620 MUlllEL KOBP]RTSON AND K. A, MINfJIIN. 



Most of the si^onges were preserved in osmic acid followed by 

 picrocarmine, a good method for showing clearly the cyto- 

 plasmic structures, especially the collar and flagellum, but 

 not suitable for demonstrating the finer details of the nuclear 

 apparatus. Some of the material, however, had been pre- 

 served in Hermann's fluid, and it is on this that we base the 

 results set forth in this memoir. Sections cut from sponges 

 preserved in this way were stained with various stains, more 

 particularly by Heidenhain's iron-hiematoxylin method, and 

 counter-stained with eosin or Lichtgriin, the latter being 

 found to be of great assistance in making out the details of 

 the collar and Hagelluui, since these parts are tinged by it. 



(1) The Re sting Collar-cell. — In Clathrina coriacea, 

 as in all sponges of the family Clathrinidae, the nucleus lies 

 invariably, in the ordinary "vegetative" or resting conflition, 

 at the base of the columnar collar-cell, that is to say, at the end 

 which is furthest from the collar and flagellum. At the apex 

 of the cell, in the centre of the area enclosed by the base of 

 the collar, lies a minute granule — the blephai'oplast — from 

 which the flagellum takes origin. These structures, no less 

 than the general form of the collar cell and its position in the 

 epithelium, of which it forms a part, give a definite orienta- 

 tion to the cell ; any direction parallel to an imaginary axis 

 continuing that of the flagellum and passing tiirough the 

 blepharoplast and nucleus may be termed vertical, while any 

 plane at right angles to the vertical axis may be termed 

 horizontal. ^y 



The form of the collar-cell and the dimensions of their 

 different pai-ts vary considerably with tiie condition of the 

 sponge, whether expanded or conti*acted, and may be different 

 also in different parts of the same sponge. In specimens in 

 which the pores are fully open, and in which all appearances 

 indicate that the collar-cells are in full functional activity, 

 the bodies of the cells are fairly broad, aud about 12-13 ju 

 in height by 5-6 n in bi-eadth ; the collar reaches a length of 

 10-11 fx, and the flagellum some 25-27 ju- When, on the other 

 hand, the pores are closed up and the sponge is partially con- 



