Tricotylous Half Races. 389 
attained its maximum in lS f /c. The latter occurred 
amongst the 300 seedlings from the seeds of a daughter 
plant of one of the two tetracotylous grandparents. 
In the last three generations the maximum ratios 
were, therefore, 12, 11 and 15%, indicating no advance 
of any significance. 
Polygomim Convolvulus. A tricotylous plant flow- 
ered in 1888, in isolation, in my garden. Its seeds gave 
rise to normal seedlings only (1889). From their seeds, 
about 4000 in all, 1450 seedlings were raised, and 12 of 
them were tricotylous, i. e., \% (1890). I harvested the 
seeds of six tricotylous plants separately, and obtained 
ratios from 1% to 2.4% in sowings of about 1000 seeds 
each. Of these, 12 seed-parents produced only about 
1% in the next generation in the spring of 1892; two 
of them, however, produced 1.5 and 2%. Seventeen 
plants were planted out. In their crop (April 1893) the 
proportion of tricotyls varied from 0.5 to 2% in lots of 
200 and 400 seedlings, and twelve tricotyls succeeded in 
flowering. The next generation (April 1894) contained, 
in the best cases, 2.8% tricotyls; the next one, to 2% 
from 8 seed-parents (1895) ; and the last, or ninth, again 
2% only. 
Silcnc conica. In 1892 I had a few tricotyls in 
flower from seed received by exchange from another 
botanical garden. Their seed gave 3 tricotyls amongst 
1000 seedlings. I planted these out, together with some 
dicotylous seedlings, and in May 1894 I had from 0.2 
to 1% tricotyls in every lot of 350 to 800 seedlings. 
Eight tricotyls were planted out and in the spring of 
1895 their harvest gave a proportion of only 2% and 
less. From these I obtained in 1895 a fourth generation, 
