SELECTION OX THE VARIABILITY AND CORRELATION OF ORGANS. 
45 
Thus we have the following values of if :— 
Pi = 1. 
gi = - 8. 
/ii= -6. 
Pi = - 4. 
Pi=-2. 
/Xi = 0. 
y= 2 
1-1600 
1-1087 
1-0642 
1-0296 
1-0076 
1 
= 10 
2-4400 
1-9779 
1-5775 
1-2662 
1-0681 
1 
= 50 
8-8400 
6-3243 
4-1448 
2-4492 
1 - 3705 
1 
Lastly, if r' = 0 : 
^15 hi/pi 2 is given by : 
1 
1 + 
21 1 , 
4 
pi = 1. 
pi= ‘B. 
/xi= -6. 
Pi= '4. 
/xi= -2. 
gi = 0. 
X-ajpu 
-1600 
-1087 
-0642 
-0296 
•0076 
0 
Thus, even if there were no correlation between the organs A and B in the general 
population, still a selection of parents in which such organs were correlated would 
lead to offspring with correlated organs A' and B'. The amount of such correlation 
would only be '1600, if the variability of the jDarent were not selected, and would 
diminish rapidly wuth stringent selection of variability. Still '1600 is cpiite sensible, 
and would, if the selection continued for a few generations, continue to increase. 
Thus we see how selection of a pair of organs in a parent may increase or even create 
correlation between the like organs in the offspring. 
The reader will find other interesting illustrations in tracing the influence of an 
absolute selection of one parent only on the correlation of the offspring, e.r/., relation 
between pairs of foals which all have a common sire, the influence of selecting an organ 
A in the sire and an organ B in the dam on the correlation of the organs A and B in 
the offspring, the influence on assortative mating of selecting parents of men of 
genius,* and in many other problems. 
(10.) It is not wuthout value to consider how arises in the case of natural or 
artificial selection. Suppose we have two organs, A and B, then we shall endeavour 
(i.) to give these definite values, say x and but we shall not be able to get all our 
individuals with such absolute values, we shall select with certain deviations from 
X and y, given by x = x -{• x' and y — y y\ say. Further, we shall endeavour to 
* This is a peculiarly interesting case, for if we select men of remarkable intellectual ability, we should 
expect to find both parents above the average of the general population, but with a negative correlation 
between them amounting at a maximum to - • 1905. 
