241 THE AMERICAN NATURALIST [You XLV 



to positively demonstrate the existence of distinct geno- 

 types in maize, because the slight variations which must 

 always be present even in the most uniform progeny, 

 can not be certainly distinguished as genotypic or fluctu- 

 ating simply by inspection. Such demonstration must 

 rest upon a combination of biometric and genetic evi- 

 dence in order to prove acceptable. Most of the differ- 

 entiating characters of my several strains of maize are 

 such that they do not lend themselves readily to bio- 

 metric methods, but the number of rows on the ear is 

 well adapted for such study and several important re- 

 sults have been derived from the consideration of this 

 character. An important proof that the self-fertilized 

 families derived from my common original stock of corn 

 are genotypically distinct, and that they do not owe their 

 different morphological and physiological qualities to 

 fluctuations within a single genotype, was quickly found 

 in the fact that two of these families selected respectively 

 to 12 and 14 rows of grains on the ears, showed a regres- 

 sion of row-number toward different centers instead of 

 toward a common center. The mean of the original popu- 

 lation was slightly above 14 rows. The selection to 14 

 rows was very near this mean and the selection to 12 

 rows was very near this mean and the selection to 12 rows 

 considerably below it. According to Galton's well-known 

 law of "regression toward mediocrity," the mean of a 

 family whose parents were selected to 12 rows should have 

 lain somewhat above 12 rows, and that selected to 14 rows 

 should have retained the mean approximately at 14 rows. 

 The actual result in the case of selection to 12 rows was 

 the production of a family having a mean row-number 

 considerably below the number of rows selected, and the 

 subsequent generations have since shown a close ap- 

 proach to an 8-rowed condition; while the family whose 

 parents were in each generation selected to 14 rows has 

 always had the mean very near to 14 rows. As these 

 families were grown under as nearly uniform conditions 

 as possible, the fact that the 14-rowed family continues 



