i 9 2o] COULTER— A LEU RONE COLOR 423 



III. If one is to regard true mottling as limited to those cases 

 in which the character appears in all the grains which receive R 

 from the male parent only, that is, descendants of C tester, then, 

 unfortunately, no data are at hand on comparative mottling ratios 

 in two ears with the same parents. Such data, however, are 

 available from descendants of R tester, where mottling (or pseudo 

 mottling, as it may prove to be) appeared in relatively small 

 percentages, as shown in table IV. The discrepancies apparent 

 in table IV may be taken to indicate that the supposed mottling 

 of the R tester line is altogether sporadic and without genetic sig- 

 nificance. Even though that may be so, the surprising case of 

 670-678 is probably worthy of further investigation. 



IV. Among the progeny of the faint colored grains only two 

 poor examples of this sort are available, as may be seen in table V. 



TABLE V 



It is highly probable that the occurrence of these faint grains is 

 sufficiently limited by local conditions so that discrepancies between 

 two such ears will prove common. This is also suggested by the 

 results of a test which need only be briefly mentioned at present. 

 It is expected that the numerical value of any ratio, which is 

 affected only by matters Mendelian, will conform with predictions 

 based on the laws of chance. With equal certainty the laws of 

 chance indicate that, where a 1 colored :i colorless ratio appears, 

 we will not find many long strings of colored grains lying together 

 in the same row. Actually we may expect that the colored grains 

 will be scattered within the rows of the ear in such a way that there 

 will be, speaking relatively, n groups of 1 colored grain each, 

 n/2 groups of 2, n/4 groups of 3, etc. These values will differ, of 

 course, with the ratios themselves. Thus, in a 3 colored : 1 colorless, 



