REVISED CLASSIFICATION FOR CERTAIN CIRRHITOID GENERA 



sJt? 



.v6' 



• 



x6 fl 



>* 



** 



4 ,15 | 



,15*1 



A? 



3*. 15*| 



O* 



7* 1 1*. 1 3* 



1 or 2,6,8,10,12,14 



Fig. 5 Tentative scheme of phyletic relationships within the Cirrhitoidei. Asterisks indicate apomorphic characters (see also Table 1). Since 

 both the cirrhilid (1) and latrid (2) condition of the urohyal are rated as pleisomorphic, that bone in the common ancestor of all lineages is 

 taken to be 1 or 2. 



derived myotomal arrangement of the sternohyoideus 

 muscles (p. 8). (iii) More than 28 vertebrae, (iv) Parapophy- 

 ses developed on the first three abdominal vertebrae, with 

 the first pleural rib associated with the second or third 

 vertebrae, (v) Predorsal bones reduced to 2. The three 

 different types of derived urohyal morphology (pp. 2-5) (ie in 

 chironemids, aplodactylids and cheilodactylids) are each 

 taken to be independently evolved apomorphies. Relation- 

 ships (Fig. ) suggested by these data are: (i) That the 

 cirrhitids are the plesiomorph sister group to the other four 

 families combined, (ii) The chironemids are the sister taxon 

 of the aplodactylids, latrids and cheilodactylids combined, 

 and that for the moment this assemblage should be treated as 

 an unresolved trichotomy since no two lineages uniquely 

 share a recognisable synapomorphic feature. For example, 

 the urohyal in the latrids is of a basal percoid type, and 

 although that bone in the cheilodactylids and aplodactylids is 

 highly derived, each is unique to the families respectively. 

 The cheilodactylids it will be noted, retain the plesiomorphic 

 suborbital shelf, whereas it is lost in the latrids and aplodac- 

 tylids (and in the chironemid lineage as well). The value of 

 this feature as an indicator of relationship, however, is 

 problematical because it involves a loss (and not an acquisi- 

 tion) in the lineages concerned. Clearly, a greater number 

 and variety of characters must be sampled and their polarity 

 determined before this hypothesis of cirrhitoid intrarelation- 

 ships can be improved and the trichotomy resolved. 



The same reservation would apply before any sister-group 

 hypothesis can be erected regarding the relationships of the 

 cirrhitoids within the Percomorpha. However, based on the 



synapomorphic features discussed (above pp. 6-7) it seems 

 reasonable to hypothesize that the five families comprising 

 the lineage, given informal ranking as the 'cirrhitiform per- 

 coids' by Regan (1911), and suprafamilial status by recent 

 authors (Nelson, 1994), should be elevated to subordinal 

 status (Cirrhitoidei) within the Perciformes. 



Acknowledgements. I am particularly indebted to Dr Colin Buxton 

 of Rhodes University's Department of Ichthyology and Fisheries 

 Science who first aroused my interest in the anatomy of cirrhitoid 

 fishes, and who provided many specimens for dissection. Professor 

 Tom Hecht of that department is to be thanked for personally 

 obtaining a specimen of Cheilodactylus fuscus from Salamander Bay, 

 New South Wales, in which operation he was aided by Bill Talbot of 

 the Brackish Water Fish Culture Research Station there. 



To Dr Phil Heemstra of the J.L.B. Smith Institute goes my 

 gratitude for the many discussions we have had about cirrhitoids, and 

 for his advice on the relevant literature, and to Tony Booth of DIFS 

 my appreciation for his painstaking preparation of several skeletons. 

 Once again it is a great pleasure to thank my old colleague Gordon 

 Howes for his elegant and accurate draftmanship, and also to thank 

 another former colleague, Oliver Crimmen of the Fish Section, The 

 Natural History Museum, London for his untiring help with matters 

 radiographic and bibliographical. Yet again, my special thanks go to 

 Huibre Tomlinson, for her patience and skills in producing the 

 typescripts, and to Robin Stobbs for producing certain of the 

 radiographs, and the photograph, used in this paper. 



