FOSSIL AND RECENT 197 



on the other, quoting similarities in primitive characters (gular plate, numerous 

 branchiostegals, large maxillaries, etc.) to support his argument. Others have 

 adopted the families Elopidae and Megalopidae (e.g. Jordan 1923 ; Berg 1940 ; 

 Danil'chenko 1964 ; Romer 1966). It is perhaps significant that these latter authors 

 considered both fossil and Recent forms. Fishes grouped here as megalopids have 

 a long history, extending back at least to the Albian. Furthermore, megalopids 

 show several specializations of long standing that would not be appreciated by 

 including the megalopids with the elopids. Therefore, in recognition of the long 

 separate history and the fact that the megalopids show several specializations not 

 seen in the Elopidae, the megalopids are accorded separate familial status. 



The Elopidae, the most primitive of the Elopiformes, are represented by Elops, 

 Davichthys and at least one species of Anaethalion (A. vidali). The primitive nature 

 of Elops has been stressed by many authors and both Saint-Seine (1949) and Nybelin 

 (1956) have suggested that Elops is a halecostome. This view has little to support it. 

 Firstly, Elops shows basic teleostean features such as perichordally ossified centra, 

 fin-rays of the lower caudal lobe that are supported by two hypurals which articulate 

 with a single centrum, modification of the ural neural arches to form uroneurals 

 (these caudal features have been used by Patterson 1967a, 1968a to define the 

 Teleostei), the development of a complete lateral commissure, and narial openings 

 which are situated close together and distinct from the nasal bone. Secondly, to 

 include Elops in the Halecostomi would result in all elopiforms, notacanthiforms and 

 anguilliforms also being considered as halecostomes and the term teleost would, in 

 consequence, have little meaning. 



The Elopidae are recognized as a group distinct from other elopiforms by the 

 absence of specializations. This is clearly an unsatisfactory situation but one that 

 is very difficult to rectify. Trends within the fishes grouped here are few and minor : 

 the snout becomes slightly longer, with a resulting increase in the length of the 

 dermethmoid ; the quadrate/mandibular articulation shifts slightly posteriorly 

 with a corresponding slight increase in mandibular length ; the first infraorbital 

 develops an ascending limb which reaches the supraorbital and thus excludes the 

 antorbital from the orbital margin ; the preopercular sensory canal moves nearer 

 to the anterior margin of the bone and there is a reduction in branching of the infra- 

 orbital and preopercular sensory canals. A reduction in the branching of the 

 sensory canals is seen in all elopiforms and the upward growth of the first infra- 

 orbital is seen in megalopids. 



Elops itself cannot be ancestral to any other elopiform since it lacks fringing fulcra 

 (present in megalopids and early albulids) and the ventro-lateral projections asso- 

 ciated with the dermethmoid (present in megalopids). Davichthys, from the 

 Cenomanian and Upper Santonian, retains these primitive features, but occurs too 

 late in time to be considered ancestral to either the megalopids or the albuloids, since 

 both the latter groups must have originated prior to the Albian. Interest thus 

 centres on Anaethalion vidali, an Upper Jurassic elopid, which does show certain 

 features normally associated with the megalopids (see p. 42). 



The Megalopidae contains six (possibly seven) genera which show distinct speciali- 

 zations. The differences between the megalopids and their closest relatives, the 



