VARIATION 23 



may differ from a sweet fruit in a qualitative way de- 

 pendent upon the chemical constitution of the fruit in 

 question. 



i. With respect to their relation to an average stand- 

 ard variations may have a fluctuating distribution 

 around an arithmetical mean, as when some of the off- 

 spring have more and some less of the parental char- 

 acter, or the variations in the progeny may all center 

 about a new average quite distinct from the parental 

 standard and consequently come under the head of 

 mutations. 



j. Finally, and most important in the present con- 

 nection, with respect to heritdbility, variations may 

 possess the power to reappear in subsequent genera- 

 tions, or they may lack that power. It is this aspect of 

 variability which bears most directly upon genetics. 



Other possible categories might be mentioned, but 

 a sufficient number have been cited to show the great 

 diversity of variations in general. 



4. METHODS OF STUDYING VARIATIONS 



Roughly stated, there are three ways of studying 

 variations: first, Darwin's method of observation and 

 the description of more or less isolated cases; second, 

 Galton's biometric method of statistical inquiry ; and 

 third, Mendel's experimental method. The second of 

 these methods will be considered in this chapter. 



5. BIOMETEY 



The science of biometry, that is, the application of 

 statistical methods to biological facts, has been de- 



