554 Length of the Fruit in Oenothera Lamarckiana. 



the end of September I had 69 plants with ripe fruits. 

 Their lengths were measured in the usual way on the 

 lowest five good fruits; and the figures thus obtained 

 are given in the 1893 column of the table on page 553. 



I now had to make a sowing without reference to 

 fruit-length. I sowed the seeds of nine plants whose 

 mean fruit-lengths were 24.2, 26.2, 26.8, 27.0, 27.4, 29.0, 

 32.4, 34. 6, 35.2 mm. The mean of them was therefore 

 28.1 mm., which is below the median of the whole 1893 

 crop (see below) whilst only the last two figures lie 

 above the upper quartile. 



Sowing, pricking out, manuring and cultivation were 

 the same in 1894 as in the previous years. At harvest 

 time I had 88 plants with ripe fruits; from 7 to 19 from 

 each parent plant, with the exception of the three parents 

 with the longest fruits which gave rise to only 4, 5 and 

 6 plants respectively, whereby their effect on the curves 

 was diminished considerably. The measurements of the 

 mean fruit-lengths were again made in the same way. 



The numbers of individuals in the various groups of 

 fruit-lengths are given in the table on page 553. 



The mean fruit-length has, it will be seen, increased 

 considerably : this is also well seen by looking at curve 

 D in Fig. 116 and comparing it with the corresponding 

 curve for 0. Lamarckiana, a perfectly legitimate pro- 

 ceeding inasmuch as both species exhibit the same length 

 of fruit under similar circumstances. The greatest length 

 attained in the experiments with 0. Lamarckiana was 

 33.4 which was reached by manuring the parent plants 

 sown in the garden, and by the selection of the longest 

 fruited plant as seed parent. 



This enormous increase holds good not only for the 

 mean but for all individuals. For the amplitude of 



