with dorsal end of latter" would seem to describe a 

 modified tympanal scale. 



The two Recent genera which lack large tympanal 

 scales, Canthidermis and Xanthichthys, are otherwise 

 very different in structure and surely are not closely 

 related. It is simply impossible to say precisely at present 

 whether the generalized condition of balistids is the 

 presence of enlarged plates which have subsequently 

 been lost by a few genera, or whether the basal balistids 

 lacked specialized tympanal scales which were sub- 

 sequently developed by most genera. I think the former 

 view to be the more likely, and that the development of 

 specialized tympanal scales aiding in and refining sound 

 production was one of the major features of the basal 

 balistids which led to their differentiation from the 

 triacanthids, concomitant with the reduction of the 

 spiny dorsal fin and development of a complex locking 

 mechanism of the first two spines, the reduction of the 

 pelvic fin and modification of the pelvis into a stout 

 rotatable shaft, and increased sturdiness of the skull to 

 support the nonprotrusible jaws with a change in den- 

 tition. 



In this view, Canthidermis lacks specialized tym- 

 panal scales simply because all of its scales have become 

 relatively small and nonoverlapping, perhaps associated 

 with the more flexible body covering necessary for lateral 

 flexion of the caudal peduncle in a strong swimming and 

 truly oceanic species. Xanthichthys, on the other hand, 

 seems closely related to Odonus, which has large 

 specialized tympanal scales, and I have no explanation 

 for the lack of them in Xanthichthys, except that in 

 Xanthichthys the scales in the tympanal region are 

 smaller than those surrounding them, but they are not 

 overlapping and are just about as flexible as those of the 

 tympanal region of other balistids. It would seem that 



Xanthichthys has not truly lost the specialized tym- 

 panal scales, but simply reduced their size. It would be of 

 interest to know if the sound productions associated with 

 the drumming of the pectoral fin against the side of the 

 body (and thereby against the anterolateral end of the 

 swim bladder) differ in Canthidermis and Xanthichthys 

 from other balistids. 



If the supposed absence of a preocular groove in the 

 fossil species of balistids, of which relatively complete 

 specimens are known only from the Oligocene, is cor- 

 rect, then that structure is a more recent acquisition by 

 all genera except BaUstapus and Rhinecanthus. The 

 latter two genera are very closely related, differing only 

 by Rhinecanthus having a lower supraoccipital crest, a 

 much smaller third dorsal spine and a more constricted 

 caudal peduncle with fewer rows of spine bearing scales. 

 In fact, BaUstapus could serve as a model for a generaliz- 

 ed balistid. The mouth is terminal, the third dorsal spine 

 well developed, the large specialized tympanal scales 

 well developed, the encasing scales at the end of the pel- 

 vis with great flexibility (although with a slightly reduc- 

 ed number of scales in one of the segments) and well- 

 cusped teeth at all sizes and the teeth of decreasing size 

 laterally in both jaws. It is possible, but hardly proven 

 here, that BaUstapus is a relative generalized balistid 

 and that its lack of a preocular groove may also represent 

 the generalized condition. 



One final character must be considered, that of the 

 degree of development and massiveness of the supraoc- 

 cipital and epiotic supports of the basal pterygiophores of 

 the spiny dorsal fin. Is a high sturdy supraoccipital crest, 

 such as found in most balistids, the generalized con- 

 dition? In the early development of balistids from 

 triacanthids, it is reasonable to expect that the process of 

 conversion of the domelike uncrested supraoccipital of 



^W 



Figure 68.— Scales just above tip of 

 pectoral fin of Balistoides viridescens , 



114 mm SL, Indonesia, to show 



the hexagonal shape often obtained by 



the scale plates in adults as the 



anterior and posterior edges of the 



rhomboid plate become flattened. 



^* 



>£ 



