Hclianthiis Aiimms Syiuolylciis. 471 



a value of 19/^ a race with \alucs ranging between dO^o 

 and 89% arose immediately, and tliat this value was 

 fairly constantly maintained under C(jntinued selection. 

 Some low values (25-50%) still occurred in the first 

 two generations which followed the original plant of 

 1887, but they ha\e not occurred since. The highest 

 values attained in the various years deviate but little ; 

 the deviations lying well within the pro])a])le error of 

 observation, or due to occasional differences Ijetween the 

 climates of the succeeding years. The hereditary values 

 of the parents selected each year were as follows : 



HARVEST OF 1887 1888 1889 1890 1891 1892 1894 1895 1896 

 Values of the parent ^^ g^ ^^ ^^ ^^ ^^ ^^ ^3 ^^ 

 plants in percentages 



In some years the parent was not the individual which 

 possessed the highest value. This v^as due to the fact 

 that when the differences in the values are small it may 

 be necessary to pay attention to other circumstances also 

 and especially to reject weak individuals, or such as 

 flower too late, or set little seed, or happen to be fas- 

 ciated. 



1die choice of a parent must further be left to chance 

 in groups of equally good plants with practically the 

 same hereditar}- values; because sunflowers are too large 

 to permit of diminishing the elements of chance by dou- 

 bling the principle of the selection by the application of 

 the method of selection by grandparents (see p. 385): 

 but, as it ha]:)pened, I was mostly fortunate in my choice, 

 with the exception of the last year in which the mean 

 sank from 85% to about 68%. 



My race is not to be regarded as a normal eversport- 

 ing variety but as an improved breed within it. Instead 

 of selecting parents with 80-90% and over I could have 



