2l6 EDMUND 13. WILSON. 



(exceptionally into a larger number), though with many varia- 

 tions. Twenty -four eggs were cut, at intervals of three to ten 

 minutes during the period between discharge and cleavage, into 

 from two to five pieces that were isolated in water containing 

 spermatozoa. Of these fragments at least one piece from each 

 egg developed in all but one case. Of the first nineteen, cov- 

 ering a period of eighty-seven minutes only one piece from each 

 egg developed (eighteen cases). These fragments divided like 

 the whole eggs, and at the same time with the latter (two and 

 one half hours), into approximately eight or sixteen blasto- 

 meres. Of the five remaining cases two or three pieces from 

 each egg developed, segmenting in such a way that the total 

 number of blastomeres formed from each egg was at least ap- 

 proximately, and probably accurately, the same as those pro- 

 duced from an entire egg. The number of the blastomeres in 

 the segmenting fragments of the respective eggs (which, owing 

 to lack of time, could not be exactly counted) were approxi- 

 mately 4 -f 12, 4 + 12, 4 + 8, 4 -f 5 + 1 6, 4 -f 12. These re- 

 sults are what would have been expected. It is evident that the 

 development of only a single piece in the earlier period is due to 

 the fact that the egg still contains only a single nucleus ; and 

 only when the nucleus has divided one or more times does the 

 possibility arise of obtaining more than one nucleated piece from 

 a single egg. The facts seem worth recording, as showing that, 

 despite the absence of a fertilization-membrane after fertilization 

 has occurred, non-nucleated pieces of the egg cannot be again 

 fertilized (which agrees with my later experiments on Ccrcbratu- 

 lns, an account of which is now in press) ; and also that the period 

 at which cleavage occurs does not depend on the number of 

 nuclei in the piece but on some other progressive change which 

 reaches the critical point at the same time whether the egg be 

 cut to pieces or remain a whole. 



Of more general interest is the fact that egg fragments not 

 more than one fourth the bulk of the egg may develop into 

 dwarf larvae, which, after swimming for the normal period (about 

 forty-eight hours), sink to the bottom, develop tentacles, and 

 produce the first pair of buds in normal fashion. Thus arise 

 dwarf colonies of various sizes that are, in every detail, minia- 

 tures of the normal colonies of the same age. Fig. I shows side 



