468 SyJicotyloits Races. 



accurate as those calculated from more extensive crops. 

 Nevertheless they fall into definite groups, so that this 

 source of error may be neglected. 



In the harvest of 1888 it was at once evident that 

 the race had been obtained in its pure state by the selec- 

 tion carried out in the foregoing year. Twelve plants 

 set an abundance of seed ; 9 of them had from 30 to 55% 

 .syncotyls, whilst the remaining three parents had 76, 81 

 and 89%, in samples of 121, 275 and 128 seedlings. 

 Of course the plant with 89% was chosen for the con- 

 tinuation of the race. 



It was important to find out whether, if these three 

 high values had not been found, the remaining parents 

 would have gi\'en the same result. For this purpose I 

 made a subsidiary culture in 1890, and chose for this 

 purpose the syncotylous seedlings of a parent with 51%. 

 They produced 23 seed-parents of which the values were 

 distributed evenly between 24 and 91% with a mean of 

 55%. The four highest numbers were 77, 79, 84 and 

 91%. The result can therefore be reached also by an 

 indirect way, as the following scheme shows at the first 

 glance. 



ORIGIN OF THE SYNCOTYLOUS EVERSPORTING VARIETY OF 



HELIANTHUS ANNUUS 



HARVEST OF 1887 1888 1889 



Hereditary' values of the best parents ,q j 89 81 



in percentages ( 51 91 



In order to consider this fact in combination with the 

 figures found for the tricotylous races, w^e must compare 

 it with the table on page 439. Into this table the experi- 

 ment with Helianthus could be inserted straight away. 

 On the one hand it confirms the expectation, which that 

 table justified, of finding a value of 55% in the second 



