128 



R.D. MOOI AND A.C. GILL 



,SC /SN SP1 



EPAX 



^PM + DI 



Fig. 5 Type 1 epaxial morphology with extreme fibre separation from the main epaxial body of the epaxial muscle slip inserting on to 



pterygiophores in Haemulon aurolineatum (MPM 23228, 64.2 mm SL). SP1, first dorsal-fin spine; SRI, first segmented dorsal-fin ray; other 

 abbreviation and methods of presentation as in Figs 1,3. Scale bar = 5 mm. 



group for so long that they have been reified; in practice, 

 most systematists regard the Percoidei as a bona fide taxon. 

 As a consequence ichthyologists have rarely examined taxa 

 from among the Percoidei as potential relatives of non- 

 percoid taxa (exceptions include Johnson, 1984, 1986, 1993; 

 Tyler et al. 1989), and few characters have been identified to 

 suggest a relationship among percoids and scorpaeniforms, at 

 least in part because few researchers have looked. These 

 same problems apply to the more inclusive Perciformes, for 

 which no satisfactory definition exists and membership is 

 often questionable; families considered perciforms are rarely 

 examined as either sister taxa or possible members of other 

 acanthomorph orders (although see Johnson & Patterson, 

 1993) because, in practice, the Perciformes is treated as a 

 monophyletic taxon. 



Several additional characters suggest that a relationship 

 between scorpaeniforms and at least some of the 'percoids' 

 with a Type 1 epaxial morphology is worthy of consideration. 

 For example, some larval serranids (particularly anthiines) 

 bear at least a superficial resemblance to larval scorpaeni- 

 forms, with suspensorial and cranial bones highly orna- 

 mented by spines and ridges (cf. Figs and descriptions in: 

 Baldwin, 1990; Johnson, 1984; Kendall, 1984; Washington et 

 al., 1984). Moreover, the general physiognomies of many 

 adult serranids bear striking resemblances to certain scor- 

 paeniforms. Although general similarities do not provide the 

 necessary evidence for relationship, they hint that there 

 might be more evidence than shared epaxial morphology; we 

 feel it is premature to dismiss these similarities as being due 

 to convergence before relationships are better understood. 



The occurrence of Type 1 epaxial morphology in few 

 non-percoid perciform taxa (blennioids, some cirrhitoids and 

 some trachinoids) suggests that these should also be included 

 in a search for a scorpaeniform sister group, or considered for 

 inclusion among scorpaeniforms (Mooi & Johnson, in prep). 

 For example, blennioids also resemble scorpaeniforms in 

 having the supratemporal sensory canal enclosed by the 

 parietal (except in most tripterygiids where the cephalic 

 sensory canals are incompletely enclosed by bone; Springer, 

 1993:487 and pers. obs.). This condition is found in several 

 other perciform taxa, including at least some zoarcoids (sensu 

 Anderson, 1984; Travers, 1984b; all 'zoarceoids' according to 

 Gosline, 1968:46), some pseudochromids (Gill, in prep.), and 

 mastacembeloid synbranchiforms (Travers, 1984a), but these 

 taxa do not have a Type 1 epaxial morphology. Champsodon- 

 tids more closely resemble scorpaeniforms in having a serrate 

 ridge overlying the canal (Johnson, 1993:14; Mooi & 

 Johnson, in prep.), as well as Type 1 epaxials. Although 

 blennioid parietals lack the serrate ridge or spine over the 

 canal, the possibility of a blennioid/scorpaeniform relation- 



ship deserves further study. Certain cottoids closely resemble 

 blennioids in dorsal gill arch morphology, notably in lacking 

 an interarcual cartilage, and in having only a single 

 infrapharyngobranchial (infrapharyngobranchial 3), which 

 articulates posteriorly with epibranchials 3 and 4 (e.g., com- 

 pare cottoids in Rosen & Patterson, 1990: figs 34A, C and 

 Yabe, 1985: figs 23, 24E with blennioids in Rosen & Patter- 

 son, 1990: figs 33A-B, 37, 38C-D and Springer, 1993: fig. 1). 

 Members of the cottoid family Liparididae further resemble 

 blennioids in lacking an uncinate process on epibranchial 1 

 (Kido, 1988: figs 12A-D). 



Johnson & Patterson (1993: 591) found no evidence to 

 indicate a 'pre-perciform' position for scorpaeniforms, and 

 considered ranking them at the subordinal level within the 

 perciforms, 'if only to stimulate the search for characters 

 justifying their individuality.' We concur with Johnson & 

 Patterson's proposal and award subordinal ranking for the 

 Scorpaeniformes, as the Scorpaenoidei, within the Perci- 

 formes. In addition to the justification provided by Johnson 

 & Patterson (1993), we believe this action will be a major step 

 forward in diagnosing a monophyletic Perciformes. There is 

 no contrary evidence for maintaining the two orders as 

 separate, and the epaxial morphology and other evidence 

 noted above suggests that the Perciformes is non- 

 monophyletic without the inclusion of the Scorpaeniformes. 



The almost universal occurrence of Type 1 epaxial muscles 

 in the Scorpaenoidei has implications for its composition. It 

 casts doubt on the inclusion of the Dactylopteridae and 

 Bathylutichthyidae within the suborder, as neither family has 

 insertions of epaxial muscle to dorsal-fin pterygiophores 

 (Table 1). Johnson (1993: 7) also raised doubts about a 

 relationship between dactylopterids and scorpaenoids based 

 on the absence of a bone-enclosed supratemporal canal and 



Dl /PM 



EPAX 



Fig. 6 Type 1 epaxial musculature in a 'primitive' scorpaeniform 

 Anoplopoma fimbriata (USNM 208296, 122.2 mm SL). Note the 

 separate slip of epaxial muscle to the third dorsal-fin 

 pterygiophore, and other insertions of epaxial to as far posteriorly 

 as the ninth pterygiophore. Abbreviations and other methods of 

 presentation as in Fig. 1. Scale bar = 5 mm. 



