186 Observation of the Origin of Varieties. 
instance, a terminal flowerhead with 56 rays had 53 on 
the periphery and 3 inside ; while one with 74 rays had 
58 on the circumference and 16 in the disc. In the 
records on which were based the table on page 180 and 
Fig. 32, both kinds of ligulate florets were counted to- 
gether. The two flowers with 99 and 101 ligulate flo- 
rets respectively were to all appearance almost entirely 
double. 
The "doubling" was also exhibited on the lateral 
branches. When these were in full flower, I selected the 
twelve best "double" plants and pulled up the rest. The 
lateral inflorescences of the rejected plants gave a curve 
whose maximum was at 47 (=13 + 34) in accordance 
with the indications referred to above and apparent in 
Fig. 32. The worst flower had only 28, the best one 
94 rays. The average of the whole lot was 47; but the 
curve, in spite of the coincidence of the mean and the 
maximum was not symmetrical. Altogether the rays of 
378 inflorescences were counted. 
As was to be expected, the selected seed-parents ex- 
hibited great differences in the degree of "doubling" in 
the lateral inflorescences. On some this was inconsider- 
able. On others the mean was from 2-5 disc-tongues 
per inflorescence whilst on two a mean of 1 1 was counted. 
One plant bore nothing but wholly double flowers. It 
had seven flowers on which 279 disc-tongues were 
counted, giving an average of 40. In consequence of 
this the plant was absolutely sterile; it bloomed well 
afterwards, but in spite of every care I could not get 
a single seed from it. But the finest specimens of C . in- 
odoritm plenissimum are also known to set no seed. Like- 
wise the two plants whose lateral flowers had on the 
average 11 disc-tongues, set no seed. 
