Peoblems of Hbtekogeneity 187 



combinations, in the physical material itself. It must be understood that these 

 terms are purely descriptive, indicating only an observed condition in the hetero- 

 geneity present in species. 



Any particular instance of determinate change may be so because it is also 

 delimited, but not necessarily so ; because if determinate, the environment may 

 determine the response, if delimited, the organism. The same kind of organisms 

 might, if undelimited, exhibit determinate variation ia different directions in 

 different portions of its natural habitat in a given character, while a delimited 

 character would show determinate variation always the same under different 

 environments. The effect which these relations produce in the population 

 responses of a species is truly important when one considers organisms in 

 nature, and is beautifully demonstrated in some of my experiments in trans- 

 planting populations into new environmental complexes. The earlier writings 

 of the Darwinian period much emphasized these variation phenomena, but made 

 no experimental test of their effects. 



With the development of modern methods of genetic analysis, attention of 

 necessity was more and more focused upon the individual and upon the char- 

 acters thereof, so that now our problems are almost entirely those of the iadivid- 

 ual and its characters and not so much the mass action of the population. 



In the development of the mutation theory, De Vries created added compli- 

 cations in the problems of heterogeneity. The familiar assertion that there are 

 really two distinct kinds of variation — mutations (sports) and fluctuations — 

 seems an attempt to separate by definition the two extremes of " variation " 

 which Darwin recognized but was unable to separate. If mutations are the salta- 

 tion-like changes by which a new unit-character is added or an old one taken 

 away or made latent or recessive, and these steps are multifarious in direction, 

 and if the ordinary " variations " are quantitative, adding to or subtracting 

 from the character in question, then all such variations are always in Hne, like 

 the swing to and fro of a pendulum, and this is of necessity made so by the 

 definition and the method of description which is mathematical in terms of 

 measure. Associated with this idea is the further assumed difference between 

 the two kinds of variations — ^that mutations are qualitative and fluctuations are 

 quantitative. 



The sharp distinctions made by De Vries have seemed to many a questionable 

 subterfuge by which the unit-character concept becomes more clearly differen- 

 tiated, and of the two kinds of heterogeneity one, fluctuation, is rendered inoper- 

 ative in evolution, while mutation bears the entire burden of evolution change 

 and is due to transmutation from present to subsequent unit-characters. Even 

 though this proposition could be shown to be true, it does not follow that fluctua- 

 tions in a imit-character are always plus and minus, and that the mutations of a 

 unit-character are always multifarious. If there are unit-characters, as con- 

 ceived by De Vries, each must have relations in space and in interactions, and it 

 is hardly to be supposed that a unit-character is so fixed, so stable and invariable 

 that no matter where placed it is only a little smaller or larger, depending upon 

 the conditions of growth present during its development. 



According to Weismann, all variation is ultimately quantitative, due to an 

 increase in the number and vigor of particular determinants, which increase is 

 due to quantitative increase in the determinants. If this be true, it follows 

 that the difference between species, or between their characters, are quantitative, 



