358 PERSONAL PREPAEATION FOR RESEARCH. 



* An experiment with distilled water gives less acid and 



alkali. 



* Therefore sometimes with less impurity we have less 



acid and alkali. 



' V. The contact of moist hands may be an additional 

 cause of the residual phenomenon. Improved syllogism, 

 to include this concurrent cause. 



4 An experiment with distilled water and apparatus kept 

 from contact with hands will admit still less im- 

 purity. 

 c An experiment, &c., results in the production of still 



less acid and alkali. 



' Therefore sometimes, with still less impurity, we have 

 still less acid and alkali. 



' VI. Amended syllogism. 

 6 A case where we use these precautions in vacuo is a 



case of no acid and alkali. 

 ' Therefore a case of no impurity is a case of no acid 



and alkali. 



' VII. Immediate inference from last conclusion. 

 ' Cases of no impurity are cases of non-production of 



acid and alkali. 

 ' Therefore all cases of production of acid and alkali 



are cases of some impurity ; 

 which was to be proved.' ! 



For additional examples of successful inference in 

 scientific research, see the Chapter on 'Accidental Disco- 

 very,' page 227. 



Indirect inference is largely employed in devising hypo- 

 theses. Whenever we infer from insufficient evidence in 

 science, we make a tentative guess, to be tested by means 



1 Thomson, Outline of the Laws of ThouyJit, p. 225. 



