The Factors of Evolution 241 



Since tlie publication of DeVries' work on the mutation 

 theory the attention of biologists has been focnssed more and 

 more on this factor of species building, and large numbers 

 of cases have been recorded in both animals and plants. 

 These cover the whole garant of variation and are both (juali- 

 tative and quantitative in kind. They include variations in 

 color, markings and size ; in form ; length and number of 

 hairs, bristles, etc. ; ditferences in shape of parts such as leaves 

 of plants, wings of insects, horns of cattle, etc. ; differences 

 in habit of plants, whether erect or procumbent, straight or 

 branching, etc. ; differences in the number and union of parts, 

 such as syndactylism and polydactylism in man and other 

 animals ; differences in presence or absence of parts, such as 

 absence of a tail in cats and chickens ; in fact, virtually every 

 character known of either animal or plant is liable at some 

 time or other to show mutation. INIutations vary moreover 

 in size, from the large "sports" already mentioned, down to 

 variations so small that it is mere hair-splitting to attempt 

 to distinguish between them and Darwin's "fortuitous" vari- 

 ations on the basis of size alone. 



What distinction then if any can be made between these 

 two classes? Darwin did not attempt to define "fortuitous" 

 variations, but speaks of the term as serving "to acknowl- 

 edge plainly our ignorance of the cause of each particular 

 variation." His distinction between them and "sports" or 

 mutations was one of size alone. More recently however we 

 have learned to recognize in the small "fortuitous" varia- 

 tions of Darwin two distinct kinds of variability, one of which 

 we call "continuous" or "fluctuating" and the other "dis- 

 continuous" or "mutating." If the sizes of any group of 

 organisms, let us say men, be plotted on a chart, as Ave 

 shall find a point which is called the mode, at which 

 fall a greater number of measurements than at any other, 

 while the remainder graduate to either side of this point 

 until the limits of the series are reached. Such a chart is 

 called a frequency polygon (or curve, if the measurements are 

 sufficiently numerous and close together). Such a chart rep- 

 resents graphically "fluctuating" variation, the variations 

 falling indiscriminately on either side of the mode. If indi- 

 viduals from either end of the series be mated, their off- 

 spring will be widely diff'erent from the average of the series, 

 but will nevertheless approach that average more nearly 

 than their parents, and no amount of selection will serve to 

 raise or lower the limits of the series, that is, to increase the 

 original variability. In the cases of Johannsen's beans and 

 Jenning's Parama?cia and similar cases above cited however, 

 selection does serve to sort out groups of individuals whose 



