372 RECORDS. 



question of preformation presents itself to us to-day. I ask you 

 to glance at the results of a few very simple experiments 

 designed to test this question. They will give apparently quite 

 contradictory results. 



Experiments on the eggs of certain animals, such as cteno- 

 phores or mollusks, seem to give an unequivocal answer to our 

 questions. If, for example, the cells of the segmenting egg of 

 the mollusk Dcntaliiivi or Patella be separated from one another, 

 at the two-cell stage or any later period, they continue to 

 develop and produce living, actively swimming structures ; but 

 these creatures are not completely formed whole embryos, but 

 monsters that in many respects resemble pieces of a single 

 embryo (Fig. I, A). It is true that the wounds usually close 

 and heal ; but these structures, nevertheless, remain monstrous 

 and defective, and if they are carefully studied it is found that 

 only when taken collectiv^ely can they be said to constitute a 

 single whole embryo. The cells are thus proved to be in some 

 measure inherently different, and to this extent the cell-mosaic is 

 shown to be a real mosaic. If we now extend our operation to 

 the undivided ^gg, a result in harmony with this is reached. 

 If certain portions of the Q.gg of Deiitalittm be artificially cut off, 

 the remaining portion, upon fertilization, regularly gives rise to a 

 defective and monstrous creature that is not a whole embryo, 

 but resembles a piece or fragment of an embryo. It is evident 

 that this experiment seems to show pretty clearly that even 

 before the Q.gg has begun to divide into cells the parts of the 

 future embryo are in some measure definitely prelocalized and 

 predetermined in its different protoplasmic regions ; and evi- 

 dently, if this be the case, we seem further to have good ground 

 for the mechanistic assumption that the undivided ^gg contains 

 some kind of structural or material configuration upon which 

 the character of the development depends. 



But let us not on this account too hastily accept a theory of 

 preformation or prelocalization. Let us first look at the results 

 of an exactly similar experiment performed on the ^gg of cer- 

 tain other species of animals, for example, Anipliioxus, a sea • 

 urchin, or a nemertine worm. Separate here the first two or 



