1904.] OSTEOLOGY OF THE ELOPID/E AND ALBULID.E. 77 



attached by two ligaments to tlie right and left liypohyal bones, 

 except in the Momiyroid fishes, in which it is I'igidly fixed beneath 

 the anterior pai-t of the copular skeleton. 



Shufeldt (Rep. U.S. Com. Fish. 1883 (85), p. 820) writes that 

 the urohyal " lies between the sternohyoid muscles, and is not 

 always present where a glossohyal exists." The latter part of the 

 quotation is, I think, open to serious question ; it is probably 

 always present in Teleostean fishes. Brooks (Proc. Roy. Dubl. 

 Soc. n. s. iv. 4, 1884, pp. 180-183), in his description of the skull 

 of the Haddock, while adopting the name basibranchiostegal for 

 this tendon-bone, misapplies the term urohyal to that cartilage 

 which lies between the last two pau-s of ceratobranchials, and 

 represents the foui'th and fifth basibranchials. 



In connection with the urohyal, it is of interest to observe that 

 in Notopterits there are, in addition to the ui-ohyal proper, a pair 

 of tendon-bones of similar character to it, but of smaller size, 

 projecting backwards from the posterior end of the ventral surface 

 of the second basibranchial. Such tendon-bones are also present 

 in Osteoglossum, Heterotis, and the Mormyridss ; but in these 

 fishes they are confluent with the reduced second hypobranchials. 

 In dealing with the homologies of tendon-bones, one must ever be 

 prepared to admit the possibility of convergence ; thus, while in 

 Notopterus there are separate tendon-bones related to the second 

 hypobranchials and the second basibi"anchial, in Diodon {Dicotyl- 

 ichthys) a pair of exactly similar bones project down from the 

 mesial ends of the third ceratobi-anchials, and repi-esent either the 

 downwardly directed and gi-eatly elongated thii-d hypobranchials, 

 or a pair of tendon-bones confluent with the third hypobranchials, 

 or simply a pair of tendon -bones, the third hypobranchials being 

 absent. In Polyptertis, again, the urohyal as a median bone is 

 wanting, but a pan- of tendon-bones project downwai'd and 

 backward from the ventral surface of the anterior end of the 

 ceratohyal. 



The hypohj^al of each side of the head is usually double, con- 

 sisting of distinct upper and lower ossifications. The right and 

 left upper hypohyals are separated by the first basibranchial or 

 the glossohyal, or both, but the two lower hypohyals are articu- 

 lated together in the median plane. The lowei- hypohyal in most 

 cases is larger than the upper, but in Hyodon, IJlops, Megalops, 

 and Alhula the upper and lower hypohyals are approximately 

 equal in size. In A rapalma, Heterotis, Osteoglossum, Petrocephahis 

 and Notopterus there is but one hypohyal on each side, and this 

 would appear to represent the u.pper one. This is rather odd 

 when considered in relation with the probability that the 

 single hypohyal of Amia represents the lowei- of the two. In 

 Mormyroids other than Petroceplialus no hypohyals are I'eco"-- 

 nisable. 



It is difficult to understand why Supino, in his recent work on 

 the skull of deep-sea Teleosfceans (Ric. Lab. Anat. Univ. Roma 

 viii.-ix. 1901-2), and Starks, in his description of the Serranoid 



