Correlation and Application of Statistics to Problems of Heredity 93 



basis of a national religion, in the sense of that word as defined by J. S. Mill, for, though it be 

 without any ultra-rational sanction, it would serve to 'direct the emotions and desires of a 

 nation towards an ideal object, recognized as rightly paramount over all selfish objects of 

 desire.'" (pp. 761-3.) 



I trust this long citation will not have wearied the reader; for his 

 biographer it contains some of the most important lines Galton ever wrote. 

 There is no reason to be afraid of plain words. Man has learnt how to breed 

 plants and most inferior forms of life that are of service to him. He has yet 

 to learn how to breed himself When he has studied heredity and environ- 

 ment in their influences on man, the application of the laws thus found to the 

 progressive evolution of the race will become the religion of each nation. Such 

 is the goal of Galtonian teaching, the conversion of the Darwinian doctrine 

 of evolution into a religious precept, a practical philosophy of life. Is this 

 more than saying that it must be the goal of every true patriot* ? 



L. Miscellaneous Papers on Evolution, Heredity, etc. We may now 

 turn to a series of short papers by Galton, chiefly published in Nature, 

 and dealing with hereditary and evolutionary topics from 1897 onwards. 



The first paper we note is entitled: "Rate of Racial Change that accom- 

 panies Different Degrees of Severity in Selection," and will be found in 

 Nature, April 29, 18.97 (Vol. lv, p. 605). This is an important paper, 

 because it deals with the effect of continued selection in modifying a variate 

 continuously distributed in a population. Galton starts with his two-thirds 

 regression of the offspring on the midparent for stature and the reduction of 

 the variability of the offspring of such midparents in the ratio of T5 to 17 

 inches. He then continues to select both parents at the 99th, 95th, 90th, 

 80th and 70th grades for 1, 2, 3, 4, 5 and an infinite number of generations 

 in order to determine the progression there would be in stature by such 

 continuous selection. Galton, unfortunately ignorant of the formulae of 

 multiple regression, makes three erroneous assumptions, namely: (i) that the 

 regression between each generation is the same, namely §, notwithstanding 

 the earlier ancestry being as we advance more and more selected; (ii) that 

 the variability of the array of selected ancestry remains for the later genera- 

 tions the same as for the first selected generation ; this is of course incorrect, 

 the variability steadily diminishing towards a finite limit; (iii) that if the 

 selected race be now left to itself, it will regress indefinitely to the old 

 general population mean : 



" It must be borne in mind, that there is no stability in a breed improved under the supposed 

 conditions ; but that as soon as selection ceases it will regress to the level of the rest of the 

 population through stages in which the deviation, at starting, sinks successively to w, w t ...w n of 

 its value f. It may, however, happen that a stable form will arise during the process of high 



* Some may question whether we have more here than in Comte's Religion of Humanity. 

 I think so, because it is freed of the ceremonialism which Comte and Gruppe demanded as 

 a factor of religion, and it is essentially based on the acquirement of knowledge in a field of 

 science, which had little if any existence in Comte's day. 



t w is Galton's regression coefficient in the case of selected midparent, with no selected 

 previous ancestry. 



