51 



lu his mouograph of the Victorian sponges, Dendy found again 

 that the flagella were »retracted". On the other hand he found 

 »rotl-like processes" on Sollas's membrane. It seems to us not im- 

 probable that those processes represent portions of shrivelled 

 flagella. 



A uew hypothesis is brought forward by Bidder. According to 

 hini the union of the collars , i. e, the existence of Sollas's mem- 

 brane depends on the state of saturation. When the cell is satiated 

 the flagellum degenerates and the eollar unites with the neigh- 

 bouring collars. When the food bas been digested the flagella 

 become again active and the collars are separated. We cannot 

 dweil long at present on Bidder's hypothesis ^), but we can only 

 say that we never saw anything of the kind in the sponges we 

 studied. Preparations of sponges which were fed with carmine or 

 with milk daring several hours, always showed separated collars 

 and long distinct flagella, quite indepently of their having taken 

 in much or little food. They were so before being fed and after 

 having been for some time back in a current of water. With 

 regard to these experiments we may say already that if the 

 sponge was studied after having been fed for a short time (one hour) 

 with carmine , particles of this were seen not only in the collum of 

 the choanocytes, but especially frequently within the collars and 

 the lumen of the flagellated chamber itself. If the sponge had 

 been for a longer time in water with carmine, particles of this 

 were not or very rarely found in the charabers and collars, but 

 chiefly in the cell-body. After a still longer time it also appeared 

 in the cells of the parenchyme, This observation strongly con- 

 tradicts Bidder and Dendy. According to these investigators Sollas's 

 membrane should hinder food-particles to enter the lumen of the 

 chamber. And still they found an argument for the existence of 

 Sollas's membrane in a supposed advantage of filtering the water 

 and retaining food. 



]) We hope to do thi3 within short in another paper. 



