GREGORY: FISH SKULLS 111 



corresponds to the maxillary component of the superior maxillary bone of mammals." 

 This type of maxillary is considered by AUis (1919, p. 386) to be not homologous either 

 with the type of maxilla that is found in the Holostei or with the type that is characteristic 

 of most Teleostei. The addition of suborbital elements bearing latero-sensory canals to the 

 true maxilla makes the complex as a whole non-homologous with the maxilla of the Holostei 

 and Teleostei. Nevertheless, the alveolar or dentigerous part of this bone appears to be 

 fully homologous with that of other fishes. Indeed the palate of the palaeoniscid Elonichthys 

 binneyi, as described by Watson (1925, p. 854) would appear to give evidence that the 

 premaxillse, maxillae and dentigerous palatines of that fish are respectively homologous with 

 those of Polypterus on the one hand and those of Amia and the teleosts on the other. The 

 hyomandibular of Polypterus bears a small "accessory hyomandibular" on its postero- 

 superior border (Allis, 1922, p. 237, PI. X), which is regarded by Allis as having arisen 

 through the fusion of the bases of the posterior row of branchial rays of the hyal arch and 

 as homologous with the posterior head of the hyomandibular of teleosts. Edgeworth 

 (1926, p. 191), however, shows "that the hyomandibular of Teleostomi is a single structure 

 and not the result of fusion of two skeletal structures. When, as in most Teleostomi, there 

 is a foramen for the passage of the tr. hyomandibularis vii, this is due to the skeletal element 

 — primarily situated only anterior to the nerve — spreading backwards. When, as in some 

 Teleostomi, the hyomandibular has two articular heads, this is due to the gradual develop- 

 ment of the primary single articular head into two heads, and is not due to the fusion of any 

 extra cartilage." 



As to the "accessory hyomandibular" of Polypterus, Edgeworth (p. 183) states that in 

 a 75 mm. specimen of Polypterus senegalensis "the (osseus) 'accessory hyomandibular' 

 fits on the back of the head of the hyosymplecticum [hyomandibular plus symplectic] and 

 the lateral surface of its posterior process. . . . It is a covering bone for the dorsal posterior 

 part of the hyosymplecticum." The head of the hyomandibular articulates with a car- 

 tilaginous socket which apparently represents the true endocranial part of the pterotic, 

 while the dermo-pterotic has fused with the parietal. 



The primordial cranium and visceral arches of a larval Polypterus have been described 

 and figured by Budgett (1902, PI. X); but a comparison with similar stages of Heterodontus, 

 Amia and Salmo, figured by De Beer (1924a, 1927, 1928), and with the larval sturgeon 

 figured by Sewertzoff, indicates that in general the chondrocranium and branchial arches in 

 the larval stage of each foreshadow in most essentials the conditions in the respective adult, 

 except where larval specializations interfere, and that neither Polypterus nor any of the 

 others harks back to remote adult ancestors in its larval stages. Thus the lower end of its 

 hyomandibular in the larval stage figured by Budgett does not project below the level of 

 the attachment of the "stylohyoid" (interhyal) so as to give rise to a symplectic and thus 

 it throws no additional light on the question whether a symplectic is absent in Polypterus 

 or whether the attachment of the "stylohyal" has simply shifted down to the symplectic 

 end of the hyomandibular. 



PALiEONISCOIDEI 



The basic pattern of the teleost skull was already established in the palaeoniscids, which 

 ranged from the Old Red Sandstone of Europe (Lower Devonian) into the Upper Jurassic. 

 These relatively very primitive forms have long been well known through the researches of 

 Traquair and others. More recently Stensio (1921), D. M. S. Watson (1925, 1928) and 



