DEVELOPMENT OF GILLS IN AMPHIBIAN EMBRYO. 165 



The regeneration that takes place after this operation is very 

 imperfect. In four cases a single fairly normal but small gill 

 was regenerated, identified as the first gill in two cases, the 

 second in one and the third in one. In three of these there 

 were, in addition, one or two rudimentary gills. In the other 

 cases either rudimentary gills alone were formed, or else no 

 regeneration of gills took place at all. Blood circulation was 

 established on the operated side in only two individuals. The 

 circulation was defective throughout the embryo in four cases, 

 as shown by the cedema that developed. 



The imperfect restitution that takes place in these experiments 

 must be ascribed to the severity of the operation rather than 

 to the removal of anything specifically essential, for when either 

 the mesoderm or the ectoderm alone is removed there is often 

 complete restoration of function and structure. 



TRANSPLANTATION OF BRANCHIAL ECTODERM. 



The branchial ectoderm was transplanted either to the flank 

 or to the anterior part of the head in a few cases, in order to 

 test the potency of this tissue in strange surroundings. In all 

 four individuals of the first group the results were essentially 

 the same. Small nodules, which might possibly be interpreted 

 as incipient gill sprouts, developed, but after a few days they 

 became obliterated. In one case of the second group, where 

 the gill ectoderm was transplanted to cover the eye and mandib- 

 ular region, no supernumerary gills were formed and even the 

 normal first gill failed to develop. On the other hand in two 

 cases in which the graft was implanted not quite so far forward 

 (Fig. I, A) there were traces of rudimentary gills in the region 

 of the hyoid and mandibular arches respectively. One of these 

 (Fig. 6) showed two supernumerary filaments quite distinctly, 

 in which, however, the circulation failed to become established. 

 In two other cases, operated in a similar way, no supernumerary 

 filaments were observed. 



The potency of the gill ectoderm to form gills in any of these 

 abnormal positions is obviously far less marked in Amblystoma 

 than in the anuran species studied by Ekman, especially in 

 Rana fusca. 



