BY THE SELECTION OF SOMATIC VARIATIONS. 69 



spontaneous development of a less vigorous virus may be very common, 

 and many cases of variegation, even of those that are apparently seed 

 constant, may be due to such a condition. 



Lindemuth (1905) determined that the variegation in Coleus was 

 not infectious like that of Abutilon. Just what types he used is not 

 clear from the data given. Conclusive evidence regarding this point 

 has not been obtained by the writer. Thus far grafting experiments 

 between green-yellow and pure green types have shown no cases of 

 development of yellow, and the writer has assumed that the variegation 

 is not at least vigorously infectious. 



We may also note that spontaneous loss of ability to produce green 

 in a part of the cells of the growing-point may result in a chimeral 

 variegation such as Baur (1909) reports in white-margined types of 

 Pelargonium zonale. His anatomical studies showed that in growing- 

 points the white-colored tissues may lie over the green, forming histo- 

 genic layers and giving remarkable permanency of the periclinal 

 chimera in vegetative propagation. 



In maintaining this relation white cells give rise to white cells and 

 green to green, but mechanical readjustments in the growing-points 

 may give branches with quite different distribution of the two kinds 

 of cells. Branches may thus arise with sectorical distribution of green 

 and white, with only white or only green cells, or even with reversed 

 positions (Baur, 1909; Stout, 1913). These readjustments give no 

 new qualities to cells, nor do they appear to involve changes of either 

 kind to the character of the other. Yet the occurrence of numerous 

 types of variegation with this chimeral relationship indicates that such 

 spontaneous loss is not infrequent. 



Certain types of bud variation in Coleus present features quite 

 similar to the readjustments that appear in Pelargonium, and raise the 

 question whether there is possibility of spatial readjustments of dis- 

 tinct tissue elements. The sudden and apparently complete loss of 

 epidermal red suggests that this layer may exist as peripheral in a more 

 or less chimeral relationship, but the development of red and non-red in 

 the adjacent cells of the epidermis of a single leaf indicates clearly that 

 these differences can arise within cells of the same immediate progeny. 

 If for any reason a part of the epidermal cells fail to develop red, the 

 red might be absent in the entire epidermis for the same reason. 



The loss of yellow giving pure green might seem to be due to the 

 exclusion of yellow cells. Also, the sharp contrast between patterns 

 of green-yellow (figs. 2, 8, and 12) and pattern .of yellow-green (figs. 6, 

 11, and 14) suggest the possibility of a spatial readjustment in the 

 growing-point of two distinct cell elements. But certain plants in 

 all these patterns have fluctuated, getting greener to a green-yellow 

 blotched pattern during winter and returning to the type pattern in the 

 next summer, and often exhibiting at one time in a single row of 



