BY THE SELECTION OF SOMATIC VARIATIONS. 



43 



Judging from the data on changes in green and yellow, there appears 

 to be no general correlation between the number of fluctuating plants 

 and the number of cases of bud variations. In the green-yellow-red 

 blotched group there were proportionally more than four times as many 

 bud variations as in the group green-yellow spotted-red blotched, but the 

 percentages of constant plants were nearly identical. Not one of the 

 plants with laciniate leaves was constant for green and yellow, but only 

 one case of bud variation occurred. In the green-red blotched group 

 there was chance only for the appearance of yellow, and this change 

 occurred in a relatively large number of cases, both in fluctuations and 

 as bud variations. 



A very marked contrast appears in a comparison of the two patterns 

 green-yellow-red blotched and yellow-green-red blotched. Both have about 

 the same proportions of green and yellow, except that the relative 

 position is reversed. Both groups agree quite closely in the per- 

 centage of constant plants. In the latter, however, bud variations 

 were 2.5 times as frequent. The position of the j^ellow in the center 

 seemed to increase bud variations involving green and yellow over 

 that in plants with the yellow at the border. 



DISTRIBUTION OF BUD VARIATIONS AMONG DIFFERENT CLONES. 



The wide range of variation both of fluctuations and of bud varia- 

 tions emphasized in the summaries already given was realized in a 

 series of plants derived by vegetative propagation from two plants 

 having the same color pattern. The records of pedigrees show that 

 marked differences appeared among the various clones with respect to 

 constancy and to the range and the frequency of bud variations. 



This is shown quite clearly when the data regarding the main clones 

 derived from plant 1 are grouped together as arranged in table 12. 



Table 12. — General summary of clones. 



The main clones 11 and 12 were derived from two branches of plant 

 1 which had the same color pattern. Although the branches were 

 identical in appearance, the two progenies were quite different. 62 

 per cent of clone 11 were constant, while 45 per cent of clone 12 were 

 constant; but in the more constant clone 11 there were proportionally 



