388 



METAMORPHOSIS 



between artificially produced races and natural species : the former are transi- 

 tory, the latter are constant. 



It has been said above that the most important point about a variation is 

 whether it is hereditary or not. Are these individual variations hereditary or 

 not ? This is a question by no means easy to answer. The beetroots, from 

 which the graphic curve given at Fig. 114 was taken, arose from seed, which 

 had been obtained from plants possessing 16-18 per cent, of sugar, and yet by 

 far the majority of them have less than 16 per cent., and individuals them- 

 selves differ very greatly. We have to deal here only with a ' partial ' in- 

 heritance, since only some of the characters of the parents reappear in the 

 offspring. On the whole, this is probably not a case of heredity. Our posi- 

 tion in relation to this question will be more secure after we have successfully 

 determined the factors which induce individual variations. Meanwhile, we may 

 suggest a hypothesis which may be considered as highly probable. Fluctuating 

 variations must arise from irregularities in growth conditions (nutrition in the 

 widest sense of the word) which must arise even in the most carefully conducted 

 experiments. [Compare Klebs, 1903.] Indeed, one cannot prevent one plant 

 from taking up more water or mineral constituents from the soil or absorbing 

 more light than another ; similar organs on the same axis must be differently 



nourished according to their position, 

 for it may be readily proved that the 

 flow of nutritive materials to a terminal 

 organ cannot be equal to that which 

 reaches a similar organ in a lateral 

 position. The lateral inflorescences of 

 the sunflower, for instance, become 

 considerably larger when the terminal 

 inflorescences are removed, and many 

 examples of like character might be 

 quoted in support of this view. 

 When one flower is better nourished 

 than another larger seeds will natur- 

 ally result, as also larger embryos 

 and more abundant reserves. Again, 

 a plant which has arisen from a large 

 seed will in turn tend to grow more 

 vigorously, produce a greater number 

 of seeds, and develop greater quantities of sugar. If individual variations 

 be really due to nutritive influences only, we cannot speak of an inheritance 

 of such peculiarities in a single individual, and, at the same time, readily 

 account for the rapid disappearance of these characters when selection is dis- 

 continued. Further, it is conceivable that selection may be aided or even re- 

 placed by good manuring. [As the result of Johannsen's (1903) researches it is 

 clear that even races which breed quite true are always the result of the mingling 

 of several forms differing from each other in minute but constant characters. 

 If we start from the offspring of a self-fertilized plant such offspring also exhibit 

 differences among themselves, which, when plotted out, give a graphic curve 

 such as that shown in Fig. 114. If from among such forms it be desired to 

 produce a breed with a certain definite character, and if we select for that 

 purpose seeds for a third generation from plants standing at the very beginning 

 or the very end of the curve, it will be found that the tivo sets of plants result- 

 ing present no differences ; in fact, selection, in this case also, is quite useless. 

 If selection appears to have a different effect in the other experiments already 

 spoken of, it must be assumed that we are dealing in these cases not with indi- 

 vidual variants but with mutation phenomena. Compare also Correns, 1904.] 



16 I 17 I 

 r, iG-o 17-. 



I li' I 



Percentage 

 of Sugar. 



Fig. 114. Percentage of sugar in 40,000 beets, ex- 

 pressed graphically. DeVries (Mutationstheorie, I). 



