176 Observation of the Origin of Varieties. 
This is seen at a glance from the two series that fol- 
low and from Fig. 31 B which relates to the second group. 
The data were obtained in the same way as in previous 
years, only the terminal inflorescence of the main stem of 
each plant being recorded. The character of the second 
generation of the 21 -rayed race in 1897 was, therefore, 
as follows : 
L. F. 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 32 
No. la 0012023 41 41200000 
No. 1 b 1 3 3 7 14 43 142 43 21 11 5 3 1 1 
Both groups are very symmetrical, a fact which can 
be immediately seen in Fig. 31 B which is even more 
regular than the corresponding figure of the 13-rayed 
race (Fig. 31 A). There were only 56 flowering off- 
spring of No. la but 298 of No. Ib. 
If I had not limited myself in the previous year to 
such a small number of seeds, I should have had to sow 
the seed either of less suitable individuals or from flowers 
on the same plants which had opened earlier, i. e., which 
had been pollinated with inferior pollen. In that case 
my race would have been just as incompletely pure in 
1897 as it had been in 1896. I have convinced myself, 
by special experiments with such seed, of the correctness 
of this view, but do not consider the details worth print- 
ing. 1 
By this result the isolation of the races supposed to 
exist in the mixture, was accomplished. Let us therefore 
once more examine Fig. 30 on page 167 and Fig. 31 on 
page 169. The first thing that we see is that the maxima 
are the same in both figures; they lie at 13 and 21. The 
1 Races differing in their number of ray-florets can be mixed by 
crossing (Ber. d. deutschen hot. Ges., Vol. XVII, p. 92). This mix- 
ture is an extremely interesting phenomenon in many respects, but 
needs a closer investigation. 
