On the Occurrence of Gelatinous Spicules. 



319 



knobs offer a striking contrast to the cells themselves in their hyaline, 

 non-granular character. They stain fairly darkly with paracarmine and 

 appear to be entirely devoid of structure. They are attached to the sclero- 

 blasts by broad bases and vary considerably in size, progressive stages of 

 growth being represented in figs. 12-16. They evidently represent an 

 extracellular secretion of the scleroblasts, possibly derived from the minute 

 refractive granules which line the inner surface of the cell-membrane. The 

 largest knobs are of just about the same size as the contracted colloscleres, 

 but, unlike the latter, they do not lie in hollow vesicles and do not swell up 

 on addition of water. This fact seems, at first sight, to negative the view 

 that the knobs are really the colloscleres in process of secretion, but I think 

 the apparent discrepancy may possibly be explained as follows. 



We have already seen that the colloscleres always exhibit an indentation 

 or notch on one side, which we may conveniently term the hilum. This may 

 be taken to represent the original attachment of the collosclere to the 

 scleroblast, an attachment which is strongly suggested by the appearance 

 represented in fig. 9. We have also seen that, on the addition of water, the 

 contracted collosclere swells up chiefly on the convex side, the concave 

 surface, which is supposed to lie next to the scleroblast^ being apparently 

 formed of much denser silica. We are therefore, I think, justified in 

 assuming, at any rate provisionally, that the secretion, when first discharged 

 from the scleroblast, is in a concentrated condition, and that it only acquires 

 the property of absorbing water and swelling up after the lapse of a longer 

 or shorter interval. It seems not unlikely that the swelling up may be 

 coincident with its complete separation from the scleroblast. 



Probably, in life, as I have already suggested, the swollen collosclere 

 completely fills the vesicle in which it lies, and the wall of the vesicle may 

 be regarded either as a concentration of the mesoglcea due to the pressure of 

 the collosclere, or as a precipitation membrane formed at the surface of contact 

 between the gelatinous collosclere on the one hand and the gelatinous 

 mesogloea on the other, the former consisting of colloidal silica and the latter 

 presumably of an albuminoid character.* 



It appears, then, that the colloscleres are gelatinous spicules of colloidal 

 silica, formed by special scleroblasts or mother-cells, but as extracellular and 

 not intracellular secretions. They are undoubtedly a normal constituent of 

 the sponges in which they have been found, and, as I have already pointed out, 

 they occur in two perfectly distinct, though related species. In the Indian 

 Ocean species they are much smaller than in the Australian, and are 

 associated with smaller scleroblasts. I have not traced their development 

 * Probably both contain small quantities of mineral salts. 



