510 The Inconstancy of Fascia ted Races. 



mentioned specimens as seed-parents and destroyed the 

 rest before they flowered. 



These three plants therefore formed the l)eginning of 

 my race. Denoting the wild specimens of 1886 as tlie 

 first generation, the second grew in 1887 and 1888, the 

 third in 1889 and 1890, and so on. This third one con- 

 sisted of 120 plants, of which 48 or about 40 ^r already 

 exhibited a comb-shaped linear growing point in tlie 

 heart of their rosettes in the winter. This comi) was in 

 some specimens as much as six centimeters long. I se- 

 lected the three finest fasciations as seed-parents and re- 

 moved the rest before flowering. The fourth generation 

 raised from this seed produced a slightly lower propor- 

 tion of fasciations, containing, as it did, only 30% of 

 them. In the fifth generation a further reduction took 

 place, viz., to 24%.^ The sixth generation (1895-1896) 

 was very rich in fasciations, and in ten plants out of the 

 40, produced a growing comb of from 4 to 7 centimeters. 

 Unfortunately the remaining fasciations of less degree 

 mostly died in the winter before they were recorded, so 

 that an exact percentage value can not be given. Never- 

 theless it was obvious that the character of the race had 

 been displayed oftener than in the previous generation. 

 In the following, viz., the seventh generation, I also re- 

 corded only the expanded rosettes before the winter, and 

 found ten of them amongst 49 plants, that is about 20%. 

 The eighth generation was not sown till 1902, and only 

 on a verv small scale. The combs became visible during 

 the winter 1902-3. 



Tf we summarize the results which we have described, 

 we see that the seed collected in the field, without selec- 

 tion, gave about 12% fasciated ofi^spring, whilst the seeds 



* Bofanisch Jaarhoek, Gent, 1894. p. 80, and 1897, P- 66. 



