2nd uroneural 



epural 



neural spines 



haemal spines 



caudal vertebrae 



procurrent fin rays 



are fully fused together in diodontids. Even accepting 

 this reasoning, the presence of Triodon-like jaw frag- 

 ments tells us nothing about what the fish behind the 

 jaws was like, there being no guarantee that the general 

 form of antiquus was at all like that of the Recent 

 Triodon macropterus . Moreover, apper jaws without fus- 

 ed premaxillaries but with small rounded dental units in 

 the matrix are what one would expect to find in the 

 Eocene ancestors of the early to late Eocene diodontids 

 such as Prodiodon and Progymnodon. 



Relationships to the Balistoidei and to the other 

 Tetraodontoidei. — The relationship of the Triodon- 

 tidae as an intermediary between the eoplectin 

 Triacanthodidae and the other more specialized 

 Tetraodontoidei is discussed under the subfamilial rela- 

 tionships of the Triacanthodidae, it being shown, in es- 

 sence, that Triodon, which is by far the most generalized 

 of the gymnodonts, clearly is derived from the eoplectins, 

 which retain many generalized triacanthodid features 

 while having highly specialized gymnodontlike jaws with 

 small rounded dental units incorporated into the matrix 

 of the premaxillaries and dentaries. Since it is highly im- 

 probable that such complex dentitional changes as found 

 in the gymnodonts have arisen independently in various 

 lines, the eoplectins are obviously ancestral to the gym- 

 nodonts, with their closest relatives among the gym- 

 nodonts being the triodontids. 



While the evolution of small rounded dental units in- 

 corporated into the matrix of the jaw bones is a highly 

 complex matter that has arisen only once among plec- 

 tognaths (although similar dentition has been in- 

 dependently developed in some of the Scaridae among 

 the perciform fishes), the same is not true of the fusion of 

 the dentaries and premaxillaries to their opposite mem- 



Figure \92.— Triodon macropterus: lateral 



view of caudal fin supporting structures 



(see T>ler 1970b:figs. 42a and 42b, 



for details), 391 mm SL, Philippines. 



bers, which, relative to the dentitional and general struc- 

 tural changes involved in developing a crushing or biting 

 beak, is a simple matter. The relative superficiality of 

 the jaw bone fusion is attested to by the fact that the 

 dentary and premaxillary of very large specimens of 

 tetraodontids occasionally fuse to their opposite mem- 

 bers and that Reuvens (1894:130) described a young 

 molid in which the dentaries were fused but the pre- 

 maxillaries were separate. Thus, the form of the dental 

 units in the biting edge of the jaws is a far more 

 anatomically complex and phylogenetically important 

 indicator than is the fusion or lack of fusion of the den- 

 taries and premaxillaries. 



The fusion of these bones to their opposite members 

 undoubtedly at least slightly increases the rigidity and 

 strength of the beak beyond that which it would possess 

 if the premaxillaries and dentaries remained separate, 

 but I doubt that the difference is functionally of great 

 magnitude, for the two halves of the upper and lower 

 jaws of tetraodontids are firmly and relatively inflexibly 

 held together medially by a combination of interlocking 

 emarginations and thick bands of tough fibrous tissue. 

 The two halves of the upper jaw in Triodon, for example, 

 are so extensively interdigitated that they are ex- 

 ceedingly difficult to force apart, and I doubt that much 

 more pressure would be needed to break apart the fused 

 dentaries than to separate the interdigitated premaxil- 

 laries. 



The most generalized dentitional condition in gym- 

 nodonts is small rounded units incorporated into the 



