22 PROGRESS OF SCIENCE IN THE CENTURY. 



edge, and projected upon nature to satisfy our desire 

 for continuity. They relate to what lies beyond the 

 range of observation, beyond the range of our sense- 

 impressions. 



An interesting method of testing the accuracy of a 

 formula is to use it as a basis for prediction. Many 

 observant people are familiar with a mild form of 

 scientific prophecy in connection with the weather. 

 After long observation they hazard a generalisation, 

 in private, if they are wise; and they test this by a 

 prediction. As this is usually wrong, they conclude 

 that their generalisation had not a sufficiently wide 

 basis. But better examples may be found in the 

 prediction of Neptune by Adams and Leverrier 

 (from calculations based on the gravitation-formula) 

 and the subsequent discovery of that planet by 

 Galle; or in the prediction of the element german- 

 ium by Mendelejeff and its discovery by Winkler. 



(e) Test Experiments and Control Experiments. 

 The distinction between an observation and an ex- 

 periment seems quite artificial, the point of contrast 

 being that in the former we study the natural course 

 of events, while in the latter we arrange for the oc- 

 currence of certain phenomena. In studying the 

 effect of electric discharges on living plants we might 

 wait for the lightning to strike trees in our vicinity ; 

 but as this would be worse than tedious, we prefer 

 to mimic the natural phenomenon in the laboratory. 

 This is obviously a distinction without a difference, 

 and instead of calling the first step (a) observation, 

 as we have done, we might equally well have used 

 the word experiment. 



On the other hand, at a later stage in the scientific 

 treatment of a problem, our opportunities for experi- 

 ment can be profitably used, not for accumulating 



