The Origin of Chrysanthciniun Scgctum Plcniun. 175 



the other maximum, without however heing symmetrical. 

 Even in this group the race was therefore still far from 

 heing pure. Below I give the curve for the offs])ring 

 of the single hest seed-parent together with the sum of 

 the curves representing the offspring of the five remain- 

 ing seed-parents (Nos. 2-6). These curves therefore 

 refer to the initial culture of the 21-raycd race (1896). 



L. F. 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 



No. 1 1 3 9 15 15 22 30 33 36 64 123 15 5 



Nos. 2-6 1 5 11 12 70 84 69 92 79 77 114 150 416 46 3 1 



All in all 370 plants were recorded for No. 1 and 

 1220 foi- Nos. 2-6. 



Only the first named group, that is to say the off- 

 s])ring of the plant numhered 1 in 1895, was used for 

 the continuation of the experiment, and from it the hest 

 seed-parents for the purification of the race were selected 

 on the hasis of an examination of their lateral hranches. 

 These were two plants the lateral floAv^rheads of which 

 gave the following curves (1896) : 



L. F. 12 13 14 15 16 17 18 19 20 21 22 



No. 1« 01143220330 

 No. 1^ 000000203 14 



Of the two. No. 1/; most ohviouslv l)elonf>-s to the 

 race I was looking for. 



I harvested only the seeds of these two plants, and 

 sowed them separately in the following year. In har- 

 vesting this seed I confined myself to flowers which had 

 hloomed after the other plants had heen removed and 

 had therefore heen pollinated with their own or similar 

 l^ollen. The result corresponded with mv expectation, 

 for in the following summer the race was pure on hoth 

 beds. 



