446 WILSON. [Vol. VI. 



spheres, different as their past history has been and indepen- 

 dent of external conditions, arrive at the period of bilateral 

 divisions practically at the same moment. When we reflect 

 upon the history of the nuclear spindles, the diverse forms and 

 mechanical relations of the cells, such an assumption makes, 

 under the theory of probabilities, a demand upon our credulity 

 which I for my part am not prepared to meet. To accept the 

 second alternative, however, is to admit that the form of the 

 individual cell-divisions is in some degree controlled by influ- 

 ences emanating from the entire cell-complex. 



That such influences exist is, however, certain in view of 

 recent researches upon embryological mechanics. Roux, in 

 one of his justly celebrated contributions to this subject (No. 

 20) showed, in the first place, that if one of the first two blasto- 

 meres of the frog-embryo be killed (by puncture with a heated 

 needle), the remaining blafetomere develops into a half-blastula, 

 followed by a half-gastrula. He succeeded also in producing 

 three-quarter embryos by killing one of the blastomeres of the 

 four-celled stage. From this he drew the following conclusion 

 {Separat-abdrtick, p. 30) : " Die Entwicklung der Froschgastrula 

 und des zunachst daraus hervorgehenden Embryo ist von der 

 zweiten Furchung an eine Mosaikarbeit und zwar aus mindes- 

 tens vier verticalen, sich selbstandig entwickelnden Stiicken." 

 In the second place, however, he showed that ultimately the 

 missing half or fourth is perfectly restored by a process of 

 "post-generation," which begins about the time of the forma- 

 tion of the medullary folds. As far, therefore, as the later 

 stages are concerned, Roux's own results contradict the mosaic 

 hypothesis; for the normal course of events in the uninjured 

 blastomere, or its products, is radically altered by the changes 

 011 the other side of the embryo. 



This significant fact is rendered still more striking by the 

 recent experiments of Driesch (No. 6) and Fiedler (No. 7) on 

 the embryos of sea-urchins, and by those of Chabry (No. 3) on 

 ascidians. Driesch shows, in his extremely interesting paper, 

 that if the two primary blastomeres of the EcJiimis embryo be 

 completely, or even partially, separated (by shaking), each blas- 

 tomere may give rise to a perfect embryo of half the normal 

 size. The cleavage of each blastomere (whether isolated or still 

 in partial contact with its fellow) at first follows the same course 



