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 num1)er 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. segetwn grandifloruin were recorded with 

 the following result : 



L. F. 8 9 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. grandiflorum 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,-^ 

 and wdiich has often raised difficulties in the earlier in- 



'See Vol. I, Part I, p. 56; and Part IT, §25, p. 430. 



