NATURE AND CAUSES OF DIFFERENTIATION 423 



place, each blastomere is set, as it were, for a half-development, but 

 not so firmly that a rearrangement is excluded. 



I have reached a nearly related result in the case of both Aniphi- 

 oxus and the echinoderms. In AmpJiioxus the isolated blastomere 

 usually segments like an entire ovum of diminished size. This is, 

 however, not invariable, for a certain number of such bla.stomeres 

 show a more or less marked tendency to divide as if still forming part 

 of an entire embryo. The sea-urchin Toxopnenstcs reverses this rule, 

 for the isolated blastomere of the two-cell stage usually shows a per- 

 fectly typical half-cleavage, as described by Driesch, but in rare cases 

 it may segment like an enth'e ovum of half-size (Fig. 1.83, />>)and give 

 rise to an entire blastula. We may interpret this to mean that in 

 Aniphioxus the differentiation of the cytoplasmic substance is at first 

 very shght, or readily alterable, so that the isolated blastomere, as a 

 rule, reverts at once to the condition of the entire ovum. In the sea- 

 urchin, the initial differentiations are more extensive or more firmlv 

 established, so that only exceptionally can they be altered. In the 

 snail and ctenophore we have the opposite extreme to Avipliioxiis, the 

 cytoplasmic conditions having been so firmly established that they can- 

 not be readjusted, and the development must, from the outset, proceed 

 within the limits thus set up. 



Through this conclusion we reconcile, as I believe, the theories of 

 cytoplasmic localization and mosaic development with the hypothesis 

 of cytoplasmic totipotence. Primarily the egg-cytoplasm is totipotent 

 in the sense that its various regions stand in no fixed relation with the 

 parts to which they respectively give rise, and the substance of each 

 of the blastomeres into which it splits up contains all of the materials 

 necessary to the formation of a complete body. Secondaril)-, how- 

 ever, development may assume more or less of a mosaic-like character 

 through differentiations of the cytoplasmic substance involving local 

 chemical and physical changes, deposits of metaplasmic material, 

 and doubtless many other unknown subtler processes. Both the ex- 

 tent and the rate of such differentiations seem to vary in different 

 cases ; and here probably lies the explanation of the fact that the 

 isolated blastomeres of different eggs vary so widely in their mode 

 of development. When the initial differentiation is of small extent 

 or is of such a kind as to be readily modified, cleavage is imictcrmi- 

 nate in character and may easily be remodelled (as in AmpJiioxus). 

 When they are more extensive or more rigid, cleavage assumes a 

 mosaic-like or determinate character,^ and qualitative division, in a 

 certain sense, becomes a fact. Conklin's ('99) interesting observa- 

 tions on the highly determinate cleavage of gasteropods {Crepitiit/a) 



1 The convenient terms iuJeterminatc and determinate cleavage were suggested by 

 Conklin ('98). 



