Correlation and Application of Statistics to Problems of Heredity 79

the case of recording disease. Much of this the reader who wishes to go farther than our pages will find more easily here than in the original papers. Certain numerical misprints in the tables require that they should be carefully examined before use.

J. Discontinuity in Evolution. In 1894 appeared William Bateson's Materials for the Study of Variation, treated with especial regard to Discontinuity in the Origin of Species. One of the strange misconceptions with regard to Galton's views and to his work lies in the fact that he has been over and over again considered as the propounder of the view that evolution has taken place by the selection of slight or continuous variations. As a matter of -fact Galton had for some years before the appearance of Bateson's book been preaching emphatically the doctrine of discontinuity in evolution. Indeed his opinions on the manner of evolution date back to 1872: see our Vol. ii, pp. 84, 170-174 and 190. They are more clearly expressed in the preface to the 1892 reprint of Hereditary Genius. There we read

"Another topic would have been treated more at length if this book were rewrittennamely the distinction between variations and sports. It would even require a remodelling of much of the existing matter. The views I have been brought to entertain since it was written, are amplifications of those which are already put forward in pp. 354-5*, but insufficiently pushed there to their logical conclusion. They are that the word variation is used indiscriminately to express two fundamentally distinct conceptions: sports and variations properly so called. It has been shown in Natural Inheritance that the distribution of faculties in a population cannot possibly remain constant if on the average the children resemble their parents. If they do so the giants (in any mental or physical particular) would become more gigantic and the dwarfs more dwarfish, in each successive generation. The counteracting tendency is what I called 'regression.' The filial centre is not the same as the parental centre but it is nearer to mediocrity; it regresses towards the racial centre. In other words the filial centre (or the fraternal if we change the point of view) is always nearer, on the average, to the racial centre than the parental centre was. There must be an average `regression' in passing from the parental to the filial centre." (pp. xvii-xviii.)

The flaw in Galton's argument is again one that we have had several times to notice, namely that he is overlooking the fact that he has clubbed together parents of all possible types of ancestry, and the " regression " of his sons is solely due to the large number of such parents who have sprung from an ancestry mediocre or below mediocrity. The amount of filial regression depends entirely on the amount of this mediocrity, and there will be no regression if two or three generations above the parents are of like deviation from mediocrity. Thus although it may still be a matter for experiment and discussion, whether evolution proceeds by variations proper or by sports, whether it be continuous or advance by jerks, the reason which made Galton the pioneer in advocating discontinuous evolution was a misinterpretation of his own discovery of " regression."

* These pages deal with Galton's idea of the stability of types : see our p. 61 and Vol. ii, p. 1l 3. It is quite reasonable to suppose that by successive selection of extreme variations proper we might reach a position of unstable equilibrium of the parts of an organism. But there does not exist experimental evidence at present to indicate that such instability would lead to a sport breeding truly rather than to non-viable forms of the organism. See our pp. 93-4.