165 



parently undifferentiated plasma, between three canals (Fig. 77, 1) ; 

 there are some green symbiotic algac to bc foimd in the plasma, 

 but by no means so many as in the ordinary amoebocytes; no 

 vacuole present. A few moments later a vacuole arises around 

 the conglomerate, while at the same time also some green sym- 

 biotic algae are enclosed. The vacuole rapidly increases 

 and protrudes more than half way into the lumen of a canal 

 (Fig. 77, 2). But it does not burst! On the contrary, the vacuole 

 withdraws into the tissue, to protrude then into another canal 

 (Fig. 77, 3). But here it does not burst either, but withdraws 

 again. Now it lasts for some time. At last it protrudes into the 

 3''d canal (Fig. 77, 4). All at once the thin vacuole wall disap- 

 pears at the outside ; the conglomerate still remains in its place, 

 it only moves a little to and fro, as if it were still kept back, 

 then it slightly moves on and again to and fro, a pull — and it 

 runs off through the canal. 



A remarkable process, that protruding into a canal, to withdraw 

 again after some time ! I have repeatedly observed it, as said 

 before. Why dit not the vacuole burst immediately? 



The only answer I can give, based on my observations, makes 

 us acquainted — if it is right ~ with a remarkable phenome- 

 non of sponge-life, which, however, is very logical, even indis- 

 pensable. Just consider : The fecal cotiglomerates must necessa- 

 7'ily be ejected exclusively into the excurrent canals of the sponge^ 

 otherwise the whole defecation-system would unavoidably fail. 

 As we saw, however, the vacuoles with the feces are not bound 

 to certain permanent places, but they are, just as the whole 

 sponge tissue is continually moving, always carried along. But how 

 does a vacuole „know" then if a canal is an incurrent one, in 

 which it may not eject its contents at all, or an excurrent one? 

 The vacuole — or better the protoplasm, to which the vacuole 

 belongs — must examine this on the spot and then it will show 

 the phenomenon which we saw it perform above: it will protude 

 into the canal. Thus the thin protoplasmic vacuole-wall will get 

 to know in some way or other the state of the canal, either, for 

 instance, by the rapidity of the current or by the pressure of 



