BY THE SELECTION OF SOMATIC VARIATIONS. 5 



The exhaustive summary of facts given by Cramer (1907) fully 

 substantiates the views of Darwin and de Vries as to the main classes 

 of variegation and the remarkable gradation in degrees of constancy 

 and inconstancy which are exhibited by the various types and by their 

 bud variations. Cramer gives the most complete and detailed sum- 

 mary of the known cases of bud variation that has been published. 

 He includes (p. 18) under the term " Knospen variation" all cases where 

 a character suddenly changes in a plant in a way that can not be attrib- 

 uted to environmental influence. He considers that there are three 

 main classes of bud variation: (1) vegetative segregations in hybrids; 

 (2) intermediate-race bud variations; and (3) vegetative mutations. 

 Recessive characters in hybrids may, he assumes, separate out by 

 vegetative cell-divisions, and dominant characters which have been 

 latent may reappear. His conception of the variability of characters 

 in an intermediate race is the same as that of de Vries. Characters 

 concerned in bud variations in intermediate races show, it is considered, 

 great irregularity in expression. The bud variations which give a seed 

 progeny quite constant for the character involved are classed as vege- 

 tative mutations. This classification emphasizes the fact that charac- 

 ters involved in hybridization may come into expression quite irregu- 

 larly, and that spontaneous and fluctuating variations are common in 

 vegetative development. 



Cramer devotes a most interesting and instructive chapter to varie- 

 gated plants, showing, especially, their wide distribution in the plant 

 kingdom, the wide fluctuation in the degree of constancy of their seed 

 progeny, the range from vegetatively constant to inconstant types, and 

 the frequency of bud variations in variegated plants. He considers (pp. 

 126, 127) that loss of variegation can occur in two ways: (1) by atavistic 

 bud variations and (2) through influence of external conditions. He 

 notes that it is very difficult to distinguish fluctuating variability from 

 mutation. 



The general variability of the characters concerned with variegation 

 is well shown in his discussion of changes that may occur in a single 

 plant, such as the following : In apparently inherited types of variega- 

 tion, the seedlings are often green at first. Seedlings may have at 

 first variegated cotyledons, then a few green leaves, and then variegated 

 leaves. Some biennial plants are pure green during the first year of 

 growth, but variegated in the second year. Various plants exhibit 

 a periodicity in their variegation, being green in spring and variegated 

 later in the summer, or vice versa. He gives numerous cases of the 

 appearance of variegation by bud variation on green plants resulting 

 in new varieties or in the duplication of types already known. These 

 exhibit various degrees of constancy. 



Various adherents of Mendelian doctrines have more recently dis- 

 cussed the transmission of characters appearing as bud variations, 



