66 AMPHIBIA 



CHAP. 



cences. It would be e(|ually wrong to allude to the absence of 

 lungs in many newts as a piscine and therefore ancestral feature. 

 The develo])inent of the typical pentadactyloid linilj, the con- 

 nexion of the pelvic girdle with the vertebral column, the 

 development of the lungs, and absolute suppression of internal 

 gills point without doubt to terrestrial creatures. What then, 

 may we ask, were the first Ampliibia like ? and how about 

 the external gills ? They were undoubtedly akin to the less 

 specialised Lepospondylous Stegocephali, in particular the gill-less 

 Microsauri, and the various stages may perhaps be reconstructed 

 as follows : — 



(1) Terrestrial, with two pairs of pentadactyloid limbs; 

 breathing by lungs only ; with a fully developed apparatus of 

 five pairs of gill-arches, which during the embryonic life perhaps 

 still carried internal gills ; with or without several pairs of gill- 

 clefts. Eeduction of the dermal armour and of the cutaneous 

 scutes had taken place. 



(2) Additional respiratory organs were developed by the 

 embryo, in the shape of external gills ; these were at first re- 

 stricted to embryonic life (as in the existing Apoda), but were 

 gradually used also during the aquatic life of the larva. These 

 external gills, together with the lungs, have superseded the 

 internal gills, of which there are now no traces either in Urodela 

 ov in Anura. 



(3a) Some Urodeles, retaking to aquatic life, retained and 

 further enlarged the external gills into more or less permanent 

 organs (cf. also Siren, p. 186). 



(o?>) The majority of Urodela hurried through the larval, 

 aquatic stage, and some — e.g. Salamandra atra — became abso- 

 lutely terrestrial. The possession of unusually limg external gills 

 by this species and by the Apoda indicate that these organs are 

 essentially embryonic, not larval, features. 



Regeneration. — Most Amphiliia possess the faculty of re- 

 generating mutilated or lost limbs. This takes place the more 

 certainly and quickly the younger the animal. Tlie amputation 

 necessary to study these phenomena need not be experimental. 

 Axolotls and other Urodelous larvae frequently maim each other 

 fearfully, by biting otf the gills or one or more limbs. The gills do 

 not even require amputation. If the larvae are kept in stagnant 

 water the gills often shrivel up or slough off and grow again. 



