176 Observation of the Oriyin 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. 1« 0012023 41 41200000 

 ^o.l 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. 1^. 



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 tliat 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. ^ 



By tin's 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 



^ Races differing in their number of ray-florets can be mixed by 

 crossing {Bcr. d. dcutschcn hot. Ges., Vol. XVJI, p. 92). This mix- 

 ture is an extremely interesting phenomenon in many respects, but 

 needs a closer investigation. 



