460 WILSON. [Vol. VI. 



etc.) produced at the fourth spiral division of the macromeres. 

 In Hydroides these four cells now arise in a left-handed spiral, 

 as in the other forms, but they are equal in size and are formed 

 simultaneously. If the analogy holds, one of these cells {d"^) 

 should be the equivalent of the second somatoblast, M., from 

 which the mesoblast arises, the other three {a^, b'^, c^) forming 

 part of the entoblast-plate. I have not yet succeeded, how- 

 ever, in tracing the cytogeny with certainty beyond this point, 

 and the origin of the mesoblast therefore still remains undeter- 

 mined. Should my surmise prove well founded, we would have 

 a very simple explanation of the difference between the equal 

 and unequal types of cleavage in annelids — a difference which 

 has hitherto seemed to involve a difficult problem with regard 

 to the mesoblast. 



Naples, April, 1892. 



POSTSCRIPT. 



While correcting the last pages of the proofs of this paper I received Oscar 

 Hertwig's extremely interesting work entitled " Urmund und Spina bifida," 

 etc. \_Arch. f. Mic. Aftat., Bd. 93, Heft III., 1892], in which the author, after 

 a critical review of the work, especially of Roux, Driesch, and Chabry, throws 

 the immense weight of his authority on the side of the same conclusions that 

 I have adopted at p. 447 regarding the dependence of the development of 

 individual blastomeres on that of the whole embryo. In opposition to Roux, 

 Hertwig maintains that, (i) "Die Entwicklung eines Organismus ist keine 

 Mosaikarbeit," and, (2) " die Theile eines Organismus entwickeln sich in 

 Beziehung zu einander oder die Entwicklung eines Theiles ist abhangig von 

 der Entwicklung des Ganzen" (p. 480). 



It seems to me, however, if I may venture the suggestion, that Hertwig 

 underestimates the importance of early differentiation among the blastomeres 

 (J.e. the hereditary element in the development of individual parts), and that 

 the mosaic view may still, i}t a modified form, be of value. In the echinoderm 

 embryo the individual blastomeres show very little morphological differentia- 

 tion until a late period. In the annelid such differentiation exists from the 

 very beginning of the cleavage, and the mosaic appearance of the development 

 cannot be overlooked. The facts seem to accord best with the hypothesis 

 that the blastomeres are capable within certain limits of pursuing their indi- 

 vidual development, yet at the same time depend in a greater or less degree 

 on that of the whole. How far this dependence goes, and how far the various 

 blastomeres may be capable of replacing one another, is a question to be deter- 

 mined not by analogy, but by direct experiment. 



Naples, May 3, 1892. 



