GREGORY, PRESENT STATUS OF ORIGIN OF TETRAPODA 365 



poda {cf. Jaekel, 1909, Figs. 2-18), the great difference being that in the 

 tetrapod the tarsals have become relatively shortened and wider while the 

 metatarsals and digits, the tibia and femnr have lengthened. 



The stont pelvic bones appear to represent the ischio-pnbis. No 

 known representative of the ilium is present in the rhizodonts ; bnt even 

 in such relatively advanced tetrapods as Fry ops, although the ilium is 

 ^.ery large, it has not yet gained a secure contact with the backbone, and 

 we must either suppose that the dorsal growth of the pelvic cartilage 

 followed very rapidly upon the change in function from fins into support- 

 ing limbs ; or possibly that the anterior process of the "ischio-pubis^' of 

 Eusthenopteron became rotated upward and gave rise to the ilium, while 

 the posterior expanded portion broadened out into the true ischiopubic 

 mass. 



STEGOCEPHALI 



Between the oldest known Amphibia of the Coal Measures and all fishes, 

 there remain profound structural differences which are as yet unbridged 

 by palaeontological discovery. Even the branchiosaurs and their still 

 more degenerate modern successors the urodeles have the dermal shoul- 

 der-girdle reduced and the epiphyses of their limb-bones cartilaginous, 

 which may indicate that instead of being primarily aquatic animals, true 

 links between fishes and terrestrial quadrupeds, they are secondarily 

 aquatic (see also B. G. Smith, 1912, pp. 547-551). And yet in the larval 

 state the branchiosaurs undoubtedly retained true piscine branchial 

 arches, while the branchial skeleton of the Permocarboniferous "urodele'^ 

 LysoropJius, as figured by Williston (1908), can leave no doubt of an 

 ultimate piscine origin. 



So wide are the differences between the various groups of Palaeozoic 

 Amphibia that one is led at first to inquire whether they may not have 

 come off from different, but allied, groups of fishes, so that for example 

 the Steresopondyli and the Temnospondyli might be related to the Osteo- 

 lepidae and Ehizodontidse, while the Branchiosauria, Microsauria, Aisto- 

 poda and Urodela might conceivably run back to some form like Tarras- 

 sius (see pp. 357-358 above). Yet in spite of these wide differences 

 among the Palaeozoic Amphibia they all agree in having one fundamen- 

 tally identical skull pattern and cheiropterygial limbs, so that present 

 evidence suggests that the transformation of fishes into* amphibians oc- 

 curred but once. 



This transformation involved first of all the abandonment of the tail 

 as the principal propeller, the loss of its dermal rays, the abortion of the 

 hypural bOnes and the assumption of the gephyrocercal form. Such a 



