22 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. I42 



snout. The evolution of these bones is closely associated with the 

 functions they serve. Originally this function was that of protecting 

 the lateral line canals of the surface of the snout. However, in many 

 lower teleosts three of these bones — the supraorbital, antorbital, and 

 lacrimal — have developed into a system for pumping water in and out 

 of the olfactory capsule. Some aspects of the sensory canal system 

 and its ossicles in certain lower teleosts will be described first. The 

 remainder of the section will be devoted to the pumping system and 

 its ossifications. 



Transverse sensory canals of the snout region in certain lower 

 teleosts. — The ethmoidal commissure of the sensory canal system has 

 a rather brief history in modern teleosts. Presumably, like the supra- 

 temporal commissure, it was originally enclosed in a series of sepa- 

 rate, flat, roofing bones, as in Amia. 



Among the teleosts the ethmoidal commissure most closely ap- 

 proaches the Amia condition in Elops (Nybelin, 1957). Here the 

 median portion of the canal passes through a bone called the rostral 

 by Nybelin (1957, p. 456, fig. 2), but more generally termed the 

 mesethmoid (cf., Starks, 1926, p. 143) ; in any event all traces of in- 

 dependent, superficial, canal-bearing plates have disappeared. Lat- 

 erally on each side, the commissure just misses the front of the 

 supraorbital canal and then passes back through two lateral rostral 

 plates to join the infraorbital canal at the front of the lacrimal. 



In a young specimen of Megalops examined, as in Tarpon (Nybelin, 

 1957, p. 457), there is also a well-developed ethmoidal commissure. 

 But unlike Elops, that of Megalops fails to connect laterally with the 

 infraorbital canal. Furthermore, there is only one lateral rostral on 

 each side instead of two. Presumably, the posterior lateral rostral 

 with its canal has dropped out, eliminating the junction between the 

 commissure and the infraorbital system. 



My efforts to find, by gross dissection, either a bone-enclosed 

 ethmoidal commissure or lateral rostral ossicles in alepocephalids, 

 clupeids, and round herrings have been unsuccessful. Presumably the 

 ethmoidal commissure described for Chipea by Wohlfahrt (1937) 

 passes entirely through the flesh of the snout. (The sensory canals 

 of the head in teleosts are by no means always bone enclosed, cf., 

 Gosline, 1949, p. 3.) 



The ethmoidal commissures described to this point are easily recog- 

 nized as such. However, certain other lower teleosts have transverse 

 canals in the snout region so peculiar as to arouse doubt whether or 

 not they are commissure derivatives. One such series has been de- 



