416 ROSS G. HARRISON 
therefore be taken into consideration. For instance, Miss 
Byrnes (’98 b), working with Rana embryos, found that if the 
region in which the hind limb develops is destroyed, a limb 
nevertheless develops out of the tissues that move in from the 
periphery and cover the wound. Braus (’06 b) found, however, 
that in Bombinator, as well as in Rana, the hind-limb region is 
much more capable of this kind of regeneration than is the fore- 
limb region.’ In the case of the latter, as Braus (’09) shows, it is 
only the shoulder girdle and some of the muscles that are formed 
after removal of the limb bud, unless the operation is done at a 
comparatively early stage, in which case regeneration of the 
whole extremity may take place. When the limb bud is itself 
transplanted it gives rise to a small shoulder girdle with some 
of the muscles in addition to the free appendage, so that in 
such an experiment certain muscles and skeletal elements de- 
velop only at the seat of origin of the bud, certain others develop 
only out of the transplanted bud itself, while again others de- 
velop in both places. The same has been found true in many 
cases in the course of the present work upon Amblystoma. 
Under certain conditions, however, the whole limb may be re- 
generated after removal of the bud, two complete limbs being 
thus derived from the single limb rudiment. The size of the 
extirpated region of body wall and the completeness with which 
all mesoderm cells within that region are removed have been 
found to be important factors in determining whether complete 
regeneration takes place or not. 
In connection with the duplicated development of the shoulder 
girdle in its original position and at the seat of implantation, 
Braus has treated in full the question of mosaic development vs. 
equipotentiality.6 He reaches the conclusion that in the Anura 
the shoulder girdle constitutes an equipotential restitution sys- 
tem, though in its original surroundings there is evidence of 
localization of materials before visible differentiation sets in. 
The results recorded in the present paper agree as to the main 
5 In R. esculenta, according to Braus (’09, footnote, p. 194), the power of re- 
generation is greater than in Bombinator. 
6 Braus, ’09, p. 271, et seq. 
