Tricotylons Races Do Not Arise by Selection. 397 
shift the variability in the half race in the phis direction 
might perhaps be the same as those which would lead 
to a mutation and thereby to the sudden origin of a 
rich intermediate race. This would, in fact, perhaps, 
constitute the smallest step amongst all the possible forms 
of mutation. 1 If a mutation were ultimately to be in- 
duced by an improvement in the external conditions and 
by the choice of individuals thus modified in the desired 
direction, 2 it should most easily occur on the limits be- 
tween the half race and the corresponding intermediate 
race. My hope was, therefore, that an intermediate race 
would suddenly arise from my improved half race and 
at first give about 50%, and then, by further culture, 
yield still higher numbers, perhaps even 80 to 90%. 
This hope was based on the analogous case of the origin 
of Chrysanthemum segctum plenum from C. segetum 
grandiflorum; for in this way tricotylous races behave 
when they happen to be found and are isolated (see 6). 
A step of this kind I have, however, not yet observed. 
In the case of both species I have determined the heredi- 
tary capacity of as many as 300 seed-parents in a single 
year, but without result. I have employed every device 
of culture and in Scrophidaria I sowed the seeds of the 
second and sometimes those of the third year of the life 
of the plant; but every advance was followed by a step 
in the reverse direction. I believe that I have tried every- 
thing that was possible to me at the time, and I have 
continued to make every conceivable effort in spite of the 
1 See p. 20 ; with regard to premntation. Vol. I, p. 490 ; and with 
regard to varieties and subspecies, p. 64 of this volume. 
2 Alimentation et selection. Volume jubilaire de la Societe de 
Biologie, 1899; and Vol. I, p. 142 of this work where the statement 
is made that "selection is the choice of the best nourished individ- 
uals." 
