vii EVOLUTIONARY ORIGIN OF LIMBS 451 



Fig. 200 illustrates how close may be the general resemblance 

 between the earliest stages of development of limb and external 

 gill. The same is brought out also by Fig. 204, representing part of 

 a frog larva on which had been grafted a piece of skin from the 

 gill-producing region of another larva. The external gills of the 

 graft have gone on developing and are remarkably limb - like in 

 appearance. 1 



The External Gill Hypothesis as to the evolutionary origin of the 

 limbs fits in well with other facts which are now known. In the breed- 

 ing male Lepidosiren the hind limb regularly and the pectoral limb 

 occasionally (Agar, 1908) take on temporarily the characters of an 

 external gill both in structure and in function (Fig. 205). This 

 remarkable fact — otherwise a morphological mystery — becomes at 

 once understandable on the hypothesis outlined above, as a simple 

 "reversion" towards an ancestral condition. Fig. 205 brings out 

 clearly a further peculiarity of these gill-like limbs of the male 

 Lepidosiren, namely that the respiratory outgrowths of the limb are 



"•"—■'-TiMiiir 'i i ■- " ; -' ; - "■^--'"■- 



FlG. 205. — Lepidosiren, breeding male showing apparent 'reversion of both pectoral and 

 pelvic limbs to the branchial condition. (From a specimen in the Zoological Museum 

 of the University of Glasgow. ) 



in the case of the pectoral limb attached to its ventral side, in the 

 case of the pelvic to its dorsal side. But it has already been shown 

 that the definitively ventral side of the pectoral limb is homologous 

 with the definitively dorsal side of the pelvic limb — the difference 

 in position being due to the rotation in different directions under- 

 gone by the limb rudiments in the course of their development. 

 This reversed position of the respiratory filaments in the two sets of 

 limbs clearly then fits in exactly with the view that they are ancient 

 morphological characteristics of the limb which have reappeared in 

 the male Lepidosiren. 



The striking resemblance between the pectoral girdle and the 

 branchial arches in some of the more ancient Fishes again finds its 

 explanation in the morphological identity of the two structures. It 

 is now established that the swim-bladder of Fishes is morphologically 

 a lung, and that the lung is to be regarded as at the least an extremely 

 ancient organ in the Vertebrate phylum. This points to the prob- 

 ability that the early Vertebrates were creatures which clambered 



1 Reference should also be made to Fig. 88 (p. 157) which brings out clearly the 

 remarkably limb-like character of the Urodele "balancers." 



Budgett (1901) mentions the case of an abnormal Protopterus larva which had 

 failed to develop the pinnae upon one of its external gills. ! ' This bare shaft so much 

 resembled the pectoral limb that the larva appeared to have two pectoral limbs on one 



