90 



9. The oildroplets appear to be more numerous in the choano- 

 cytes, the collar-cells lining the flagellated chambers, than in the 

 amoebocytes. See Table 14. Perhaps this might make one suggest, 

 that all oildroplets in the sponge have been taken from the sur- 

 rounding water by means of the choanocytes, the food-capturers 

 par excellence. This suggestion was the more justified, because 

 my sponges were mostly gathered in a by-channel of the lake, 

 along which were numbers of houses which emptied their refuse 

 from the drain-pipes in that canal ; even so, that the sponges were 

 often in the middle of the dirt. This supposition, however, proved 

 untenable; for the sponges gathered in the lake itself, in very 

 clean water — far from houses — , contained about as many oil- 

 droplets in their choanocytes as those which had grown in the 

 dirty canal water; while this also appeared to count for sponges 

 which had been cultivated for some time in flowing water from 

 the conduit (while no reduction occurred). See Table 14. 



10. As we saw above under 7, that the number of oil-droplets 

 in the amoebocytes was in some relation with the number of 

 colourless algae, we might also expect here for the choanocytes 

 some relation between the number of their oil-droplets and that 

 of their colourless algae. That proves however not to be the case, 

 as follows from this table: 



choanocytic layer 



11. I have already mentioned on p. 25, that I have never 

 been able to show any carbohydrate within the symbiotic algae 

 — except the cell-wall, which of course will consist of cellu- 



