Double Flowers and Flowcrheads. 197 
the ligulate florets are female, and inasmuch as very 
often all the flowerheads on a single plant attain to this 
degree of doubling the best variants cannot serve as 
seed-parents. But two further types are always found 
with yellow discs which are either uniform (Fig. 34A) 
or contain scattered ligulate florets amongst the tubular 
ones, as is so often seen in Chrysanthemum indiciun and 
Zinnia clcgans. The double Bcllis perennis also, if grown 
from seeds, is highly variable in this respect. These 
two types are fertile and therefore constitute the seed- 
parents of the variety; if the plants with central ligulate 
florets (see p. 185) furnish sufficient seed the harvest 
is saved exclusively from them ; but they often set little 
or hardly any seed. 
This unavoidable restriction in the choice of the seed- 
parents and the frequent difficulties of selection depen- 
dent on it account for the fact that bought samples of 
the seed of double composites often give rise to only 
a relatively small proportion of the desired type, as has 
long been known 1 to be, and still is, the case (Chrysan- 
themum coronanum sometimes only 50%, Ccntaurca 
Cyanns 40-50%, Tagctcs africana with rare exceptions 
double etc.). 2 
Many double varieties of composites seem to be al- 
most as old as horticulture itself (See Vol. I, p. 183). 
According to the oldest accounts the degree of doubling 
and the range of its variation were formerly the same as 
now. 
Finally I have to mention the fact that bud- and sec- 
torial variations are found in this case as well as in others. 
*E. g., Pyrethrum roseum, Dahlia, Chrysanthemum indicum, ac- 
cording to VERLOT, Production ct fixation des varictcs, 1865, P- 83. 
2 See the catalogues of BENARY, and HAAG-E-& SCHMIDT of Er- 
furt, VEITCH & SONS of London and SUTTON & Co. 
