462 



ONTOGENY AND SYSTEMATICS OF FISHES-AHLSTROM SYMPOSIUM 



13.21c 



2.IOb,llb.l5.2lb 



TETRAODONTOIDEA 

 6. 8.10a. I la. 12. 17c, 20. 21a. 22 



TRIACANTHOIDEI 



TETRAODONTOIDEI 



TETRAODONTIFORMES 



Fig. 252. Phylogeny of tetraodontiform fishes based on early life history characters (excluding Aracanidae and Triodontidae for which no 

 information is available). Numbers refer to characters (see text) and are located on the branch which possesses the derived state. Characters 2, 6 

 and 21a would be deleted if the Zeiformes is accepted as the sister group of the Tetraodontiformes. Note that character 1 occurs m two places 

 indicating conflict with the accepted classification. 



forming dorsal spines (character 16) is considered primitive 

 because it is closer to the presumed ancestral (i.e., acanthuroid 

 and zeiform) condition of early-forming dorsal and pelvic spines 

 than is late formation oftheanlage. (19) Pelvic fin — Acanthuroid 

 (and zeiform) fishes have a pelvic fin formula of at least I, 3. 

 Presence of a pelvic fin is primitive, and its absence is derived. 

 (20) Pelvis— The pelvis is present in acanthuroid (and zeiform) 

 fishes, and its absence is considered derived. (21) Caudal fin 

 rays- Acanthuroids have 16-17 principal caudal fin rays. The 

 maximum number (i.e., 1 2) in tetraodontiform fishes is consid- 

 ered primitive. The intermediate derived condition is < 1 1 rays 

 (2 1 a). The next most advanced condition is < 1 rays (21b), and 

 the most advanced condition is the complete absence of the 

 caudal fin (21c). Zeiform fishes have 1 1-15 principal caudal fin 

 rays, so if zeiforms are accepted as the outgroup, polarity of 2 la 

 cannot be determined, while 21b and 21c would not change. 

 (22) Pectoral fin development— The pectoral fin in acanthuroid 

 (and probably zeiform) fishes develops after or simultaneously 

 with the dorsal and anal fin soft rays, and this is considered 

 primitive. (23) Body width — The condition of body width > 

 body depth found in the Diodontidae is unique in the Tetra- 

 odontiformes and is considered derived. 



Phylogenetic Analysis 



Relationships within the Tetraodontiformes based on ELH 

 characters are presented in Fig. 252. In the following section, I 



will contrast the present phylogeny with three phylogenies based 

 on adult characters (Fig. 253): myology (Winterbottom, 1 974a), 

 external and internal characters (Tyler, 1980) and osteology 

 (Rosen, pers. comm.). Lauder and Liem's (1983) review of in- 

 terrelationships of tetraodontiform fishes depends heavily on 

 Winterbottom's (1974a) work and, for my purposes here, is 

 identical to his phylogeny. Therefore, Lauder and Liem's (1983) 

 phylogeny will not be considered separately. The ELH-based 

 phylogeny exactly matches none of the three adult-based schemes, 

 but is closest to Rosen's (pers. comm.), differing only in place- 

 ment of the Tetraodontidae. Two cautions should be kept in 

 mind: 1) Rosen's (pers. comm.) study is primarily concerned 

 with inter-ordinal relationships, and the portion dealing with 

 intra-ordinal relationships of the Tetraodontiformes is based on 

 relatively few characters; and 2) the present phylogeny has lim- 

 itations flowing from exclusion of two families and many 

 subfamilial taxa due to lack of information. 



There is most agreement between the four phylogenies in the 

 question of the relationship of the triacanthodids (Figs. 252, 

 253). The present phylogeny and those of Winterbottom (1974a) 

 and Rosen (pers. comm.) agree in the erection of the suborder 

 Triacanthoidei as the sister group to all other tetraodontiform 

 fishes. Tyler (1980) includes the triacanthodids in the balistoid 

 line, but this is a result of philosophy of classification more than 

 anything else (Tyler, pers. comm.). 



The four phylogenies are evenly divided on the question of 



