200 ELOPIFORM FISHES 



epaxial fin-rays (Gardiner 1970), raylets (Hollister 1936; this term applied to those 

 rays which are not articulated), accessory rays (Goody 1969b), procurrent rays 

 (Patterson 1968b) or basal fulcra (Patterson 1968a). Irrespective of terminology, 

 these rays, which are supported by epurals or neural spines, appear to represent 

 fringing fulcra that have migrated downwards (Gardiner 1970). In support of this 

 interpretation is the fact that, in most lineages, fringing fulcra occur prior to epaxial 

 rays, the exception being the pholidopleurid Australosomus. 



It is hypothesized that in the primitive actinopterygian tail there were no fin-rays 

 above the axis of the notochord. All rays now found in this position are derived 

 from fringing fulcra that have moved down and forwards. The downward move- 

 ment of fringing fulcra has occurred in the lineages leading to the pholidopleurids, 

 pycnodonts, perleidids, pachycormids, amioids and teleosts, and in all instances it is 

 associated with tail shortening. Having moved down, the fringing fulcra become 

 basal fulcra (epaxial fin-rays, etc....), supported by endochondral elements. 

 Subsequent elongation of the basal fulcra produces the need for articulation. Mega- 

 lops and Tarpon are unusual among teleosts retaining fringing fulcra in having 

 articulated basal fulcra, although a great many teleosts without fringing fulcra have 

 articulated basal fulcra. 



A more vexatious problem concerns the origin of the fringing fulcra themselves. 

 Reasonable suggestions as to their origin fall into three categories : firstly, derivation 

 from the median scale row ; secondly, derivation from the repeated unilateral 

 dichotomy of the leading fin-ray ; thirdly, breaking up of the articulated leading ray. 

 For the first hypothesis there is little evidence. A median scale row would form an 

 equally effective cutwater as a fulcral element formed of two lateral halves, and 

 there is no reason to believe that scales broke into lateral halves. 



The second hypothesis, that fringing fulcra arose by unilateral dichotomy of the 

 leading fin-ray, was suggested by Gardiner (1970). Such a development followed 

 loss of the cutwater previously provided by the median scale row and the scaled 

 body lobe of the tail. That fringing fulcra are not found in early actinopterygians 

 with unbranched fin-rays supports this view. Furthermore, the first occurrence of 

 caudal fringing fulcra is concurrent with the first occurrence of caudal ray branching 

 (Stegotrachelus finlayi, upper Middle Devonian of Scotland) . However, if fringing 

 fulcra were formed by branching of the leading ray, numerical correspondence would 

 be expected between the fulcra and the segments of the supporting ray. Ideally, 

 one fulcrum would be associated with one articulation. This is certainly not the 

 case in the Elonichthyidae (Gardiner 1970), and in a range of genera examined there 

 was never any constant numerical relationship between the fulcra and the articula- 

 tions of the supporting ray. One final point is that the number of fulcra would be 

 fixed at an early ontogenetic stage, when, although not necessarily ossified, the rays 

 exhibit the adult branched condition (observations on young Tarpon). 



In Megalops (and to a lesser extent in Tarpon) the number of caudal fringing fulcra 

 increases throughout the life of the individual (Text-fig. 34), long after the caudal 

 rays have ossified. Furthermore, the ray immediately anterior to the uppermost 

 principal in Megalops shows articulations which lie at an increasingly oblique angle 

 distally ; it is possible that the posterior articulations become separated as fringing 



