ON THE DEVELOPMENT OP THE POND-SNAIL. 379 



external shape. I am not prepared to give a detailed account 

 of those changes, but can only draw attention to some general 

 features. 



The invaginated cells or segmentation-products, which 

 form the endoderm or primitive alimentary sac of the gas- 

 trula, are not at first distinguishable through the walls of the 

 widely excavated pyramidal embryo. But as the wide orifice 

 narrows to a slit, the two sets of cells become clearly dis- 

 tinguishable, a result due, not merely to the clearing-up of 

 the outer cells, but also to the gradual assumption of a 

 specific character — globular form, dark granulation, and high 

 refrangibility — by the invaginated cells. During the whole 

 of the later development, as far as I have watched it in 

 Lymnceus, the gastrula'sendoderm-cells are undergoing modifi- 

 cation, resulting at last in the separation of a pellucid material 

 from a more superficial granular matter, which appears ulti- 

 mately to give rise to a cellular network (PI. XVII, fig. 22). 

 The minute history of the changes in these cells would be an 

 important matter to determine, since it appears that the history 

 of other invaginated gastrula-endoderms is not so simple as 

 one might suppose beforehand. They are by no means 

 simple masses of formative protoplasm, which merely multiply 

 by division, but appear in many cases to contain other 

 elements analogous to the nutritive yelk (whence obtained is 

 not clear), which in earlier stages have accumulated in each 

 endodermal cell. Each endodermal cell then appears to play 

 a part analogous to that of a whole ovum in its early stages, 

 segregating and giving rise to new cells by endogenous forma- 

 tion. A process of this sort appears to go on in the gastrula- 

 endoderm of Pisidium as well as in that of Lymnceus, and 

 probably also in the " residual yelk-spheres " of Pleuro- 

 branchidium. An important histiological arrangement seen 

 in the specimen (PI. XVI, fig. 14) is the connection of the 

 endodermal mass of cells with those forming the body-wall 

 by means of long processes. This is seen again in a later 

 phase in PL XVII, fig. 19, The processes appear to be 

 actual filaments of the cell-substance of the endodermal 

 cells. 



In the trochosphere so far the shape is nearly spherical, 

 excepting for the raised ciliated ridges, which together make 

 a heart-shaped outline on the surface of the embryo, the 

 indentation of the heart being occupied by the mouth. 



The Veliger-phase. — In fig. 1 of PI. XVII we have a some- 

 what more advanced stage, and in figs. 2, 3, 4, 5, 6, the 

 definite Veliger-phase is attained. In the Veliger the area 

 of the velum has a definite development, occupying relatively 



