BY THE SELECTION OF SOMATIC VARIATIONS. 71 



in large measure to inner conditions. Cramer (1907, p. 128) sum- 

 marizes numerous cases of those showing marked differences in beha- 

 vior. Certain varieties of Quercus have pure-green foliage each spring, 

 but later produce variegated leaves. Ulmus scabra var. viminalis is 

 yellow during midsummer, but pure green at close of the summer. 

 Linaria biennis is pure green during the first year of growth, but varie- 

 gated in the second. 



These phenomena illustrate again that cells of the same lineage may 

 fluctuate in development from a maximum of green to various degrees 

 of loss of chlorophyl, with often development of yellow coloration, and 

 that such phenomena may reappear with marked constancy in progeny. 



Aside from these classes there is a wide range of types, including 

 many cases in which the variegation reappears more or less generally 

 and constantly in the individual plants and in the variety. The 

 variegation appears to be inherited, at least in a certain degree. 



In the propagation of variegated plants, much general data has been 

 obtained regarding the degree of constancy both in vegetation and seed 

 propagation. Cramer (1907, p. 129) gives a summary of numerous 

 cases where the variegation disappears in certain methods of vegetative 

 propagation. In Cornus mascula variegata (T. M., Gard. Chron., 

 32: 952), root-cuttings give pure-green individuals, but plants from 

 layers retain the variegation. Such cases, however, should be investi- 

 gated with special regard to the nature of the variegation. In many of 

 the cases noted, especially those in Pelargonium, the variegation may be 

 chimeral, and when root-cuttings are made the green cells have greater 

 power of regeneration. The inconstancy exhibited in vegetative propa- 

 gation, however, is no greater or more marked than that which develops 

 on a single individual. Exact evidence of the inheritance and con- 

 stancy of variegation through pedigreed vegetative progenies seems to 

 be lacking. While in some varieties the variegation appears to be 

 quite constant, in others it is widely variable. 



The Mendelian studies that have been made of the seed progeny of 

 variegated plants and of the bud variations which involve changes of 

 pattern show likewise a wide range of behavior. Hybridization studies 

 involving variegation, as has been pointed out in the introduction, 

 indicate clearly the wide range of variability, and what, from the 

 Mendelian viewpoint of unit characters or unit factors, is most erratic 

 inheritance. These studies have contributed interesting and valuable 

 data on the sort of variations one may espect in hybridization studies 

 of this kind, but they indicate very clearly that the assumed factors are 

 themselves fluctuating. In these studies the variability of plants used 

 as seed parents has not been determined by vegetative propagation. 

 This is not possible in all cases, but whenever it is possible the emphasis 

 should be placed on this line of investigation if one is to speak with 

 certainty regarding the nature of the inheritance. 



