ioo INHERITANCE IN ANIMALS 



element taken at random ; this addition is roughly analo- 

 gous to the addition of elements during fertilization ; so 

 that the relation I have established between my second 

 result and my first is roughly analogous to that which 

 Weismann's theory leads us to expect between a child and 

 one of its parents. It is still more closely analogous to 

 the relation indicated as a result of direct statistical study 

 of inheritance by Mr. Francis Galton. 



I mention this to show you that the illustration of 

 heredity by such a rough experiment as I have described 

 has more justification than some of you may realize at 

 first. 



The experiment itself must be repeated many times, 

 and on a larger scale, in order to illustrate the points I 

 want to show you. With my wife's help, this has been 

 done. She and I took not two only but twelve dice, and 

 painted six of these red. Tossing all the twelve, we 

 counted the number which showed more than three points, 

 and so obtained a first result The six red dice were 

 then left undisturbed, the white dice were thrown again, 

 and we had a second set of twelve, six carried over 

 unchanged from the first throw, and six in a condition 

 independent of that observed after the first throw : the 

 number of dice with more than three points in this set 

 of twelve gave us our second result. In this way we 

 obtained a pair of results, half the elements which deter- 

 mined either member of the pair being common to it and 

 to the other member. When we had obtained such a pair 

 of correlated results, we tossed all the dice thoroughly, so 

 as to destroy any effect of their previous arrangements, 

 and went through the whole process again, until we had a 

 second pair. We did this 2 12 = 4,096 times, until we had 

 4,096 pairs of results, each result determined by some 

 elements common to itself and to the other member of 



