The Isolation of Tricotylous Intermediate Races. 425 
seedlings were counted from each lot. The result was a 
very remarkable one. With a single exception the fig- 
ures formed a close series of which the mean was 7%, 
whilst the great majority of figures lay between 2 and 
12%. (Two parents with 0.5 and 1.0 and three with 
13, 14, and 17%. ) Besides these there was a single 
plant which stood far from the others in the series. It 
had produced 56% tricotyls, i. e., more than half of its 
seeds were tricotylous. For the sake of greater certainty 
this value was determined twice. The percentage value 
calculated from a lot of 768 seedlings was 58% ; from 
another of 657 seedlings 54% ; with a mean of 56% for 
1492 seedlings. 
This one plant, therefore, had a hereditary value 
which corresponded closely with the mean value which 
we should expect the intermediate race sought for to 
possess. Of course the experiment was continued from 
the seeds of this plant only. The seedlings of all the 
other parents, tricotyls as well as dicotyls, were thrown 
away. 
Two courses were now open to me, either simply to 
maintain the race, or to endeavor to improve it by further 
selection. In the former case, selection would have to 
be avoided as much as possible, and we should have to 
harvest and sow the seeds of all the specimens mixed 
together, and in planting out the seedlings be careful not 
to give preference to the tricotyls. I have not done this ; 
but as is my custom, I have harvested and recorded the 
seeds of each individual separately and only planted out 
the seedlings of the best parents, that is to say the parents 
with the highest hereditary values, for the continuation of 
the race. 
In the summer of 1897, 37 tricotylous offspring of the 
