Progress and future directions in archosaur phylogenetics

Journal of Paleontology, Nov 2001 by Brochu, Christopher A

Bennett (1996) argued that most of the characters supporting a close pterosaur-dinosauromorph relationship were hindlimb features correlated with locomotion and thus not independent. Removal of these characters resulted in a tree with pterosaurs lying outside the clade including Archosauria, Erythrosuchidae, and Proterochampsidae. Although his arguments of functional relationship between some characters have merit, there is a difference between functional linkage and phylogenetic dependence. Characters can be phylogenetically independent and not covary on a cladogram, and yet still be functionally correlated. Most of the characters correlated with flight in birds, for example, did not arise at the same phylogenetic level-they can be regarded as functionally related, but phylogenetically independent. There is also the problem of determining functional linkage, especially in groups such as pterosaurs that are extinct (and thus unable to function). Although we must remain vigilant in our efforts to apply independent characters in phylogenetic analysis, nonindependence should be established on the basis of character covariation on a tree.

A more recent analysis by Peters (2000) is interesting because it draws pterosaurs out of Archosauriformes entirely and supports a close relationship between pterosaurs and prolacertiforms, a group of small Triassic reptiles. He argued that many of the characters linking pterosaurs and archosauriforms in other analyses were actually present (convergently) in prolacertiforms or absent in pterosaurs. He also reconsidered the morphology of several Triassic prolacertiforms-most notably Cosesaurus, Longisquama, and Sharovipteryx. Cosesaurus and Longisquama have been implicated in the origin of birds by some authors, especially those who continue to argue against the dinosaurian origin of the group (Feduccia, 1996). Most recently, the unusual long scales preserved on Longisquama have been interpreted as actual feathers (Jones et al., 2000), though most authorities dispute that interpretation (Reisz and Sues, 2000; Prum et al., 2001). Sharovipteryx is a strange animal with what appears to be a membrane of skin stretched between the hindlimbs and tail. Some pterosaurs may have had a similar structure (Unwin and Bakhurina, 1994).

Peters' hypothesis is interesting, but is limited by the quality of preservation of some of the prolacertiforms forming the basis of his argument. Many are flattened or preserved as natural molds, and in many cases they are best studied with low-angle light casting shadows on the slab. Many of the characters discussed by Peters can be given multiple interpretations. Not all are convinced, for example, that Longisquama had an antorbital fenestra (Unwin et al., 2000; Unwin and Benton, 2001). Resolution of this conflict awaits collection of fossils that allow less ambiguous interpretation.

Slender-Snouted Crocodyliforms.-Snout shape varies systematically within Crocodylia, and a handful of basic head shapes appear to have arisen multiple times independently within Crocodylia and throughout Crocodyliformes (Busbey, 1994; Brochu, 2001). Five living species-Gavialis gangeticus, Tomistoma schlegelii, and three species of Crocodylus-have extremely slender snouts. A slender snout is usually viewed as an adaptation for catching fish (Langston, 1973), although ecological information for some of the extant forms with this snout morphology is lacking.