326 ANNALS NEW YORK AlCADEMY OF SCIENCES 



Crossopter3'gii had a more primitive brain structure. In Polypterus the 

 scales and dermal plates have become like" those of the primitive Actinop- 

 terygii (Goodrich 1909, p. 292), the cosmine layer being covered exter- 

 nally by stratified ganoine layers; but for several reasons I regard this 

 condition as secondary. N'or can I accept Goodrich's view (1909, p. 300) 

 that Polypterus should be removed altogether from the Crossopterygii. 

 In any. event it is further removed from relationship with the Tetrapoda 

 than were its Devonian predecessors. 



The only crossopterygians that can claim even remote relationships 

 with the Amphibia are the Devonian Ehipidistia, especially the Osteole- 

 pidas and the nearly allied Ehizodontidse. Dr. A. S. Woodward's com- 

 parison (1898, pp. 24, 25) of the skull-roof of Rliizodopsis, representing 

 the Ehizodontidse, with that of Felosaurus, representing the Stegoce- 

 phalia, serves, however, to emphasize the remoteness of this relationship 

 and to raise the question whether the supposedly homologous elements, 

 as the frontals, parietals etc., in the two groups, may not after all be 

 analogous rather than truly homogenous. Dr. Moodie's comparison 

 (1908) of the sensory canals of the skull-roof of stegocephalians with 

 those of Amia and Polypterus offered some reinforcing testimony, which 

 would have been strengthened if Pander's superb figures showing the sen- 

 sory pits and skull elements of Osteolepis, Diplopterus, Dipterus and 

 other genera had been taken into consideration. Watson's comparisons 

 (1912) of the skull-base of Rhizodus with that of certain Carboniferous 

 stegocephalians {Pteroplax etc.) having a single median occipital condyle 

 and a continuous or non-fenestrated palate, further strengthened the case. 



A wider basis of comparison appearing desirable, I have made during 

 the last few years repeated comparison of the skull patterns of various 

 Pala9ozoic and recent fishes with each other and with those of stegoce- 

 phalians of all known orders. Original material of Osteolepis, Megalicii- 

 thys, Dipterus and of many other fossil and recent Osteichthyes lias been 

 studied, and for the skull patterns of mauy of the Palaeozoic fishes I have 

 also had recourse to the excellent figures of Pander, Huxley, Traquair, 

 Smith Woodward, Goodrich, Wellburn (Megalichthys), Jaekel (Diplop- 

 terus) and others. Similarly among the Stegocephali the American 

 Museum collections have furnished for study typical examples of the 

 principal groups ; and in making comparisons with the fishes I have also 

 had before me the figures of Credner, Fritsch, Moodie, Fraas, Embleton 

 and Atthey {Loxomma) , "Watson, Broom, Williston and others. 



Before considering in detail the skull-roof in primitive fishes and Tet- 

 rapoda, it may be appropriate to ask whether any general adaptational 

 reasons can be assigned provisionally for the several characteristic ar- 



