EARLY DEVELOPMENT OF GRANTIA COJIPRESSA. 271 



11. Formation of Oogonia in situ from Collar-cells. 



In PI. 22. figs. 23, 25, and 26, I have drawn with the camera lucida parts 

 of the gastral epithelium to show the metamorphosis in situ of collar-cells 

 into oogonia. The first stage is in PI. 22. fig. 26 ; at PC. a collar-cell has 

 sunk below or become partly covered by its immediate neighbours. That 

 this cell, which is larger than ordinary collar-cells, is really a lately metamor- 

 phosed gastral epithelial cell is indicated, firstly because its nucleus resembles 

 that of the collar-cell, and secondly because its cytoplasmic inclusions are 

 closely similar to those of its neighbours. In PI. 22. fig. 25, the cell marked 

 OG has a cytoplasm just like that of the collar-cell except that its size is 

 greater ; the nucleus of this cell is like the nucleus of the collar-cell, except 

 for the size of the karyosome. In PI. 22. fig. 2.3 at PC and OG are two 

 stages which are similar to those already described. The cell OG is now an 

 oogonium, and is about at the same stage as that drawn in PI. 19. fig. 8. 

 In PI. 22. fig. 27 at OTE is a still later stage. 



The figures 23-26 of Plate 22 were drawn carefully with the camera 

 lucida, from the edge of a sponge where a large number of young oocytes 

 were found. All stages up to the formation of larger oocytes were present, 

 and I believe there can be no doubt that such a cell as that in PL 22. 

 fig. 25 is a stage in oogenesis, and is derived from a collar-cell. 



12. Position of Germ-cells in Grantia compressa. 



In PI. 19. fig. 5 is part of a sponge showing an egg (OTE) on the left 

 and a group of spermatids (SPT) on the right. These lie beneath the collar- 

 cells, and tend to abut into the gastral cavity (GCAV). Jorgensen (8aj 

 mentions that oogonial divisions may take place in the gastral cavity 

 beneath the collared epithelium, but more often in the mesoglea. The 

 growing oocyte alwaj's occupies a position beneath the basement-membrane 

 of the collar-epithelium, and I believe that it may come to lie in this position, 

 in many cases by a simple sinking inwards as shown in PI. 22. figs. 25 and 26. 

 Between the stages drawn in the latter figures and in PI. 22. fig. 27, OTE, 

 Jorgensen finds oogonial divisions and the prophases of the first maturation 

 division. These latter have occurred in PI. 22. fig. 27, OTE, which is well 

 advanced in growth. Jorgensen would possibly interpret the cells marked 

 PC and OGA in PI. 22. figs. 23. 25, and 26, as secondary oogonia which had 

 migrated from the mesoglea to their present position ; this I would dispute, 

 for I find all stages intermediate between such cells and the collar-epithelium 

 cell. The migrations of the oogonia are very peculiar, and I cannot claim 

 that this account is at all exhaustive. The problem is one which will need 

 special study and a larger amount of material than at my disposal ; the main 

 point is to note that collar-cells do not always remain as such, bul may grow 

 and metamorphose into other elements. 



