NO. 3 TELEOSTEAN FISHES — GOSLINE 21 



formes {Aphredodorus and Percopsis examined), the Beryci formes 

 {Holocentrus examined) and the Beloniformes (Chriodorus and 

 Athlennes examined). 



In view of the variability of the radials and the paucity of the ma- 

 terial studied it seems inadvisable to push too far the inference that 

 can be drawn from the inner radial regarding genetic relationships. 

 Suffice it then to summarize by suggesting that the Notacanthi formes 

 and Cypriniformes seem to show an isospondylous type of inner radial 

 organization, and the Percopsi formes, Beryciformes, and Beloni- 

 formes, an iniomous type. 



The history of the outermost radial may be dealt with even more 

 briefly. In a fish like Salmo (fig. 6A) or Synodus, this radial is a 

 flattish, pebblelike nodule lying between the bases of the halves of the 

 three outermost pelvic rays. In the beloniform Athlennes and the 

 syngnathiform Fistularia the author has been unable to find any in- 

 dependent radial ossifications. In the beryci form genera Holocentrus 

 and Myripristis the outer soft ray has been transformed into a rather 

 formidable spine. This spine has a complicated articulation with the 

 pelvic girdle which effectively restricts movement to one plane. (The 

 same thing is true of the holocentrid dorsal spines, which seem to have 

 essentially the same system of basal articulation.) In Holocentrus and 

 Myripristis there is a small, movable nodule of bone (fig. 5D), flattish 

 except for a curved projection that penetrates a hole in the base of the 

 pelvic spine, that presumably represents the outer radial. In the zei- 

 form genus Antigonia and several percoids examined there is no trace 

 of a separate ossification between the base of the pelvic spine and the 

 girdle. 



Concerning the pelvic spines formed from soft rays (and not from 

 the outer, curved splint discussed earlier), it has already been noted 

 that there are several of these in notacanthids. Aside from this group, 

 a true pelvic spine seems to occur first in the berycoids and zeoids. 

 At least, the outer rays of the gadid Lota, the lampridiform Velifer, 

 and others all proved to be of the divided and segmented type (lepido- 

 trichia) found in lower forms. There is, however, no guarantee that a 

 fish with an outer soft pelvic ray has not merely lost the spine as has 

 indeed happened in the majority of the flatfishes (Norman, 1934; 

 Hubbs, 1945). 



III. SOME BONES OF THE SNOUT REGION IN MODERN 

 TELEOSTEAN FISHES 



The bones discussed in the present section are the supraorbital and 

 certain of the ossifications surrounding the sensory canals of the 



