XIII VARIATION AND EVOLUTION 135 



transmitted to the offspring. Some of these would 

 possess it in a less and some in a greater degree 

 than the parent. If the variation were a useful one, 

 those possessing to a rather greater extent would be 

 favoured through the action of natural selection at 

 the expense of their less fortunate brethren, and 

 would leave a greater number of offspring, of whom 

 some possessed it in an even more marked degree 

 than themselves. And so it would go on. The 

 process was a cumulative one. The slightest variation 

 in a favourable direction gave natural selection a 

 starting-point to work on. Through the continued 

 action of natural selection on each successive genera- 

 tion the useful variation was gradually worked up, 

 until at last it reached the magnitude of a specific 

 distinction. Were it possible in such a case to 

 have all the forms before us, they would present the 

 appearance of a long series imperceptibly grading 

 from one extreme to the other. 



Upon this view are made two assumptions not 

 unnatural in the absence of any exact knowledge of 

 the nature of heredity and variation. It was assumed, 

 in the first place, that variation was a continuous 

 process, and, second, that any variation could be 

 transmitted to the offspring. Both of these assump- 

 tions have since been shown to be unjustified. Even 

 before Mendel's work became known Bateson had 

 begun to call attention to the prevalence of dis- 

 continuity in variation, and a few years later this 

 was emphasised by the Dutch botanist Hugo de 

 Vries in his great work on The Mutation Theory. 

 The ferment of new ideas was already working in 

 the solution, and under the stimulus of Mendel's 



