AND PLANTS 97 



concerning inheritance which have been reached by its 

 means. 



All experience, which we are obliged to deal with stat- 

 istically, is experience of results which depend upon a 

 great number of complicated conditions, so many and so 

 difficult to observe that we cannot tell in any given case 

 what their effect will be. Consider the height of a man. 

 This is a result depending on the length of each of a very 

 large number of bones, cartilages, and other structures, 

 and the length of each of these is determined in its turn 

 by a large number of conditions, among which are the 

 length of the corresponding part in the father and in the 

 mother. When you know the height of a man's father, 

 you know one out of a number of sets of conditions which 

 have helped to determine the height of the man himself, 

 but you are still in ignorance of the rest : and the student 

 of inheritance is always in this position, that he is trying 

 to find out how much he can learn to predict about results, 

 when he knows only a few of the conditions on which 

 these results depend. 



In order to show you the kind of knowledge one can 

 obtain in this way, and the limits to its exactness, I must 

 ask you to let me trouble you once more with dice. I 

 have here two dice, one white and the other red. Sup- 

 pose I want to get more than three points with each when 

 I throw them. I throw them, and try. I get a certain 

 result. If I throw them both a second time, I am no 

 more and no less likely to get more than three points with 

 each of them than I was the first time, no matter what 

 the result of the first throw may have been. The result 

 of the second throw is quite independent of the first 

 result. 



Now I want to make the result of the second throw 

 partly dependent on the first, and I do so in this way. 



- H 



