208 GENERAL CONDITIONS OF SCIENTIFIC RESEARCH. 



classification will best enable us to draw particular con- 

 clusions, and others to draw different ones, in an ex- 

 haustive research we classify the substances, phenomena, 

 or results, not only according to their most intrinsic or 

 fundamental characters so far as these are known, but 

 in every possible way that occurs to us, in order to 

 extract from the new data the utmost amount of new 

 knowledge. 



Classification is not only a means of discovery, but 

 one of the effects of scientific progress. Science is con- 

 tinually simplifying our knowledge and placing it more 

 at our command, by classifying truths under the headings 

 of different causes and relations. Truths of science can 

 only be classified in proportion as they are known, and the 

 perfection of classification depends upon the extent of our 

 knowledge. The most perfectly philosophical classifica- 

 tion of scientific truths can only be made when the most 

 essential and fundamental characters of them are dis- 

 covered, and these are probably the most difficult to find, 

 and doubtless will be nearly the last to be evolved. 



Classification and discovery of scientific truths mutu- 

 ally act and react to further each other's progress. When 

 we discover a new truth, it is almost always one of a class 

 already known, and we at once add it to that class, and 

 thereby render the new truth more fit to assist mental 

 progress and future discovery ; and when it is not of a 

 class already known, we form a new class for it. If we 

 discover a new substance, we naturally expect to find it 

 more or less similar to each member of the class to which 

 it belongs ; for instance, if we find a new alkali-metal, we 

 expect to find its carbonate undecomposable by heat, and 

 so on. 



A chemical investigator requires not only to classify 

 his knowledge, but also to arrange his substances systema- 



