ORGAN-FORMING MATERIALS IN FROG S EMBRYO. 133 



Whether the inwandering, or inpulling of some of the cells of 

 the upper hemisphere is only due to the shifting of the cells as 

 the) r divide, so that the pressure relations are better adjusted, or 

 whether there is a further meaning to be attached to this process 

 cannot be stated. One might be tempted to assign to the cells 

 the function of producing the mesoderm, and the later pushing 

 outwards at the sides of much of the material derived from these 

 cells into a position where the mesoderm appears might be made 

 to give color to this interpretation. However this may be, the 

 experiment of removing the upper four cells shows that some, 

 at least, of the mesoderm comes from other parts of the egg. 

 Whether the same amount forms under these circumstances is 

 too difficult to determine. In the later stages of gastrulation, 

 and at the time also when the neural plate is forming, the meso- 

 derm on the ventral surface increases at the expense of the yolk- 

 cells. These are the yolk-cells that form the tongue of cells 

 that pushes upward into the blastoccel. 



In my last paper dealing with the gastrulation of the frog's- 

 egg, I discussed the " mechanics " of the downward movement 

 of the material that forms the embryonic ring. Certain points 

 that bear on this question have been noted in studying the three 

 species here described, and may be briefly mentioned. There 

 seems to be a good deal of variation in the extent to which the 

 blastocoel enlarges, not only in different species, but in the same 

 species when eggs from different bunches are compared, and 

 even to a slight extent in eggs from the same bunch. The 

 segmentation cavity appears to be formed as a result of the 

 secretion of material from the surrounding cells, as shown by 

 its occurrence in eggs developing out of water. The egg as a 

 whole increases in size as the segmentation cavity grows larger, 

 indicating that the water absorbed from outside more than com- 

 pensates for the amount of fluid secreted into the central space. 



The thinning of the roof of the segmentation cavity takes 

 place at the time of enlargement of the segmentation cavity and 

 it may appear that this enlargement is the cause of the down- 

 ward movement of the material ; but that this is not the correct 

 interpretation of the mechanics of the process is shown by the 

 following experiment. The top of the segmentation cavity was 



