INHERITANCE IN PURE LINES 



377 



4-000. 



.,75,000 



2000 _ 



in the two species. Hence a comparison of variation in two 

 cases, where the modes are dissimilar, can only be made by 

 taking such dissimilarity into consideration, a comparison of 

 the curves alone being misleading. The position of the mode 

 can often be to some extent changed by modifying the environ- 

 mental conditions, 

 although the range of 

 variation remains 

 practically unaltered. 

 It appears that 

 individual variations 

 are not inherited, and 

 that in a pure line 

 of descent the average 

 of the race is main- 

 tained. ^ This has 

 been established by 

 sowing seeds from a 

 single individual of 

 pure descent which 

 has been self-ferti- 

 lised, the seed pro- 

 duced by the result- 

 ing plants (likewise 

 self-fertilised) being 

 sown in separate 

 groups, and the same 

 procedure followed 

 for several successive 

 generations. It was 

 thus found that the 

 average size of in- 

 dividuals derived from small parents is practically identical 

 with that of individuals growai from tall parents ; also that 

 heavy seeds do not beget heavier-seeded offspring than those 

 derived from light seeds. The following data, which serve 



J 



4 56 7 5^10 

 Perianth se&? 



(HlMEMOlME) 

 G. U. YULE 



Fig 



10 \5 20 



RPvYS OF STIG^AR 



(PnPRVEFk) K.TEHRSON 



2 20. — Variation curves, symmetrical oa 

 the left and asymmetrical on the right. 



' By a pure line is understood a pure-bred strain produced by self- 

 fertilisation from a single individual. 



