172 Observation of the Origin of Varieties. 
cases very short, in others very long; in some cases so 
narrow that they did not touch one another, in others 
more than twice as broad as those of the wild form. 
The color varied between golden and straw yellow, the 
tips of the florets were entire or indented, and so forth. 
This was sufficient to indicate the presence of several 
races. With regard to the number of ray-florets the 
differences were not so great, as in the mixtures we have 
already dealt with. There was only one perfectly dis- 
tinct apex, that at 21. The other at 13 was more or 
less obscured. It was obvious that the commercial race 
was the 21 -rayed one, and that it had been adulterated 
by admixture with the other only as much as is unavoid- 
able and therefore admitted in all cultivation on a large 
scale. 
The terminal inflorescences of the 282 plants of this 
culture of C. scgetuin grandiflonnn were recorded with 
the following result : 
L. F. 89 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 
Individuals 10 1 1 27 27 24 31 30 21 29 24 58 7 1 
These figures confirm what I have said and show that 
the 21 -rayed race of C. s. grandiflonnn contains an ad- 
mixture of a relatively small number (which probably 
fluctuates from year to year) of the 13-rayed race. 
My next task was to isolate from this mixture the 
21 -rayed race, whose existence had so far been merely 
inferred. I devoted the two following years to this in- 
quiry, and in the summer of 1895 selected the necessary 
seed-parents from the mixed crop. 
We here encounter an obstacle in the shape of trans- 
gressive variability, to which we have already referred, 1 
and which has often raised difficulties in the earlier in- 
1 See Vol. I, Part I, p. 56; and Part II, 25, p. 430. 
