440 THE MAINTENANCE OF SPECIES 



The observational method has been practiced from time immemorial, 

 and to it is due most of the accumulations of our general knowledge 

 concerning heredity up to about the turn of the present century in 

 1900. The phrase "like produces like" expresses the general impres- 

 sion that is gained from observation, although there are plenty of 

 exceptions to the apparent rule. We say that children in a general 

 way "take after" their parents, although there are conspicuous in- 

 stances when it becomes necessary for parents to "take after" their 

 children, in order that they may be made to conform to a family 

 tradition, whatever it may be. It is repeatedly observed that not 

 only individuals of one generation may be in general like their pred- 

 ecessors, but that certain noticeable characteristics in the make-up 

 of an individual may occur more often in some family lines, breeds of 

 animals, or strains of plants than in the general population of which 

 they are a part. Whenever this is so we are led to suspect, even when 

 we may not be entirely convinced, that such characteristics are 

 hereditary. General but more or less vague observations of this sort, 

 while useful in establishing the simple fact of inheritance, do not go 

 very far in determining and analyzing the causes of heredity and the 

 laws of procedure that underlie the mechanism of inheritance, which 

 it is necessary to know in order to establish a real science of genetics. 



The statistical method recognizes the desirability of arranging quali- 

 tative data in quantitative terms, as a necessary process in reducing 

 random observations and guesses to definite scientific form. Recourse 

 must always be made to mathematical treatment in formulating any 

 science, and genetics is no exception. Mathematics, however, is 

 simply a useful tool to be employed in arranging the facts and in 

 bringing them together in convenient form for interpretation. There 

 are repeated occasions when it is not only desirable but indispensable 

 to focus isolated and scattered facts into a single comprehensive pic- 

 ture which can only be accomplished by statistical treatment. Statis- 

 tics, however, to be of value in solving problems of heredity, must be 

 based upon careful observations and accurate measurements pre- 

 viously obtained. Biometry, the science of measurement when ap- 

 plied to biological data, is powerless to extract true conclusions out 

 of faulty observations or findings. 



The biometrical approach is about the only way available in which 

 to investigate the problems of heredity as applied to mankind. It is 

 obviously not feasible, even if it were desirable, to plan and execute 

 controlled experiments in human breeding, of sufl&cient magnitude 



