458 



THE GRAMMAR OF SCIENCE 



individual, his knowledge is of the group containing great 

 numbers. 



Lastly, consider the variability, the standard-deviation, 

 of an array of sons corresponding to a father of given 

 height. According to the statement on p. 397, we find 

 this standard - deviation by multiplying the standard- 

 deviation of the whole population of sons by the square 

 root of I — r^, or in the present case by.9183. Thus 

 we see that a selection of fathers reduces the variability of 

 sons by about 8 per cent only. To select a type of 

 parent thus alters the type of the offspring and reduces 

 their variability by quite definite quantitative amounts. 

 Biologists ought no longer to speak as if the amount of 

 change in type and variability due to selection had un- 

 known values, and to reason as if they could be anything 

 whatever under the sun. 



The following table, extracted from a larger mass of 

 material, may give some idea of the strength of heredity 

 for different organs and relationships : — 



