318 



Prof. A. Dendy. 



I regard as their mother-cells or scleroblasts (silicoblasts). These scleroblasts 

 occur in groups, sometimes so closely packed together as to become polygonal 

 from mutual pressure (fig. 4), sometimes more loosely arranged (figs. 5-8). 

 Each one is about - 02 mm. in diameter and provided with a very thin cell- 

 membrane. When darkly stained (figs. 5 and 6) the whole cell appears rather 

 coarsely granular and no nucleus is visible. When more lightly stained 

 (fig. 4) it is seen that the cell is vesicular and contains a small nucleus, 

 apparently suspended in the middle by threads of cytoplasm, while the highly 

 refringent granules lie against the inner surface of the cell-membrane. 

 Except for the presence of these granules the scleroblasts agree very closely 

 with the well-known form of sponge-tissue termed by Sollas cystenchyme. 



Amongst the scleroblasts lie the colloscleres ; each is commonly enclosed 

 in an oval, thin-walled vesicle very much larger than itself (figs. 5, 6, 8), 

 though it seems probable that in life it may have completely filled the vesicle. 

 The wall of the vesicle stains very distinctly and in much the same way as 

 does the collosclere itself. 



It is by no means easy to determine the exact relations between the 

 colloscleres, their vesicles, and the scleroblasts. In teased preparations the 

 vesicles, with their contained colloscleres, sometimes appear quite separate 

 and isolated, but there is, I think, no question of the vesicle itself being the 

 wall of a mother-cell. Not infrequently one vesicle can clearly be seen 

 attached sideways to a single scleroblast, as shown in figs. 9 and 10, and in 

 such cases it is hard to resist the conviction that the vesicle, or, at any rate, 

 its contents, is in some way or other the product of the scleroblast. In such 

 cases the collosclere usually lies at the side of the vesicle remote from the 

 scleroblast, with its concave surface turned towards the latter. 



In other cases a vesicle may appear to be attached end-on to a scleroblast, 

 as shown in fig. 8, but I am not satisfied that the association in such cases is 

 not accidental, and that the scleroblasts that really belong to the vesicles in 

 q\iestion have not been removed in the course of preparation. In fig. 11 a 

 vesicle is shown which has quite evidently been torn partially away from a 

 scleroblast, but whether its own or not it is again impossible to say. 



The strongest evidence that the colloscleres really are secreted by the cells 

 which I have ventured, perhaps somewhat prematurely, to call the sclero- 

 blasts, is as follows. A large proportion of these cells are found to exhibit 

 a well-defined rounded knob, attached to the outer surface of the cell 

 membrane at one pole, as shown in fig. 3, a, and fig. 7. In fig. 7 the four 

 cells represented do not seem to have been disarranged in the course of 

 preparation, and each one shows the characteristic knob, all the knobs, 

 curiously enough, pointing in approximately the same direction. These 



