THE POSITION OF SPONGES, ETC. 455 



gemmules " represent simply portions of the collar cell layer 

 set free. Few, I think, will dispute the correctness of this 

 interpretation, and it is easy to confirm Schulze's statement 

 that such appearances can be readily obtained by teazing 

 up living sponges, when also true larvae may be set free 

 with plates of collared epithelium sticking to the posterior 

 pole, as figured by Kent (pi. ix., fig. 26). Schulze further 

 interprets the hyaline and refractile margin figured by Barrois 

 in the ciliated cells of Oscarella lobularis, not as a collar, but 

 as the solid non-granulated terminal portion of the cell. 



Shortly after Schulze's very thorough refutation of 

 Kent's statements, Heider (1886), in his studies upon the 

 development of Oscarella lobiilaris resuscitated the collars 

 already described in the ciliated cells of the larva of this 

 form. He found that each fiagellum arose from a cup- 

 shaped depression of the hyaline margin of the cell, and 

 regarded this as the rudiment of a collar. The appearances 

 figured by Barrois as the " collier des cellules " were ex- 

 plained as the rudimentary collar together with the hyaline 

 cell margin. But he considered the occurrence of collar 

 cells in sponge blastulae as of no phyletic significance. 



Heider's description of the structure of the ciliated cells 

 of the Oscarella is very clear and convincing, but at the 

 same time it is an open question whether such a shallow 

 depression can really be regarded as a true collar. In any 

 case I am unable to agree with him that the occurrence 

 of collar cells in the blastula would be of no phyletic 

 significance. At the same time I find myself unable to 

 endorse Schulze's opinion, that the theory of descent from 

 the Choanoflagellata should have as a necessary conse- 

 quence the occurrence of choanocytes in the larva. For 

 though the blastula no doubt represents a protozoan colony 

 in phylogeny, it only represents it morphologically, and not 

 physiologically ; that is to say, the cells are not functionally 

 active, except as regards locomotion, since they do not take 

 in food, or at least have never been observed to do so. 

 The collar, however, is always a sign of functional, and 

 especially nutritive activity, most developed when the 

 sponge is most expanded, and always to be found more 



