324, E. A. MINCHIN. 
radiates in sections. On many of the gastral rays I found, 
in addition to the two formative cells one or two conspicuous 
rounded cells (figs. 64—67, ew. c.), packed full of coarse 
granules which take the carmine-stain more or less distinctly. 
In young gastral rays these cells were absent, and they were 
frequently absent also in full-sized rays; when present there 
were usually two, but in sections one of them may, of course, 
be cut away. ‘The most striking peculiarity of these cells 1s 
that their coarse granules are always arranged in a layer at 
the surface of the cell, and in many cases appear to be in the 
act of being cast off from the cell (fig. 64—67, 69). The 
appearances suggest strongly, in short, that the coarse 
granules represent some kind of excretory me ieretal which 
the cells are producing and throwing off. 
After finding these cells on the gastral rays I searched for 
them in other places, and soon discovered that they were 
fairly common among the collar-cells, from which they could 
easily be distinguished by the characters of their protoplasm 
(figs. 43, 58, 57, ew. c.). Lalso found them frequently in the 
oscular rim, at or close to the uppermost limit of the gastral 
collared epithelium (figs. 56, 57). Their presence in the 
oscular rim shows that they must be derived from the in- 
turned layer of dermal epithelium in this region, and that 
consequently they are cells of the same class as the poro-. 
cytes, which they resemble in their characters. I take this 
opportunity of remarking that similar cells occur abun- 
dantly between the collar-cells of Ascandra falcata, pro- 
jecting into the gastral cavity between the collar-cells, and 
that in Clathrina coriacea similar cells form the endo- 
gastral network described by me formerly (1899, p. 128, 
figs. 42c and 46; 1900, p. 48). 
As regards the existence of the excretory cells, as I believe 
them to be, upon the gastral rays, I was unable to decide to 
my own satisfaction whether they are formed where they 
are found, from the actinoblasts themselves, or whether they 
migrate on to the rays from their position between the 
collar-cells. If I am correct in regarding both the excretory 
