other tetraodontids are the exceptionally well-developed 

 neural prezygapophyses as anterolaterally projecting 

 prongs found in lagocephalus and scleratus, which, while 

 not as large as in Triodon, at least approach those of 

 Triodon in size. However, these prongs and the keels on 

 the neural arches seem to be specialized structures 

 associated with the musculature of a strongly powered 

 caudal peduncle for rapid swimming and are probably 

 independently evolved in Triodon and in a few of the 

 species of Lagocephalus. 



Lagocephalus cannot be reasonably considered as an 

 especially generalized tetraodontid. It possesses many 

 unique features not found elsewhere among the tetra- 

 odontids or the ancestral triodontids and triacanthodids 

 (most importantly the posterolateral wings of the fron- 

 tals partially or completely enclosing a temporal fossa, 

 the horizontal epural, especially short first dorsal and 

 anal fin rays, posterior prolongation of the distal ends of 

 the last dorsal and anal fin basal pterygiophores) and 

 other features that, while not unique to it, can still be 

 considered specialized (highly folded olfactory 

 epithelium, two lateral lines). 



It seems more reasonable to me to consider Lago- 



Fifcure 262.— Dorsal views of skulls of: 



left, Guentheridia formoaa, 175mmSL, Panama 



(Pacific); right, Colomeeus peittacus, 179 mm SL, 



Surinam. 



cephalus a specialized derivative of a Sp/ioerotdes-like 

 group, for the following reasons. Lagocephalus scleratus 

 must be considered, overall, the most generalized species 

 of the genus, because, in spite of a few specializations 

 probably mostly associated with its mode of swimming 

 (depressed body and caudal peduncle, 17 vertebrae, 

 single hypohyal, exceptionally well-developed neural 

 prezygapophyses and last basal pterygiophore prongs) it 

 has a lesser degree of development of the specialized 

 features that characterize the genus (especially the even- 

 ly tapered frontals of moderate width and the relatively 

 poorly developed posterolateral wings which do not meet 

 the pterotic and thus only partially enclose the temporal 

 fossa) as well as many other generalized features (first 

 pharyngobranchials with numerous teeth, no dorsal 

 flange on the parasphenoid contacting the frontals in the 

 orbit, well-developed prootic prongs representing the 

 remains of the dorsal roof of the myodome, moderate 



325 



