12 CHOICE OF METHODS 



establish the percentage within the Hmits of probability. It would not do to 

 compare the amount of pigment in the specimen, the normality of wliich is in 

 question, with that in too few others. This is the statistical definition of nor- 

 mality which is not universally accepted but which is useful and easily under- 

 stood. 



Only a few samples of the various kinds of technique have been mentioned in 

 this survey as a kind of menu from wliich to make a selection or to obtain clues 

 to other methods that may fit the case. Many of them are very ingenious and 

 were only discovered after Avisely conceived attempts to overcome practical 

 difficulties. This overcoming of obstacles is a pleasant experience. It calls for 

 actual work and experiment and appeals to many of our best minds. The 

 techniques may be regarded as keys by which scientific treasure can be unlocked. 

 Unused they are worthless. 



