(OMKVSIKI) < IIAHVCIKKS IN AN OVT 



At the left is a seed of llic \aricly /•'iciitel Momitaiu with long hairs on the callus at its base; 

 at the right is a short-haired specimen of the same variety. Dr. Fniwirth tried, by select- 

 ing the extreme long-haired seeds and the extreme short-haired seeds in each generation, 

 planting them separately and self-fertilizing them, to change the proportions in which the 

 two lengths were ])roduced; but he failed. He therefore concludes that long-continued 

 selection does not affect this character. Photomicrograph from the Imperial Station for 

 (Iraphic Arts, Vienna. (Fig. 1,?.) 



bred one jjarl (branch) of a ptire line of 

 oats for selection to intensify the mean 

 jier cent of bristled otiter-kernels, and 

 similarly to au*.';ment the other char- 

 acters; I bred another i)art (branch) 

 of the same ptire line for selection in the 

 reverse direction. It was not ]jractic- 

 able to l)a}.i the flowers, but the rarit>- 

 of cross-pollination in oats, and the 

 failure to obser\'e any eflfects of cross- 

 pollination in any character in any 

 {generation of the ex])eriment, warrant 

 me in c(jnsidcring my trials to rej)re- 

 sent ])ure lines. The wiriety tised is 

 that called Fichtel Mountain Oat. 

 The mean (M) and the standard devia- 

 tion (a) used in the trials and given in 

 the accompanying tables, are deter- 

 mined in the way outlined l>y Daven- 

 jKjrt," the same methods having been 

 used l)y Johannsen. who has made 

 them well known in luiro])e."* I have 

 further followed Johannsen's example 



in tising the mean error of the mean 

 (Mm) and mean error of the standard 

 deviation (Ma) instead of the probable 

 error tised b>' Dax'cnport (probable 

 error = mean error X 0.6745). 



The results of selection on the bristled 

 outer-kernel are given in full in Table I. 



CHANGES NOT HEREDITARY 



Of course, one may well think that he 

 sees in my table an elTect of selection. 

 The mean in the first generation, a 

 generation not selected, is 5.11''^; the 

 mean after eight minus-selections is 

 2.65, after eight plus-selections is 3.14. 

 But I reckon such an effect as not a 

 hereditary elTect. A change in the 

 hereditary factors, if one were produced 

 through selection, ought to augment 

 from year to year if the selection is 

 continued; at least, it shotild not drop 

 back. As in IIoj)kins and Smith's trials 

 with open-fertilized corn, the numbers of 



' ///. Af>ric. Exp, Sla. Hull. 110, 1907. 



'« Elementc der cxakten Krblichkeitslehrc, 2d ed., 191.^. 



92 



