triacanthids into the flatter and crested supraoccipital of 

 balistids passed through a low crest stage such as is 

 found today in Rhinecanthus before achieving the higher 

 and sturdier buttressing crest found in most other Recent 

 balistids. However, the low crest of Rhinecanthus could 

 well be a secondary reduction of the crest from a high 

 crested ancestral stock and simply be associated with the 

 low nape of Rhinecanthus. 



In short, the generic relationships within the 

 Balistidae are not clear on the basis of the osteological 

 data presented here, nor are many of them clear on the 

 basis of external features. However, Xanthichthys and, 

 especially, Odonus, are two of the most specialized 

 genera and are probably closely related, the former an- 

 cestral to the latter. Canthidermis is also specialized, in 

 a much different way, for an oceanic existence, the scales 

 all being reduced in size and nonoverlapping and the 

 body thus more flexible. Canthidermis is also specialized 

 in having the full ossification of the skeleton much 

 delayed, as well as in having the most rudimentary pel- 

 vic apparatus among the balistids. Instead of the firm os- 

 sification and sturdiness of the skeleton of all other 

 balistids, the bones of specimens of Canthidermis under 

 about 100 mm remain poorly ossifed and slightly spongy, 

 although not to the extent found in the molids Mola and 

 Masturus, but less ossifed than in the molid Ranzania. 

 Only in fully adult specimens of Canthidermis is the 

 skeleton fully ossified. The pelvic fm ray of Canthidermis 

 is discussed below. 



Balistapus and Rhinecanthus are also obviously close- 

 ly related, as explained above, the latter probably being 

 a specialized offshoot of the former. Likewise, Batistes 

 (including Nematobalistes and Verrunculus, properly 

 synonymized with Batistes by Berry and Baldwin 

 1966:435) and Pseudobalistes seem closely related, the 

 latter a derivitive of the former with reduced cheek scales 

 and preocular groove. Sufftamen may be related to 

 Batistes, and Metichthys to Xanthichthys, hnt Abatistes 

 and Batistoides are not clearly (to me) related closely to 

 any of the groups of genera discussed above. I suppose 

 that the basic question is whether the Batistes-like 

 genera are more generalized than the Batistapus -like 

 genera. I suspect the latter are the more generalized, but 

 have no firm evidence of it. 



Fraser-Brunner (1935a:660-661) thought Canthi- 

 dermis to be the most primitive genus, "since there is no 

 evidence that it is derived from any of the forms having 

 ossified plates behind the gill opening." With this I do 

 not agree, Canthidermis simply having lost the overlap- 

 ping body scales found in all other genera, but from what 

 stock Canthidermis is derived I cannot say, although I 

 would suspect a Batistes-like origin. 



Fraser-Brunner thought Abatistes to be a slightly 

 modified derivative of Batistes on a line leading to 

 Batistoides, Sufftamen, and Batistapus, with Rhine- 

 canthus a derivative of the latter. He also thought 

 Batistoides to be ancestral to Pseudobalistes on the one 

 hand and to Metichthys and hence Xanthichthys on the 

 other, with Odonus a derivative of Metichthys. I can add 

 nothing to these unproven speculations. 



Relationship to the Monacanthidae.— Balistids are 



clearly derived from triacanthids, with the major 

 changes involving the sturdier, stronger skull structure 

 associated with the nonprotrusible jaws with less mas- 

 sive crushing teeth, and heavier scales with less body 

 flexibility but more delicacy of swimming movements 

 associated with the sacrifice of speed for defensive 

 features, including the development of the complex 

 locking mechanism of the first two spines and a rotatable 

 pelvis. Balistids are likewise clearly ancestral to mona- 

 canthids, the major changes involving a reduction 

 in massiveness of the skull associated with the thinner, 

 more sharply edged teeth serving a more nibbling 

 function with finer food and thinner, less regularly over- 

 lapping scales associated with a sacrifice in defensive 

 features such as the reduction of the spiny dorsal fin and 

 less swimming strength for a more secretive and often 

 cryptic existence, blending into the environment rather 

 than standing out conspicuously from it. The more 

 secretive monacanthids also tend to be much smaller in 

 size than balistids, which are among the largest of plec- 

 tognaths (exclusive of molids), except that the four 

 species oi Atutera all reach sizes in excess of 200 mm; 

 one, A. monoceros, reaching over 5(X) mm SL (Berry 

 and Vogele 1961). 



The more important osteological differences between 

 the balistids and monacanthids, as pointed out in the 

 comparative diagnosis of each, are mostly concerned 

 with the reduction in size and forward displacement of 

 the spiny dorsal fin and its basal pterygiophores and the 

 features associated with the more delicate teeth and jaws 

 and their supporting structures. 



The heavy crested supraoccipital of balistids sup- 

 porting the front end of the first basal pterygiophore of 

 the spiny dorsal fin becomes, at least posteriorly, a 

 delicate flattened plate in monacanthids upon which the 

 ventral surface of the basal pterygiophore rests, at least 

 in part. However, a few genera of monacanthids retain a 

 vestige of the large vertical crest of balistids as a smaller 

 crest on an anterior prolongation of the bone above the 

 orbit, the crest supporting the ventral edge of the ver- 

 tical platelike anterior region of the basal pterygiophore. 

 The deep hole in the balistid skull surrounded by the 

 supraoccipital and epiotics in which is held the short 

 anteroventral shaft of the first basal pterygiophore is lost 

 in monacanthids, and the epiotics in the latter no longer 

 have the dorsal extension that in balistids helps buttress 

 the apex of the skull for the support of the anterior end of 

 the first basal pterygiophore. 



The second basal pterygiophore of balistids sup- 

 porting the third spine is lost by monacanthids along 

 with the third spine and the supraneural strut that in 

 balistids supports the posterior end of the carina against 

 the vertebral column. The second dorsal spine of 

 balistids becomes much reduced in size in 

 monacanthids, but the locking mechanism, involving 

 only the basal regions of the two spines, remains the 

 same. The massive first dorsal spine of balistids becomes 

 more slender in monacanthids, and usually more highly 

 ornamented with spinules and barbs. However, the first 



