200 CERTAIN LAWS OF VARIATION. 



seed, when developing on the maternal plant, has at 

 least very often a greater influence on the variability 

 than nutrition during its germination and later growth. 



De Vries does not quote any concrete instance in sup- 

 port of the first of these conclusions, but in proof of the 

 second, he adduces some observations on CEnothera La- 

 marckiana.* On sowing some seed of this plant in 

 well-manured ground, he found that the seed capsules 

 of the plants obtained were slightly smaller than those 

 from seeds grown in unmanured ground (the mean 

 lengths being respectively 25.2 and 27.2 mm.). The 

 seeds of three of the manured plants, when sown next 

 year, yielded plants with seed capsules, on an average, 

 29.9 to 33.4 mm. in length. This considerable in- 

 crease seems to have been almost entirely due to 

 the manure treatment received by the plants from 

 which the seeds had been derived, and in compari- 

 son with it, artificial selection was not nearly so 

 effective. Thus the (well-manured) plants arising 

 from seeds taken from long seed capsules (32.6 to 43.0 

 mm. in length) had themselves capsules of only 31.6 to 

 33.4 mm., whilst those from the seeds of short seed cap- 

 sules (15.6 to 23.4 mm.), had capsules of 24.2 to 29.9 

 mm. 



Doubtless the numerical measure of the effect of en- 

 vironment on growth varies enormously in different 

 organisms, and the extent of its variation can be deter- 

 mined only by experiment. Fortunately, however, 

 numerous data which have been collected as to the rate 

 of growth in certain animals, more especially in man and 

 the guinea-pig, can be readily applied to the partial 

 *Loc. cit., p. 383. 



