Rìvista Italiana di Paleontologia e Stratigrafia Dicembre 2000 I--r4-""l*-I-."-''* 1 A REEXAMINATION OF FOUR PROLACERTIFORMS \TITH IMPLICATIONS FOR PTEROSAUR PHYLOGENESIS DAVID PETERS ReceìterJ October 23, 1999; accepted October 20, 200A Kqt uorcls: Pterosauria, Prolacertiiormes (Reprilia, Diapsida), Traditionally the answer has been rhat prerosaurs Phyìogeny, Cladisric an:ìy.is. are archosaurs (Romer 1956); the sister group of the Riassunto . Tradizionalmente gli prerosauri venir.ano considerati Dinosauria, ScleromochÌus a.nd Lagosuclcws/Maraswchus come appartenenti agli Archosaurifomes e molti specìalistì contempo_ (Benton 1985, 1990, 1999; Padian 1984; Gauthter 1984, ranei considerano gli pterosauri quali sisrer groups di Lagosuchus, 1986; Sereno 1991, 1994; Kellner 1996); or perhaps Schleromochlus e dei Dinosauria. La nuova analisi filogenerica qui pro- archosauriformes close posta merte in discussione queste affinirà jn quanto tutte le presunte to prorerosuchids and eryrhro- sinapomorfie che collegherebbero gli Pterosauria con gli Archosauri_ suchids (Bennett 1996a), chiefly because prerosaurs formes o con gli pterosaurìa, Ornìthodira mancano in realtà negli have a prominent anrorbiral fenestra and a suite of other oppure sono condivise anche da alcuni taxa di prolacertiformi. ll archosaur-like characrers almosr entirely recente riesame degli olotipi dt confined to the Cosesaurus a,Liceps, Longisquama ìnsig_ hind nis e di Sharovipteryx mìrabi/ìs suggeriscono che molti caratteri potreb- limb (Bennert 1996a). Although Benton (1982, bero venire interpretati in maniera diversa rispetto alle precedenti L984) initially indicated that the prerosauria are descrìzioni. I risultati di molteplici analisì cladistjche suggeriscono che archosauromorphs and the sister-group ro all other questi tre prolacertìformi enigmatici, uniramente a Langobardìsawrws, archosauromorphs, later work (Benton 1985, 1.990, recentemente descritto, costituirebbero i sister taxa degli prerosauri, in base ad un insieme di sinapomorfie di nuova identificazione. 1999) supported the traditional view. Previous ro the present work, the hypothesis that pterosaurs are Abstract. Traditionalll', pterosaurs have been included within archosauriformes has been challenged only by \íild the Archosauriformes and many contemporarv workers consider the (1978,1984a) and Peters (1997). Pterosauria the sìster group ro Lagosuchus, Scleromocblus and the Dìnosauria. New anall'5ss cast doubts on those relationships because FIere a phylogenetic framework for determining nearly all presumed ;rrchosauriform or ornithodire "synapomorphies" the position of the Pterosauria within the Archosauro- are either not present within the Pterosauria or are also present with- morpha is provided l:y analyzing a broad range of char- in certain prolacertrform taxa. Recent examinations of the holotypes acter data across an exrended range of taxa including the of Cosesaurus ariceps, Longisquama insignis and .gharooiptery,, mirabí/is following suggest that manv characters may be interpreted differently than pre_ key proiacertiforms: Langobarclìsaurus pan- viousì1' reported. Results of ser.eral subsequent cladistic analyses sug- dolfii (Renesro 1994), L. tonelloi (Muscio 1996, Frg. gest that these three "enigmatìc" prolacertiforms, along with the newly 1C), Cosesaurus aeiceps (Ellenberger and de Villalta described Langobard,isaurzs, are sister taxa to the pterosauria based on 1974, Frg. 3-6), Longisquama a suite of newly identified svnapomorphies. insignis (Sharov 1,970, Ftg. 7) and Sharooipteryx mirabilis (Sharov 1971, Cowen 1981, Fig. 8-10). As a group these four genera have not Introduction. been compared to each orher or ro pterosaurs in cladis- tic analyses, although lahl (1997) did include Cosesaurws Pterosaurs have been known since Collini (1784) and Langobardisaurus in his .:ladistic analysis of published a description of a small strange fossil he ten- lesairosaurws. Renesro (1994) regarded Langobardis- tatively regarded as a marine amphibian preserved in aurus panclolfii as a prolacertiform close to Cosesaurus Solnhofen limestone. Cuvier (1S01, 1809) classified the and the Tanysrropheidae but did not provide an analysis. creature among the reptiles and coined the term "Ptero- Ellenberger and de Villalta (1974) and Ellenberger dactyle" from the Greek "wing finger." Vorkers today (1977, 1978, 1993) regarded Cosesaurws as a proto-bird, agree that pterosaurs are archosauromorph reptiles. The but Sanz and López-Marr;nez I 1984) nnd orh.rs lOlsen question posed by this study is: where within the 1,979, Evans 1988) considered it a prolacertiform close Archosauromorpha cladogram do prerosaurs belong? to Macrocnemus (Fig. 1A) or Thnystropheu.s. Sharov Davìd Peters Studio - 1247 Highland Terrace - St. Louis, Mo ^3112, uSA - [email protected] 294 D. Peters 1197 11 lefr tr,. .rffinities of Sbaroaipteryx in question. is hoped that by proposing a phylogenetic framework at Gans, et a1. (1987) tentatively identified it as a "smal1 this time a new forum for discussion will emerge to primitive diapsid, perhaps a lepidosaur or close to the enable more concentrated anatomical and phylogenetic Archosauriformes . or perhaps adjacent to the 'Pro- studies in the future. torosauria'." Tatarinov (1989) and Benton (1999) regarded it as a derived prolacertiform and analysis of the holotype confirms rhis placement. Sharov (1970) Historical background. and Haubold and Buffetaur (1.987) regarded Longisqua- Bassani (1886) and Nopcsa (1922) both consid- ma às a pseudosuchian, Benton (1999) could not assign ered the jumbled remains of a long-necked, juvenile pro- it more precisely than within the Archosauria and Jones, lacertiform, Tanystropheus, to be those of a primitive Tri- et :r1. 12OOO) considered it a distant relative of birds. The assic pterosaur, Tribelesodoz. Although the elongate cer- present anaÌysis of the holot,vpe places it within the Pro- vicals were mistaken for wing phalanges, other aspects lacertiformes close fo Cosesaurus. the anatomy s,'ere pterosaurian to validate Cosesaurus, Longisquama and Sharooipteryx are of sufficiently the distinctive dilficult taxà to rtudv and interpret. ForLunately. all the identification, especially considering elongate pedal digit. The error was not recognized three holotypes are well articulated, but each one is rep- fifth unttl 1.929 when a superior Tanystropheus fossil was dis- resenred b1- a single specimen. Each one has strange corered.rnd Peyer was able to make a positive and/ or unexpected anatomical features. Al1 three are ilgJlaT hoiotype Tribelesoclon. No other crushed or compressed. Often only the impressions of comparison to the of relationship Tanystro- bone are observable. All three must be viewed under published research into the ol pheus, to pterosaurs is known. magnification from multiple angles employing a varietl' or other prolace rtiforms, of illumination to gather sufficient observational data to make positive identifications. A few rdentifications are Traditional ( Pre-Cladìstic) Classification inferred from vague impressions o{ the bone in the Romer (1956) expressed the generally accepted "thecodont" descent matrix and b1. anatomical placement. Inference mav be hypothesis that pterosaurs were of influenced by personal bias: whether "seeing" what one with no known transitional forms. Thus they were has been told to see, "seeing" what one is hoping to see. archosaurs in the traditional sense (Bennett 1996a). \flild or "seeing" something altogether different. To counter (1978,1984a) reviewed the various sorts of animals that the influence of bias, I had the advantage of access to the had been proposed as ancestors and argued that it would observations of previous workers and to recent finds of be "r'ery unlikely that the Thecodontia are the ancestors importance not available to earlier paleontologists. In of pterosaurs" (translated from German). Instead he my opinion all three taxa appear to be incompletely or argued that the Pterosauria are "rooted separately in the improperly described and figured in the literature Eosuchia or went through intermediate forms between (Ellenberger and de Villalta 1974, Ellenberger 1977, Eosuchia and Prolacertilia." He argued that the antor- 1978, 1993, Sharor' 1920, Haubold and Buffetaut 1987, bital fenestra could have arisen separately in pterosaurs Sharov 1971, Gans, et aL 1987, Tatarinov 1989 and Ben- and traditional archosaurs. \fild (I978) illustrated a ton 1999) which is one reason they have remained the hypothetical ancestral form based on Heleosawrus (Car- "enigmrtic" reptiles of the Triassic. Although each is roll 1976) provided with an elongate flight digit. worthy of a separate paper in which its description and classification should be given with some confidence Recent Stwdies before attempting a cladistic anall'sis, the key to under- The following cladistic studies focused on the standing them is to consider them as a group. I do this interrelationships of archosauriforms. Benton (1982, for three reasons: 1) Al1 three display characters also 1984) indicated that the Pterosauria are archosauro- seen in higher archosaurs, which might be confusing if morphs, but the sister-group to all other archosauro- one were not able to readily argue that these characters morphs. Later Benton (1985, Fig. 11A) suggested that are better considered synapomorphies of this small "Pterosaurs are archosaurs and a close sister-group of clade; 2) Many characters are difficult to observe due to the Dinosauria." Benton (1990) presented