166 TRANSACTIONS OF THE AMERICAN PHILOSOPHICAL SOCIETY 



St.) erroneously called the supratemporal." In the Eocene species of the genera Megalops 

 and Notelops, Smith Woodward states that the "operculum is subdivided by a transverse 

 suture" (1901, pp. 24, 27) and he notes the presence in a certain specimen of Notelops 

 brama of a "separate plate above the operculum" (p. 28). Hence it seems not improbable 

 that in primitive isospondyls a "subtemporal" was present either as a separate element or 

 as the lower and anterior moiety of the "supratemporal" scale bone or tabular. 



With regard to the relationships of the family with other Cretaceous isospondyls, 

 Smith Woodward concluded that the "Osteoglossidae with a curiously thickened skull, also 

 seem to be closely related to the early Albulidae" (1901, p. vii). Several Cretaceous and 

 Eocene genera having small teeth on the parasphenoid and other bones within the mouth 

 may possibly mark the transition from the Albulidae to the Osteoglossidae. The several 

 species of Plethodus, for example, bore concave dental plates on the base of the cranium, 

 which were opposed to a similar but convex plate supported probably by the basihyal bone. 

 This at least suggests the prominent patch of teeth on the parasphenoid of Arapaima, 

 which were opposed to the lingual teeth. Bridge (1895) inferred that in Osteoglossum 

 there was a lateral movement of the teeth on the mesial edges of the two entopterygoids, 

 but Ridewood (1905a), referring to the fact that although these upper teeth are obliquely 

 set, their tips point downward, concludes that "There can be no question that these teeth 

 act in a vertical direction and are opposed to the lingual teeth borne upon the bone that 

 covers the glossohyal cartilage and the basibranchials." Thus in certain Cretaceous 

 Albulidae and Osteoglossidae, as well as in their modern relatives, there was a dental appa- 

 ratus within the mouth analogous in some respects with those which have been developed 

 independently and in various ways, in pycnodonts, wrasses, synentognaths, etc. 



Returning to the consideration of Ridewood's analysis of the skull characters of the 

 Osteoglossidae, he did not find any very conspicuous specializations peculiar to osteoglossids 

 and albulids apart from the tendency to develop a masticatory apparatus within the mouth, 

 as above noted. Nevertheless he remarks, in discussing the various arrangements of the 

 families of malacopterygian fishes proposed by Giinther, Gill, Smith Woodward, Boulenger, 

 that "I should be disposed to associate the Osteoglossidae with the Pantodontidae for the 

 reasons given on page 276, and to regard the next nearest family to be the Albulidae. The 

 conclusion is arrived at by a consideration of the craniological features mainly, but the 

 characters of the other parts of the skeleton and of the soft parts of the body, so far as they 

 are known to me, do not militate against the suggestion that the Osteoglossidae and the 

 Albulidae have descended from a common stock." 



Garstang (1932, p. 245) considers that the "feeding mechanism of parasphenoidal and 

 hyoidean teeth" in the osteoglossoids constitutes an important link with the elopines, 

 that their skull and jaw characters entitle them to membership in the "archicraniote" 

 section of the "Otophysi" (pp. 253, 256), and that the presence of air vesicles in the tem- 

 poral fossae link them with the Ostariophysi in the grand division Otophysi. I, however, 

 am now rather more impressed by the differences from the albulids and elopines, even in 

 essential details of the median teeth on the floor of the branchial region; while the total 

 absence of Weberian ossicles, together with the lack of essential resemblances to the 

 characins even in the Eocene osteoglossids, indicate that the connection with the Ostario- 

 physi is, at best, very remote. 



