20 Ar". 1. — X. Annandale and T. Kawamura: 



passing from the one to the other, on the same pillar. Phase I 

 may, therefore, be no more than a young form of Phase II, but in 

 places where a strong current passes the point of attachment it is 

 probable that the thinner phase persists. 



A remarkable variation in colour, correlated to some extent 

 at any rate with environment, occurs in these two phases. On the 

 stone pillars of Seta railway-bridge, which is largely built of iron, 

 the sponge has invariably a yellowish tinge, while on the wooden 

 pillars of a landing-stage only a few hundred yards away it is dark 

 grey or almost black. The colour in the latter case is due to 

 minute blackish particles in the cells of the parenchyma, but 

 their nature has not been investigated. On the wooden piers of 

 Katata landing-stage, on the other hand, the sponge varies from 

 almost white to bright green in colour and is sometimes greyish 

 and sometimes brownish. The green colour is due to symbiotic 

 algae and may be brought about by infection from S. lacustris or S. 

 semispongilla. It is noteworthy that the algae are confined in the 

 more massive specimen of 8. dementis to the superficial parts of the 

 sponge, whereas in thin encrusting examples they permeate the 

 whole organism except the dermal layer. 



The correlation between external form and environment is 

 much more definite in Phase III. of S. dementis. Here the small 

 but massive growths are produced only on bodies too small to ad- 

 mit of a spreading out of the sponge, and for the most part capable 

 of movement in comparatively deep water. In most cases these 

 bodies are the shells of living molluscs, but sponges of the same 

 type are also found on small stones. We will discuss the precise 

 nature of the relation between the sponge and the molluscs later. 

 Here we may point out again (see page 10) that the strong devel- 

 opment of the exhalant system and the large calibre of the oscula 

 and the main exhalant canals is correlated with the comparatively 

 large size of the particles of silt or gravel by which the sponge is 

 liable to be overwhelmed, and is exactly comparable to the same 

 phenomenon in the Palestine species Cortispongilla barroisi 

 (Topsent). ° This species lives in a lake (the Lake of Tiberias) the 



1) Journ. As. Soc. Bengal, (n.s.) IX, p. 67, PI. II, Fig. 1, PI. IV, Fig. 4 (1913j- 



