DR. E. B. WILSON ON THE DEVELOPMENT OF RENILLA. 741 
however (fig. 97), the ectoderm becomes pretty well defined as a single layer of large 
cuboidal cells. The central mass is composed of large rounded polygonal entoderm 
cells, which differ little in structure from those of the ectoderm. 
§ 3. General considerations and comparison with other forms. 
With the formation of the germ-layers the segmentation may be regarded as 
finished, and it may be useful to review the facts in comparison with other forms, in 
order to appreciate their significance. 
Examples of the continued division of the segmentation nucleus before cleavage of 
the vitellus are very common, but in most cases the nuclei become far more numerous 
before cleavage occurs than in the ovum of Renilla. In the case of the Isopod Asellus 
(Van BenepeEn) the segmentation is entirely similar to one of the forms observed in 
Renilla, the nuclei multiplying to the number of eight, and the vitellus then dividing 
at once into eight spheres. In view of the total dissimilarity of the adult forms, this 
identity in segmentation is a striking instance of the independence of the yolk- 
cleavage from the adult structure; and it would be clear, from this case alone, that 
the particular form of the segmentation may be wholly determined by secondary or 
adaptive causes. 
This fact is rendered especially conspicuous from the astonishing amount of variation 
shown in the Renilla segmentation. This variation concerns not only small details, 
but also features which are usually held to be characteristic of quite different types of 
development. Hence we can see how readily the form of segmentation might be 
acted on by natural selection, for advantageous variations would certainly tend to be 
preserved and harmful ones destroyed. It must, however, be admitted that the 
action of heredity appears to have little precision in this case, for the most unlike 
variations appear in the eggs of the same parent, and I have not observed that any 
particular variation occurs more frequently in the eggs of particular individuals. 
We may now inquire, What is the direct cause of the variations in the 
yolk-cleavage? As we have seen, the nuclei divide, so far as can be determined, 
in the ordinary course, and sooner or later the vitellus follows. It is highly 
probable that the division of the nuclei is in all cases nearly regular, and the 
variations of the yolk-cleavage depend upon the varying activity of the vitellus, either 
as a whole or in its various parts. There seems to be always a tendency to the 
cleavage of the vitellus simultaneously with the division of the nuclei, but this 
tendency varies in force or meets with varying resistance. As described above, the 
vitellus seems sometimes to make abortive attempts to divide simritaneously with 
the nuclei, these efforts being expressed in temporary changes of form in the vitellus, 
but not resulting in complete cleavages. In other cases the attempt is partially 
successful, as where the vitellus divides incompletely into eight (fig. 45), Sooner or 
later the tendency gathers energy enough to carry out a complete segmentation. The 
MDCCCLXXXII, 5 Cc 
