In all plectognaths except the triacanthoids, the upper 

 jaw is nonprotractile and the premaxillary immovably 

 articulates, usually by extensive interdigitation, with the 

 maxillary. In triacanthoids the premaxillary and maxil- 

 lary movably articulate with one another and the upper 

 jaw is slightly protractile. The premaxillary of triacan- 

 thoids is more or less L-shaped, with the long arm of the 

 L forming the posteriorly directed ascending process 

 which slides over the dorsal surface of the ethmoid in the 

 process of protracting and retracting the upper jaw. In 

 balistoids and ostracioids the posteriorly directed process 

 of the premaxillary is essentially absent, and the postero- 

 medial end of the premaxillary articulates with the an- 

 terior edges of the ethmoid and vomer, particularly with 

 the former. Along its anterior edge the premaxillary of 

 the scleroderms bears a variable, but usually small, 

 number of discrete teeth. The premaxillary of gym- 

 nodonts is modified, in conjunction with the teeth, into a 

 crushing plate which is often and aptly referred to as a 

 parrotlike beak. The posteromedial end of the premaxil- 

 lary articulates with the ethmoid, and to a lesser extent 

 with the vomer, in Triodon, but in most tetraodontids 

 this articulation is primarily with the vomer. In Triodon 

 and in the tetraodontids the medial edges of the two pre- 

 maxillaries are closely apposed and articulate with one 

 another by fibrous tissue. Emarginations from the 

 medial surfaces of each of the premaxillaries in tetra- 

 odontids alternate with one another and strengthen the 

 articulation between the two elements. In diodontids and 

 molids the two premaxillaries are indistinguishably fus- 

 ed together into a large U-shaped bone. 



Hollard (1857b:308) did not believe that fusion of the 

 premaxillaries, or dentaries, was a character of much sys- 

 tematic importance, possibly because it was known that 

 the premaxillaries and dentaries of extremely large 

 specimens of tetraodontids occasionally fuse together in 

 the same way as do those of diodontids. More interesting, 

 however, is Reuven's (1894:130) observation of a 

 phenomenon in Mola of just the opposite nature to that 

 occasionally observed in tetraodontoids. In his descrip- 

 tion of a young Mola, Reuven said that he "found that 

 the maxilla superior is splitted up in the middle, the in- 

 ferior is not cleft," and a figure was presented (pi. 5) to 

 show this "tetraodont" condition of the upper jaw. 

 Hollard's view that too much importance has been at- 

 tributed to the fusion of the premaxillaries, or dentaries, 

 as a major indicator of phylogenetic affinity seems well 

 taken. It might be mentioned here that the premaxillary 

 and maxillary are often said to be fused or coalesced, 

 when in actuality they, at least in the great majority of 

 instances, only extensively interdigitate with one 

 another. Only very rarely is an extremely large specimen 

 of a tetraodontid or diodontid found in which the pre- 

 maxillary and maxillary are for all practical purposes 

 fused. One of the more extreme statements about the fu- 

 sion of the various jaw bones in plectognaths is that of 

 Wagner (1845:190) who said that the upper jaw bone 

 "coalesces with the vomerine, palatal and nasal bones to 

 form a single bone, which unites, however, with that of 

 the other side by suture, as in Orthagoriscus." Such a 



condition would certainly be the death of any plectog- 

 nath and whether Wagner had ever actually examined a 

 molid is problematical, although he had (1841) once 

 presented a figure of the skeleton of a balistid. 



Only in the triacanthoids is the maxillary movably ar- 

 ticulated and not closely apposed to the premaxillary, for 

 in all other plectognaths the maxillary firmly articu- 

 lates, usually by extensive interdigitation, with the pre- 

 maxillary. In triacanthoids the lateral surface of the 

 rounded dorsal end of the maxillary movably articulates 

 with the anterior process of the palatine. In balistids the 

 palatine articulates with a slight concavity on the sur- 

 face of the maxillary and premaxillary, but in monacan- 

 thids the maxillary scarcely makes contact with the 

 palatine or is even entirely excluded from articulation 

 with it. With reduction of the palatine in ostracioids the 

 maxillary articulates with only the anterior edges of the 

 vomer and ethmoid. In gymnodonts the maxillary al- 

 ways articulates over a broad, and usually slightly con- 

 cave, area with the anterodorsal end of the palatine, and 

 it is this articulation which is the major source of sup- 

 port for the upper jaw in gymnodonts. In triacanthoids 

 the support of the upper jaw is shared about equally 

 between the articulation of the maxillary with the 

 palatine and the articulation of the premaxillary with 

 the ethmoid and vomer, but in balistoids the pala- 

 tine-maxillary articulation becomes progressively less 

 important and in ostracioids it is completely absent. The 

 posterior edge of the maxillary of most balistoids and ostra- 

 cioids bears an indentation which is characteristic of 

 these two groups, but not of the other plectognaths. The 

 lateral surface of the maxillary is relatively flat in all 

 plectognaths except the diodontids, in which this sur- 

 face is upraised into a stout flange for muscle attach- 

 ment. The lower end of the maxillary forms the lower 

 part of the anterior edge of the upper jaw in all plec- 

 tognaths. Among Shufeldt's (1917) nearly totally 

 erroneous observations on the osteology of Diodon, there 

 occurs the statement that an "admaxillary" is present on 

 the dorsal edge of each maxillary, but since no descrip- 

 tion or illustration was given of this structure, that 

 author's statement is best forgotten, as is his other work 

 (1926) on plectognaths. 



In scleroderms the teeth are usually few in number but 

 they are always individually recognizable units. In the 

 gymnodonts, however, the teeth are so highly modified 

 that they are scarcely recognizable as such. The number 

 and shape of the teeth in the jaws are discussed in the 

 diagnoses of the familial groupings recognized here and 

 are summarized below. 



The number and shape of the teeth vary most in the 

 triacanthodids, and Gunther (1870), Myers (1934), 

 Fraser-Brunner (1941a), and Tyler (1968) have all made 

 prominent use of dental formulae in their revisions of the 

 family. The teeth of Recent triacanthodids are usually 

 conical and usually occur in a single series. This outer 

 series contains the majority of teeth, which usually vary 

 in number from about 10 to 50, but with reduced num- 

 bers (sometimes absent) in the long-snouted genera. 

 When teeth in an inner series are present, in the more 



