492 APPENDIX 



chorda-mesoderm, the ectoderm is induced to form brain with eyes, 

 nasal pits, and ear-vesicles. If placed on the trunk chorda-meso- 

 derm, it produces a normal spinal cord, which becomes displaced 

 below the surface. And if the piece of ectoderm is placed on the 

 protruded tail-region, the chorda-mesoderm grows into the 

 ectoderm, induces a neural tube, and then the two tissues in co- 

 operation grow out as a typical tail. The regional differentiation of 

 the chorda-mesoderm into head-organiser, trunk-organiser, and 

 tail-organiser, is here clearly seen (see p. 147). 



An interesting point concerns the behaviour of any endoderm 

 which happens to be overlain by such a piece of ectoderm. Instead 

 of the epithelium being polarised with the free or distal ends of its 

 cells facing the outer medium, as over the rest of the surface of the 

 exogastrula, its polarity is directed internally, as in a normal 

 embryo ; and in a number of cases a miniature gut-lumen is pro- 

 duced. Thus, in normal development, the topographical arrange- 

 ment of the germ-layers determines in the gut the normal polarity 

 of its epithelium, and the formation of its lumen, although as noted 

 above (p. 483) the determination appears to be purely mechanical 

 in its nature. 



But perhaps the most important of the facts revealed by these 

 exogastrulation experiments concern the absence of neural differen- 

 tiation in the ectoderm. This is all the more striking in view of the 

 indications which previous work has given (see pp. 50, 136, 203) of 

 the existence of a labile determination of neural folds, as evidenced 

 by experiments of removal or inactivation of the organiser-region, 

 and of explantation and interplantation of portions of blastulae (see 

 figs. 18, 62, 63). 



It might be held, and is held by Holtfreter {loc. cit.), that these 

 new results show that an invaginated organiser is indispensable for 

 the determination of neural folds ; that these results dispose of the 

 hypothesis of a labile determination (and therefore of a *' double 

 assurance") of the neural folds; and that the conclusions drawn 

 from previous experiments are erroneous. While realising the 

 strength of this argument, it is as well to consider the possibilities 

 that the non-appearance of neural differentiations in the exo- 

 gastrulation experiments may be due to other causes. In this 

 connexion, four points may be called to mind: 



