198 GENERAL CONDITIONS OF SCIENTIFIC RESEARCH. 



current was really determined by the molecular structure 

 of the liquid. 



Superficial explanations often appear to agree with 

 ordinary instances, and we are sometimes thus led to 

 mistake a coincident phenomenon for a true cause and 

 explanation (as in the above case), until an exceptional or 

 anomalous instance is met with which can only be explained 

 by means of a wider truth capable of including both the 

 anomalous as well as the ordinary cases ; for example, 

 previous to the discovery of certain anomalous cases of 

 the effect of combined pressure and heat upon the boiling 

 points of liquids (first observed by Baron Cagniard de 

 Latour), the theory of definite boiling points sufficiently 

 agreed with what was then known of the effect of heat 

 upon liquids ; but an investigation of those anomalous 

 instances led Dr. Andrews to the important discovery of 

 the continuity of the liquid and vaporous states of matter. 



The method of discovering exceptional cases is simple 

 enough. It usually consists in examining a sufficient 

 variety and number of instances, or, in other words, in 

 making a sufficiently exhaustive research, and it is only 

 the great amount of labour required in carrying out this 

 method which has caused the discovery of anomalous and 

 exceptional instances to be apparently surrounded by 

 mystery. If we make a sufficiently exhaustive research, 

 we are in many cases almost certain to meet with an 

 exceptional instance. As also the small proportion of 

 cases in which we are able to predict results successfully 

 proves that many new laws probably remain undis- 

 covered, 1 we may reasonably expect the occasional dis- 

 covery of anomalous phenomena and entirely new laws. 



By observing in what respects exceptional cases differ 



1 See page 28. 



