IT. i CELL-DIVISION 45 



sphere, when placed in a fluid medium, and it is also a familiar 

 fact that after the first (and subsequent) divisions the blasto- 

 meres are flattened against one another (Cytarme, to use Roux's 

 term), and that whether they are compressed by an egg 

 membrane or not (examples of the second alternative are to be 

 found in Utiio, Dreissensia, Umbrella, Crepiilnla, Aplysia linia- 

 chia, Asferias), though the surface of contact is not always 

 curved when the cells are unequal. The two cells, however, 

 often become rounded off and partially separated from one 

 another prior to the next division. Such a separation (Cyto- 

 chorismus) has also been observed by Roux in the case of cells 

 of the Frog's egg, which, having been isolated in albumen or 

 salt solution, have subsequently reunited. 



That the cells flatten against instead of repelling one another, 

 as free oil-drops would do, suggests that they, like soap-bubbles, 

 are provided with an insoluble coating-film, while their subse- 

 quent separation may be provisionally explained by supposing 

 that this coating-film becomes temporarily dissolved under the 

 action of some substance formed in the cell. This idea is borne 

 out by a striking experiment of Herbst's, who found that in 

 sea-water deprived of its calcium the blastomeres of the sea- 

 urchin egg came apart and resumed their spherical shape. At 

 the same time the surface membrane underwent a visible altera- 

 tion, becoming radially striated. It seems reasonable to conclude 

 that there is a membrane by which contact is normally effected, 

 and that this is soluble in sea-water devoid of calcium. On the 

 addition of calcium the cells cohere again. 



It may be mentioned that when systems of drops of jelly, 

 floating in a medium of oil and united by their coating-films 

 of water, are removed to alcohol, in which both oil and water 

 are soluble, the films disappear and the drops separate. 



In the next stage (four cells) the type of segmentation in which 

 the laws of capillarity are most strictly obeyed is obviously 

 that which we have distinguished above as the spiral or tetra- 

 hedral type, and Robert has been able to show that successful 

 imitations of the four-, eight-, twelve-, and sixteen-celled stages 

 of the egg of Trochus may be made with soap-bubbles. 



Four equal bubbles were placed in a porcelain cup, which held 

 them together in the same way that the actual cells are held 



