350 AXXALS XEW YORK ACADEMY OF i<CIEXCES 



rhombic scales and heterocercal tail, has a short wide fleshy lobe in the 

 pectorals while the pelvics according to Traquair's restoration (Good- 

 rich, 1909, p. 283) have still shorter and relatively wider lobes. Thus 

 the pelvics in turn are only a little more advanced than the anal and 

 posterior dorsal. In Glyptopomus the pelvic base is narrower and the 

 axis more protruded, but the resemblance to the widely based dorsal and 

 anal fins is still obvious. But the pectoral of this genus has become 

 almost full}' archipterygial. Coming to the Dipnoi, in Dipterus the 

 process is carried further and the resemblance between the pelvic and the 

 anal is progressively effaced. In the latter Dipnoi the ^'archipterygiiun'" 

 is perfected by the protrusion of the mesopterygial axis to the extreme 

 tip of the fins; the pelvic fin also is finally as fully developed as the 

 pectoral. 



Turning to the evidence of comparative anatomy we find that in 

 Ceratodus the plexuses of nerves that supply the pectoral and pelvic fijis 

 are brought together by the coalescence of many segmental nerves (Braus, 

 ]901) and by analog}' with the conditions in sharks (cf. Goodrich, 1909, 

 pp. 72, T8), where both the paired and median fijis are formed in the 

 same manner, through the concrescence of myotomic elements, it appears 

 highly probable that in Ceratodus also the narrow-based fins arose through 

 the concentration of metameric elements, the base becoming constricted, 

 as the power of rotating the fin increased. From the cone-in-cone ar- 

 rangement of the myomeres of the Ceratodus paired fins it seems also 

 probable that the outgrowth of the mesopterygial axis to the very tip of 

 the fin was due to a sort of apical budding, or repetition of similar seg- 

 ments, on the part of the mesopterygium and its radials. That the cen- 

 tral axis of the Ceratodus fins does represent an outgrowth of the mesop- 

 terygium was held by Huxley in opposition to Gegenbaur, and was 

 supported, with strong evidence, by Howes (1887). 



From these and similar considerations I reject the traditional view 

 that the ''^archipterj-gia*' of Dipnoi are primitive structures and I regard 

 the imperfect archipterygia of the Devonian Osteolepis as more primitive 

 than the perfected archypterygia of the modern Ceratodus. Xor can I 

 accept the views of Watson, Smith Woodward and others that the paired 

 limbs of the Ehipidistia are ^'reduced archipterygia'' ; they seem rather 

 to be incompletely evolved or primitive ^'archipterygia,'' with less ex- 

 tended mesopterygia and an asymmetrical arrangement of the radials 

 (parameres), those on the postaxial border being fully developed, the 

 preaxial parameres being small or wanting. 



The pectoral and pelvic fins of existing dipnoans have this important 

 character in common ^vith the paired limbs of Tetrapoda, namely, that 



