teleosts as a group is supported by the presence of a double 

 myodome (stage 3). 



The compound (epi- and laterohyal) hyomandibula of 

 the teleost is like that of the palaeoniscoid or crossoptery- 

 gian and is associated with a symplectic and an interhyal 

 (or stylohyal). In its development the symplectic is the 

 distal part of the epihyal. The interhyal appears to be a 

 part of the epihyal or to have an independent origin be- 

 tween epihyal and ceratohyal. It involves two tissues in its 

 origin, a medial part, which is continuous with the hyoid 

 arch blastema, and a lateral part, which may represent the 

 bases of hyoid rays. 



The palaeoniscoid lacks the interhyal, but Aafienser ap- 

 pears to have it. Polypterus has a single element in the posi- 

 tion of an interhyal but lacks the symplectic, suggesting an 

 "either-or" situation in primitive fishes and an evolutionary 

 trend toward elements to support both lower jaw and hyoid. 

 Lepuosteus and Arma agree in having cartilaginous inter- 

 hyals, whereas the teleost represents the acme with an os- 

 sified interhyal. 



Although quite different, the palaeoniscoid and teleost 

 types are generally thought of as extremes; that is, the 

 teleost is a product of an evolutionary sequence having the 

 holostean as an intermediate stage. 



The nature of the holostean type can be considered in 

 terms of the fossil record. This type is first observed in the 

 Upper Permian genus Acenlroplwrus (Figure 5-13). The small 

 species of this genus are like those of the later Family Semi- 

 onotidae. The endocranium o[ Acentrophorus is not know but 

 the suspensorium is quite holostean. An internasal is pre- 

 sent but occurs in other holosteans as well. The premaxilla 

 was probably fused to the rostrals. This genus already 

 showed modification away from the ancestral pattern in 

 that the parietals were fused at the midline. 



Following Acentrophorus, holosteans are lacking in the fos- 

 sil record until the Middle Triassic. In the Upper Triassic 

 they appear in numbers. Their sudden blossoming in a 

 variety of forms has been interpreted as indicating that 

 several lines of palaeoniscoids, through functional changes 

 in the jaw apparatus, suddenly assumed the guise of the 

 holostean. 



In the early Triassic, the palaeoniscoids are represented 

 mainly by the subholosteans. This group is defined as hav- 

 ing the body and fins essentially holostean, while the skull 

 femains palaeoniscoid. The name itself implies that there 

 were several lines of fishes undergoing parallel modification 

 leading to the holostean stage. Of the subholosteans, two 

 families are particularly suggestive of the holostean in that 

 the nasals meet at the midline and in the general pattern 

 of the cranial roof: these are the Perleididae and the Ospi- 

 idae. The former is quite palaeoniscoid but transitional in 

 scale structure; the latter is transitional in the form of the 

 lower jaw, the freeing of the maxilla, and the presence of 

 a large dorsal branchiostegal ray in the position of an inter- 

 opercle. An enlarged branchiostegal ray is also observed in 

 the perleidids. 



More like the holostean are the species of the subholo- 

 stean Family Paraseniionotidae. These resemble the holo- 

 stean semionotids in some features but retain the massively 

 ossified endocranium of the palaeoniscoid. Contemporary 

 with the parasemionotids and included with them are forms 

 having the preopercle reduced to a canal bone and the cheek 

 covered by several irregularly shaped squamosoids. It is 

 possible that their endocrania were subdivided into the sev- 

 eral separate bones of the holostean. 



The critical feature of this subholostean-holostean transi- 

 tion was the acquisition of an interopercle — an interopercle 

 of constant form and relationships. It seems unlikely, even 

 assuming that it was a response to a mechanical need, that 

 the interopercle was arrived at by the conversion of a 

 branchiostegal ray— and in several lines of fishes independ- 

 ently. Its origin in development and its overlapping of the 

 subopercle in the adult suggest that the interopercle is. a 

 fragmentation product of the subopercle produced by the 

 mechanical changes in the jaws. These can be summed up 

 as the forward movement of the point of articulation of the 

 mandible, to a position in which the axis of the suspensorium 

 IS vertical or even directed ventroanteriorly. accompanied 

 by a jaw angle that moves laterally as well as forward when 

 the mouth is opened. The first of these has involved the 

 acquisition of the J-shaped preopercle, which necessitated 

 attenuation of the ventral part of the subopercle; whereas 

 the second, the lateral movement of this area, would be 

 aided by an oblique (posteriorly and down) joint across the 

 subopercle. 



Lepisosleus (Figures 5-14, 5-15) has acquired a J-shaped 

 preopercle with the forward shift of the suspensorium, but 

 the cheek is not flexible and no interopercle is present nor 

 is it likely that one was present in the ancestry of this type. 

 It is probable that several details were interrelated in the 

 formation of the interopercle and that only some of the lines 

 of subholosteans acquired the interopercle. It is just as pos- 

 sible that the interopercle was acquired by a single line that 

 produced Acentrophorus and the much later Triassic para- 

 semionotids. Comparison of the living "holosteans," /,<■/;/- 

 sosteus 2ind Amia (Figures 5-14, 5-15), suggests that two quite 

 distinct types of fishes are involved. Differences exist that go 

 much deeper than the obvious specializations of the jaws. 

 Differences such as the presence of the dermohyal in 

 Lepisosleus, a cheek covered by squamosoids, and a large 

 quadratojugal suggest that this fish, like the sturgeon, is a 

 palaeoniscoid, or subholostean. Amia on the other hand is a 

 typical holostean in every feature. 



The systematic position of Polypterus is an undecided ques- 

 tion. This type differs in its lack of a dermohyal, myodome. 

 posterior ceratohyal, symplectic, and fifth branchial arch. 

 The cheek region is sufficient in itself to cause doubt as to 

 the place of the brachiopterygian; the presence of a spiracle 

 suggests a very primitive type. It is quite probable that this 

 line of fishes radiated from the ancestral stem at about the 

 same time as the several known palaeoniscoid lines. To in- 

 clude it as a palaeoniscoid does violence to the definition of 



GNATHOSTOME FISHES • 115 



