The Isohiliun of Tricotylous Intcnncdiatc Races. 429 



respectively. The values tor the two groups did not 

 differ essentially, and were (listril)uted hctween 51 and 

 92^0, and the mean of the 2? plants (the rest had heen 

 male) was 73%. The race had, therefore, in compari- 

 son with the previous year, undergone a further im- 

 provement. 



It was continued one year more in the same way 

 (1897). 12 female and several male tricotylous off- 

 spring of the parent with 92% were planted out, and the 

 values calculated from these were found to be distributed 

 between 65 and 91%, with a mean of 78%. 



We see, therefore, that after the figure 55% had 

 been reached in the harvest of 1894 the mean value rose 

 in the three following years of the experiment to 67, 73 

 and 78%. 



Clarkia pulchella, Fig. 85. It was in the spring of 

 1895 that I made the extensive sowings of horticultural 

 seeds to which I have already referred, for the purpose 

 of isolating tricotylous intermediate races. The seeds 

 of Clarkia pulchella alba produced about 1% tricotyls. 

 v30 of these flowered, but only 18 of them produced suf- 

 ficient seed. Two of them had hereditary values of 14 

 and 16%, the rest from to 7%, with a mean of 4%. 

 In 1896 only the tricotyls of the parent with 16% were 

 planted out. There were 39 of these, and for all of them 

 a value could be calculated. These values have been oiven 

 on page 422, and were above 50% for eight plants. The 

 intermediate race, therefore, was already represented by 

 several specimens. 



In 1897 I planted out only the tricotylous seedlings 

 of the ])lant with 64%, and saved the seeds of 39 

 of them. Their hereditary values were distributed be- 



