): i51- 184, 1970b, The relationships of the skinks referred to the genus Dasia, Breviora (345): 1-30. 1974. The generic relationships of the scincid lizard genus Leiolopisma and its relatives. Aust. J, Zool, Suppl. Ser, (31): 1-67. 1976. On the evolution of the giant Cape Verde scincid lizard Macroscirtcws coctei. J, nat. His 10(6): 691-712. 1977. The systematics and evolutionary relation- ships of the scincid lizard genus Lygosoma. J. Nat. Hist. 11(5); 515-540, 1979a, A phylogenetic subdivision of Australian skinks. Rec. Aust. Mus. 32(8): 339-371. 1979b, Eremiascincus, a new generic name for some Australian sand swimming skinks (Lacer- lilia; Scincidae), Rec. Aus!, Mus. 32(7): 321- 338, 1952. A new species of Geomyersia (Scincidac) from the Admiralty Islands, with a summary of the genus, J. Herpetol. 16(1): 61-66, 1983. The Australian scincid lizard genus Calyp- otis De Vis: Resurrection of the name, descrtip- tion of four new species, and discussion of Telationships. Rec, Aust. Mus. 35(1-2): 29-59, 1985. A new species of Sphenomorphus from northeastern Queensland. J. Herpetol. 19(4); 469-473. 1986a, Lygosomine (Scincidae) monophyly: A third, corroborating character and a reply to critics. J. Herpetol. 20(1): 123- 126. 1986b. Diagnosis of the Lerista bipes species- group (Lacertilia: Scincidae), with a description of anew species and an updated diagnosis of the genus. Rec. West. Aust. Mus. 13(1): 121-127, 1987. Limb reduction in the lizard genus Lerista. 1. Variation in the number of phalanges and presacral vertebrae. J. Herpetol. 21(4): 267-276. GREER, A.E. AND COGGER, H.G. 1985. Sys- tematics of the reduce- Jimbed and limbless skinks currently assigned to the genus Anomalopus {Lacertilia: Scincidae). Rec. Aust Mus. 37(1): 11-54, GUNTHER, A. 1867. Additions to the knowledge of Australian reptiles and fishes. Ann. Mag. nat, Hist, (3)20(115): 45- 68. HEYER, W.R. 1972, A new limbless skink (Reptilia: Scincidae) from Thailand with comments on the genericstatus of the limbless skinks of southeust Asia, Fieldiana, Zool. 58(10): 109-129. HIKIDA, T, 1978. Postembryonic development of the skull of the Japanese skink, Eumeces Jaris- cutaiis (Scincidae). Jap. J. Herpetol, 7(3); 56-72, 318 HOFFSTETTER, R. 1949, Les reptiles subfossiles de Vile Maurice I. - Les Scincidae, Ann, Paléont. 35: 45-72. HOFFSTETTER, R. AND GASC, J.-P. 1969. Ver- lebrae and ribs of modern reptiles. pp. 201-310. Jn Gans, C.. Bellairs, A.d’A. and Parsons, T.S. (eds), “Biology of the Reptilia. Vol. 1. Morphol- ogy A’, (Academic Press: London). NORTON, D.R, 1972, Evolution in the genus Egernia (Lucertilia: Scincidae). J. Herpet. 6(2); 101-109. HUTCHINSON, M.N. 1981. The systematic relation- ships of the genera Egernia and Tiliqua (Lacer- lilia: Scincidae). A review and immunological reassessment. pp. 176-193, Jn Banks, C.B. and Martin, A.A. (eds). ‘Proceedings of [he Mel- bourne Herpetological Symposium’. (Zoalogi- cal Board of Victoria: Melbourne). HUTCHINSON, M.R. AND MAXSON, LR. 1987, Albumin evolution, paraphyly and the relation- ships of the Australian frag Megistealotis lig- narius (Anura: Limnodynastinue). Syst. Zool. 36(4): 387-394, KING, M. 1973a. Chromosomes of two Australian lizards of the families Scincidae and Gek- Konidae, Cytologia 38(2): 205- 210. 1973b. Kuryolypic studies of some Australian Sein- cidae (Reptilia), Aust J. Zool. 21(1): 21-32, KINGMAN, R.H. 1932. A Comparative study of the skull in the genus Fumeces of the Scincidae (a preliminary paper). Uniy, Kansas, Sci. Bull, _ 20015): 273-295. LECURU, S. 1968, Remarques sur Je scapulo- coracotde des lacertiliens. Ann. Sci. Nat., Zool. (12) 1(4): 475-510. LUTKEN, C. 1863. Nogle nye Krybdyr og Padder. Vidensk, Meddr dansk naturh. Foren, 1863; 292- Sit. McCOY, F. 1885. “Prodromus of the Zoology of Victoria; or figures and descriptions of the living species of all classes of the Victorian indigenous animals’. (RS. Brain. Government Printer: Mel- bourne). MERTENS, R, 1958. Neue Eidechsen aus Australien, Senckenberg. biol. 39(1/2): 51-56, MITCHELL, F.J. 1950. The seincid genera Bgernia and Tiligua (Lacertilia). Ree. S$. Aust. Mus. 9(3): 275- 308. NASH, D.F. AND TANNER, W,W. 1970. A com- parative study of the head and thoracic osteology and myology of the skinks Eumeces gilberri van Denburgh and Eumeces skiltonianus (Baird and Girard). Brigham Young Univ. Sci. Bull. Biol. Ser. 12(2): 1-31. PARKER, W.K. 1868. “A monograph on the structure and development of the shoulder-girdle and ster- MEMOIRS OF THE QUEENSLAND MUSEUM num in the Vertebrata’. (Kobert Hardwicke tor the Ray Society: London), PERRET, J.-L. 1975. La différenciation dans le genre Panaspis Cope (Reptilia, Scincidae), Bull. Soc, neuchateloise Sci, nat. 98: 5-16. PETERS, W. 1864. Ubersicht der van Hrn, Richard Schomburgk an das zoologische Museum ein- gesandten Amphibien, aus Buchsfelde bei Adelaide in Siidaustralien, Mber. K, Preuss. Akad. Wiss, Berlin 1863; 228-236, 1866, Eine Mitleilung tber neue Amphibien (Am- phibalurus, Lygasoma, Cyclodus, Masticaphis, Crotaphopeltis) und Fische (Diagramma: Hap- alagenys) des Kgl, zoologischen Museums. Mber. K, Preuss. Akad, Wiss, Berlin 1866: 86-96. 1867, Herpetologische Notizen. Mber, K. Preuss. Akad. Wiss. Berlin 1847: 13-37. 1872. Uber neve oder weniger bekannte Saurier, Mber. K, Preuss. Akad. Wiss. Berlin 1872: 774- 776. QUOY, J.R.C. AND GAIMARD, J.P. 1824. Zoologie. /n de Freycinet, L, “Vayage autour du monde, Entrepris par ordre du Roi, sous le ministére et confarmeément aux instructions des. exc, m.lé vicomte du Bouchage, secrétaire d'état au Département de la Marine, Exécuté sur les corvettes de S.M. !Uranie et la Physicienne, pendant Jes années 1817, 1818, 1819 et 1820". (Piller Ainé, Imprimeur- Libraire: Paris). RAWLINSON, P.A. 1966, Reptiles of the Victorian Mallee. Proc. R. Soc. Vict. (ns)82(1); 113-128. 1974. Biogeography and ecology of the reptiles of Tasmania and the Bass Strait area. pp. 291-338. Jn Williams, W.D. (ed.), ‘Biogeography and Ecology in Tasmania’, (W, Junk: The Hague). RENOUS-LECURU, S. 1973. Morphologie comparée du carpe chez les Lepidosauriens ac- tuels (Rhynehocéphales, Lacertiliens, Amphisbéniens). Gegenbaurs morph. Jahrb. 119(5); 727-766, ROMER, AS. 1956, ‘Osteology of the reptiles’, (University of Chicago Press, Chicago). ROSE, S. 1985. Captive breeding of Cunningham's skink, The Vipera 1(8): 14-22. SADLIER, R.A. 1987. A Review of the scincid lizards of New Caledonia Rec. Aust. Mus, 39(1): L-66. SHAW, G, (1790), f1 White, J. Journal of a Vayage to New South Wales, with sixty five plates of non-deseript animals, birds, lizards, serpents, curious cones of irees and other natural productions’. (Debrett: London). SHEA, G.M, 1988. On the diet of the Sheoak Skink, Cyclodomorphus casharinee. Herpetofauna IR(1): 7-8. SHEA, G.M. AND WELLS, BR. W. 1985. New iccords GENERA TILIQUA AND CYCLODOMORPHUS of a skink and a gecko from western New South Wales. Herpetofauna 15(1): 1- 4, SIEBENROCK, F. 1892. Zur Kenntnis des Kopfskelettes der Scincoiden, Anguiden und Gerrhosauriden. Ann. K. K. Naturh. Hofmus. Wien 7(3): 163-196. 1895. Zur Kenntnis des Rumpfskelettes der Scin- coiden, Anguiden und Gerrhosauriden. Ann. K. K. Naturh. Hofmus. Wien 10(1): 17-40. SMITH, A. 1834. [untitled]. S. Afr. Quart. J. 2(2): 143- 144, SMITH, M.A. 1937. A review of the genus Lygosoma (Scincidae: Reptilia) and its allies. Rec. Ind. Mus. 39(3): 213- 234. SMITH, M.J, 1976. Small fossil vertebrates from Vic- toria Cave, Naracoorte, South Australia [V. Rep- liles. Trans. R. Soc. S, Aust. 100(1): 39-51. 1982. Reptiles from Late Pleistocene deposits on Kangaroo Island, South Australia. Trans. R. Soc. S. Aust. 106(2): 61-66. STERNFELD, R. 1919, Neue Schlangen und Echsen aus Zentralaustralien. Senckenbergiana 1(3): 76-83. STORR, G.M. 1964. Some aspects of the geography of Australian reptiles. Senckenberg. biol. 45(3/5): 577-589. 1965. The physiography, vegetation and vertebrate fauna of the Wallabi Group, Houtmans Abrol- hos. J. R. Soc. W. Aust. 48(1): 1-14. 1976, The genus Omolepida (Lacertilia, Scincidae) in Western Australia, Rec. West. Aust. Mus. 4(2): 163-170. 1978. The genus Egernia in Western Australia. Rec. West. Aust. Mus. 6(2): 147-187. STRAUCH, A, 1866, Uber die Arten der Eidechsen- 519 gattung Cyclodus Wagl. Bull. Acad. Imp. Sci. St- Pétersbourg 10: 449-462. TAYLOR, E.H. 1935. A taxonomic study of the cos- mopolitan scincoid lizards of the genus Eumeces with an account of the distribution and relation- ships of its species. Kansas Univ. Sci. Bull. 23(14): 1-643. WAGLER, J. 1828. ‘Descriptiones et icones am- phibiorum. Part I’. (J.G. Cottae: Stuttgartiae, Tubingae, Monachii), 1830. ‘Natiirliches System der Amphibien mit vorangehender Classification der Sdiugthiere und Vdégel*. (J.G. Cotta: Miinchen, Stuttgarten, Tubingen). WAITE, E.R. 1929. ‘The reptiles and amphibians of South Australia’. (Harrison Weir, Government Printer: Adelaide). WELLS, R.W. AND WELLINGTON, C.R. 1984. A synopsis of the class Reptilia in Australia, Aust. J. Herpetol. 1(3-4): 73-129. 1985, A Classification of the Amphibia and Rep- tilia of Australia. Aust. J. Herpetol. Suppl. Ser. (1): 1-61. WIEGMANN, A.F.A. 1834. ‘Herpetologica Mexicana; seu descriptio Amphibiorum Novae Hispaniae quae itineribus Comitis de Sack, F: Deppe et C.G, Schiede in Museum Zoologicum Berolinense perienerunt. Pars prima, Saurorum Species amplectens. Adjecto systemalis saurorum Prodromo, additisque multis in hunc amphibiorum ordinem observationibus’. (C.G. Luderitz: Berolini). WORRELL, E. 1963. ‘Reptiles of Australia’. (Angus & Robertson: Sydney). MEMOIRS OF THE (QUEENSLAND MUSEUM BRISBANE © Queensland Museum PO Box 3300, South Brisbane 4101, Australia Phone 06 7 3840 7555 Fax 06 7 3846 1226 Email qmlib@qm.qld.gov.au Website www.qm.qld.gov.au National Library of Australia card number ISSN 0079-8835 NOTE Papers published in this volume and in all previous volumes of the Afemoirs of the Queensland Museum maybe teproduced for scientific research, individual study or other educational purposes. Properly acknowledged quotations may be made but queries regarding the republication of any papers should be addressed to the Editor in Chief. Copies of the journal can be purchased from the Queensland Museum Shop. A Guide to Authors is displayed at the Queensland Museum web site A Queensland Government Project Typeset at the Queensland Museum BARU DARROWI GEN. ET SP. NOV., ALARGE, BROAD-SNOUTED CROCODYLINE (EUSUCHIA: CROCODYLIDAE) FROM MID-TERTIARY FRESHWATER LIMESTONES IN NORTHERN AUSTRALIA PAUL WILLIS, PETER MURRAY AND DIRK MEGIRIAN Willis, P., Murray, P. and Megirian, D. 1990 09 20: Baru darrowi gen. et sp. nov., a large, broad-snouted crocodyline (Eusuchia: Crocodylidae) from mid-Tertiary freshwater lime- stones in Northern Australia. Memoirs of the Queensland Museum 29(2): 521-540, Bris- bane. ISSN 0079-8835. Baru darrowi gen, et sp. nov., is acommon element in limestones of late Oligocene to late Miocene age on Riversleigh Station in northwestern Queensland and at Bullock Creek in the Northern Territory. Although Baru is a member of the Crocodylinae and appears to have many features in common with certain early Tertiary crocodiles such as the North American Brachyuranochampsa eversolei Zangerl, it also resembles sebecosuchian and pristichampsine crocodiles in having ziphodont (serrated, laterally compressed) teeth similar to those of flesh-eating dinosaurs. The Australian ziphodont crocodile Quinkana fortirostrum Molnar, was previously considered to be closely related to the Pris- tichampsinae on the basis of its cranial profile and highly developed ziphodonty. Quinkana fortirostrum and Baru darrowi share characters not present in pristichampsine crocodiles and they appear to be more closely related to one another than to any other ziphodont taxa. Because Baru darrowi is clearly a member of the Crocodylinae, we propose that Quinkana and Baru represent a new crocodyline ziphodont clade and that these two forms, together with Pallimnarchus pollens, form a monophyletic endemic Australian radiation. 1 Crocodylidae, Eusuchia, systematics, Tertiary, Ziphodont, Baru. Paul Willis, University of New South Wales, GPO Box 1, Kensington, New South Wales 2033, Australia; Peter Murray and Dirk Megirian, Northern Territory Museum of Arts and Sciences, GPO Box 4646, Darwin, Northern Territory 0801, Australia; 22 August, 1990. An unusually complete assemblage of fossil crocodile material has been recovered from fluviolacustrine sediments of middle to late Miocene age at Bullock Creek in the Northern Territory and late Oligocene to early Miocene age on Riversleigh Station, Queensland. The material provides clear evidence of a member of the subfamily Crocodylinae possessing ziphodont teeth. Previous finds of Australian ziphodont crocodiles have not been complete enough to determine their subfamilial affinity with confidence (Hecht and Archer, 1977; Mol- nar, 1981, 1982). The material described here is referred to the new genus and species Baru dar- rowi, a large crocodilian with many distinctive features. Its broad, short snout, robust propor- tions and deeply festooned jaws set it apart from any living Crocodylus species. Its dentition con- sists of posteriorly inclined, slightly recurved, laterally compressed crowns of greatly varying dimensions, bearing well-developed anterior and posterior crests (carinae). In some Northern Territory specimens, these carinae are finely ser- tated like the teeth of South American sebecosuchian and Northern Hemisphere early Tertiary pristichampsine crocodiles. Oddly, no specimens of Baru from the Riversleigh deposits have serrated carinae. Large, slightly com- pressed carinate teeth with fine serrations are also known from the Alcoota Local Fauna of the Northern Territory (P.M., pers. obs.) and from other middle to late Tertiary localities throughout the interior of Australia. These have been variously assigned to the genus Pallimnar- chus (Molnar, 1982) or to unidentified sebeco- suchians (Hecht and Archer, 1977). It can now be demonstrated that at least some of the ziphodont crocodile teeth found in Australia belong to a crocodyline genus. The proposition that pristichampsine and sebecosuchian ziphodont crocodiles may have been present in Australia is therefore re-examined. Quinkana fortirostrum (Molnar, 1981), the first Australian crocodile formally described as a ziphodont, is known primarily from a snout. Although sufficiently well represented to sug- gesta closer affinity with Baru darrowi than with Pristichampsus, its principle features are dominated by trophic specialisations. Because crocodilians are otherwise structurally conserva- tive, there are few character states suitable for a cladistic evaluation. We are therefore conlined to uw few observations strongly supporting the more parsimonious hypothesis that Australian ziphodont crocodiles represent a monophyletic radiation with Gondwana as its likely origin.. [Interpretation of the polarity of character states, and basic concepts of crocodilian phylogeny used in this study, are based on Mol- nar (1981), Benton and Clark (1988) and Langston (1973); nomenclature follows Steel (1973) and lordansky (1973). Prefixes used to indicate the source of specimens are as follows: NTM P, Northern Territory Museum, Palacon- tological Collections, NTM R, Northern Ter- ritory Museum, Reptile Collections; QM F, Queensland Museum, Fossil Collections; SAM P, South Australian Museum, Palacontological Collections, Order CROCODILIA Gmelin, 1700 Suborder EUSUCHIA Huxley, 1875 Family CROCODYLIDAE Cuvier, 1807 Subfamily CROCODYLINAE Cuvier, 1807 Baru darrowi gen. ct sp. noy. CGIFNOTYPIC SPECIES Baru darrawi sp. nov, (Fig. la-c). DIAGNOSIS Species of Baru differ from all other cracodylines in the following combination of features: Broad moderately deep snout contain- ing thirteen maxillary teeth; five premaxillary tecth present in juveniles and four in adults owing to loss of the second tooth; premaxillary and anterior six maxillary qeeth directed posteriorly: tooth crowns moderately com- pressed buceo-lingually with carinac on the anterior and posterior margins; tooth crown and socket dimensions highly differentiated along both upper and lower tooth rows with cor- respondingly wide, deep alveolar processes; conspicuaus maxillary receptian pits, cor- responding to dentary tooth crowns, situated lin- gual to the upper tooth row; anterior margin of the palatal fenestrae extending to the level of the seventh maxillary tooth; anterior palatine process absent; mandibular symphysis extends posteriorly to between the sixth and seventh dentary teeth; splemal terminates anteriorly at the level of the seventh dentary tooth and does MEMOIRS OF THE QUEENSALND MUSEUM not enter symphysis; internal nares with raised rim; external nares terminal; distinctive bony crest arches posteriorly from the maxillae and jugals, extending to the quadratojugals. ETYMOLOGY ‘Baru’ is the Dreamtime Crocodile Man from the Aboriginal mythological lore of Eastern Arnhem Land (Groger-Wurm, 1973). The specific name honours British actor Paul Dar- row, best known for his role in the television series ‘Blake's Seven’, in recognition of his sup- port of continuing palaeontological investiga- tions of the Riversleigh deposits. SPECIFIC DIAGNOSIS ‘That of the genus until additional species are known. MATERIAL EXAMINED HOLOTYPE. NTM P8695-8, 4 nearly complete cranilim missing the skull roof (frontals, parietals, postorbitals and squamosals) and basicranium posterior to the orbits. ParATyres, From D-Site, Riversleigh: NTM P8778- (1-5), right posterior mandible lragment, right posterior skull fragment preserving the lateral emparal fenestra, right pterygoid, ectopterygoid and posterior region of the maxilla; NTM P8681-14, left mandible lacking the articular and adjacent angular and surangular posterior tothe lateral foramen and asmall portion of the dentary at the Jevel of the third tooth; NTM P8738-1, right jugal, plerygoid, ectopterygoid and posterior maxilla and an associated dentary fragment; QM F16822, premaxilla and anterior portion of left maxilla retaining fourteen teeth; from Pancake Sile, Riversleigh: SAM P27866, right premaxilla; from Blast Site, Bullock Creek. NTM P87103-11, left squamosal, quadrate and opisthotic (juvenile), REFERRED SPECIMENS. From D-Site, Riversleigh; OM F16823, jugal fragment: OM F16824, premaxil- lary fragments; QM F16825, right dentary; OM F16826, right dentary; From Site Y, Bullock Creek, NTM P87105-1, right mandible fragments, Fram Blast Site, Bullock Creek, NTM P87103-12, juvenile right maxilla; NTM P8697-2, right jugal. TyPe LOCALITY. Blast Site, Camfield Beds, located ‘16 miles southeast of Camfield Homestead in north central Northern Territory’ (Plane and Gatehouse, 1968), AGE Late Oligocene to mid Miocene, A NEW LARGE BROAD-SNOUTED CROCODYLINE 52 TABLE |. Snout proportions of Bara avd other crocouilitns, wo Srachyuranochampsa eversolei Sebecus icaeorhinus Prisuichampsus vorax Ouinkana fortirostrum Osteolaemus tetraspis tetraspis Osteolaemus tetraspis osborni Paleosuchus palpebrosus Crocodylus porosus Alligator mississippiensis Gavialis gangeticus Burts darrowr L. is the distance from the anterior extremity of the orbit to the pasterior extremity of the external nares, H is the maximum depth of the snout at the filth alveolus, and W is the maximum breadth of the snout atthe fifth alveolus. Values for the first seven taxa are fram Molnar (1981.p,809), Values for Crocodylus, Alligator and Gavialis are from Australian Museum Specimens (AM R32646, AM R130772 and AM R131340 respectively) Values for Baruate from NTM P8695-8. STRATIGRAPHY Vertebrate thanatocoenoses often occur as geographically or stratigraphically discrete as- semblages in the middle Tertiary limestones of northern Australia. Because of uncertainty about the relationships (temporal and ecological) of these assemblages, it has become common prac- tice lo treal each as a separate local fauna (sensu Tedford, 1970), Archer et al (1989) suggest three significant jime periods are represented at Riversleigh be- tween the Oligocene and the Miocene. Wood- burne et al. (1985) suggest a mid to late Miocene age for the Bullock Creek Local Fauna. How- ever, ifthe more derived Aleoota Local Fauna is also considered late Miocene, the Bullock Creck Local Fauna is more appropriately designated as mid Miocene, The specimens of Baru darrowi from Bullock Creek were collected from the Blast Site and nearby Site Y, approximately 17°S, 131°30'E. It is not yet clear that any particular Bullock Creck site assemblage is signiticantly different fram any other and all have been tentatively referred to the Bullock Creck Local Fauna (Murray etal., In prep.). Consequently, the age range of Baru darrowi probably extends from late Oligocene (Riversleigh) to mid Miocene (Bullock Creek). ed DESCRIPTION Because of the limited comparative material available, Baru darrowi is compared with the living saltwater crocodile, Crocodylus porosus. However, Baru has much in common with more archaic crocodylines (¢.g, a wide incisive foramen and overlapping bile (Langston, 1973)), conditions apparently lost among the more derived living genera. Large triangular palatal fenestrae were also characteristic of many early Tertiary crocodylines, A comparison of the snout proportions of Baru darrowi with other erocodilians is given in Table 1. Table 2 provides a classification of snout proportions according to Molnar (1981). Table 3 Jists specific features of Baru darrowt and compures them with other crocodilians. Cranium. The cranium of Baru darrowi is triangular in dorsal profile and trapezoidal in section at the level of the maxillo-jugal suture. Compared to C, porosus the cranium of Baru is much deeper and broader in proportion to its length (Table 1). fn lateral profile (Fig. 1A) the cranium is deep, slightly wedge-shaped and nearly as high immediately posterior to the nanial aperture as it is just anterior to the orbits. The dorsal outline of the snout is concave. The profile of the premaxilla is distinctive in its shortness and depth. The anterior margin is a vertical sur- face, rounded ventrally and demarcated posteriorly by a wide notch for the caninifarm fourth mandibular tooth. In dorsal view (Fig, 1B) the premaxillae describe 4 broad, D-shaped sur- face immediately anterior to the tooth notches, Posterior to the constriction, the maxillae widen over laterally swollen alveolar festoons, Posteriorly the maxillac become more steep sided, slub-like and shallowly concave. Dorsally the nasomaxillary junction ts accentuated by a low crest. The alveolar process (sensu Molnar, 1981) is a wall of interconnected, buttressed alveoli. Anteriorly the alveolar process is strong- ly festooned bul pesteriorly it is more uniform, The jugal extends deeper ventrally and the maxillo-jugal suture is longer than in C. porosus of comparable size. The subtemporal ramus of the jugal widens laterally. Lateral to the lateral temporal fenestra this process is dorsoventrally flattened gradually becoming more rounded in cross section lateral to the quadratojugal. The lateral edge of the subtemporal ramus extends anteriorly as a ridge onto the broad anterior face of the jugal. In lateral view, the shape and size of the orbit is similar to and no less dorsally situated than in C. porosus. A well preserved 524 MEMOIRS OF THE QUEENSALND MUSEUM TABLE 2. Classification of snout proportions. A. SNouT DEPTH (H/W) B, SNOUT BREADTH (W/L)* Low x<0.5 Broad x2 0.66 Moderately deep, 0.5s x<1.0 Moderately Narrow 0.66> x>0.33 Deep x 21.0 Narrow xs0.33 * Molnar (1981,p.817) states that this ratio is L/W. This contradicts discussion of snout width ratios elsewhere in that paper. Molnar (pers. comm.) reveals that this ratio was intended to be W/L (not L/W). Classification of values for snout depth and snout breadth ratios according to Molnar (1981,p.817). portion of the jugal and quadratojugal indicates that the lateral temporal fenestra was both longer and wider than in any living crocodyline species. In dorsal aspect, the posterior of the cranium is about one-third broader than a C. porosus of equivalent length and the anterior is broader by a quarter. The premaxillae are wider relative to their length than in the Saltwater Crocodile and the narial aperture is shorter and broader. It ex- tends to the anterior margin of the premaxillae. Reception sockets for the first dentary teeth do not breach the outer surface of the snout as in C. porosus. Sutural relations on the dorsal surface of the cranium are essentially like those of C. porosus (Fig. 2). Well-developed dorsal processes of the premaxillae project posteriorly alongside the nasals. The premaxillae join in the midline anterior to the nasals, excluding them from the external nares. The paired nasals are elliptical in shape and slightly expanded posteriorly, shorter and less wedge-like than in C. porasus. The maxillae are greatly expanded laterally into deep, steep-sided lobes, which flatten out posteriorly before expanding outwards again at the base of the jugals. In the large mature specimen, the sutural pat- tern of the upper facial region is party obscured by age-related fusion and elaborate bony or- namentation. The basic pattern is like that of Crocodylus spp. The lachrymal extends anterior- ly to meet the nasal bone, so excluding the prefrontal from contacting the maxilla. The prefrontal forms the anteromedial orbital mar- gin. The posteromedial half of the orbit is formed by the orbital process of the frontal (Fig. 2). The position of the orbits, their shape and the mor- phology of the interorbital area are essentially the same as in C. porosus. The shape of the orbits of Baru differ from those of C. porosus only in being slightly longer, wider posteriorly and also more pointed anteriorly (Fig. 1B). The quad- ratojugals and jugals form a wide shelf bounding the comparatively large, triangular lateral tem- poral fenestra (approximately 52.0mm wide by 86.0mm long in NTM P8778-4). This opening is about twice the length of that of a C. porosus of equivalent size. A portion of the superior tem- poral fenestra is also preserved on that fragment and on a fragment of the skull roof of a much smaller individual, NTM P87103-11. These in- dicate that the proportions of the superior tem- poral fenestrae were similar to those of C. porosus. The auditory meatus is more anteriorly placed in Baru than it is in C. porosus. The portion of squamosal preserved on NTM P87103-11 indicates that the skull roof of Baru was flat and wider posteriorly than in C. porosus. The ventral surface of the cranium is dominated by the broad, flat maxillary palate with its wide alveolar processes and by the large triangular, anteriorly located palatal fenestrae (Fig. 1C). The premaxillae are penetrated by a large, oval incisive foramen recessed within a deep fossa, the anterior portion of which is con- fluent with a pair of reception pits for the first dentary teeth. In combination with the wide, deep alveolar process containing four large tooth sockets on each side, the premaxillary palate is distinctly vaulted in contrast to the relatively flat premaxillary palate of C. porosus. The maxillary palate is broad and short and is elevated above the alveolar margins. A row of small nutrient foramina clearly define the maxillary palate. The maxillary alveolar process is greatly expanded to accommodate the enlarged fourth and fifth max- illary caniniform teeth. Like Caiman and Os- teolaemus, Baru darrowi has fewer maxillary teeth than C. porosus and the size range of the tooth sockets and their corresponding teeth is greater than in any living crocodile. Baru specimens have a consistent number of thirteen maxillary teeth, as in the broad-snouted caimans. The moderate lateral compression of the tooth crowns of Baru is not clearly reflected in the shape of the alveoli which are predominantly round (Figs 1C,2,5A). The anterior maxillary teeth are posteriorly directed. The genus also A NEW LARGE BROAD-SNOUTED CROCODYLINE FIG. |. Baru darrawi holotype, NTM P8495-8: (A) lateral view, (B) dorsal view; (C) ventral view. MEMOIRS OF THE QUEENSALND MUSEUM 526 “eaysouay peroduiay so1adns ‘4.1.5 ‘yesowenbs Og {yI00} Je;Nqrpuew 103 yd uoNdasas ‘gy ‘yeSnfoyespenb “CO ‘prodAraqd ‘pg {equosaid “aud {Jeiquoisod ‘Od ‘ounjns Aseqyixew-oypixewaid ‘Sw ‘eyixewaid “Ww ‘esysauay auneyed ‘4g ‘ounryed “Ty ‘qi0 ‘YO ‘Seo yeseu ‘ON ‘feseu “VN SBypxem “Kw euAsyory] ‘Wy sqednf ‘ap Sesjsouoy pesoduigy [esoye] “YL I :Soseu [BUIOIUT “NI ‘UsWwesOy SAISIOUL “yf PeIUOA “YZ ‘prlosAsa9}doy99 “Og fginqiode [Bue ‘Nyy :SUONPIAIIQGY "8-S698d W.LN ad Ayojoy oy) uo ‘poyuasoidas AyJOod 10 ‘pajuasaidai jou |]Ny¥s 9Y} JO suo1od ayeorpur (xtjaid WLLN YIM []e) SIaquinu anBoyeyeD ‘soinjonzs TurmOYs IMOMEP MLE JO “SMAIA [B1JUIA PUP [PSIOP ‘uonRs0jsai ayisodwog (q) ‘UONR}UaWeUIO pue UOIPI9}11G0 ade Aq painosqo Aqjeried ase sainjns jesiop oy) Jo Auew uawiseds sty) Ul “8-S698d W.ALN ‘edAjojoy mossnp navg ay) JO Sainjns [esiop ay) JO uonryaidiaquy (Y) ‘Z “Old és 06 ar here a VW a oe 7 y-BLL8a] , oa dul ad rant at o nt A NEW LARGE BROAD-SNOUTED CROCODYLINE differs from Crocody/us in thal the crowns of the lower dentition occlude inside the upper tooth row, the longer dentary teeth having reception pits between and mesial to maxillary tecth four (hrough eight. The palatal fenestrae of Baru darrowi, in ad- dition to their large size and distinctive shape. invade the maxillary palate anteriorly to the level of the seventh maxillary tooth (Figs 1C,28). Tn C. porosus the fenestrae extend only to the ninth maxillary tooth. In this respect also, Baru is similar to short-faced crocodyline Os/eolaemus fe(raspis and the equally short-faced al- ligatorine, Paleasuchus trigonatus, \t differs from all living and most extinet crocadylids in lacking anterior palatine processes, The course of the maxillo-palatine is a wide chevron be- tween the anteromedial margins of the palatal fenestrae. The palatine bones ate concave medially to accommodate the long, posteriorly wide palatal fenestrae, In contrast fo C, porosus but like C. novaeguineae, the posterior margins of the palatal fenestrae are formed mainly by the pterygoids. In keeping with the width of the back of the cranium, the pterygoids are browd. In lateral profile, the ectopterygoids are longer and project ventrally at a somewhat different angle than in C, porosus. The posteroventral process of the eetopterygoids appears to be slightly longer than that of C, porosus and the anterior (palatal) process is decidedly more robust. Overall dimen- sions of the holotype are given in Figs 3A-C, Mandible. NTM P8681-14 comprises an al- most complete left mandible lacking only suran- gular posterior to the external lenestra, the coronoid, and a short length of the dentary bear- ing the third tooth. The first, second, fourth, tifth, sixth, fourteenth and fifteenth tecth are preserved. In gencral proportions the mandible is slightly larger than that of an approximately four-metre-long C. poroasus (NTM K13748), In oeclusal view (Fig. 4a) the symphysis extends posteriorly to just beyond the level of the sixth tooth, In C. poresus it ends level with the filth woth, The angle between the axis of the man- dibular ramus and the symphyseal plane is similar to that of C. porosus. The lateral surface of the manuible and the tooth row are concave laterally in contrast to a gentle convexity in C. porosus (Fig, 4a), The caniniform fourth tooth and its broad alveolus protrude laterally. NTM PSosSI-14 shows.a slightly greater variation in alveolar size than C. porosas although the pat- tern of tooth differentiation is basically the same. lan Ww 4 A pronounced difference in alveolus shape: is exhibited by the confluence of the tenth and cleventh, whereas in C. porasus the alveoli ate usually ee oho by between five and ten mil. limetres of bone. The greater degree of festooning in Barte reflects the enlargement of the eaniniform fourth tooth and the laterally compressed tenth and eleventh teeth, As in -C. porosus, Baru has a conspicuous excavation on the lateral surface of the dentary to accommodate the upper fourth and fifth maxillary teeth (Fig. 4b), The coronoid is not preserved on any specimen nor are there any examples of a complete Meck- clian fossa. Incomplete specimens indicate that the Meckehan fossa was similar in size to C. porosus. In the Bullock Creck specimen NTM P87105-1, the long axis of the Meckelian fossa is aligned ata rclatively high angle (c, 25°) to the inferior border of the mandibular ramus; in Riversleigh specimens and C, porosus itis nearly parallel to the inferior border (Figs. 4b, 5c). The lateral mandibular ramus 18 more heavily sculptured than that of C. porosus. In NIM PS779-2 and NTM P&7105-1, the seulptured region on the angular and surangular is delineated from the adjacent smooth bone by a prominent margin, In C. porosus, the wo surface textures in this region grade into.vach other, The external mandibular fenestra is narrower dor- soventrally and the posterior upward inflection of the inferior border of the mandibular ramus is: greater in Bartedarrowi(Figs 4b, Sb-c). On NTM PS681-14, a 5.0mm wide longitudinal sulcus originates from a small foramen located about 45,0mm from the last looth, A similar sulcus ts not evident on any Crocodylus specimens in our possession. A damaged articular ts preserved on NTM P&778-2 (Fig. 4b). It has a somewhat longer anterior process entering the adductor fossa than that of C. porosus, thus providing a relatively larger sutural contact with the angular, Sutural relations between the mandibular ele- ments show only minor differences from those of C. paresus. In NTM P&7105-1 and NTM P8778-2 the angular and surangular butt against cach other within the adducter fossa and ter- minale antenorly high on the posteroventrel margin of the external fenestra. In C. porasus they meet al a lap joint and the suture contacts the external fenestra in the mid-region of the posteroventral border. Varianilirty On the basis of a limited Selection of material, $28 Baru darrowi has been described as a variable species that existed over a considerable span of geological time. The possibility that more than one Baru species was present between Riversleigh and Bullock Creek times has been considered. At present, there is insufficient evidence to support a specific separation of the two populations due to lack of information about sexual dimorphism, ontogenetic changes and al- lometry in these extinct crocodiles. By analogy with living crocodiles, at least some observed differences between the Riversleigh and Bullock Creek specimens could be attributed to these factors. One of the more intriguing differences be- tween the two populations is the absence of serrations on the carinae of the teeth in the Riversleigh Baru sample. Riversleigh Baru specimens also have a more pointed premaxilla when viewed from the dorsal aspect. This may also relate to the apparently longer span between the fourth and the first dentary teeth observed in Riversleigh mandibles. While attempting to reconstruct Baru for an illustration, one of us (P.M.) was unable to match the length of the Riversleigh symphysis to the Bullock Creek premaxilla, although the remainder of the man- dible seemed to fit reasonably well in terms of size and shape. A single specimen of the dentary symphysis from Bullock Creek is proportionally shorter and fits the contours of the premaxilla of the holotype. Differences in the angle of the long axis of the mandibular fenestra to the inferior border of the mandibular ramus are noted above. Given the current state of resolution, we consider the definition of a single chronospecies subsum- ing these variations to be adequate for the time being. COMPARISONS WITH OTHER CROCODYLIDS Wider comparison of Baru darrowi em- phasises some of its more unusual features. This comparison is unavoidably brief and incomplete due to our limited comparative material. We confine our observations to crocodilians which MEMOIRS OF THE QUEENSALND MUSEUM have certain obvious similarities to B. darrowi either in terms of the dentition or cranial mor- phology. Living Crocodylids. Of living crocodylid species, Baru darrowi most closely resembles such broad-snouted forms as Osteolaemus tetraspis among the crocodylines and Paleosuchus trigonatus among the alligatorines (Table 1). Similarities include the number of maxillary teeth (13) and the size and position of the palatal fenestrae (Table 3). Paleosuchus spp. also possesses the alligatorine overbite which is similar to the condition in B. darrowi. Some caiman species have marked differences in tooth size, festooning and large caniniforms, whereas Osteolaemus tetraspis appears to have mildly durophagous specialisations. Although the pala- tal fenestrae of both forms are very large and end at the level of the seventh maxillary tooth, as in B. darrowi, they are differently shaped and have somewhat different proportional contributions to their margins from the surrounding palatal complex. A conspicuous difference is the presence in both living forms of a well defined anterior palatine process, absent in B. darrowi. These striking proportional similarities indicate that a substantial portion of Baru’s rostral mor- phology is trophically dedicated, derived and the result of parallel evolution. Similar remarkable parallel developments within various croco- dilian lineages are discussed by Langston (1973). The extent to which B. darrowi’s rostral proportions differ from Crocodylus porosus depends largely upon the state of maturity of the specimens compared. The holotype is obviously an adult and bears little resemblance to young specimens of C. porosus. However, when com- pared to a very large C. porosus, the width to length proportions (Webb and Messel, 1978) of Baru appear to differ very little (Fig. 6). This brings us to the peculiar case of the ‘Lansdowne snout’ (QM F1752), a Pliocene crocodile rostrum recovered from Lansdowne Station, Queensland, It was originally described as Pal- limnarchus pollens (Longman, 1925) but was FIG. 3. (A) Lateral view of Baru darrowi holotype NTM P8695-8 showing structures and dimensions. All measurements in millimetres. Abbreviations: EC, ectopterygoid; JU, jugal; LAC, lachrymal; NA, nasal; PRF, prefrontal; PT, pterygoid. (B) Dorsal view of the Baru darrowi holotype NTM P8695-8 showing structures and dimensions. The two circular structures on the premaxilla are artefacts produced by the breaching of the dorsal surface by the alveoli of the upper teeth. All measurements in millimetres. Abbreviations: AC, antorbital crest; CN, canine notch; JS, jugal sulcus; JU, jugal; NC, nasal crest; PMS, premaxillo-maxillary suture; PO, postorbital. (C) Ventral view of the Baru darrowi holotype NTM P8695-8 showing structures and dimensions. Abbreviations: MPS, maxillo-palatine suture; PES, pterygoid-ectopterygoid suture; PMS, premaxillo-maxi- llary suture; PPS, palatine-pterygoid suture. 529 A NEW LARGE BROAD-SNOUTED CROCODYLINE 530 later assigned to C. porosus (Molnar, 1982). The Lansdowne snout is proportionally shorter and broader than that of the B. darrowit holotype. Its ventral profile, moreover, closely resembles that of B. darrowi in its exaggerated maxillary swell- ings, short, broad premaxillary outline, its over- bite, and, so far as can be determined from MEMOIRS OF THE QUEENSALND MUSEUM illustrations (viz Molnar, 1982; fig. 5), its large and anteriorly positioned palatal fenestrae. Work in preparation by one of us (P.W.) and Molnar suggests that the Lansdowne snout should be referred to Pallimnarchus after all, but perhaps not P. pollens. Fossil Crocodylines. \n addition to its short FIG. 4. Baru darrowi, NTM P8681-14, left mandible from ‘D-Site’, Riversleigh: (a) occlusal view; (b) lateral view; (c) medial view. 531 A NEW LARGE BROAD-SNOUTED CROCODYLINE “FALL {s89001d plodXsajd rowaysod ‘gd ‘e4ysauay aunejed uowdeay |[Nys e1a}R]O191s0d YSU “p-9/ / 8d WLLN (P): Mala “prjsauay stuedui Ay “Ad {jiqio ‘YO fessoj jesodwi9] pesaqzey ‘ALI Seveu [RUIAIU! ‘NJ :SUOTIBIADIQGY “MATA [BSIOP jes91R] uawsesy sey Nqipuew JOLZa}Sod yYyBtI ‘Z-g/. L8d WLLN (9) :B|]]xewW Jo UO!sal JOLa}sod pure pioSAsajdoya ‘prod A1ayd yySis “¢-8/. 8d WLLN ‘MAA [es27e] (q) [MLA [BUDA (B) :YBIO|SIDAIY *,AUS-C, WO [eNPLAIPUl 878d W.LN ‘Mo4dnp Nang “COLA $32 MEMOIRS GF THE QUEENSALND MUSEUM TABLE 3. Characters of Baru darrawi and their distribution. CHARACTER A B Cc L. Procoelous vertebrae ? 2, Internal nares 3. Tooth enlargement sequence 4. Tooth notch 5. Lacrymal/nasal contact 6, Palatal fenestrae position (Mx, tooth number) 7. Palatal process 8. Occlusion 9, Jugal ridge 10, Pseudoheterodonty 11, Festooning 12. Snout width 13, Snout depth 14, Tooth compression 15. Serrated carinae 16, Teeth inclined to posterior ican ovwU bad NDDUO CS cueuuunuroak d =| a pUPyvUBo Se auaUvU Ss Teepe a F G H |! J a Pp Pp p p Pp ? P/F Po Pt RF Pr Pe P <— ot c al« eG Pp P Pp a P Pp P p Pp Pp a P Pp p m 10 9 12 7-8 7 8-9 Pp oP p Pp ap a/p a 0 oD i re) a i a 7? %p p p %?p Dp 2% a a P Pp Pp P Pp wt a, Pp P P p n m ob b b b ™ d md | J md | ? p p a a a a 9 P p a a a a 7 a a a a a a ? Key (o species: A, Baru darrowi, B, Quinkana fortirastram;, C, Pallimnarchus pollens; D, Sebecus icaeor- hinus; E, Pristichampsus vorax; F, Crocedylus perasus; G, Alligator mississippiensis, H, Paleosuchus osborni: 1, Osteolaemus tetraspis; J}, Brachyuranochampsa eversolei, Key to character states: a, absent; al, alligatorine; b, broad; ¢, crocodyline; d, deep: i, interlocking; |, low; m, moderately narrow; md, moderately deep; n, narrow; 0, overbite; p, present; Pt, pterygoid only; Pt/P, palatine and pterygoid contact. Interpretations from the following sources: A, B, C, F and G from specimens; D from Colbert (1946); E from Langston (1975); H and T from Mook (1921); J trom Zangerl (1944); all interpretations were compared and completed from Molnar (1981). snout, Baru darrowi has a distinctive broadly triangular cranium, great width of the jugals lateral to the orbits, elongation of the inferior temporal fenestra, large triangular palatal fenestrae that constrict the palatines posteriorly, absence of the anterior palatine processes and the elliptical shape of the nasal bones. Crocodylines with similar features were widespread in North America in the early Ter- tiary. One of the best preserved of these crocodilians is Brachyuranochampsa eversolei Zangerl, 1944, from the Washakie Eocene of Wyoming, U.S.A, Like Baru, Brachyurano- champsa combines the presence of a crocodyline notch for the fourth dentary tooth with an al- ligatorine-like overbite denoted by a series of reception pits medial to the upper tooth row. Although Brachyuranochampsa is a moderately narrow-snouted crocodyline (Table 1) it is heterodont and the alveoli are closely ap- proximated. The jugals are broad and everted, nasals are elliptical, inferior temporal! fenestra are large, the quadrates and quadratojugals are broad. The nasal aperture, although damaged anteriorly appears to have been terminal or near- ly so and trapezoidal in shape, like that of Baru, in contrast to the elliptical nares of Crecodylus spp. The palatal structure resembles Baru in its lack of an anterior palatine process and large, triangular palatal fenestrae and palatines that natrow posteriorly rather than widen as in Crocodylus. However, unlike Baru its dentition is not ziphodont and fourteen rather than thirteen maxillary teeth are present. The palatal fenestrae extend anteriorly only to a level between the eighth and ninth maxillary teeth. This is consis- tent with the observation that short-snouted crocodylines have more anteriorly-positioned palatal fenestrae. With our present state of knowledge it would be imprudent to force Baru into a phyletic relationship with this particular American genus, which may be expressing a symplesiomorphic character complex widely distributed among primitive early Tertiary crocodylines. However, given the dearth of other living and fossil forms that lack the anterior palatine processes combined with the broad similarities previously mentioned, the likelihood of an entirely parallel development of these fea- tures seems fairly remote, A NEW LARGE BROAD-SNOUTED CROCODYLINE 533 if ‘ ‘0 pB7104-2 FIG. 6, Comparison of the lateral profiles of the skulls of (A) Baru darrowi and (B) an extant saltwater crocodile, Crocodylus porosus, of approximately the same length. Among the contrasts with Crocodylus porosus, Baru posses deeper jaws with correspondingly exaggerated festoons, more anteriorly situated external nares, a conspicuous jugal crest and posteriorly slanted pseudoheterodont teeth. These features reflect significant differences in the manner of dispatching, and perhaps in its preference of, prey animals. Sebecosuchian Ziphodonts. Although clearly eusuchian, Baru is compared to sebecosuchian crocodiles because of its convergent ziphodont features. With the exception of its laterally com- pressed, serrated dentition, Sebecus shows few similarities with Baru. This is of some impor- tance because the concept of ziphodonty is often broadened to imply a specialised terrestrial predatory complex. The laterally directed orbits, high, narrow, convex snout profile and modifica- 25 20 Anterior-posterior (mm) 0 = 4 6 8 MEMOIRS OF THE QUEENSALND MUSEUM Plesiomorphs Apomorphs Saru darrawi (AR 9114) Saru darrawi (NTM P8736-1) 10 12 Thickness (mm) FIG 7. Scatter diagram showing tooth compression in various crocodilians. This shows the teeth of Baru darrowi lo be more compressed than those of Crocodylus porosus and Alligator mississippiensis but not as compressed as in the ziphodont forms Pristichampsus rollinati, Sebecus icaeorhinus and Sebecus sp. Measurements for ziphodont forms fram Langston (1956). Measurements for C. porosus and A. mississippiensis from unnum- bered specimens in the Australian Museum reptile collection. tions of the trochlear surface of the quadrate in relation to specialised jaw mechanics (Colbert, 1946; Langston, 1973) suggest that Sebecus was an active predaceous carnivore capable of purs- ing prey on land. Although the depth of Baru's snout appears to most closely approach that of Sebecus (Table 1) this is a somewhat misleading comparison because the convention of measur- ing the depth of the snout at the level of the fifth tooth includes the marked alveolar festoon. The proportions of the snouts of the two forms are actually very different; that of Sebecus is high and narrow and virtually triangular in section. Baru’s snout has a broad-based trapezoidal cross-section and is short and broad. Its lateral profile is strongly concave as opposed to the convex, narrow bridge of Sebecus. More impor- tantly , however, is the typical crocodyline dorsal orientation of Baru's orbits and its nares being sufficiently elevated, despite their terminal posi- tion, to allow the head to lie cryptically sub- merged. Pristichampsine Ziphedonts. Baru darrowi shows a greater degree of overall similarity with the early Tertiary Eurasian eusuchian ziphodonts of the subfamily Pristichampsinae than to the sebecosuchians. However, Pristichampsine crocodiles, known from several species of the genus Pristichampsus, are strikingly convergent with the sebecosuchians, not only in their pos- session of double-serrated and compressed teeth, but in the lateral position of the orbits, the narrow A NEW LARGE BROAD-SNOUTED CROCODYLINE snout and the similarly specialised cranioman- dibular joint (Langston, 1973), Although the trochlear surface of Baru's quadrate is tmper- fectly known, the shape of the yuadratojugal and the distal surface immediately preceding the jaw jaml are more reminiscent of Prislichampsus than Crocadylus. The articular of Baru indicates (hat its craniomandibular joint could be modified ulong the lines of the sebecosuchian and pris- tichampsine ziphodonts. Pristichampsus has a vaulled palate and the skull is narrow, as opposed tn broad, across the base of the jugals and through the orbijal region, The primary resemblance between Baru and Pristicnampsus isin the lateral View of the snout where the dorsal vulline is deeply concave, although the premaxilla of Pristichampsus is less bulbous. The teeth are moderately differentiated, at least in some species of Pristichampsuys, closely ap- praniinaiee and are directed slightly backwards rom the maxillary festoon (see Langston, 1973, fig. 4d),The notch for the fourth dentary tooth is weakly developed, particularly when viewed from below and the dentition is considerably less robust than in Baru. The teeth of Barw are moderately compressed, nol as compressed as in the ziphodont crocodilians (ic. Pristichampsus and Sebecus) but more so than in less derived crocodilians (Fig. 7), The palatal morphology. of Pristichampsus differs from that of Baru in possessing well- developed anterior palatine processes, propor- tionatcly similar to those of Crocadylus. The anterior palatine processes persist among the living short-snouted crocodylids and therefore their presence or absence Goes not appear ta be conditioned by the relative anteroposterior length of the palatal fenestrae. Apparently the resemblances between Baru and Pristichampsus ire largely plesiomorphic for crocodylids but include some clements of the ziphodont trophic complex. Australasian Endemic Crocedilians. The two endemic Crockd\lus species, C. jofinsani and C. novaeguineae are subsumed under the remarks previously made for Crocadylus. Besides the formally deseribed endemicaily Australian aenera, Pallimmarchus and Quinkania, there are other extinct species that are [oo poorly repre- sented to merit systematic designation, The ap- parent distinetiun of (he Australian Crocodylus species aad the remaining known Australian penera mikes it improbable that a direct ances- torUescendant relationship between them will be found en this continent. A compelling alter- 535 nalive, therefore, is to consider the possibility of a close relationship among the endemic genera not affiliated with Crocedylus. Quinkana fortirostrum Molnat, 1981, isa highly distinctive crocodilian characterised by a broad, short snout with a deep, convex profile, large anteriorly positioned palatal fenestrae and doubly-serrated laterally compressed teeth showing only moderate differentiation along the tooth row. Its short snout and palatal morphology is unlike that of either sebecosuchian or pris- tichampsine ziphodonts (Table 1), but its denti- tion is morphologically similar to members of thase groups. The type specimen, AM F,57844. is a fragment of the snout broken immediately anterior to the orbits, but including the.anterior margins of the palatal fenestrae and the anterior palatine suture. In section, the rostrum is \rapezoidal, with well developed alveolar processes. Due to the pasi- tion of the break, Molnar (1981) was able to examine the internal structure of the snot cavily, He observed that “A high, posteriorly concave partition dorsally bounds the anterior margin of the palatal fenestra. A similar bul Icss developed partition is found in Crocodylus johnseni, C, novaeguineae and C. porpsus, where, however, it is placed Well anterior to the margin of the fenestra, and is restricted to the lateral portion of the snoul cavity. In Q. for- firostrien the maxilla ts excavated anterior ta this partition, forming lateral chambers thal open posteriosly, Above the junction of the palatal processes of the two maxillac rise two thin, near- ly vertical Manges, which together form a narrow trough along the floor of the snout cavity’ (Mol- nar, 1981), Itis therefore of some importance to note that a similar arrangement occurs in Barw. Howover, as this condition is regarded only in contrast ta the typical Crocedylus condition, we are unsure of the marphology of the same region in other short snouted crocodylids such as Os- teolaemus and Paleasuchus in which (he more anterior position of the palatal fenestrae might also determine similar relations. Although desenbed by lordansky (1973), be does not com- pare this region in various genera. Q@uinkana and Baru lack the anterior palatine processes, Which is unusual among cmendilians and apparently not conditioned by jhe anteriar disposition of the palatal fenestrae, or by the proportions of the intertenestral laminae of the palatines. Quinkana js othenyise very different from Baru, but has few specific similarities wilh any other group of crocndilians. Thus Quinkana has a combinanan 3 of characters; some ziphodont features {imply- ing a terrestrial or scmi-terrestrial predaceous existence) others unique and a few, very specific and rather compelling features, suggestive of a close relationship with Baru, To date, specimens of Pallimnarchus pollens have been fragments and no complete skulls are known (Molnar, 1982). However, nearly com- plete snouls referable to Pallimnarchus have recently become available for study (Willis and Molnar, in prep.). This more complete material reveals that Pailimnarchus has anteriorly located palatal fenestrae (anterior leve] with the seventh alveoli) and lack anterior palatine processes. Tecth referred to Pallimnarchus (Motnar, 1982) are distinguished by serrate carinae on a broadly conical form. A more complete comparison with Baru will have to wait until the new Pallimnar- chus material is properly described, We are unable to fully support the hypothesis that the three known Australian endemic crocodiles represent a monophyletic group be- cause of limited comparative material and in- complete fossils. There is, however, sufficient evidence jo indicate that this is a solid alternative to the notion of sebecosuchian and/or pris- lichampsine ancestry of the group. The possible relationship between these forms are considered in the following section. CHARACTER ANALYSIS The following examination of crocodilian character states is based on Molnar (1981), Norell(1989) and Benton and Clark (1988). Mol- nar used character frequency to determine char- acier polarities where as Benton and Clark, and Norell, used the outgroup method proposed by Maddison et a). (1984). We have accepted the polarity of characters as determined by Molnar, Benton and Clark, and Norell. The polarity of new characters intraduced into this study have been determmed by their distribution among the ten taxa indicated tn Tuble 3. SUBORDINAL CHARACTERS |) Proceolous vertebrae have been found with specimens altributed ta B. darrowi No am- phicoclous vertebrae are known from deposits from which B. darrowi has been found, It is therefore a reasonable assumption that Barw had proceolous vertebrae, which is recognised as a cusuchian character (Steel, 1973; Kuhn, 1968). Benton and Clark (1988) recognise praceolous vertebrae as an apomorphy of a group that in- MEMOIRS OF THE QUEENSALND MUSEUM cludes Eusuchia and an undescribed early Cretaceous crocodile from North America. 2) The movement of the internal nares posteriorly in advanced crocodiles was recog- nised by Huxley (1875). The internal nares are completely surrounded by the pterygoids in Baru. This is regarded as a eusuchian character state (Steel, 1973; Kuhn, 1968; Benton and Clark, 1988), SUBFAMILIAL CHARACTERS 3) The pattern of tooth enlargement in the crocodilian skull has been used to distinguish members of the Alligatorinae from the Crocodylinae (e.g. Steel, 1973). In alligatorines the fourth maxillary tooth is usually the largest; in crocodylines it is the fifth, In Bara the fifth tooth is largest, 4) The presence of a notch between the premaxilla and maxilla can be used to distin- guish alligatorines from other crocodilians (Steel, 1973), In crocodilians the fourth dentary tooth fits into this notch when the jaws are closed. In alligatorines this tooth usually fits into a pit in the palate medial to the upper tooth row. Bart conforms to the plesiomorphic condition. 5) In Alligater and many fossil alligatorines the fachry'mal is separated from the nasal bone by the maxilla (a derived condition), whereas in crocodylines and the caimanoid alligatorines the lachrymal contacts the nasals, This may also be expressed as the prefrontals lacking any contact with the maxilla in crocodylines. Barw is crocodyline in this respect, FEATURES OF AUSTRALIAN FORMS 6) Baru, Quinkana and Pallimnarchus have large anteriorly placed palatal fenestrae, This condition appears to be part of a functional com- plex related to short, broad snouts, 7) Most crocodiles possess an anterior palatine process. Baru, Quinkana and Pallimaarchus lack this process. The only other cusuchian crocodiles for which descriptions are available that lack these structures are those of the American Eocene genus Brachyuranochampsa. This is a moderately narrow snouted form with ore posteriorly situated palatal fenestrae. Therefore it appears that the lack of the anterior palatine processes is independent of the position of the palatal fenestrae. 8) Molnar (1981) determined interlocking teeth to be a derived crocodilian state. However, Norell (1989) determined that an overbite, as seen in Baru and Quinkana, is the derived state. A NEW LARGE BROAD-SNOUTED CROCODYLINE Norell’s determination is accepted here because of his use of the outgroup method of Maddison etal. (1984). 9) The conspicuous jugal ridge observed in Baru appears to be an unique feature among crocodiles. Its presence in Baru is taken to be autapomorphic. ZiPHODONT CHARACTERS 10) Molnar (1981) considers highly differen (lated crocodyline dentitions to be plesiomorphic and more uniform dentitions of longirostrine and ziphodont crocodiles to be derived. 11) Festooning is a plesiomorphic feature. The derived condition is a straight tooth raw (Molnar, 1981). These conditions are clearly associated with the degree of size differentiation of the dentition. 12) Extremely narrow snouted eusuchians are derived. Moderately narrow to moderately broad snouts are plesiomorphic. Extremely broad, short snouts are also derived (Molnar, 1981). Quinkana and Baru are unusual ziphedonis in having short, broad snouts. Pelresawrnus (Gasparini, 1982) appears to be a ziphodont with a moderately broad or broad snout. 13) According to Molnar's classification, ziphodont crocodilians have ave to moderately deep snouts. He proposes that this is a derived state. Both Baru and Quinkana have moderately deep snouts. The plesiomorphic condition is a Jow snout form. 14) Laterally compressed teeth are considered to be derived. The plesiomorphic condition ts a tooth of circular or broadly oval cross section. The teeth of Quinkana ate decidedly compressed whercas those of Baru retain the plesiomorphic conical shape towards the base, becoming progressively flattened on the lingual side towards the lip of the crown, 15) Serrations are not known to occur on any round conical crocadilian teeth (with the excep- tion of some teeth attributed to Pallimnarcius, Molnar, 1982), they are invariably associated with some degree of transverse compression of the crown. Laterally compressed teeth with ser- rated edges are termed ziphodont. The ziphodant condition is a derived character state, 16) Posterior inclination of the teeth appears to be an unusual feature in crocodilians. The con- dition may be present in the sebecosuchian Baurusuchus and perhaps to some extent in Pris- tichampsus. The condition is probably a derived one, tn La J DiscuSSION The most complete cladistic analysis of the Crocodilia is that of Benton and Clark (1988), They left the crocodylids (including pavials, al- ligators and crocodylines) as an unresolved crown group. Norell (1989) analysed this crown group based on twelve characters and defined the relationship between the gavialinae, crocodylids and alligatorids, Unfortunately, Norell’s work was published after this paper had been reviewed and his characters have not been fully incor- porated in this analysis. However, Baru has all three apomorphies that Norell has used to separate crocodylids from gavials and alligators, Baru retains many plesiomorphic crocodyline features, [ts principle derivations are related to s Specialised trophic complex which involves some elements of the so-called ziphodont condi- tion, As is often the case with ancient surviving groups, they are exceptionally conservative in their basic morphology and many lineages have paralleled and converged in their trophic mechanisms. It is under these circumstances that the phylogenetic methodology becomes severe- ly strained. Most spomorphic features are dedi- cated to trophic adaptations and the field of relevant character states (discrete or exclusive characters mdependent of functional require- ments) are few and difficult to substantiate. In terms of phylogenetics, therefore, we are con- fined to a single possible synapomorphy, the absence of the anterior palatine process, in wnit- ing the three extinct Australian genera under consideration. Ziphodont teeth have evolved convergently and in parallel, und anteriorly placed palatal fenestrae have evolved indc- pendently in the caiman and Osteolaemus, We are unable to verify the uniqueness of the similarity of the internal rostral partitioning in Barwand Quinkana at this time duc to lack of the necessary specimens, The absence of the anterior palatine processes appears to be the least trophically related apomorphic character uniting Baru, Quinkana and Patlimnarchus with another group (e.g. Brachyuranechampse). We consider this possible relationship to be a more par- simonious one than basing a relationship with the Pristichampsinae, on the assumption that the anterior palatine process was bost in parallel. The ingroup interrelationship of Baru, Pallim- narchus and Quinkana are little closer to resolu- tion. Bera and Pallinmarchus are more plesiomorphic than Quintana according to the character polarities used here. However, Pallim- narchus is not sufficiently well known to deter- mine its phyletic position relative to Baru, WW appears, however, that Baru darrowi is too specialised to have given rise to Quinkana. We are therefore unable to build a connected se- quence and must assume that another clade for which we have no information is involved, PALAEOBIOLDGY A detailed functianal anulysis of Baru’s cranial analomy must preclude any definite con- clusions as ta the nature of its traphic specialisa- (roms, However, its distinctive dentition and robust proportions justify some speculittion on the nature of its habits. The prominent upper and lower festonns bear- ing large, posteriorly-directed and closely spaced Wweth constitute a specialised cleaver-like biting mechanism, designed to deliver an jmme- diale incapacitating, blow to its prey. The upper and lower festoons and their dentitions comple- ment ope another so as to produce a fulcrum ubove which the lower caniniforms drive into the prey, The resultant is a combined shearing and tissue deforming (tearing and breaking) uction cipuble of breaching tough, flexible material (thick hides, as well as more durable materials such as armoured skin and perhaps bony carapaces), Because of the fulcrum-like siruc- lure af the interposed maxillary festoon, tissues are stretched against and severely deformed by i trangle of forces, The large pasteriorly-angled teeth restrain the prey-object during the curly phases uf jaw closure, when resultant forces exeried by the jaw tend to drive the object for- Ward, The purpose of the serrations in Haru appears to be a secondary refinement in which the strug- gling movements. of the prey combined with small movements of the jaws and perhaps equal- ly importantly, elevation and depression of the head at the craniocervical joint can continue to sever Lissucs in the grasping period during which the adducted jaws are restricted in (heir move- ment, Baru's dental specialisations are therefore in- terpreted as a mechanism tor rapid immobi- lisation of relatively Jarge prey.Judging from the dimensions of the type, Baru was capable of killing animals up to 300 kg in weight based on unalugous feats by the saltwater crocodile, Baru would therefore have been @ likely predator of mammals aiid other large crocodiles, as its dental com pia and powerful adductar mass was capable of breaching armoured hides. MEMOIRS OF TITE QUBENSALND MUSEUM The remains of Baru are consistently as- sociated with fluvio-lacustrine sediments. Its short, broad heavy cranium and the morphology of its atlas-axis complex indicate that it had no greater head mobility than C. porosus, which would have limited an active terrestrial predator. Unlike sebecosuchians, pristichampsines and Australia’s ziphodont Quinkana, Barw has dor- sally oriented orbits like aquatic crocodiles which spend the majority of their lives partially submerged. We conclude that Baru was. an aquati¢ crocodile adapted to shallow, inland freshwater lakes and small streams in which the saltwater crocodile habit of dragging its larger prey into deep water may not have heen pussible In shallow water and narrow streams the prey has an opportunily lo continue lo Struggle, whereas the saltwater crocodile is often able to release tls half-drowned prey to effect a new grasp. Barte probably ambushed large mammals from the edge of streams and shallow lakes relying on its powerful bite to incapaciiate its prey through shock-inducing trauma. If Baru were a terrestrial or semi-terrestrial form, the risky and energy- consuming action of immediately immobilising its prey would be unnceessary. Large terrestrial tepuiles, best known from the studics of the Komodo Dragon by Auffenberg (1982), initially injure large prey by hamstringiny i from behind, then follow ituntil itexpires from exhaustionand bleeding. This pattern appears more appropriate to species of Pristichampsus and Quinkana. The postulated predatory behaviour of Baris too specialised to suggest a preadaptation to lerrestrialily for Quinkana, However, terrestrial predation in crocodiles probably had its roots in behaviour in which prey was ambushed from the water, then followed onto the land, CONCLUSIONS Functional complexes reflecting trophic udap- talions, superimposed on a general morphologt- cal conservatism have produced numerous convergences. within the Crocodilia, Conse- quently, (axovomic relationships are difficult to Unravel, The contribution of the pterygoids to the secondary palate, posterior intra-plerygoidal position of the internal nares, confluent external nares, subdermal postorbital bar, small superior temporal fenestra, well developed mandibular fenestra and associated proccolous, keeled cer- vical vertebrae (NTM P9778) place Baru in the Eusuchia, Family Crocodylidae (sensu Romer, ANEW LARGE BROAD-SNOUTED CROCODYLINE 1956), The diagnostic enlarged fifth maxillary tooth and the lateral notch at the maxillo- premaxillary suture lo accommodate the fourth mandibular tooth align Baru with the Crocodylinae. Baru shares a number of character states with two other Australian endemic fossil genera, [ls incipient ziphodonty, broad snout, presence of a similar arrangement of the internal partition of (he maxilla and similar palatal proportions may support a phylogenetic relationship with Quinkana and Pallimnarchus. The absence of the anterior palatine process in all three of these egencra may link them to the Eocene North American taxon Arachyuranochanipsa, and dis- tinguishes the Australian crocodiles from the pristichampsine ziphodonts. In Baru, the shorter, wider and deeper rostrum, hypertrophied festooning, greatly differentiated jaoth size and laterally compressed serrated teeth are a functionally related complex and as such are notreliable taxonomic indicators, Ziphodont teeth have evolved independently in several crocodilian lineages and have been associated with highly specialised, perhaps terrestrial predatory habits. Adaptations apparently suited to a terrestrial predatory mode include a convex deep, dorsal snout profile and dorsolaterally directed orbits and external nares. In contrast, Baru has elevated premaxillae, high anterodor- sal placement of the external nares, concave dorsal snout profile and dorsally oriented orbits. These features indicate that Baru was an aquatic crocodile, ACKNOWLEDGMENTS We would like to thank M. Archer, R, Molnar, R, Sadlier and N, Pledge for access io material used in this paper. Ralph Molnar, Suc Hand, John Scanlan and M, Archer provided advice and help in technical matters associated with this paper and assisted with constructive commients on early drafts. Thanks are also extended to Paul Darrow and The Blake’s Seven Society of Australia (and in particular Elaine Clark) and Dr Tony Wicken for financially assisting this work. Janine Willis and Heather Bender helped type ihe final draft, General support for the Riversleigh Project which enabled some of the material studied here to be obtained was provided by: The Australian Research Grant Scheme (Grant PG A3 851506P); The National Estates Grants Scheme (Queensland); The Department of Arts, Sport, 539 the Environment, Tourism and Territories; Wang Australia Pty Ltd; L.C.1, Australia Pty Leds the Queensland Museun); the Australian Museum; the Australian Geographic Socicty; Mount Isa Mines Pty Ltd; the Royal Australian Air Force; the Australian Defence Force; the Riversleigh Consortium (Riversleigh being a privately owned station); the Mount Isa Shire: the Riversleigh Society; the Friends of Riversleigh; Probe and the many volunteer workers and colleagues. A Northern Territory Heritage Grant provided funds for documentia- tion of the Camfield Locality. We thank Wayne Brown, Karl Roth and Jan Archibald for their help in excavation. LITERATURE CITED ARCHER, M., GODTHELP, H.. HAND, S.J, AND MEGIRIAN, D. 1989. Fossil mammals of Riversteigh, northwestern Queensland: Prelimi- nary overview of biosiratigraphy, correlatan and environmental change, Aust, Zool, 25 (2): 29-65, AUFFENBERG, W. 1982. ‘The behavioral ecology of the Komodo Dragon’. (University Press of Florida: Gainesville, Florida), BENTON, M.J. AND CLARK, J.M. 1988, Archosaue phylogeny and the relationships of the Crocodilia, pp 295-338, Jn Benton, M, J. (ed,), ‘The phylogeny and classification of the tetrapods, Vol. 1; Amphibians, reptiles, birds’. Systematics Association Special Volume No, 35A, (Clarendon Press; Oxford), COLBERT, E.H, 1946. Sebecus, representatives of a peculiar suborder of fossil Crocodilia from Patagonia. Bull. Amer, Mus. Nat. Hist. 87: 2L7- 270. GASPARINI, Z. 1982. Una nueva familia de crocudrilps. zifodontes Cretacicos de America del sur, Quinto Congreso. Latinamericano de Geolagia Argentina 3: 317-329, GROGER-WURM, H.M, 1973, ‘Australian aboriginal bark paintings and (heir mythological interpretation”, (ASIS; Canberra), HECHT, M.K. AND ARCHER, M. 1977. Presence of xiphodont crocodilians in the Tertiary and Pleis- tocen¢ of Australia. Alchesinga 1; 383-385, HUXLEY, T.H, 1875, On Stagonalepis rabertsoni, and the evolution of the Crocodilia. Quart. J - Geol. Soc. London 31:.423-438. IORDANSKY, N.N. 1973. The skull of the Crocodilia, pp.201-262. fn Gans, C. and Par- sons, ‘T.S. (eds), ‘Beology of the Reptilia. 4. Morphology DB.’ (Academic Press: London), 540 KUHN, O. 1968. ‘Die Vorzeitlichen Krokodile.’ (Oeben: Munich). LANGSTON, W. 1956. The Sebecosuchia; cos- mopolitan crocodiles? Amer. J. Sci. 254: 605- 614. 1973. The crocodilian skull in historical perspec- tive. pp. 263- 284. In Gans, C. and Parsons, T.S. (eds), ‘Biology of the Reptilia. 4. Morphology D.’ (Academic Press: London). 1975. Ziphodont Crocodilia. Fieldiana, Geol. 33: 291-314. LONGMAN, H.A. 1925. A crocodilian fossil from Lansdowne Station, Mem. Qd Mus. 8: 103-8. MADDISON, W.P., DONOGHUE, M.J. AND MAD- DISON, D.R. 1984. Outgroup analysis and par- simony. Syst. Zool. 33: 83-103. MOLNAR, R.E. 1981. Pleistocene ziphodont crocodilians of Queensland, Rec. Aust. Mus. 33: 803-834. 1982. Pallimnarchus and other Cenozoic crocodiles of Queensland. Mem. Qd Mus. 20: 657-673. MOOK, C.C. 1921. Skull characters of recent Crocodilia with notes on the affinities of the MEMOIRS OF THE QUEENSALND MUSEUM recent genera. Bull. Amer. Mus. Nat. Hist. 44: 123-268. NORELL, M.A. 1989. The higher level relationships of the extant crocodilia. J. Herp. 23 (4): 325-335. PLANE, M. AND GATEHOUSE, C.G. 1968. A new vertebrate fauna from the Tertiary of northern Australia. Aust. J. Sci. 30: 272-273 ROMER, A.S. 1956. ‘Osteology of the reptiles’. (Uni. Chicago Press: Chicago). STEEL, R. 1973. Crocodylia. Jn Kuhn, O. (ed.), ‘Handbuch der Palaoherpetologie, En- cyclopedia of Paleoherpetology.’ Vol. 16. 116 pp. (Gustav Fisher Verlag: Stuttgart). TEDFORD, R.H. 1970. Principles and practices of mammalian geochronology in North America. Proc. N. Amer. Paleon. Conv., Sept., 1969: 666- 703. WEBB, G. AND MESSEL, H. 1978. Morphometric analysis of Crocodylus porosus from the north coast of Arnhem Land, northern Australia. Aust. J. Zool. 26: 1-27. ZANGERL, R. 1944. Brachyuranochampsa ever- solei, gen. et sp. nov., a new crocodilian from the Washakie Eocene of Wyoming. Ann. Carn. Mus. 30: 77-84 MEMOIRS OF THE (QUEENSLAND MUSEUM BRISBANE © Queensland Museum PO Box 3300, South Brisbane 4101, Australia Phone 06 7 3840 7555 Fax 06 7 3846 1226 Email qmlib@qm.qld.gov.au Website www.qm.qld.gov.au National Library of Australia card number ISSN 0079-8835 NOTE Papers published in this volume and in all previous volumes of the Afemoirs of the Queensland Museum maybe teproduced for scientific research, individual study or other educational purposes. Properly acknowledged quotations may be made but queries regarding the republication of any papers should be addressed to the Editor in Chief. Copies of the journal can be purchased from the Queensland Museum Shop. A Guide to Authors is displayed at the Queensland Museum web site A Queensland Government Project Typeset at the Queensland Museum CONTENTS fy Ce eA A River it eee Aleta? SOS eG ce ae, eae eY A Oa ee som Ser ae ee Cree Conferciice photorraph -/ re oa. eens ee be else de bat A ae een ae A ary PAPERS BAuER, A.M. AND RUSSELL, A.P. Alternative digital scansor design in the New Caledonian gekkonid genera Bavayia and Eurydactylodes . . . Bauer, A.M. AND RUSSELL, A.P. Dentitional diversity in Rhacodactylus (Reptilia: Gekkonidae) ... 2... ... 02.0000 2 0c eee eee ae BaVeRSTOCK, P.R. AND DONNELLAN, S.C. Molecular evolution in Australian dragons and skinks: a progressreport .........2+- Carter, D.B. Courtship and mating in wild Varanus varius (Varanidae: Australia) .........0+5- Covacevich, J., Couper, P., MOLNAR, R.E., WITTEN, G. AND YOuNG, W. Miocene dragons from Riversleigh: new data on the history of the family Agamidae (Reptilia: Squamata) in Australia ....... A ey 2 ee ee ihe tac A > oe fee CzecuurA, G.V. AND INGRAM, G. Taudactylus diurnus and the case of the disappearing frogs .... . Lee eeLr a” Onno aeemes- ane ren Eee et nae DanieLs, C.B. The relative importance of host behaviour, method of transmission and longevity on the establishment of an acanthocephalan population in two reptilian hosts .......... DaniELs, C.B. AND HEATWOLE, H. Factors affecting the escape behaviour of ariparian lizard .............. ioe a 245 tt oy HamLey, T. An inexpensive force platform for use with small animals: design and application .. 2.0.5. .005+004 Hutcuinson, M.N. The generic classification of the Australian terrestrial elapid snakes INGRAM, G.J. AND CZECHURA, G.V. Four new species of striped skinks from Queensland ........-.2 2000s eee eeer ence ‘ Jenkins, R.W.G. AND Fores, M.A. Elements i in the process of recovery by Crocodylus porosus (Reptilia: Crocodilidae) in the East Penis th River apdlaasopiaion wellness geri cae tees Meee ads adenes Pur ek ey che tree ct se ay Og ; MILteER, J.D. AND Jones, M.E, Growth and calcium metabolism of embryos of the long-necked tortoise, Chelodina logicollis (Shaw) Mo nak, R. New cranial elements of a giant varanid from Queensland 2.0.1... eee eee ee ee ee Movnakr, R.E. AND CZECHURA, G.V, Putative Lower Cretaceous Australian lizard jaw likely afish . . NAYLOR, L. Treatment of cloacal prolapse in the Estuarine Crocodile... 2. 66s see eee eee RusseLt, A.P, AND BAUER, A.M. Digit I in pad-beating gekkonine geckos: alternate designs and the potential constraints of phalangeal number . RUSSELL, A.P. AND BAUER, A.M. Oedura and Afroedura (Reptilia: Gekkonidae) revisited: similarities of digital design, and constraints on the development of multiscansorial subdigital pads? ......... a ee on vk ne ets SADLIER, R.A, The scincid lizard genus Nannoscincus Giinther: arevaluation .. 1... 6. ee ee ee Suea, G.M. The genera Tiliqua and Cyclodomorphus (Lacertilia: Scincidae): generic diagnoses and systematic relationships WILLIs, P., MURRAY, P., AND MEGIRIAN, D. Baru darrowi gen. et sp. nov. a large, broad-snouted crocodyline (Eusuchia: Crocodylidae) from mid-Tertiary freshwater limestones in northern Australia .......... SRG ARTE AB LS Cec k ERAS, tha eee ie NOTES Covacevich, J., Couper, P., MCDONALD, K. AND TriGceR, D. Walnunarra, bungarra mali and the gangalidda at Old Doomadgee ... . . es Ses cig Met et cos Ee) CzecuuRA, G.V. Further evidence of ophiophagy in an Australian falcon... 1... ee ee eee CzeEcuurRA, G.V. AND INGRAM, G.J Notes on the distribution and occurrence of some striped skinks (genus Cfenotus) in Queensland ......... INGRAM, G.J, AND Covacevicn, J. Tropidonotus mairii vs Bufomarinus ........4.5. een ata! it Set PEARN, J. Herpetologists and snake-bite ...........+.--04- Sak RAVEN, R.J. Spider predators of reptiles and amphibia 2.1... et eee ee ee ee ee ee eee 453 473 487 . 495 521 322 332 366 396 406 . 448