30 



flicting opinion, these being, according to various authors, 

 amoebocytes (3), archeocj^tes (4), choanocytes (5 and 6). 



In the material of June, 1916, the earhest stages in the develop- 

 ment of the germ cells were found. These varied considerably 

 in size, and were situated behind the choanocytes of the flageilate 

 chambers — the smallest being actually amongst them — that is 

 to say, on the same side of the gastral epithelium. This at first 

 sight suggests their having originated here, but on referring to 

 Professor Dend3^'s work (3) I found that m G. labyrinthica '" the 

 ova are obviously derived from amoeboid wandering cells." 

 Although the j^oung germ cells themselves appeared in many cases 

 to be amoeboid it was not at all obvious from my preparations that 

 they originated from amoebocytes, wandering in the mesoglea, 

 for the smallest cells recognised as future ova were not only 

 invariably situated among the choanocytes but were at first no 

 larger than these, and differed from them only in being less 

 elongated, being without the characteristic collar and in taking 

 the stain more deeply (Figs. 12 and 14). In not a few cases these 

 very young oogonia appeared to retain traces of a collar (Fig. 12), 

 but omng to their smallness and the crowding of the cells together, 

 it is difficult to be absolutely certain. The observations of Carter 

 and Saviile-Kent also bear out this view that the odgonia are 

 derived from the collar-cells. 



In his last work on Grantia, Dendy (1) also comes to the con- 

 clusion that though the ova originated as amoebocytes, these 

 amoeboid cells are themselves derived from the collar-cells. 



Affinities. — This being the case we are led back to the old 

 theory of the close relationship of the Porifera (or at least Calcarea 

 Hke Grantia) to the Choanoflagehates. The fact that the choano- 

 cytes are responsible for practically all the essential functions of 

 the sponge, including reproduction, indicates that the latter is 

 little more than a colony of choanoflagellate Protozoa. Such a 

 colony would originally be reproduced by the setting free of a 

 collared cell which by fixation on a new site, followed by repeated 

 division, would give rise to a new colony. If we imagine that, 

 in the course of evolution, such a cell becomes provided with 

 larger quantities of nutritive material for the free-swimming period, 

 causing an increase in size until the Umit of growth is reached 

 and division takes place before the cell is set free, we have exactly 



