B RE VI O R A M mseuinri mi p <a r a 1 1 v t ©logy US ISSN 0006-9698 Cambridge, Mass. 19 October 2009 Number 518 AN ARCHOSAUR-LIKE LATEROSPHENOID IN EARLY TURTLES (REPTILIA: PANTESTUDINES) Bhart-Anjan S. Bhullar 1 and Gabe S. Bever 2 Abstract. Turtles are placed with increasing consistency by molecular phylogenetic studies within Diapsida as sister to Archosauria, but published gross morphology-based phylogenetic analyses do not recover this position. Here, we present a previously unrecognized unique morphological character offering support for this hypothesis: the presence in stem turtles of a laterosphenoid ossification identical to that in Archosauriformes. The laterosphenoid is a tripartite chondrocranial ossification, consisting of an ossified pila antotica, pila metoptica, and taenia medialis + planum supraseptale. It forms the anterior border of the exit for the trigeminal nerve (V) and partially encloses the exits for cranial nerves III, IV, and II. This ossification is unique to turtles and Archosauriformes within Vertebrata. It has been mistakenly dismissed as anatomically dissimilar in these two groups in the past, so we provide a complete description and detailed analysis of correspondence between turtles and Archosauriformes in each of its embryologically distinct components. A preliminary phylogenetic analysis suggests other potential synapomorphies of turtles and archosaurs, including a row or rows of mid-dorsal dermal ossifications. Key words: Archosauria; Archosauriformes; Diapsida; turtle origins; chondrocranium; Proganochelys; Kayentachelys; fossil; braincase; interorbital ossification; Testudines Turtles (Pantestudines; Joyce et al., 2004) have traditionally been classified as "ana- psid" reptiles owing to their lack of the lateral and dorsal fenestration of the skull 1 Department of Organismic and Evolutionary Biology, Harvard University, Biolabs room 4110, 16 Divinity Avenue, Cambridge, Massachusetts 02138, U.S.A.; e- mail: firstname.lastname@example.org 2 Division of Paleontology, American Museum of Natural History, Central Park West at 79th Street, New York, New York 10024, U.S.A. that is ancestral for diapsid reptiles, includ- ing tuatara, lizards, crocodiles, and birds (Gauthier et al., 1988, and references there- in). Early gross morphology-based phyloge- netic analyses suggested that turtles are the sister taxon to Diapsida and thus one of the two branches of the initial reptilian diver- gence (Gauthier et al., 1988). Most subse- quent morphological analyses have either supported this position (Brochu, 2001; Laurin and Reisz, 1995; Lee, 1997) or have ® The President and Fellows of Harvard College 2009. BREVIORA No. 518 placed turtles close to the marine Euryapsida along the stem of the lizard/tuatara clade Lepidosauria (Li et al., 2008; Rieppel and Reisz, 1999), thus suggesting that they are highly modified diapsids. One analysis (Merck, 1997) similarly indicated affinities to the Euryapsida but recovered a novel result because the included characters of non-turtle euryapsids placed the entire turtle + euryapsid clade as sister to the archosaur lineage (Brochu, 2001). In contrast, a growing body of molecular phylogenetic work strongly supports a posi- tion of turtles within Diapsida as sister to the crocodile/bird clade Archosauria (Cao et al., 2000; Iwabe et al., 2005; Kumazawa and Nishida, 1999; Organ et al., 2008). Until now, no unique gross morphological support has been reported for archosaur affinities of turtles (Rieppel, 2000), and in particular, no morphological evidence has been forthcom- ing that would help place turtles along the archosaur stem. However, such evidence has existed, largely overlooked, since the further preparation and monographic description by Gaffney of the best-preserved stem turtle, Proganochelys quenstedti, from the Late Triassic (Norian) of Germany (Gaffney, 1990). A single specimen of P. quenstedti (SMNS 15759) preserves the region anterior to the braincase. In this region, which would in life have been occupied by the membranous anterior braincase, a pair of dorsoventrally tall, flat ossifications articulate with the prootic and basisphenoid on each side (Fig. 1A). The initial description of this region by Gaffney (1990) documented the form of these bones but did not treat the detailed morphology of each of their pro- cesses. It was noted that they are similar to a pair of ossifications synapomorphic for the clade Archosauriformes, the pleurosphe- noids (Fig. IB), which are now usually called laterosphenoids (Clark et al., 1993). Howev- er, the general consensus at the time, including the hypothesis presented by Gaff- ney (1990), was that turtles were sister to all other extant reptiles. These elements were thus termed "pleurosphenoids," with the quotation marks indicating probable non- homology with those of Archosauriformes. We posit, in contrast, that they are in fact homologous to the laterosphenoids of Arch- osauriformes. MATERIALS AND METHODS All specimens examined are from the collections of the Museum of Comparative Zoology, Harvard University. The following specimens from the Herpetology collection were examined: Alligator mississippiensis MCZ 17711, 34323; Caiman crocodilus MCZ 5031; Crocodylus cataphractus MCZ 13985, 175004; C niloticus MCZ 4372; C porosus MCZ 72937; Gavialis gangeticus MCZ 33950; Osteolaemus tetraspis MCZ 22913; Paleosuchus palpebrosus MCZ 84030; Tomistoma schlegeli MCZ 12459. From the Ornithology collection: Tinamus major MCZ 342723, 342774. From the Vertebrate Paleontology collection: Eothyris parkeyiMCZ 1161. Phylogenetic analyses, as described below, used a modified version of the matrix from Dilkes (1998). Both used parsimony searches in PAUP* v4.0bl0 (Swofford, 2001) with the branch-and-bound search option (1,000 rep- licates), specifying Petrolacosaurus as the outgroup as in Dilkes (1998). The con- strained search used the monophyly con- straint option to unite Proganochelys with the archosauriform clade, including Eupar- keria and Proterosuchus. DESCRIPTION OF THE LATEROSPHE- NOID IN PROGANOCHELYS Following is a more complete description of the left laterosphenoid in Proganochelys 2009 ARCHOSAUR-LIKE LATEROSPHENOID IN EARLY TURTLES fenestra epiotica pila antotica fenestra prootica (V) pila metoptica fenestra metoptica (III) ' fenestra prootica (V) fenestra epiotica ;'!."H»MiM<pr.r,<^3k' / ^taenia marginalis .taenia medialis pila metoptica ^^^ K^\>C / lv -pila antotica fenestra prootica (V) Figure 1 . (A) Left laterosphenoid of Proganochelys quenstedti SMNS 1 5759 in lateral view, after Gaffney (1990). (B) Right laterosphenoid of Proterosuchus fergusi NMQR 1484 in lateral view, reflected, after Clark et al. (1993). (C) Chondrocranium of Crocodylus porosus after ref 1 with region ossified as laterosphenoid filled in. BS, basisphenoid; FR, frontal; EP, epipterygoid; LS, laterosphenoid; OP, opisthotic; PA, parietal; PF, postfrontal; PO, postorbital; PR, prootic; Q, quadrate. BREVIORA No. 518 than was offered in the original monograph. Our goal is to elucidate the developmental origins of the turtle laterosphenoid and thus demonstrate its exact correspondence to the archosauriform morphology. Mediolaterally, the laterosphenoid is thin, especially near its periphery, and it is inclined ventromedially, reflecting the angulation of the wall of the membranous braincase within which it ossified (Fig. 1A). It has three major components. The first is a strut that extends anterodorsally from the clinoid process of the basisphenoid, but whose posterodorsal portion forms a small contact with the anterodorsal portion of the prootic. The posterior margin of the strut forms the anterior half of the border of the trigeminal (prootic) foramen transmitting cranial nerve V, which is fully encircled by virtue of its dual contacts — the ventral, broad contact with the clinoid process of the basisphenoid and the dorsal, attenuate contact with the prootic. The anterior margin of the strut forms the posterior border of a ventrally incomplete aperture that in life would have been formed around cranial nerves III and IV. Topologically and morphologically, this strut corresponds exactly to the pila antotica of the embryonic amniote chondrocranium (Fig. 1C; Bellairs and Kamal, 1981), as suggested but not fully explicated in the description by Gaffney (1990). The meeting with the prootic and thus closure of the trigeminal foramen, however, is a unique feature of laterosphenoids. The second major component of the laterosphenoid is a broad, dorsoventrally oriented strut whose posterior margin arches over to form the anterior half of the aperture for cranial nerves III and IV (Fig. 1A). The strut becomes anteroposteriorly wider at its base and then ends, presumably where it would have sat upon the unossified fused trabeculae cranii. Its anterior margin forms the lower portion of an emargination that would have formed around cranial nerve II (optic nerve) and its associated neurovascu- lar structures. Thus, topologically and mor- phologically, this strut corresponds to the pila metoptica of the embryonic amniote chondrocranium (Fig. 1C), an observation not made in the description by Gaffney (1990). The third major component of the latero- sphenoid is an anterodorsally directed, ter- minally expanded lobe (Fig. 1A) connected basally to both of the other two components whose broadly curved anteroventral margin forms the majority of the emargination for cranial nerve II and whose posterodorsal margin borders an aperture that might represent the fenestra epioptica of the diapsid embryo (Bellairs and Kamal, 1981). As noted in the description by Gaffney (1990), the dorsal and anterior margins of the lobe appear unfinished. This morphology could represent breakage, but considering the general completeness of the surrounding elements, we think it more plausible that it is instead the border between the ossified and cartilaginous portions of the structure. The rough but not jagged texture of the surfaces supports this interpretation. The original description emphasizes that there are no signs on the parietal of a bony suture with the laterosphenoid. Topologically and mor- phologically, the lobe corresponds to the taenia medialis and perhaps a portion of the planum supraseptale of the chondrocranium (Fig. 1C) — not the planum supraseptale ex- clusively as suggested by Gaffney (1990). COMPARATIVE NOTES The laterosphenoid of Proganochelys is identical to the laterosphenoid present in the clade Archosauriformes (Clark et al., 1993) with the sole exception that it retains an open suture with the skull roof. The stem turtle Kayentachelys aprix, closer to the crown than 2009 ARCHOSAUR-LIKE LATEROSPHENOID IN EARLY TURTLES Proganochelys, also possesses laterosphe- noids (again described as "pleurosphe- noids"), though in existing specimens their detailed morphology is not discernable (Sterli and Joyce, 2007). This distribution suggests their ancestral presence in the turtle lineage. Unfortunately, specimens of the oldest known stem turtle, Odontochelys semitestacea, are dorsoventrally crushed, obscuring the relevant region (Li et al., 2008). In modern turtles, ventral down- growths of the parietal articulate directly with the prootic and have thus obliterated any remnant of the laterosphenoids. The archosauriform laterosphenoid shows the three components listed above in every case where it is known, though there is some variation in their relative prominence (Fig. 1C). In crocodylians, for instance, and particularly Alligator, the pila metoptica component is reduced but present. The morphology and topology of the latero- sphenoid in turtles and Archosauriformes are unique among all vertebrates. It appears that the identity of the turtle laterosphenoid has simply been overlooked. The element is not mentioned in comprehensive reviews of the archosaur condition (Clark et al., 1993). As noted in the original description of Proganochelys and in subsequent works, stem reptiles had a more anterior spheneth- moid ossification, Y-shaped or V-shaped in transverse section. This ossification is lost in diapsids (de Braga and Rieppel, 1997). There is limited overlap between the region of ossification of the stem reptile sphenethmoid and the turtle/archosauriform laterosphe- noid. The posteriormost interorbital region, notably the base of the pila antotica, is only ossified in Archosauriformes and turtles (de Braga and Rieppel, 1997; Gaffney, 1990). The stem reptile sphenethmoid and turtle laterosphenoid were confounded and claimed to be homologous in some recent work, suggesting pareiasaur affinities for turtles (Lee, 1993, 1995, 1997). Interestingly, a rebuttal of many of the conclusions of that work (de Braga and Rieppel, 1997) asserted that both stem turtles and pareiasaurs have sphenethmoids. However, the anatomical criteria they set out for a sphenethmoid (e.g., complete enclosure of the optic nerve foramen) do not describe the structure in Proganochelys, although it does fit the structure in pareiasaurs. Simultaneously, their criteria for a true laterosphenoid ("pleurosphenoid ,, ) precisely describe the structure in Proganochelys. The only plausi- ble explanation for this oversight is that the authors of that paper accepted the homology assessments of the study they were rebutting (Lee, 1995) without referring to the descrip- tion of Proganochelys by Gaffney (1990). CHARACTER DISTRIBUTION ON THE ARCHOSAUR STEM WITH TURTLES INCLUDED Not all stem archosaurs have a latero- sphenoid — as described earlier, the bone is a synapomorphy of Archosauriformes (Clarke et al., 1993), which excludes protorosaurs, rhynchosaurs, and Trilophosaurus (Dilkes, 1998; Modesto and Sues, 2004; Sues, 2003). The presence of a laterosphenoid in turtles suggests a close relationship to Archosaur- iformes to the exclusion of non-archosauri- form archosauromorphs. Additionally, the presence of a tight suture of the latero- sphenoid to the parietal might unite Arch- osauriformes to the exclusion of turtles, suggesting, on the basis of this character, a sister-group relationship between the two. Unfortunately, the highly derived nature of the remainder of the turtle skull and post- cranium results in widely inconsistent results when turtles are included in morphological character matrices taken from other studies of reptilian relationships that did not initially include turtles. Typically, these analyses have B REV 10 R A No. 518 not included a large number of characters within Archosauromorpha that would allow the precise placement of turtles within that clade (e.g., Miiller and Reisz, 2006, and references therein). A full analysis of rela- tionships will require considerable additional work. As a preliminary exercise, we scored P. quenstedti using the 144-character matrix by Dilkes (1998), the most comprehensive arch- osauromorph matrix in the literature. The characters listed by Dilkes (1998) as candi- dates for ordering were ordered. To his matrix, we added three characters: 145. Laterosphenoid (0) not sutured to parietal or (1) sutured to parietal. 146. Skull (0) broadly wedge-shaped or (1) tall and mediolaterally narrow. 147. Mid-dorsal region dermal ossifications (0) absent or (1) present. We briefly discuss each of these in turn. See the Appendix for individual character scores. The skull of Euparkeria and archosaurs is tall and mediolaterally compressed com- pared with that of non-Archosauriformes and to an extent Proterosuchus. Progano- chelys shows what appears to be the primi- tive condition. Scoring of this character does not affect the current analysis, but it is a codification of this basic observation on skull proportions and will be useful as more taxa within Archosauria are added to the analyses. A row of ossifications close to the midline of the back is another overlooked potential synapomorphy of turtles and Arch- osauriformes. It is especially interesting because Odontochelys has only the mid- dorsal ossifications, the rest of the carapace remaining unossified (Li et al, 2008). If this is the primitive condition in the turtle lineage, it would be even more similar to the state in Archosauriformes, which have a pair of rows of osteoderms running down the center of the back (Gauthier et al., 1988). It is true that turtles appear to have a single row of discrete ossifications, whereas Archosauriformes have two, but despite this difference, they share the presence of a longitudinal series of dermal bone elements in the mid-dorsal region. In addition to the synapomorphies includ- ed in the matrix, P. quenstedti has what appears to be a typical diapsid infraorbital foramen, despite the lack of a separate ectopterygoid. This infraorbital foramen becomes progressively smaller along the lineage to extant turtles and is given the name "foramen palatinum posterius" (Joyce, 2007). This terminology implies homology to a very small vascular foramen present in stem reptiles (Gaffney, 1990), despite the greater resemblance of the large foramen of plesiomorphic stem turtles to the diapsid infraorbital foramen. Only more crown-ward turtles have a very small foramen. The first, unconstrained parsimony analy- sis yielded a single most parsimonious tree of 397 steps and recovered P. quenstedti as sister to Archosauromorpha (Fig. 2), sug- gesting archosaurian affinities for turtles, but a dual origin of the laterosphenoid. Synapo- morphies supporting this placement are: 36(1), quadrate exposed laterally; 47(1), crista prootica present; 107(1), entepicondy- lar foramen absent; 122(1), fifth metatarsal hooked without deflection. Unambiguous synapomorphies along the lineage leading to Archosauriformes, but lacking in Proga- nochelys (requiring reversal if Proganochelys is allied to Archosauriformes), are: 2(1), snout greater than or equal to 50% of skull length; 5(1), antorbital fenestra present; 8(1), maxillary ramus of premaxilla extends as posterodorsal process to form caudal border of naris; 18(1), ratio of lengths of nasal and frontal greater than 1.0; 29(0), postparietal present; 37(1), quadrate emargination pre- sent with conch; 43(1), orientation of basip- terygoid processes lateral; 45(1), internal 2009 ARCHOSAUR-LIKE LATEROSPHENOID IN EARLY TURTLES Archosauromorpha Prolacerta Euparkeria Proterosuchus Mesosuchus Howesia Rhynchosaurus Scaphonyx Hyperodapedon Stenaulorhynchus Trilophosaurus Macrocnemus Langobardisaurus Tanystropheus Megalancosaurus Drepanosaurus Protorosaurus Proganochelys Squamata Gephyrosaurus Cteniogenys Champsosaurus Lazurussuchus Youngina Petrolacosaurus Figure 2. Single most parsimonious tree resulting from unconstrained phylogenetic analysis with the use of modified matrix from Dilkes (1998). carotid foramina on ventral surface of parasphenoid; 53(1), post-temporal fenestra small; 75(1), upturned retroarticular process; 76(1), lateral mandibular fenestra; 79(0), postaxial cervical intercentra present; 87(2), second sacral rib bifurcate with caudal process truncated sharply; 88(2), proximal caudal neural spies very tall; 96(0), inter- clavicle broad diamond; 97(1), notch in interclavicle between clavicles; 104(1), ante- rior apron of pubis present; 109(1), medial centrale of carpus absent; 116(1), lateral tuber of calcaneum; 126(1), pterygoids re- main separate cranially. For the second analysis, we constrained P. quenstedti to be sister to Archosauriformes to determine potential synapomorphies in the case of a single origin of the laterosphenoid. A single most parsimonious tree of 413 steps was recovered (Fig. 3). In this tree, the ProganochelyslKrchos&unfovmQS clade was sister to the remaining archosauromorphs. Synapomorphies supporting a sister-group relationship between P. quenstedti and Arch- osauriformes are: 14(1), septomaxilla absent; 50(1), laterosphenoid present; 74(2), retro- articular process present, large, and formed by articular; 77(1), slender and tapering cervical ribs at low angle to vetebrae present; 83(1), notochordal canal absent in adult; 89(1), ratio of lengths of caudal transverse processes and centra greater than 1.0; 102(1), BREVIORA No. 518 Archosauromorpha Proganochelys Euparkeria Proterosuchus Macrocnemus Langobardisaurus Tanystropheus Megalancosaurus Drepanosaurus Protorosaurus Prolacerta Trilophosaurus Mesosuchus Howesia Rhynchosaurus Scaphonyx Hyperodapedon Stenaulorhynchus Cteniogenys Champsosaurus Lazurussuchus Squamata Gephyrosaurus Youngina Petrolacosaurus Figure 3. Single most parsimonious tree resulting from phylogenetic analysis with Proganochelys quenstedti constrained as sister to Archosauriformes with the use of modified matrix from Dilkes (1998). dorsal margin of ilium with large posterior process and smaller anterior process; 143(1), distal ends of cervical neural spines expanded in form of flat table; 147(1), mid-dorsal region dermal ossifications present. DISCUSSION The tree recovered by our first (uncon- strained) analysis agrees in its general topol- ogy with the preferred tree discussed by Dilkes (1998). This topology suggests a dual origin of the laterosphenoid; note, however, the caveats below about the overall topology of the tree. Nevertheless, Proganochelys does emerge on the basis of this dataset both as a diapsid and as part of the archosaur stem lineage. Constraining Proganochelys as sister to Archosauriformes (and therefore forcing a single origin of the laterosphenoid) pulls that clade into a sister-taxon relationship with the remaining archosauromorphs. That Proga- nochelys would exert a pull toward the archosauromorph base is unsurprising given that the apparently primitive reptilian char- acters of turtles generally place them as the sister taxon to the remaining reptiles in morphological phylogenetic analyses (Gau- thier et al., 1988). Additionally, the positions of Trilophosaurus and Prolacerta are labile, 2009 ARCHOSAUR-LIKE LATEROSPHENOID IN EARLY TURTLES with Prolacerta jumping from an affinity with Archosauriformes in the unconstrained tree to a more traditional position allied with other "primitive" archosauromorphs in the constrained tree. Trilophosaurus is highly autapomorphic and jumps from a sister- taxon relationship to a "higher" archosaur- omorph clade, including Archosauriformes, in the unconstrained analysis to a position sister to the "primitive" archosauromorph clade in the constrained analysis. Note that the new characters we added did not affect the broad-scale topology of the tree exclusive of Proganochelys. Because of the lability of the trees recov- ered using the matrix from Dilkes (1998) and the incongruence among various hypotheses of diapsid relationships, we consider that a good deal of additional work is required to create a truly comprehensive character list allowing a robust placement of turtles among fossil and extant taxa. The exercise described above is directed only at examining, in a preliminary way, the distribution of poten- tially interesting characters within Archo- sauromorpha if turtles have archosaur affin- ities. The continued lack of consensus about relationships within archosauromorphs is why we are careful to distinguish between physical identity between the laterosphenoids of turtles and archosauriforms, which we have shown, and homology between the structures. We subscribe to the "historical" homology concept, elegantly stated by Van Valen (1982) as "continuity of information" from ancestor to descendant. Thus, a con- clusive homology statement depends on a robust phylogenetic tree. The laterosphenoids in turtles and arch- osauriforms fulfill the requirements for a hypothesis of homology as set forth by Patterson (1982), including topology and ontogeny. Ontogeny, however, has since been discredited as a separate, special crite- rion for homology or character polarity determination (de Queiroz, 1985). Rather, characters from different times in an organ- ism's existence simply represent additional points of identity between putatively homol- ogous structures. The total existence of every organism in time consists of a series of "frames" or semaphoronts (sensu Hennig, 1966), and points of identity that might be homology relations can be sought between any semaphoronts, no matter their relative sequence. Interestingly, Owen (1848) already understood, as stated explicitly in the intro- duction to the cited work, that different modes of development (early semaphoronts) do not preclude homology of later struc- tures. Although the debate on turtle origins and the evolution of their unique anatomy remains unresolved, molecular studies over- whelmingly indicate archosaurian affinities for turtles. The preliminary analyses we ran identified a number of interesting characters that might represent synapomorphies of turtles and various archosauromorph clades. Yet, the laterosphenoid alone is a character shared between turtles and a monophyletic group within archosauromorphs that does not appear elsewhere among vertebrates. It represents potential morphological support for the hypothesis that turtles are part of a major stem archosaur radiation and another example of the immense variety of the archosaur lineage. ACKNOWLEDGMENTS We thank Jacques Gauthier, Tyler Lyson, and Farish Jenkins for discussion of the structure in question. Jonathan Losos, Jose Rosado, Farish Jenkins, and Bill Amaral permitted access to comparative specimens. Reviews by Chris Brochu, Michael Lee, and Randy Irmis were uniformly insightful and constructive. 10 BREVIORA No. 518 APPENDIX 1: ADDITIONS TO DILKES (1998) CHAR- ACTER MATRIX For new characters, order is: Pe, Y, G, Sq, Pr, Ma, Ta, Tr, Ho, Me, R, Sc, St, Hy, Ph, E, Ch, Ct, L, Po, Mg, Ln, D. See Dilkes (1998) for key to abbreviations. Character 145: ?????????????? 1 1 ??????? Character 146: 0000000000000001000000? Character 147: 00000000000000710000000 Proganochelys quenstedti: 1077000010 1111 200000 00070? 1011 0000? 121?? 7701011101 2100?????? ????0002?0 0012001011 0117770110 1???010021 0100111101 007010???? 1112000000 000007000? 0710001 LITERATURE CITED Bellairs, A. D'A., and A. M. Kamal. 1981. The chondrocranium and the development of the skull in Recent reptiles, pp. 1-262. In C. Gans and T. S. Parsons (eds.), Biology of the Reptilia, Volume 11: Morphology F. London, Academic Press. Brochu, C. A. 2001. Progress and future directions in archosaur phylogenetics. Journal of Paleontology, 75: 1185-1201. Clark, J. M., J. A. Welman, J. Gauthier, and M. Parrish. 1993. The laterosphenoid bone of early archosauriforms. Journal of Vertebrate Paleontol- ogy, 13: 48-57. Cao, Y., M. D. Sorensen, Y. Kumazawa, D. P. Mindell, and M. Hasegawa. 2000. Phylogenetic position of turtles among amniotes: evidence from mitochondrial and nuclear genes. Gene, 259: 139-148. De Queiroz, K. 1985. The ontogenetic method for determining character polarity and its relevance to phylogenetic systematics. Systematic Zoology, 34: 280-299. Debraga, M., and O. Rieppel. 1997. Reptile phylogeny and the interrelationships of turtles. Zoological Journal of the Linnean Society, 120: 281— 354. Dilk.es, D. W. 1998. The Early Triassic rhynchosaur Mesosuchus browni and the interrelationships of basal archosauromorph reptiles. Philosophical Transactions of the Royal Society of London B, 353: 501-541. Gaffney, E. S. 1990. The comparative osteology of the Triassic turtle Proganochelys. Bulletin of the Amer- ican Museum of Natural History, 194: 1-263. Gauthier, J., A. G. Kluge, and T. Rowe. 1988. Amniote phylogeny and the importance of fossils. Cladistics, 4: 105-209. Hennig, W. 1966. Phylogenetic Systematics. Urbana, University of Illinois Press. Iwabe, N., Y. Hara, Y. Kumazawa, K. Shibamoto, Y. Saito, T. Miyata, and K. Katoh. 2005. Sister group relationship of turtles to the bird-crocodilian clade revealed by nuclear DNA-coded proteins. Molecular Biology and Evolution, 22: 810-813. Joyce, W. G. 2007. Phylogenetic relationships of Mesozoic turtles. Bulletin of the Peabody Museum of Natural History, 48: 3-102. , J. F. Parham, and J. A. Gauthier. 2004. Developing a protocol for the conversion of rank- based taxon names to phylogenetically defined clade names, as exemplified by turtles. Journal of Paleontology, 78: 989-1013. Kumazawa, Y., and M. Nishida. 1999. Complete mitochondrial DNA sequences of the green turtle and blue-tailed mole skink: statistical evidence for archosaurian affinities of turtles. Molecular Biology and Evolution, 16: 784-792. Laurin, M., and R. R. Reisz. 1995. A reevaluation of early amniote phylogeny. Zoological Journal of the Linnean Society, 113: 165-223. Lee, M. S. Y. 1993. The origin of the turtle body plan: bridging a famous morphological gap. Science, 261: 1716-1720. . 1995. Historical burden in systematics and the interrelationships of "parareptiles." Proceedings of the Royal Society B, 263: 111-117. . 1997. Pareiasaur phylogeny and the origin of turtles. Zoological Journal of the Linnean Society, 120: 197-280. Li, C, X. Wu, O. Rieppel, L. Wang, and L. Zhao. 2008. An ancestral turtle from the Late Triassic of southwestern China. Nature, 456: 497-501. Merck, J. W. 1997. A phylogenetic analysis of the Euryapsid reptiles. Ph.D. Dissertation. The Univer- sity of Texas at Austin. 785 pp. Modesto, S. P., and H.-D. Sues. 2004. The skull of the Early Triassic archosauromorph reptile Prolacerta broomi and its phylogenetic significance. Zoolo- gical Journal of the Linnean Society, 140: 335- 351. MUller, J., and R. R. Reisz. 2006. The phylogeny of early eureptiles: comparing parsimony and Bayes- ian approaches in the investigation of a basal fossil clade. Systematic Biology, 55: 503-511. 2009 ARCHOSAUR-LIKE LATEROSPHENOID IN EARLY TURTLES II Organ, C. L., R. G. Moreno, and S. V. Edwards. 2008. Three tiers of genome evolution in reptiles. Inte- grative and Comparative Biology, 48: 494-504. Owen, R. 1848. On the Archetype and Homologies of the Vertebrate Skeleton. London, John van Voorst. Patterson, C. 1982. Morphological characters and homology, pp. 21-74. In K. A. Joysey and A. E. Friday (eds.), Problems of Phylogenetic Recon- struction. London and New York, Academic Press. Rieppel, O. 2000. Turtles as diapsid reptiles. Zoologica Scripta, 29: 199-212. — , and R. R. Reisz. 1999. The origin and early evolution of turtles. Annual Review of Ecology and Systematics, 30: 1-22. Sterli, J., and W. G. Joyce. 2007. The cranial ana- tomy of the Early Jurassic turtle Kayentachelys aprix. Acta Palaeontologca Polonica, 52: 675- 694. Sues, H.-D. 2003. An unusual new archosauromorph reptile from the Upper Triassic Wolfville Forma- tion of Nova Scotia. Canadian Journal of Earth Sciences, 40: 635-649. Swofford, D. L. 2001. PAUP*: Phylogenetic Analysis Using Parsimony (*and Other Methods), Version 4.0b 10. Sunderland, Massachusetts, Sinauer Associ- ates. Van Valen, L. 1982. Homology and causes. Journal of Morphology, 173: 305-312.