196 J. S. NICHOLAS 



SO far of what to expect when the action of the endoderm as a molding 

 structure is absent, but when the ectoderm and the nervous system are 

 serving in place of the endoderm. The process of regional localization 

 responsible for limb formation has been able to manifest itself because 

 of the absence of endodermal structures. The endoderm has not been in 

 place for the molding of the internal layer of the splanchnopleure and 

 in consequence the manifestation of limb formation has been expressed 

 in the splanchnopleuric component of the mesoderm. Is this an inhibi- 

 tion brought about by the endoderm, or is it an induction which has 

 moved regionally across in the mesoderm from what was formerly the 

 somatopleure to the splanchnopleure? 



In Harrison's ('25) experiments in which the mediolateral axis of 

 the limb is reversed, we undoubtedly have the counterpart to this series 

 of experiments ; for Harrison, when he removed the mesoderm under- 

 neath the ectoderm comprising the limb bud, undoubtedly removed and 

 reversed the splanchnopleure, leaving the somatopleuric portion inter- 

 nally while the splanchnopleuric was on the outside. From this, perfect 

 limbs developed so that the splanchnopleuric part, already regionally 

 localized when placed in contact with reversed surroundings, cannot 

 effect its normal polarization of limb structures. Whether the somato- 

 pleuric layer of the mesoderm has influenced the splanchnopleuric or 

 not cannot be said, but the somatopleure and the splanchnopleure are, in 

 this experimental series, reversible. 



That qualities do exist within groups of inducing cells which are 

 comparable to the factors responsible for regional determination in the 

 amphibian has been definitely demonstrated by Eakin ('39) for Salmo. 

 He has divided the invaginated archenteron roof or middle gastrulae 

 of Salmo into an anterior, a middle, and a posterior portion, wrapped 

 each in a tube consisting of extraembryonic half of a late gastrula, and 

 implanted the compound graft of the yolk sac epithelium of an older 

 Salmo embryo. In each case a control graft of the extraembryonic half 

 of a late gastrula was implanted in the same stage as the compounded 

 graft. The control pieces differentiated only into epidermis. When the 

 anterior portion of the archenteron roof was implanted together with 

 indifferent ectoderm, it exhibited little or no power of induction and 

 differentiated for the most part into digestive tube. When the middle 

 portion of the roof was transplanted, it differentiated notochord, so- 

 mites, gut, and pronephric ducts, inducing the accompanying ectoderm 

 to differentiate neural structures, primarily brain-like in character with 

 auditory vesicles. The posterior portion of the roof, when grafted, also 



