194 Observation of the Origin of 1 \irieties. 
excluded by the sterility of the more perfectly double 
plants. 
Let us now briefly summarize the results of this ex- 
periment. There is, on the market, a 21 -rayed race of 
the normally 13-rayed Chrysanthemum segetnm. It is 
not strictly pure, but can easily be made so; it bears the 
name C. segetnin grand iflonmi. From a plant which, 
in 1895, caught my attention by a few 22-rayed lateral 
flowers, I succeeded in raising, by a process of selection, 
a hitherto unknown race with double flowerheads, the 
new C. segctwn plenum (Plate II). The course of this 
process is exhibited in Fig. 32, p. 181, in which the 
X X X X indicate the individuals selected as seed- 
parents for each succeeding generation. C. segetuin 
plenum behaves with regard to its double character, 
exactly like the double commercial varieties of other 
species of the same genus (C. inodonun, C. indiciun etc.). 
The new variety was therefore obtained by bringing 
to light a character latent in C. segetnin grandiflorum. 
19. DOUBLE FLOWERS AND FLOWERHEADS. 
The experiment described in the foregoing section 
( 18) justifies an attempt to form some conception of 
the manner in which this phenomenon of "doubling," 
which is widely distributed among cultivated composites, 
may have arisen in other cases. If we examine the facts 
closely we shall discover in the majority of cases an 
extraordinarily close agreement with our own specimen, 
at least so far as the absence of scientific observations 
admits the possibility of a comparison. 
There are, it is true, certain abnormal types of "doub- 
ling." such as the development of secondary flowerheads 
