1892.] 



Note on Excretion in Sponges. 



479 



the neighbouring collars to prevent the water that is already filtered 

 and already foul from returning past the inactive area to pollute 

 the afferent water supply. When the food has been digested, the cells 

 elongate and become closely pressed together ; the separation of their 

 basal parts takes place in the manner already described, and the 

 distal parts start on a new cycle with, hungry protoplasm, active 

 flagella, and separated collars. 



In the Homoccela I have been unable to make sure of these 

 changes other than as regards the habitual association of Sollas's 

 membrane with the absence of flagella, and of the presence of flagella 

 with separated collars. (This was noticed by Sollas, 1888, in the 

 Tetractinellida of the ' Challenger.') 



The nucleus in the collar cells of the Homoccela is generally basal, 

 whereas in the Heteroccela, contrary to the current statements, it is 

 almost always distal, the basal spherules having been commonly 

 mistaken for it. The idea rather suggests itself that in sponges with 

 basal nuclei in the collar cells the fission into column and plinth does 

 not take place. 



Of the Metschnikoff cells in A. clathrus the following is the his- 

 tory. Under certain circumstances the Sollas's membrane and the 

 gastral surface of the collar cells become replaced by a continuous 

 hyaline film charged with the yellow granules characteristic of the 

 sponge {vide fig. 2). Certain collar cells — or, rather, quondam collar 

 cells — become charged throughout their whole substance with these 

 granules ; they are now Metschnikoff cells, and are the cells described 

 (p. 360) and figured by that author as mesoderm cells in his well- 

 known paper (" Spong. Studien ;" ' Zeits. wiss. Zool.,' vol. 32), 

 They are almost always excavated by a cavity or duct,* frequently 

 taking the form of a capillary and sometimes branching tube ; they 

 push through towards the ectodermal surface, with which they 

 become connected; the granular film covering the general gastral 

 surface disappears. It is now to be noted that the afferent pores of 

 all the Homoccela I have examined consist each of a single, nucleate, 

 perforated cell;f the film of protoplasm, though very thin compared 

 with the size of the lumen, is charged with granules similar to those 

 found in the ectoderm cells. A longitudinal section through part of 



* [May 16. — " In the next stage, which I regard as a more advanced condition, 

 they form a ring surrounding an empty discoid or sometimes crescentic space." 

 (Dendy, Monograph, p. 18.)- — The " peculiar structures " met with by this author in 

 Grantia labyrinthica (p. 17), I interpret as column -and-plinth chambers violently 

 contracted in alcohol ; the lateral coherence of the cells renders them subject to 

 this. The " nerve-cells " in fig. 31 of the same paper are my " glandular epithelium. 1 ' 

 The " gland-cells " under the gastral surface in fig. 26 I believe in S. raphanus to 

 be sperm mother-cells, having the remarkable peculiarity that the spermatozoa 

 come to maturity singly.] 



f Cf. Minchin, ' Quart. Jl. Micr. Sci.,' vol. 33, PI. 11, fig. 21. 



