EVOLUTION 



393 



I take a third capsule from a third poppy, its most prob- 

 able number of stigmatic bands will be lo, for that is 

 the mode of wild poppies in general. But if this third 

 poppy happened to be identical with the first or the second 

 selected, the most probable value would not be lo, but 

 some number nearer to 7 in the first case or to 13 in the 

 second case, owing to the resemblance of parts in the indi- 

 vidual. How much nearer is answered by the numerical 

 theory of correlation. 



Now we shall see best the basis of this theory if we 

 take an actual case. I take 5513 pairs of wild poppy 

 capsules gathered on the top of the Chilterns, each pair 

 belonging to the same plant. Then I select out of these 

 pairs those in which one member has 5 stigmatic bands : 

 there are 1 2 such cases ; in 8 of them there were 6 

 stigmatic bands, and in 7 of them 4 stigmatic bands on 

 the other member of the pair. Now I take out of my 

 5513 pairs all in which there were 6 stigmatic bands in 

 one member ; there were 133 of these arranged as 

 follows : — 



Here those having 6 and 6 in the pair are reckoned 

 twice, once for the first member and once for the second, 

 as either might have been taken first. 



Similarly if we take out the pairs with one member 

 with 7 bands we find 661 of them, thus arranged : — 



Now a very slight inspection shows us that as we 

 increase the number of stigmatic bands in one member of 

 a pair, the mode of the other member also increases, for 

 5 it was 6, for 6 it was 7, but the frequency of 6's is 

 greater than that of 8's ; for 7 it was 7, but the frequency 

 of 8's is much greater than that of 6's. In other words. 



