DR. B. B. WILSON ON THE DEVELOPMENT OF REN ILL A. 
743 
homogeneous yolk. Penceus (Haeckel) undergoes a similar development, but the 
fusion of the inner ends of the spheres occurs at a far earlier stage when only four 
spheres are formed. Whether these are at first distinct was not determined. 
In Eupagurus and a number of other Decapods studied by Mayer (Jenaische 
Zeitschrift, Bd. xi., 1877) a curious condition exists which is intermediate between 
the preceding forms and Renilla. The nucleus divides regularly into two, four, and 
eight, but without a concomitant cleavage of the vitellus. After the formation of the 
eight nuclei, however, the vitellus divides into two, four, and eight complete spheres, 
each of which contains one of the nuclei. In the next stage sixteen spheres are 
formed, but their inner ends no longer extend to the middle of the egg, the spheres 
having fused to form a yolk-mass as in Penceus or Palcemon. Finally in Asellus, 
already referred to, the nuclei multiply as in Eupagurus to the number of eight, but 
the vitellus then divides at once into eight partially-formed spheres, without under¬ 
going the previous divisions into two and four. This condition is characteristic of 
about one-third of the Renilla eggs, but in most cases the division of the vitellus is 
retarded until sixteen nuclei are formed. 
In rare cases cleavage is delayed unti] thirty-two nuclei are formed, and here again 
we find that this condition, though a rare variation in Renilla, is permanent and 
normal in another group of animals, namely, the Araneina (Hub. Ludwig, Zeitscli. 
Wiss. Zool., Bd. xxvi., 1876). In this well-known case the nuclei multiply to the 
number of thirty-two before the vitellus actually divides, though a partial segregation 
of its material is effected. From this condition the step is not great to the eggs of 
the Insecta and Acarina, where a still larger number of nuclei are formed before 
cleavage begins. 
There can be no doubt that the regular division of the vitellus in geometrical 
progression into two, four, &c., spheres, is in general to be regarded as the most 
primitive mode of development, the process being only a special case of cell-division. 
In a number of polyps, both of the Ajcyonarian and Actinarian types, as in Clavularia, 
or some species of Actinia (Kowalevsky), this primitive mode of development is still 
retained. Hence, if we regard the most frequent mode of segmentation in Renilla — 
namely, direct division into sixteen spheres—as the normal mode, the occasional divi¬ 
sion into eight, four, or two may be regarded as cases of reversion to conditions which 
were once the prevalent modes of development. On the other hand, the single observed 
case of division into thirty-two spheres shows that while the sixteen-sphere cleavage 
has been pretty well established, a tendency to further abbreviation still exists. We 
cannot doubt that if any change of condition should render a further concentration of 
development advantageous, this tendency or capability would come into play, and a 
segmentation like that of the Insecta might be produced. 
As the various forms of regular cleavage may be explained as the result of variations 
in the amount of equally distributed deutoplasm (or in the activity of the protoplasm), 
so we may in part explain the various forms of unequal segmentation as due to 
5 C 2 
