440 SCIENCE PROGRESS. 



cavity of the collar. In sponges in addition to the currents 

 of this kind produced by each fiagellum, there is the main 

 current flowing through the canals and produced by the united 

 action of all the flagellaactingtogether. If this general current 

 be considered alone in relation to the collar cells, it is evident 

 that as it flows past the openings of the collars, each of the 

 cup-shaped cavities which they contain forms a side eddy 

 to the general current, and particles passing near the collars 

 will tend to be carried into them and whirled about 

 inside them by the eddy in their interior, until they come 

 into contact with, and stick to, the protoplasm either of the 

 collar itself, or of the cell at the base. Hence, taking into 

 consideration both the effect of the collars in forming side 

 eddies to the general current, and the manner in which the 

 movements of the flagellum tend to sweep particles into the 

 cavity of the collar, it is impossible to agree with Polejaeff 

 that the collar cells are not suited for ingesting food ; on 

 the contrary they must be regarded as eminently well-fitted 

 for a function which repeated experiments have shown 

 them to exercise.^ 



While there is thus a close agreement between collar 

 cells and Choanoflagellata in their general structure and 

 functions, some details are still open to dispute. In the 

 first place there is the question of contractile vacuoles ; 

 those authors who especially maintain the Protozoan nature 

 of sponges, namely. Carter, Kent, and James-Clark, all 

 assert the existence of pulsating vacuoles like those of 

 Protozoa, and usually two in number, in the collar cell. 

 These statements are not. however, confirmed by other 

 authors, and though the protoplasm of the collar cells is 

 undoubtedly very vacuolated in a number of instances, it is 

 not certain that any of these vacuoles are contractile. 

 Schulze (1885, p. ]8o) found the contractile vacuoles to be 

 by no means so regular in their occurrence as usually 

 alleged. Bidder (1895) makes no mention of them in his 

 detailed account of the collar cells of Heterocoela. And the 

 figure given of them by Kent in ClatJiriiia coriacea (1881, 

 pi. X., fig. 2) is certainly wrong, since he figures two 



^Compare von Lendenfeld, Zeitschr. f. iviss. Zool, xlviii., p. 674. 



