398 Tricotylous Races. 
fact that the goal became ever more obviously unattain- 
able; but nothing I did was of any avail. The half races 
remained half races, and the prospect of attaining an 
intermediate race is no greater now than it was at the 
beginning. 1 
I shall proceed now to the description of my two ex- 
periments. 
Amarantus spcciosiis (Fig. 83). In 1889 I had a 
tricotylous and a hemi-tricotylous specimen of this fine 
garden plant, which usually attains a height of from 
1 1 /2 to 2 meters. They flowered together, but far removed 
from any other specimens of the same genus. 2 I har- 
vested the seeds separately, but only sowed those of the 
hemi-tricotylous plant. This had a small comb-shaped 
extension of the summit of the raceme as is shown in 
Fig. 83, and as it afterwards often occurred in this race. 
Its seeds gave rise to four tricotyls and one hemi-tricotyl 
amongst 110 seedlings, that is, a proportion of about 
4.5%. The degree of inheritance therefore proved to 
be greater than in most of the other half races investi- 
gated (4). The culture was, however, unfavorable, 
since only one hemi-tricotylous plant could be brought 
to flower, and since therefore a number of atavists had 
to be cultivated with it as a precautionary measure ; but 
T only saved the seed of the hemi-tricotylous plant. This 
produced 6 tricotylous and 5 hemi-tricotylous seedlings 
amongst 250, that is, about 4.5% or the same amount 
as in the previous generation. This time, however, I 
could plant out the aberrant forms only and I managed 
to keep the majority of them alive. Only one, however, 
1 Mutations must, nevertheless, have external causes, and these 
must be found some day, but perhaps by some other means. 
2 Amarantus sficciosus is regarded by some authors as a sub- 
species of A. paniculatus. 
