interhyal and only one hypohyal, the first pharyngo- 

 branchial has small to minute teeth and there are no 

 trituration teeth in either jaw. A sesamoid articular was 

 not found in the lower jaw of any of the specimens of 

 Chonerhinos and Xenopterus examined, and these 

 genera may be at least unusual, if not unique, among the 

 tetraodontids in the loss of this element. 



In short, the skull in the more generalized of these two 

 specialized genera bears its greatest similarity to that of 

 some of the species of Monotreta, Chelonodon, and 

 Tetraodon, perhaps especially to the latter, but the pre- 

 cise relationship of Chonerhinos to any of them is un- 

 clear. One would expect Chonerhinos to have arisen from 

 a line having a nasal sac with a single nostril, a well- 

 developed lateral line system with at least one and a half 

 or two lines on the body, a tendency to increase the num- 

 ber of dorsal and anal fin rays above the perhaps 

 generalized number of about 10 to 12 and a tendency to 

 increase the number of vertebrae above the generalized 

 number of 20, all of which features are found in varying 

 degrees among one or the other of Monotreta, 

 Chelonodon, and Tetraodon, and the ancestry of 

 Chonerhinos is probably shared at one point with that of 

 one or more of those three closely related genera. 



Chonerhinos and, progressively more so, Xenopterus are 

 probably derived from the ancestry of the Monotreta- 

 Chelonodon-Tetraodon group, but their more precise 

 relationships remain unknown. 



Canthigaster is deemed to be sufficiently anatomi- 

 cally distinct from the other tetraodontids to be recog- 

 nized as subfamilially distinct (Canthigasterinae) from 

 them (Tetraodontinae), although the author is biased 

 toward the conservative approach of only subfamilial 

 recognition for Canthigaster, since recently it usually is 

 given full familial rank. Almost as much anatomical dis- 

 tinctiveness is present between Chonerhinos-Xenop- 

 terus and the other tetraodontins as there is between the 

 canthigasterins and tetraodontins. Moreover, there is no 

 genus known as closely intermediate between 

 Chonerhinos-Xenopterus and the other tetraodontins as 

 Carinotetraodon is between the canthigasterins and the 

 other tetraodontins. Thus, there is no compelling reason 

 to recognize Canthigaster as even subfamilially distinct, 

 and it is done here more as a matter of personal 

 preference than on the basis of persuasive anatomical 

 evidence. 



The argument against subfamilial recognition of 

 Canthigaster could also take into account the situation 



Summary of generic relationships and intra- 

 familial classiflcation. — Because of the complexity of 

 the external and osteological diversity of the tetraodon- 

 tids, most of the analyses of the generic relationships 

 within the family are given in the preceding section on 

 anatomical diversity, which need be only summarized 

 here. 



The genera having a nasal sac with two nostrils seem to 

 form a natural group whose more generalized represen- 

 tatives are more generalized than most of those genera in 

 which a single nostril is present at the end of a tube or in 

 which there is a tentacle or an open cup nasal ap- 

 paratus. Among the genera with two nostrils, 

 Sphoeroides is the only one with a single lateral line, and 

 its osteology seems to be overall the most generalized, 

 with Lagocephalus and Colomesus derived from a 

 Sp/ioeroides-like ancestral group, but from rather dif- 

 ferent lines of radiation within that group. Ambly- 

 rhynchotes, Fugu, and Torquigener may also have been 

 derived from an early Sphoeroides-like group, and are 

 probably more closely related to one another than to any 

 of the other genera with two nostrils, but their more 

 precise relationships are not clear on the basis of the pre- 

 sent work. 



The genera in which the nasal sac has opened up by 

 the loss of the separation between the two nostrils are 

 probably derived from an ancestral group with two nos- 

 trils and a single lateral line. A Carinotetraodon-Vike 

 form is probably ancestral to Canthigaster, and 

 Carinotetraodon itself is probably most closely related to 

 Monotreta. Ephippion is probably a close derivative of a 

 Tetraodon-like form, and the relationship between 

 Monotreta, Chelonodon, and Tetraodon undoubtedly is 

 close but the details of those relationships are not clear 

 on the basis of the present work. The highly specialized 



Canthigaster 



Carinotetraodon 



Ephippion 



Figure 278.— Hypothesized 

 phylogenetic relationships of 

 the genera of Tetraodontidae. 



341 



