FOSSIL AND RECENT 211 



(e) Possible albuloid derivatives 



A suggestion that the Notacanthiformes were derived from early Elopiformes was 

 put forward by Greenwood et al. (1966 : 355). This suggestion was based upon the 

 common possession of a leptocephalus larva and the presence of a bone-enclosed 

 ethmoid commissure. Those authors further suggested that the halosaurid ethmoid 

 commissure agreed with that of albulid fishes, stressing the presence in both of an 

 enlarged canal system in this region. The development of such a canal system is 

 dealt with in the general discussion of the elopomorph snout (p. 190). All that need 

 be said here is that the development of the albuloid-notacanth snout canal system 

 is unique, peculiar, and provides good evidence of a common origin. 



Aside from the canal system there are two other specializations shared by albuloids 

 and notacanths : the snout is slightly elongate and fragile, and the small mouth is 

 inferior ; the maxilla has a slender inturned head and, with the premaxilla, is capable 

 of lateral movement but very little vertical movement. 



Beyond this, the notacanths are very specialized in comparison with the albuloids. 

 Some of the specializations of notacanths are probably correlated with their deep-sea 

 habitat. Thus the cranial bones are exceedingly thin and several are lost (auto- 

 palatine, dermosphenotic, epiotic, basisphenoid, orbitosphenoid, supraorbital, 

 lateral ethmoid, pterosphenoid and, in the Notacanthidae, the interhyal). Other 

 elements are reduced (interhyal, vomer). 



The trend towards an anguilliform body has resulted in a very reduced caudal 

 skeleton and a freeing of the pectoral girdle (at least in Notacanthidae) from the 

 cranium. These and other specializations are dealt with at length by McDowell (in 

 press). All that is intended here is to convey an idea of the extreme divergence of 

 the notacanthiforms. 



Fossil notacanthiforms are rare (in part this may be a result of their deep-sea 

 habit) but halosaurs are known from the Campanian (Echidnocephalus) and are 

 already so specialized that their halosaurid affinities are undoubted. Thus the origin 

 of the group must have been some time previously, and the features shared with the 

 Albulidae and the Pterothrissidae suggest that a Cenomanian age would be a 

 reasonable estimate. The pterothrissids seem the best candidates for the ancestors 

 of the notacanths since this group retains maxillary teeth (found in halosaurs) and 

 the palatal dentition is not particularly specialized. The albulid dentition on the 

 other hand is already specialized by Cenomanian times. 



Cenomanian pterothrissids are rare, being questionably represented by Hajulia. 

 This, together with the apparent absence in the fossil record of earlier, less specialized 

 yet identifiable notacanthiforms, obscures the origin of the ' Heteromi '. Neverthe- 

 less, a close association with the albuloids (in particular the Pterothrissidae) is strongly 

 suspected. 



(f ) Comments on the family Phyllodontidae 



The Phyllodontidae Dartevelle & Casier (1943) have been referred to the suborder 

 Albuloidei by Estes (1969a) on the basis of similarity to Albula in the basibranchial 

 tooth plate (in particular, the basibranchial scars) and the occlusion pattern with the 

 parasphenoid. 



