100 



PH. GREENWOOD 



from both taxa, nothing can be said about its possible 

 homology in the two species. 



The concept that a mandibulohyoid connection (usually 

 referred to as a ligament) is essentially a feature of pre- and 

 lower teleost actinopterygians, has influenced theories relat- 

 ing to the evolution of feeding machanisms in teleosts. For 

 example, Lauder (1982: 279, also fig. 1) postulated that 'The 

 first specialization involves a shift of insertion of the man- 

 dibulohyoid ligament to the interoperculum. The interoper- 

 culohyoid ligament characterizes the feeding mechanism of 

 eurypterygian fishes (=Aulopiformes + Myctophiformes + 

 Paracanthopterygii + Acanthopterygii; Rosen, 1973) and 

 effectively shifts the action of the hyoid and opercular cou- 

 pling onto the interoperculum. Only the interoperculoman- 

 dibular ligament transmits posterodorsal hyoid and opercular 

 movements to the mandible in the Eurypterygii, while other 

 teleosts retain the primitive two-coupling system of the 

 halecostomes' (ie both a mandibulohyoid and an 

 interopercular-mandibular linkage). Verraes' (1977) studies 

 on the development of Oncorhynchus mykiss show unequivo- 

 cally that in this teleost there is no ontogenetic shift of the 

 mandibulohyoid ligament's mandibular insertion onto the 

 interoperculum. Indeed, the interopercular-mandibular liga- 

 ment develops independently (and later than the mandibulo- 

 hyoid ligament) with both connections persisting in adults 

 (Verraes, 1977; pers. obs.); neither is there any ontogenetic 

 evidence to show that the epi- (or inter-) hyal to interopercu- 

 lum ligament is the result of a preexisting mandibulohyoid 

 ligament shifting its mandibular insertion onto the interoper- 

 culum. Interestingly in that context, the latrid Acantholatris , 

 which has what appears to be a genuine mandibulohyoid 

 ligament (see p. 97) also has an epihyal-interopercular liga- 

 ment. 



Thus, pace Lauder (1982), it would seem that cirrhitoids 

 (and other teleosts) with both a mandibulohyoid connection 

 and an interopercular-mandibular ligament have either 

 retained the primitive halecostome condition or, as seems 

 more likely, re-evolved it through some other form of con- 

 nective tissue linkage between the hyoid arch and the man- 

 dible. 



Parenthetically, it may be noted that the importance of an 

 interopercular-mandibular linkage in the jaw-opening mecha- 

 nism of teleosts, stressed by Lauder op.cit. and other authors 

 (see for example Liem, 1978 & 1991; Aerts et al., 1987, and 

 references therein) is underlined, albeit indirectly, by the 

 condition in three of the cirrhitoid taxa examined. In Cheilo- 

 dactylus (Cheilodactylidae) and in Mendosoma and Acantho- 

 latris (Latridae) there is, in addition to the interopercular- 

 mandibular ligament a second such linkage effected through 

 an extension of the Aw muscle's aponeurotic system onto the 

 interoperculum (see pp. 94 & 97 and Figs 2-4). 



If, as suggested above, certain teleosts have re-evolved a 

 mandibulohyoid connection, it may have arisen in different 

 ways. This seems probable even within the cirrhitoids {viz. 

 cirrhitid and latrid types; see pp. 93 & 94), and in other 

 groups as well. In the ostariophysan Hydrocynus vittatus 

 (Characidae) for example, the mandibulohyoid connection 

 appears to be an extension of the epihyal-interopercular 

 ligament which, after its insertion on the dorsal margin of the 

 interoperculum, continues forward to bridge the small gap 

 between that bone and the retroarticular (pers. obs.). The 

 salmonid Oncorhynchus mykiss, by contrast, has no obvious 

 association of the mandibulohyoid connection with the 

 epihyal-interopercular ligament. Both are discrete entities 



throughout their lengths despite having insertion points close 

 together on the epihyal (pers. obs.). The clupeid Etrumeus, 

 unlike the preceding examples, has no readily discernible 

 mandibulohyoid connection. However, the geniohyoideus 

 muscle has a thickened and tendinous dorsal margin which is 

 macroscopically continuous with the muscle from the latter's 

 origin near the dentary symphysis to its insertion immediately 

 over the epi-ceratohyal suture (pers. obs). Superficially at 

 least, the situation in this clupeid shares certain similarities 

 with the mandibulohyoid link in the perciform cichlid Astato- 

 tilapia elegans (see Aerts et al., 1987, and p. 99 above). In the 

 clupeid, however, the differention of the linkage from the 

 associated muscle is at a somewhat lower level of develop- 

 ment than that in the cichlid. 



Verraes (1977) highlighted the functional importance of 

 the mandibulohyoid connection in immediately post-hatching 

 stages of the salmonid Oncorhynchus mykiss. This apparently 

 ligamentous connection develops earlier than the 

 interopercular-mandibular ligament. Thus at this point in the 

 fish's life-history it is an essential element in bringing about 

 jaw depression, and consequently it plays a major role in the 

 creation of the trans-buccal water current involved in respira- 

 tion and feeding (see also Lauder & Liem [1989] for a 

 discussion of this ligament in the feeding mechanism of 

 another salmonid, Salvelinus fontinalis). Recently, Aerts et 

 al., (1987), working with the cichlid Astatotilapia elegans, 

 postulated that a mandibulohyoid connection is also of crucial 

 importance in the feeding mechanism in adults of that spe- 

 cies. 



Regretably, no experiental work has been carried out on 

 the feeding mechanisms of cirrhitoid fishes, nor is there 

 enough critical information on their feeding habits to deter- 

 mine what correlations may or may not exist between species 

 with or without a mandibulohyoid connection. It would be 

 interesting to know in what way the mandibulohyoid connec- 

 tion functions in cirrhitids such as Cyprinocirrhites polyactis. 

 Judging from preserved specimens it would seems to block 

 the sinking of the lower jaw when the hyoid is pulled 

 posteriorly by the contracting sternohyoideus muscle - a 

 somewhat anomalous situation, but possibly one that may be 

 associated with a specialized suction mode of feeding on small 

 crustacean zooplankters, apparently the principal food of this 

 species in South African waters. 



As yet, the intrabuccal tendon and ligament systems are 

 known from too few cirrhitoid taxa to test its usefulness in the 

 intragroup taxonomy and phyletic relationships of those 

 fishes. However, the tendon system in the Cheilodactylidae 

 examined, when compared with that in the latrid Acanthola- 

 tris monodactylus (cf. Figs 4 & 3) supports the latter taxon's 

 removal (see Greenwood, 1995) from the genus Cheilodacty- 

 lus and the family Cheilodactylidae in which it had been 

 placed previously. Those differences also provide an addi- 

 tional character complex for distinguishing the Latridae from 

 the Cheilodactylidae. 



Acknowledgements. My thanks go to Professor Tom Hecht and Dr 

 Colin Buxton of Rhodes University's Department of Ichthyology and 

 Fisheries Science, as well as to their students, for giving me access to 

 that Department's collections of preserved material, and for collect- 

 ing other specimens when needed. To Dr Phil Heemstra of this 

 Institute, my thanks for information on, and discussions about, 

 cirrhitoid fishes. For her patience, forbearance and skill, it is a 

 pleasure to thank Huibre Tomlinson who once again has turned my 



