number of dorsal and anal fin rays, least highly folded 

 olfactory epithelium, minimal surface sculpturing on 

 top of the skull, spines on the back as well as the bel- 

 ly), a greater combination of generalized features than 

 any of the other species. 



A scleratus-like stock, minus its specializations, prob- 

 ably gave rise to inermis, which alone among Lago- 

 cephalus shares a few of the generalized features of 

 scleratus (first pharyngobranchial with numerous teeth, 

 no dorsal flange on parasphenoid) and has the second 

 most generalized configuration of the cranium, with the 

 posterolateral wing of the frontal similar to scleratus and 

 the anterior region of the frontal only slightly more 

 specialized than in scleratus by its greater width and 

 more abrupt tapering. It is probable that this line leading 

 to inermis also gave rise, by further specializations and 

 before the scales were totally lost, to lunaris, and that a 

 lunaris-like stock is ancestral to spadiceus, laevigatus, 

 and lagocephalus. I would guess that in general 

 spadiceus has remained closer to this ancestral line than 

 have laevigatus and lagocephalus, with lagocephalus a 

 specialized more open water derivative of a laevigatus- 

 like stock. 



If scleratus can be accepted as the most generalized 

 species of Lagocephalus, then it is probable that Lago- 

 cephalus is an originally deepwater derivative of a 

 Sp/ioeroides-like ancestral group, for the cranium of 



Figure 263.— Dorsal views of skulls of: 



A, Lagocephalus inermis, 52.2 mm SL, Bay of Bengal; 



B, L. scleratus, 80.9 mm SL, Philippines; 



C, L. lunaris, 62.3 mm SL, Bay of Bengal. 



scleratus is not markedly different from that of several 

 species of Sphoeroides, including that of pachygaster, 

 which alone among Sphoeroides has a deepwater habitat, 

 a highly folded olfactory epithelium, and a general 

 countenance similar to that of the deep, chunky bodied 

 species of Lagocephalus, such as inermis, the closest 

 relative of scleratus and a species which may have 

 retained the general appearance of the ancestral line 

 leading to it and scleratus, which scleratus lost as it 

 became specialized for more rapid swimming. In S. 

 pachygaster the frontals are about as wide as in L. 

 scleratus, and they are similarly evenly tapered anteri- 

 orly, and the ethmoid-vomerine regions are remarkably 

 similar. Moreover, while not prolonged posterolaterally, 

 the broad posterior end of the frontal projects out over 

 the underlying epiotic and pterotic as a short roof over 

 this region. In many other species of Sphoeroides, 

 both specialized (e.g., dorsalis and testudineus) and 

 generalized (e.g., spengleri and maculatus) the posterior 

 end of the frontal is slightly prolonged posterolaterally 

 over the underlying epiotic and pterotic as a short roof. 

 An only slightly increased posterolateral prolongation of 



