Growth Measurements 173 



substance and completed on July 13, 1939, gave a 

 numerical index of 58.6 mm., the other strain, in a 

 similar experiment completed on December 7, 

 1939, gave a numerical index of only 21.8 mm., 

 but the percentage values for the two were 60 and 

 62, respectively. While the use of percentages 

 may be objected to on the grounds that it involves 

 a subjective choice of a standard, it is the only 

 way in which non-simultaneous experiments can 

 be compared and has proved very satisfactory as 

 a method of presenting results. Three examples 

 of graphical representations will serve to show 

 the types found most useful (Figs. 45^7). 



Interpretation of results. While the results 

 themselves must, wherever possible, be presented 

 in numerical form with a definite idea of the de- 

 gree of accuracy and significance which can be 

 attached to them, the interpretation of results is 

 the duty of the observer. It is necessary for him 

 to take into account not merely the numerical data 

 but also many qualitative features of the results 

 which cannot be set down in numerical form. In- 

 terpretation involves the integration of many 

 sorts of information which only the person who 

 actually handles the cultures can have available. 

 How this is to be done is a matter of personal 

 inclination and cannot be presented in any hand- 

 book. Yet it is in this, just as much as in the 



