243 THE AMERICAN NATURALIST [You XLV 



Consequently, when the F a shows the same variability as 

 the pure lines which entered into it we must conclude 

 that there was at least approximate equality among the 

 sperms which came from the one self -fertilized strain, 

 and among the eggs which came from the other. In the 

 F 2 , on the other hand, genotypic differences appear, 

 owing to the segregation of the different characteristics 

 into the. different germ-cells, and to this fact may be 

 ascribed the increased variability in the F 2 . 



While other characters have not been studied by the 

 same methods that have been used in the investigation of 

 the number of rows on the ears, several features asso- 

 ciated with the physiological vigor of the various pedi- 

 grees have given evidence which appears to me to be 

 strongly corroboratory of the uniformity of the germ- 

 cells produced by plants which have become pure-bred 

 through continued self-fertilization. The smaller size 

 and less vigor of the offspring of self -fertilized plants as 

 compared with those from a normally cross-bred plant 

 were formerly taken to indicate that self-fertilization is 

 injurious, and Darwin's "Effects of Cross and Self-fertil- 

 ization in the Vegetable Kingdom " strongly impressed 

 this point of view. I have been able to demonstrate, 

 however, that this supposedly injurious effect of self- 

 fertilization is only apparent and not real; or at least 

 that if there is such injurious effect, it is relatively in- 

 significant as compared with the increased vigor due to 

 heterozygosis. The most important evidence of this is 

 found in the fact that the continuation of self-fertiliza- 

 tion in any pedigree does not produce a corresponding 

 decrease in vitality and size. The decrease resulting 

 from a second year of self-fertilization is not as great as 

 that from the first year. The third year of self-fertiliza- 

 tion produces still less deterioration, and as this process 

 is continued a limit is approached in such manner as to 

 justify the inference that when complete purity is at- 

 tained no further deterioration is to be expected, thus 

 proving that self-fertilization is not in itself injurious. 



