440 Tricotylous Races. 
cotylous seedlings from seed which had been bought or 
obtained from some other source, or with plants which 
had been found by chance (Oenothera, Silenc). Where 
necessary, this year is denoted in the table by I (first 
generation). The numbers which succeed each other in 
the row to the right of it refer to the first and the follow- 
ing generations. Thus for instance in Clarkia there were 
as many as 16% tricotyls in the harvest of the first gen- 
eration, 64 in those of the second and 79 in the third. 
The point which this table is intended to illustrate 
is best brought out by a comparison with the series of 
figures given on page 392 for the half races. In that 
case, a selection continued from four to six years, did 
not bring this value, as a rule, further than 2 to 4%, and 
only in exceptional cases attained 15 to 20%. In this 
case, on the other hand, 55% is attained in two or three 
generations. In the half race a continuation of the 
selection would presumably not have led to any con- 
siderable increase, a fact which is demonstrated by the 
experiments with Amarantus and Scrophularia which 
were continued over a longer period of time. In this 
case, however, selection is as a rule very effective, inas- 
much as it can increase the best representatives of the 
race, in a very short time, to a hereditary capacity of 
80 to 90%. 
Therefore there can be no doubt that entirely differ- 
ft 
ent factors are at work in these two cases. In the former 
there occurred only races with half curves on which se- 
lection has little effect. In this case, however, there oc- 
curred, besides these, the highly variable intermediate 
races which are extraordinarily susceptible to selection 
and to external conditions of life. They were easily 
isolated, either because one or more examples of them 
