October 6, 1916] 



SCIENCE 



505 



The data of the table indicate that as the 

 percentage of agar in the gelatine is increased 

 the mixture swells more in distilled water and 

 less in acid or alkali, thus approaching the 

 behavior of pure agar. Concerning the rela- 

 tive effects of acid and alkali, assured con- 

 clusions are not now possible but the data sug- 

 gest that acid tends to increase imbibition at 

 the ends of the series, that is as pure agar and 

 pure gelatine are approached, while alkali 

 tends to increase it in the middle mixtures 

 containing the two colloids in more nearly- 

 equal proportions. D. T. MaoDougal 



Desekt Laboratory, 

 Tucson, Arizona 



the theory of autonomous folding in 

 embryogenesis 1 



The experiments of Roux, 2 carried out on 

 the embryonic chick, prove conclusively that 

 the folding of a neural plate into a neural 

 tube is not dependent, as His 3 had supposed, 

 on the mechanical effect of one tissue upon 

 another, but is autonomous. Self-differentia- 

 tion in this instance is identical with self- 

 folding. The question therefore arises : How 

 can the neural plate fold itself? 



Our reply must necessarily bear on all cases 

 of autonomous folding, and reciprocally any 

 one of them might serve as the basis for this 

 analysis. The nervous system, however, is by 

 far the largest, most easily studied, and, in 

 addition, the most familiar of all the embryonic 

 tissues in which self-folding occurs. More- 

 over, in its simpler forms, it indicates so 

 clearly the direction in which an explanation 

 of its autonomous transformations is to be 

 sought, that for the present it seems best to 

 limit the discussion to what may be justified 

 as a type case. 



1 Bead at the joint meeting of the American 

 Society of Zoologists and Section E of the Ameri- 

 can Association for the Advancement of Science, 

 in Columbus, December, 1915. 



2 ' ' Die Entwicklungsmechanik, ' ' W. Engelmann, 

 Heft 1, Leipzig, 1905. 



8 ' ' TJnsere Korperf orm, und das Physiologische 

 Problem Ihrer Entstehung, " E. C. W. Vogel, Leip- 

 zig, 1874. 



For our immediate purposes, the neural plate 

 of Cryptobranchus allegheniensis is especially 

 suitable. Not only is it unusually large, as 

 neural plates go, but wherever cell-boundaries 

 are distinct, it is, without question, unicellular 

 in thickness. The first problem to be solved 

 is the role of cell-multiplication. 



In a neural plate in which the cells are irreg- 

 ular in position and dovetailed into one an- 

 other as they are in crowded columnar 

 epithelia, inequalities in the rate of division 

 and protoplasmic synthesis at or near the two 

 surfaces might lead to folding, but in the 

 Cryptobranchus embryo, in which the plate is 

 partly syncytial and in which the visible cell- 

 walls are continuous from one surface to the 

 other, and remain so during the entire period 

 of folding, it is difficult to conceive how cell- 

 multiplication could result in anything ex- 

 cept uniform enlargement. The exclusion of 

 this factor from participation in the process of 

 involution, however, does not depend on mere 

 argumentation, for comparison of the number 

 of nuclei in comparable regions of the flat, 

 half -folded, and completely folded plate, shows 

 that the number of cells per section actually 

 does not increase 4 (Table I.). Indeed in less 



TABLE I 



Number of Nuclei in Comparable Sections 



simple material, such as the neural plate of the 

 mammal, in which the number of cells does 

 increase during folding, the restriction of the 

 mitoses to the concave surface must, if effective 

 at all, exert a force opposed to the forces that 

 bring about the curvature. In this instance, 



4 Eor the validity of these comparisons see 

 Glaser, Anatomical Record, Vol. 8, pp. 528-530. 



