Tricotylous Races Do Not Arise by Selection. 405 



spending number of offspring which manifested these 

 numbers amongst their seeds). 



P 1 2 3 4 5 6 7 8 9 10 11 12 13 U 15 16 17 21 25 

 A 16 27 30 18 18 11 6 4 2 2 2 3 2 1 2 1 2 1 l 



This series only relates to the tricotyls selected for 

 my experiment. If I had planted out the whole progeny 

 of the parent in question without selection, the number 

 of low values would most certainly have been somewhat 

 larger; and the apex of the curve would perhaps have 

 stood over the instead of over 3%. But the chief point 

 in this series is that from 3% onwards the figures regu- 

 larly and continually diminish in such a way that the 

 two extreme variants with 21% and 25% round off the 

 series continuously; i. e., in the ordinary manner pecu- 

 liar to physiological curves. It is obvious that they do 

 not stand in discontinuous relations to it. 



In the estimation of the higher individual values a 

 latitude of 5% must be allowed, i. e., the figure 15 de- 

 notes a real capacity of from 10 to 20%, the figure 2? 

 one of from 20 to 30%. In the case of extreme devia- 

 tions it is always well to assume that these errors may 

 have operated in the same direction. We may only state, 

 therefore, that by means of a cultivation through eight 

 generations, a ratio of 20% has been attained. But 

 whether or not this is the limit, we do not know. Also, 

 it is uncertain whether the parent with 25% was reallv 

 better than those with 15 to 21%. But it is certain that 

 the best seed-plant, as judged by its hereditarv value 

 must be one of this group. 



Therefore, at this stage in the development of our 

 race, the process of selection has become much more 

 certain than before and less dependent on chance. It is 

 only necessary not to limit our choice to those which 



