TIIE EMBRYO-SAC IN AXGIOSPERMS. 523 



the periphery, curving outwards so as to cut the epidermis at 

 right angles, or nearly so. This is plainly seen in PL XVII. 

 fig. 10, and PL XVIII. figs. 1 and 3 ; and comparison with younger 

 stages shows transitional states. 



Thirdly we may note the constancy of volume of any two 

 sister cells, again excepting the axial series. It appears as if, when 

 a cell has become of a certain size, it were in some way impelled to 

 divide by a wall perpendicular to its side walls ; each sister cell 

 then repeats ; and thus the increasing volumes of outer layers are 

 filled up. In the older stages this is less evident, and we see 



more cubical cells at the base and narrower elongated ones higher 

 up. 



Can we in any way explain the form of the ovule by the cell 

 distribution, or vice versa ? On looking at an active young ovule 

 at a stage such as in PL XVII. fig. 7, one is struck by the vigo- 

 rous appearance of all the cells, and those of the nucellus itself 

 not less than of the integuments : the cells of the epidermal layer 

 of the nucellus are becoming divided by walls in perpendicular 

 planes at right angles to the whole surface; i. e. the surface of the 

 nucellus is becoming increased. Such an increase means reduced 

 pressure on the cells immediately underneath, and not only that, 

 but also a positive stretching in a direction parallel to the surface 

 of the nucellus ; hence they have room to grow laterally and for- 

 ward (i. e. in a tangential and radial direction) ; and if one com- 

 pares PL XVII. figs. 8 to 10, and PL XVIII. figs. 1 and 3, the 

 correspondence between some such explanation and the facts 

 appears at least probable*. 



Following any of the radiating groups of cells in an advanced 

 ovule (sucb as in PL XVIII. fig. 3), we see that as it readies the 

 surface of the nucellus there are more cells to cover the greater 

 area, and that these cells are very definitely arranged: it would 

 appear in fact as if the cells could not increase beyond a certain 

 size without a division-wall appearing; and having more room 

 towards the periphery of the ovule, we therefore have more cells. 

 But if we examine the axis of the ovule, it appears that the 

 same applies here also ; probably the diminished pressure at the 

 apex has a direct bearing on the elongated form of the axial cells. 

 Of course it is not suggested that the process of formation of an 



* In all these considerations the integuments may be neglected for simpli- 

 city. 



