The Isolation of Tricotylous Intermediate Races. 427 
ure. Such a task I have, therefore, only undertaken 
with a single species, Antirrhinum uiajus. 
Only some varieties of hemp seem to include tricotyl- 
ous intermediate races. Amongst those which I have 
tested I have found such a one in the giant hemp only. 
In the spring of 1893 I sowed a large quantity of seed of 
this species, but was only able to bring 7 tricotylous plants 
to maturity. Unfortunately the majority of them were 
male, and there was but a single female plant. This, how- 
ever, produced four tricotyls amongst 126 seedlings. In 
1894 two of these were female and two male. The seeds 
of the former were saved separately and yielded 15 and 
9% tricotyls amongst 400 and 600 seedlings respectively. 
For the continuation of the race, onlv the tricotylous 
./ . 
seedlings of the parent with 15% were planted out 
(1895). There were 29 plants, of which 10 bore seed. 
Their values were 1931384043474850 
52 and 63%. As we see they attained the mean value of 
the intermediate race and even exceeded it in one case, 
(63% amongst 316 seedlings). The tricotylous offspring 
of this individual alone were planted out (1896). Of 
these, 38 specimens set seed, and from this the values of 
the individual parents were calculated. I reduced them, 
as usual to groups with 40 45 and 50% etc., as a mean 
value, and found the 38 offspring of the parent with 63% 
to be distributed as folows : 
Proportion of tricotyls 40 45 50 55 60 65 ' 70 75 80 
Number of seed -parents 553 10 55221 
The mean of the series is at about 55%, and the 
series therefore constitutes a good instance of a young, 
isolated race which, however, has not yet been improved 
to anv considerable extent bv selection. 
