FOSSIL AND RECENT 199 



megalopid otophysic connection was probably derived from a basic type (see foot- 

 note, p. 189) common to the other groups mentioned, its subsequent development 

 has followed a different and restricted path. There is never any direct connection 

 between the swimbladder and the endocranial cavity and the association of the 

 precoelomic diverticulum with the intercalar is unique. 



There are several minor osteological and soft anatomical differences between the 

 otophysic connection of Megalops and Tarpon (Greenwood 1970a), but there is 

 nothing to suggest that both types were not derived from a common ancestral type 

 very much like that of Tarpon. The Eocene megalopids exhibit a simple type of 

 otophysic connection in which the lateral cranial wall has a single large depression 

 for the swollen end of the swimbladder, which presumably lay against the skull at 

 this point as in Tarpon. The diverticulum could not have been as closely associated 

 with the cranium as it is in the Recent megalopids since the vagus foramen, the 

 glossopharyngeal foramen and the posterior opening of the jugular canal still occupy 

 primitive positions. In the Recent megalopids the nerve foramina have moved to 

 allow the swimbladder to fit tightly under the intercalar. Thus in both Tarpon and 

 Megalops the glossopharyngeal leaves the neurocranium immediately beneath the 

 vagus and slightly behind the level of the intercalar. The posterior opening of the 

 jugular canal has moved upwards, so much so in Megalops that it opens directly 

 into the subtemporal fossa, and the head vein has acquired a different spatial rela- 

 tionship to the intercalar (see p. 67). 



The intercalar of the Recent megalopids has been described in detail by Greenwood 

 (1970a) and shown to be more complex in Megalops than in Tarpon. However, in 

 both there is a lower intercalar limb associated with the swimbladder diverticulum. 

 Specimens of the Eocene megalopids do not show a lower intercalar limb yet it is 

 assumed that one was present and that it was developed much as in young Tarpon 

 (see Greenwood 1970a) . If the lower intercalar limb was absent in the Eocene species 

 one would be forced to recognize that an identical intercalar development had taken 

 place twice, a view with little to recommend it. A lower intercalar limb must have 

 been present in Protarpon as well as in Promegalops since the latter has a neurocranium 

 more like that of the more specialized Megalops than Tarpon. The apparent absence 

 of the lower intercalar limb in the fossil representatives is not surprising in view of 

 the fragility of the pedicel connecting the lower limb to the main body of the inter- 

 calar. 



The caudal anatomy of the Megalopidae is primitive and essentially similar to the 

 Elops-type. Loss of the urodermal is derived relative to the Elops condition, but 

 the retention of fringing fulcra is primitive. 



Fringing fulcra are rare among teleosts. Besides the Megalopidae, fringing fulcra 

 are known in Leptolepis (Leptolepididae), Lebonichthys (Albulidae), Anaethalion 

 (? Elopidae) and Davichthys (Elopidae). All members of the Megalopidae in which 

 the caudal fin is known show fringing fulcra. Fringing fulcra in teleosts are con- 

 fined to the region above the uppermost principal caudal raj', and, except in Seden- 

 horstia and older specimens of Megalops, only one or two are usually present. 



In the teleosts mentioned above, the upper principal caudal ray is preceded by a 

 number of fringing fulcra in front of which lie a number of fin-rays variously termed 



