lengthen by posterior elongation as less reliance was 

 placed on the soft dorsal and anal fins for locomotion and 

 greater emphasis was placed on the caudal fin as it 

 gradually became forked and more heavily muscled in 

 the tapering peduncular region. Perhaps as it became a 

 more sustained rapid swimmer there was less advantage 

 to having a large defensive spiny dorsal fin, which even- 

 tually was reduced to a rudimentary structure, such as in 

 Triodon, in which it is sometimes absent altogether. 



The spiny dorsal fin of Eoplectus has one important 

 characteristic which may shed light on the phylogeny of 

 the other scleroderras, and that is the placement of the 

 second basal pterygiophore, perhaps seemingly trivial. 

 The long shaft of this pterygiophore is oriented distinctly 

 anteroventrally and articulates either between the neural 

 spines of the first and second vertebrae or between the 

 neural spine of the first vertebra and the base of the 

 skull just behind the shaft of the first pterygiophore. This 

 is a decidedly nontriacanthodid arrangement, in which 

 the shaft of the second pterygiophore is otherwise always 

 directed ventrally or posteroventrally and articulates 

 with the neural spine of the fourth vertebra. In direct 

 contrast to the triacanthodids, the short shaft of the sec- 

 ond pterygiophore of triacanthids is directed slightly to 

 distinctly anteroventrally toward the region between the 

 neural spines of the first and second vertebrae. This is 

 similar to the situation in Eoplectus, except that the 

 shaft as well as the rest of the pterygiophore is of much 

 reduced size, as could be expected by the much smaller 

 size of the spine it bears in triacanthids relative to 

 Eoplectus. A forward shift in the orientation of the shaft 

 of the second pterygiophore (and of those posterior to it) 

 occurred somewhere in the line of evolution between the 

 ancestral triacanthodids and the derived triacanthids, 

 and Eoplectus shows such a shift of the second pterygio- 

 phore (but not of those posterior to it). 



It is entirely possible that a pre-£op/ectus-like fish, 

 before the jaws and dentition had become 7>wdon-like, 

 was on the line of hollardiin triacanthodids that gave rise 

 to the triacanthids, undoubtedly through a line of evolu- 

 tion including the Eocene Protacanthodes, the most 

 generalized or triacanthodidlike of all the triacanthids. 

 Protacanthodes has a reduced spiny dorsal fin whose 

 base is much shorter than that of the long-based soft dor- 

 sal fin, an apparently shaftlike pelvis with a well- 

 developed spine and other features typical of triacan- 

 thids, while retaining from its triacanthodid ancestry a 

 deep and only slightly tapered caudal peduncle and a 

 large well-rounded caudal fin. 



A pre-Eoplectus-like fish with: 1) a hollardiin domed 

 posterodorsal region of the skull; 2) shaftlike pelvis; 

 3) generalized triacanthodid jaws with well-developed 

 ascending premaxillary processes and discrete conical 

 teeth in sockets; 4) approximately one to one ratio of dor- 

 sal and anal fin rays to their basal pterygiophores; 5) uro- 

 neurals; 6) full-scale covering; 7) only moderately 

 developed parasphenoid ventral flange; 8) 8 -f 12 ar- 

 rangement of vertebrae; etc., could be on a line ancestral 

 to the beak -jawed Eoplectus and hence to Triodon and 

 the other gymnodonts on the one hand, and on the other 



hand to the line ancestral to the Recent hoUardiins, 

 from which latter line early arose the Eocene Protacan- 

 thodes and its decendent Recent triacanthids. 



This hypothetical generalized triacanthodid ancestral 

 line of pre-Eoplectus-like configuration could well have 

 diverged into two distinct radiations: one retained dis- 

 tinct well-developed individual teeth protruding from 

 sockets in the jaws, as well as relatively well-developed 

 spiny dorsal and pelvic fins, and the 8-1-12 vertebral 

 arrangement; the other specialized the jaws and denti- 

 tion into a crushing beak, while eventually losing the 

 spiny dorsal and pelvic fins and converting to a 9 -(- 11 

 vertebral arrangement. 



The evolution of Protacanthodes from a conically 

 toothed pre-£op/ectus-like line would involve: 1) an in- 

 crease in the number of soft dorsal fin rays and of their 

 basal pterygiophores; 2) an elongation of the soft dorsal 

 fin base concomitant with a reduction in the size of the 

 second and subsequent dorsal spines and of their basal 

 pterygiophores; 3) a shortening of the spiny dorsal fin 

 base; 4) a slight elongation of the caudal peduncle; 5) a 

 reduction in the number of pelvic fin rays and of their 

 size. All of this is well within the range of possibility. 

 Somewhere along the line between the pre-Eoplectus- 

 like form, Protacanthodes, and the Recent triacanthids, 

 the enlargement of the prefrontal and its anterior exten- 

 sion alongside the ethmoid and vomer, and other such 

 typical triacanthid specializations, could have taken 

 place, while the caudal peduncle structure and caudal fin 

 shape would not have changed until between the 

 Protacanthodes and Recent triacanthid stage. 



The configuration of the first three basal pterygio- 

 phores of the spiny dorsal fin in Eoplectus is also of in- 

 terest to the phylogeny of the balistids, for it in some 

 ways represents what one might expect a balistid 

 ancestral group to have, at least in this one respect. That 

 is, there is a large open space between the second 

 (anteroventrally directed shaft) and third (postero- 

 ventrally directed shaft) pterygiophores. The third 

 pterygiophore already seems to be acting partially as a 

 strut supporting the posterodorsal end of the second 

 pterygiophore and has the ventral end of its shaft in con- 

 tact with the neural spine of the fifth vertebra, as in 

 balistids. If the first two pterygiophores in Eoplectus had 

 greatly reduced ventral shafts so that the first pterygio- 

 phore articulated high on the rear of the skull and the 

 second pterygiophore to the posterior end of the first, 

 with the latter concurrently becoming anteriorly elongate 

 to reach the rear of the skull, then a prototype of the 

 balistid structure is achieved. The spine of the third 

 pterygiophore would become lost and the sole function of 

 this pterygiophore would be to brace the developing 

 specialized carina accommodating the complex locking 

 mechanism of the first two spines, with the third spine, 

 borne on the small second pterygiophore forming the 

 rear end of the carina, becoming reduced in size. At the 

 same time that these transformations would be taking 

 place, the fourth and fifth pterygiophores and their 

 spines would also become reduced in size and eventually 

 entirely lost. Such is a reasonable scenario for the de- 



