JOHNSON: PERCOIDEI 



483 



Table 120. Continued. Extended. 



pharyngobranchials, and an interarcual cartilage between the 

 uncinate process of epibranchial 1 and pharyngobranchial 2. 

 The first pharyngobranchial is rod-like and serves to suspend 

 the dorsal gill arches from the neurocranium. The fourth phar- 

 yngobranchial is reduced and cartilaginous, but consistently bears 

 a well-developed dermal tooth plate, as do the second and third 

 pharyngobranchials and the fifth ceratobranchials. Small tooth- 

 plates on the second and third epibranchials are variously pres- 

 ent or absent. 



Reductive departures from the primitive branchial complex 

 are few and involve only the basihyal, first pharyngobranchial 

 or interarcual cartilage. The basihyal is reduced or absent in 

 ephippidids. Pseudochromids lack a first pharyngobranchial 

 (Springer et al., 1977). Of 88 percoid groups examined for it, 

 only 1 3 lack a well-developed interarcual cartilage and at least 

 three of these (Cirrhitidae, Emmelichthyidaeand Nemipteridae) 

 may have a vestigial element. The remaining eleven groups 

 completely lack the interarcual cartilage, but most have an un- 

 cinate process with the cartilaginous tip separated by a decided 

 gap from the second pharyngobranchial and frequently pointing 

 away from it. This condition differs from the primitive state (as 

 represented in most beryciforms) wherein the uncinate process 

 of the first epibranchial directly contacts that of the second 

 pharyngobranchial, and suggests that these percoids have sec- 

 ondarily lost the interarcual cartilage. A condition resembling 

 that of the beryciforms was observed among percoids only in 

 some anthiin serranids, where it must be secondary. In eche- 

 neidids the uncinate process of the first epibranchial also artic- 

 ulates directly with that of the second, but there is a concom- 

 itant extreme reduction of the main arm of the first epibranchial 

 not seen in beryciforms. Again this condition must be derived 

 if the relationships of the echeneidids are as postulated here (see 

 discussion on utility of larval morphology). 



3/3/0-2/1 

 l-2-(pH); 3-4-UR 



5/3/2/2 



5/3/2/2 



5/3/2/2 



5/3/2/2 



3/2-3/0/0-1/ 

 (pH)- 1-2; 3-4-UR 



+ (-) 



Scales.— The unpublished work of McCully (1961) on compar- 

 ative anatomy of serranid scales provides an excellent illustra- 

 tion of the wealth of information available in the scales of per- 

 coid fishes that has largely been ignored in systematic studies. 

 More recent work on ctenoid scales of other groups (DeLamater 

 and Courtenay, 1973a, b, 1974; Hughes, 1981) using scanning 

 electron microscopy also demonstrates the systematic value of 

 ctenoid scales. Details of the scale morphology of most percoids 

 are unknown. On a gross level, three basic scale types (Ct, Ct' 

 and Cy in Table 120) are found among percoids. Although be- 

 ryciforms and some myctophids are said to have ctenoid scales, 

 these scales (Ct') differ from the type possessed by most percoids 

 and other perciforms (Ct). In beryciforms and myctophids the 

 "ctenii" are continuous spinous projections from the lateral sur- 

 face and posterior margin of the scale. A few percoids (Bramidae, 

 Epigonidae, Howella. Pomacanthidae, Priacanthidae, Ostraco- 

 berycidae and Scatophagidae) possess similar scales that may 

 represent retention of the plesiomorphic beryciform condition, 

 or may have been secondarily derived. In the "true" ctenoid 

 scale that characterizes most percoids (59 groups), the ctenii are 

 separate bony plates, or scalelets (McCully, 1961, 1970), that 

 are continually added in the posterior field as the scale grows. 

 In most groups the posterior field becomes filled with remnants 

 of old ctenii, the tips of which are amputated (or, more likely, 

 resorbed), as each new row of ctenii is added. In a few groups, 

 however (e.g., anthiine serranids and callanthiids), only a pri- 

 mary and secondary row of marginal ctenii are evident. This 

 second variation of "true" ctenoid scale also characterizes 



