758 DR. E. B. WILSON ON THE DEVELOPMENT OF RENILLA. 
wide on account of the shrinkage of the central mass, but even in this case is some- 
times nearly filled with the granular matter. The appearance of the granular mass 
varies greatly in different specimens and in different parts of the same section. It may 
be very abundant and of a loose flocculent character in one part, while elsewhere it 
gradually disappears and is replaced by a definite membrane lying between the 
ectoderm and entoderm, which is unmistakably the supporting lamella. In favourable 
specimens the granular mass may be traced around the section, becoming more and 
more closely compacted until it passes directly into the supporting lamella. 
These facts leave no doubt that the supporting lamella is derived from the 
granular mass, which becomes compacted together to form a definite membrane. The 
granular mass probably never has naturally any considerable thickness, being com- 
pacted into the membrane as soon as it is formed. The loose flocculent character is 
probably produced by the action of the reagents which causes the material of the 
supporting lamella to swell up, while the central mass at the same time shrinks away 
from the ectoderm, forming the cavity in which the granular mass lies. 
In fig. 133, which will illustrate the appearance of a section at this period, there 
are parts of the section where neither granules nor lamella appear, other parts where 
the outer ends of the entoderm cells are covered only by their own cell-membranes 
outside of which is a small quantity of granular matter ; while in other portions a 
pretty distinct lamella is formed with abundant granular matter outside of it. The 
entoderm cells show no change at any time during the formation of the lamella. 
This indicates that the ectoderm alone is concerned in the production of the granular 
matter which forms the lamella, and this conclusion is confirmed by a study of 
the ectoderm cells. The inner ends of the cells are rounded and swollen and 
often terminate in knob-like swellings (fig. 132, b.), attached to the bodies of the 
cells by narrow necks. These swollen inner ends then separate from the bodies of 
the cells and lie in the deeper parts of the ectoderm or in the space between the 
two layers. In some specimens, of which fig. 132 is a good example, the lower 
part of the ectoderm is closely packed with these rounded bodies, of which some are 
still attached to the cells, but most are free. The substance of which these balls are 
composed is quite like the granular substance, and the balls may be seen in various 
stages of disintegration. 
Hence we may conclude that the material of the supporting lamella is derived from 
the disintegrated granular balls which have separated from the ectoderm cells. It is 
possible that the swollen inner ends do not normally separate bodily from the ectoderm 
cells and that this is accidentally done in making the sections. This is hardly probable, 
however, since the outlines of the granular masses are usually regular, and they are 
found free in large numbers. The granular bodies appear in some cases to discharge their 
contents without breaking down and losing their form. I conclude this from the 
occasional presence of clear, rounded bodies in the granular mass of the same form and 
size as the granular bodies. These clear bodies (several of which are shown in fig. 133) 
