FOSSIL AND RECENT 79 



detailed comparison). The jaw structure is similar to that seen in the Recent forms, 

 in particular the deep maxilla with the palatine and ethmoid processes situated 

 close together. The convex oral margin shows a similar dentition. In the opercular 

 series, similarities are seen in the preopercular sensory canal. Further points of 

 similarity include primitive features such as small premaxillae, the large gular plate 

 and high number of branchiostegal rays. These last points cannot, of course, be 

 held to indicate relationship, but merely suggest that Protarpon has not become 

 specialized in other directions from Recent megalopids. 



On the other hand, there are several features suggesting that Protarpon should not 

 be included in a Recent genus. Protarpon was included by Woodward (1901) in 

 the genus Megalops (which included both Recent species) . In recent years, workers 

 (Hollister 1939 ; Greenwood 1970a) have qualified Jordan and Evermann's (1896) 

 allocation of the two Recent species to separate genera. The basis of this separation 

 has been centred upon the swimbladder/ear linkage (Greenwood 1970a) and certain 

 other details of neurocranial anatomy (see diagnosis of Recent forms and the dis- 

 cussion of M. cyprinoides included here). The Eocene species differ from both 

 Recent forms in such details and thus must be accorded generic status. The name 

 Protarpon has been chosen in the hope that it conveys the idea of greater anatomical 

 similarity to Tarpon than to Megalops. These differences are noted and discussed 

 below. 



The neurocranial roof of Protarpon is relatively flat, in contrast to the convex 

 form in Recent species. In these latter forms the cranial convexity is due to a 

 deepening of the post-temporal fossae, which meet above the endocranium. Such 

 a deepening is not apparent in Protarpon and may indicate that the post-temporal 

 fossae of either side remained separate from one another. As in the Recent species, 

 the post-temporal fossae probably extended forward to the orbitosphenoids, 

 since these latter bones exhibit a wide dorsal separation. The autosphenotic spine 

 of Protarpon is weakly developed and the dilatator fossa small and roofed 

 by the pterotic. This is unlike the condition in extant species where the 

 dilatator fossa is deep. The weak autosphenotic spine is probably correlated with 

 the flat skull roof. 



In the otic region of the neurocranium there is a single shallow depression formed 

 by the exoccipital, basioccipital and prootic. This depression may have contained 

 a simple diverticulum of the swimbladder which must have been only slightly more 

 complex than that seen in large specimens of Elops lacerta (Greenwood 1970a). 

 The swimbladder in Recent megalopids is received in complex depressions in the 

 lateral cranial wall. That the swimbladder projection was simple in Protarpon is 

 suggested by the primitive position of the jugular canal opening, the position of the 

 vagus and glossopharyngeal foramina and the lack of any supporting crista upon the 

 basioccipital. In Recent megalopids the swimbladder has pushed so far forward as 

 to shift the vagus and glossopharyngeal foramina to a higher and more posterior 

 position. In Megalops the jugular canal opening has also shifted. 



The premaxilla of Protarpon is considerably narrower than that of extant megalo- 

 pids. The mandible is relatively shallow, without a very large coronoid process, 

 and the symphysis does not protrude to any great extent. The dorsal margin of 



