A cladistic analysis of chordates is presented, based on some 320 nested characters. All the principal higher taxa are defined by synapomorphies, including extinct acanthodians and placoderms. The data base draws broadly from adult anatomy (including osteological data for Recent and fossil taxa), embryology, physiology, and biochemistry. A conventional sequence of chordate higher taxa is generated (hemichordates, urochordates, cephalochordates, craniates). Among the craniates, cyclostomes are considered paraphyletic. Gnathostomes are monophyletic, but two fossil "agnathan" groups (galeaspids, osteostracans) are regarded as stem gnathostomes. Chondrichthyans and osteichthyans are monophyletic. New arguments for osteichthyan affinity of acanthodians are presented. The phylogenetic position of placoderms is still problematic, but they can no longer be perceived as stem chondrichthyans or even as "elasmobranchiomorphs." Recent dipnoans and tetrapods are sister groups, but new paleontological discoveries refute many of their supposed osteological synapomorphies, thereby reopening the possibility of a closer relationship between tetrapods and osteolepiform rhipidistians.
CLADISTICS[VOL. 2
204
CLADISTICS[VOL. 2 of motor nerves in the ventral "roots" of cephalochordates (Flood, 1966) are probably primitive conditions. Contrary to Jefferies (1979), I do not think muscle innervation patterns in tunicates, echinoderms, and cephalochordates offer any clear indication of their interrelationships, and it may be that motor nerve end plates arose independently in adult thaliacean tunicates and in craniates, whereas the remaining similarities between tunicates and craniates are most parsimoniously regarded as primitive.The supposed presence of chitin in the cephalochordate branchial endoskeleton suggested to Ldvtrup (1977:94) that chordates are not monophyletic. R&r (1982) presented convincing evidence that although the branchial skeleton of Amphioxus is not collagenous, neither is it chitinous; instead the branchial rods probably consist of acid mucopolysaccharides and proteins. Ldvtrup's (1977) hypothesis of chordate non-monophyly is not substantially helped whether cephalochordate branchial bars consist of chitin or mucopolysaccharides; their structure in the first case could be primitive; in the second case it is most economically regarded as a synapomorphy of cephalochordates. Other possible cephalochordate synapomorphies include the tentacular oral languets, ciliated wheel organ, and Hatschek's nephridium. Hatscheck's pit may be the homolog of the craniate pituitary (Goodrich, 1917;De Beer, 1925, 1926, but nothing is known of cephalochordate endocrine function (Barrington, 1968).Although accepting cephalochordates as the sister group of craniates, Janvier (1981: 122) stated that they "show hardly any precise and surely homologous synapomorphy . . . sufficient to reject comparisons between the Chordata and other groups."However, cephalochordates and craniates are united by the following characters that have not been identi...