FOSSIL AND RECENT 201 



fulcra. This is the third hypothesis mentioned above. It is interesting to note 

 that in some pholidophorids (Pholidophorus bechei and Pholidolepis dorsetensis, see 

 Patterson 1968a) there is a reduced ray which lies at the base of the fringing fulcral 

 series and which perhaps represents the basal articulation of a ray which has frag- 

 mented distally into many fringing fulcra. 



A distinctive feature of both Megalops and Tarpon is the elongate last ray of the 

 dorsal fin. This feature is of little use in indicating relationships since it has 

 developed in Dorosoma and Opisthonema (Clupeidae) and also in the albulid Dixonina. 

 In Tarpon and Megalops the growth of this terminal filament shows positive allo- 

 metry. The independent development of the terminal filament in the three groups 

 of fishes mentioned may have functional significance. The dorsal fin-rays of Mega- 

 lops and Tarpon are relatively immobile, with a weak basal musculature. Most of 

 the rays are oval in cross-section but in the last ray the filament is ' T '-shaped 

 with the stem of the ' T ' directed anteriorly. The wings of the ' T ' are broad, 

 particularly at the base of the ray, and extend laterally from the plane of the fin. 

 When erect, this filament will produce turbulence resulting in drag opposing the 

 forward momentum of the fish. I suggest that the flexible filament is forced to a 

 nearly horizontal position when the fish is swimming rapidly, with the water pressure 

 acting against the erector muscles. In this horizontal position the filament would 

 produce minimum drag. As the speed of the fish decreases the filament will become 

 more vertical, producing a drag component which is located behind the centre of 

 gravity. In this position the drag would be useful in aiding the fish to decelerate 

 before the flexure of the body and caudal fin allow the animal to turn. 



Lineages within the Megalopidae. Sedenhorstia stands apart from other mega- 

 lopids in both primitive and advanced characters. The primitive characters include 

 the small scales, simple dorsal fin without an elongated terminal filament, and 

 shallow lower jaw without a prominent coronoid process. Specialized features 

 include the dentition, which is represented by a single row of teeth in the upper and 

 lower jaws, the ossifications within the dorsal ligament, and fusion of the neural 

 arches associated with the first preural and first ural centra. Such specializations, 

 although few and minor, divorce Sedenhorstia from the ancestry of Eocene mega- 

 lopids. Furthermore, Sedenhorstia occurs too late (Cenomanian-Campanian) to be 

 ancestral to a more ' typical ' megalopid such as Elopoides. 



The Albian Elopoides exhibits no specializations that would preclude it from con- 

 sideration as an ancestor to the later megalopids. The depth of the cranium and 

 the steep profile of the skull roof suggest that the post-temporal fossae were already 

 large in this form, and the megalopid nature of the dentary, maxilla and scales is 

 evident. Mention was made above (p. 85) of the apparent resemblance of Elopoides 

 to Megalops rather than Tarpon, particularly with respect to the depth of the cranium 

 and the large orbit. These features may be interrelated as a function of the absolute 

 size of the fish. A small fish may be expected to have a relatively large eye, and 

 hence a greater bulk of eye musculature accommodated within a deepened myodome. 

 This, in turn, would effectively reduce the depth of the branchial chamber. To 

 compensate, the suspensorium became deeper. Thus the apparent resemblance to 

 Megalops may not have phyletic significance in this particular case. In other 



