sphenotic visible on the dorsal surface of the skull. 

 Completion of the process would require only the coming 

 together in close proximity of the distal ends of the two 

 posterolateral wings of the frontal (those that had replac- 

 ed the two wings of the sphenotic) to form a more com- 

 plete well around the muscle to the operculum. 



In Carinotetraodon the supraoccipital crest is not as 

 high and laterally compressed as in Canthigaster, but it 

 is higher than in most other tetraodontids, and such a 

 condition could easily give rise to that of Canthigaster. 

 The parasphenoid of Carinotetraodon has a well- 

 developed ventral flange and is arched dbrsally in front of 

 the orbit much as in Canthigaster. In Carinotetraodon 

 the prefrontal is relatively posterior in position, being 

 only slightly forward of about the middle of the skull, 

 and while it does not have the characteristic smooth flat 

 dorsal surface gently curved downward anterolaterally as 

 found in Canthigaster, it is in contact anteriorly with the 

 vomer only by a thin but wide anterior prolongation. If 

 this anterior prolongation was reduced in size and then 

 lost, the prefrontal of Carinotetraodon would have the 

 Canthigaster condition of being far removed from the 

 vomer. The ethmoid of Carinotetraodon is not especially 

 long in comparison to many other tetraodontids, but it is 

 long enough for one to envision its elongation in the 

 process of lengthening the snout in Canthigaster and the 

 conversion into a T-shaped cross section. A few other 

 features that would be necessary to the conversion of a 

 Carinotetraodon-\ike fish into a Canthigaster-Vike one 



are the development of a high anterodorsal wing on the 

 suboperculum and a straightening of the dorsal edge of 

 the operculum. 



One of the important distinguishing features of 

 Canthigaster is the single nostril at the end of a very 

 small short tube, and it is reasonable to suppose that 

 Canthigaster was derived from a tetraodontid line with a 

 single nostril at the end of a better developed nasal tube. 

 Among tetraodontids a nasal tube with a single nostril is 

 found only in a few species of Monotreta (a few other 

 species of the genus have the opening bilobed) and in the 

 closely related Carinotetraodon. 



Only Carinotetraodon and Canthigaster are known to 

 have the ability to lift up a ridge of skin along the 

 midline, and the osteology of Carinotetraodon shows 

 striking similarities to that of Canthigaster, more so than 

 does that of any other tetraodontid. It seems highly likely 

 that Carinotetraodon and Canthigaster share a close 

 common ancestry, from which Carinotetraodon has 

 diverged little but from which Canthigaster has become 

 highly enough modified to warrant subfamilial recog- 

 nition. It would be of interest to know if the common 

 ancestral group was marine, fresh water, or euryhaline. 



Monotreta and Chelonodon, closely related to Carino- 

 tetraodon, contain species which show a few of the same 

 features (arched vertebral column, well-developed 

 haemal spines on the abdominal vertebrae) of similarity 

 between Carinotetraodon and Canthigaster, but, to a 

 much lesser degree, than in Carinotetraodon. Neverthe- 



parasphenoid 

 prefrontal 



Figure 245.— Spfcoeroide* mat 



(atturlateral view of head, 



97.5 mm SL, Virginia. 



