BY THE SELECTION OF SOMATIC VARIATIONS. 61 



be adjacent in the same leaf. Likewise, the differences between 

 solid-red epidermis and the non-red epidermis is the same that exists 

 between red and non-red cells in the epidermis of a single leaf having a 

 red-blotched pattern. In the blotched patterns the cells of the epi- 

 dermis are either red or non-red. The number of similar cells that 

 are adjacent to each other determine the size of the blotches. 



The facts regarding development and distribution of the colors raise 

 sharply the question whether the possibility for development of red, 

 yellow, and green is possessed by every cell that is formed or whether 

 these possibilities have been sorted out by qualitative divisions in a 

 Weismannian sense. A sharp distinction should be made between a 

 character that is metidentical and one that can only belong to a group 

 of cells. In Detto's (1907) sense, power to produce green, yellow, and 

 red seems to be metidentical; that is, this abiUty may be strictly a 

 property of all cells, while the different patterns appearing so strikingly 

 as characters of entire leaves can not in any sense be metidentical. 



Most important evidence regarding the development of leaves with 

 red-blotched patterns was obtained by the cytological studies conducted 

 by Mr. E. G. Arzberger. In the early stages of leaves that later will 

 become red-blotched, all the epidermal cells possess red pigmentation, 

 forming a complete periclinal layer of red cells, and in this respect the 

 early stages of leaves with solid-red and red-blotched patterns are alike. 

 In solid-red patterns all cells continue to possess red sap color, while 

 in blotched types the red disappears in certain cells. The evidence is 

 clear in these cases that all the epidermal cells arise from cells having 

 red cell sap and in this respect are potentially alike. 



The differences in patterns in respect to red are largely those of 

 quantity; the total of red cell sap in a leaf is comparatively large 

 in solid-red patterns and much less in blotched patterns, of which there 

 is every grade to the no-red epidermal patterns. Differences in dis- 

 tribution are also involved. There is the tendency for red to be in the 

 epidermis rather than in the subepidermal tissues. Usually it is in 

 both upper and lower surface, but in one pattern (fig. 10) the red is 

 almost entirely massed in the upper epidermis and in the center of the 

 leaf. In the blotched types the number, size, and degree of coalescing 

 of the blotches present every gradation from a finely blotched condition 

 to a sohd red. 



While the total amoimt of red and its distribution in the epidermis 

 determines quite definite and constant patterns, it should be said that no 

 plant has been obtained, either in seed or vegetative propagation, that 

 was entirely devoid of red. Subepidermal cells may also possess red 

 cell sap; red sap is quite pronounced in veins of leaves (see figs. 13 

 and 14), and varying amounts can be seen in stems. In some plants, 

 especially those of seed origin, no trace of red can be seen except at 

 the nodes, and here the amount for different plants varies from faint 



