80 CHILD. [VOL. I. 



difference in the size of the cells (i.e., the first two cells of 

 cleavage). The two cells may be equal in size when the organs 

 in question are not precociously developed. The same prin- 

 ciples suffice to explain unequal divisions throughout the 

 cytogeny." 



My work on Arenicola and Sternaspis, together with a com- 

 parison of previous work, has led me to somewhat different 

 conclusions. The oblique cleavage does not appear to be so 

 strictly or so simply a mosaic as has been supposed. 



First, corresponding cells differ greatly in size and structure 

 in different forms without any corresponding differences in 

 time of differentiation. To mention a few examples: In 

 Arenicola, where the unsegmented egg contains much yolk, 

 equally distributed, all the cells of several generations, includ- 

 ing ectomeres, first somatoblast, and mesoblast, acquire yolk 

 granules. The ectomeres are very large, and the first 

 somatoblast and mesoblast are the largest cells in the egg, 

 leaving the entomeres quite small. In Sternaspis, where the 

 egg is even more closely packed with yolk, it is all retained in 

 the entomeres, and these are enormous, leaving the ectomeres 

 and the mesoblast as small, protoplasmic cells. 



Yet in both these cases the cleavage is cell for cell the same 

 up to a late stage, and the differences in the order of cleavage 

 are not great. In general, the cells have the same fate, except 

 that Sternaspis has no prototroch. 



Comparisons along this line may easily be carried further, 

 but this is perhaps sufficient to show that the appearance of a 

 large blastomere does not necessarily imply that it is packed 

 with precociously differentiated material. 



Secondly, as the number of studies of the oblique cleav- 

 age increases, it becomes more and more evident that cell 

 homology is not a true homology. Sternaspis never possesses 

 a prototroch, but large cells corresponding to the primary 

 trochoblasts of Arenicola are formed, and every division up to 

 the differentiation of the trochoblasts in Arenicola is exactly 

 paralleled by Sternaspis. Again, Arenicola and Amphitrite 

 both possess a paratroch, but in Arenicola the cells forming it 

 arise by a different series of divisions and pass through several 



