10 



Life and Letters of Francis Galton 



falling vertically through a forest of horizontal pins. He next, starting with 

 a normal distribution of variability or,, reduces the variability to r<r 1 by 

 sloping his discharge tubes towards the type (see Fig. 2). This restriction 

 of the tubes has the same effect as giving a uniform horizontal "squeeze" 

 to his original distribution ; he thus reaches his population of " reverted 

 parents." If he now opens any single one of his tubes he will get a normal 

 distribution, about the reverted parent character as type, which will have 



Fig. 3. Galton's Quincunx illustrating the effect of Natural Selection. 



the proper variability a^Vl — r 2 if a suitable choice be made of the extent 

 of " pin-forest " through which the pellets fall. Since this variability is the 

 same for all parentages, the extent is constant, and if all the tubes be opened, 

 all the " reverted " parentages contribute their share to building up again 

 the population from which we started. 



Those who hold the hypothesis of the pure line to be true, apparently 

 overlook the fact that while the gametic distribution might be stable, they 

 must appeal to a stringent natural selection, or a differential fertility, to 



