TESTING A SUPPOSED CAUoE. 427 



of substances, and of individual substances, and selecting 

 the most likely. Use different quantities and proportions 

 of the substances. Change the materials of the appa- 

 ratus. Consider the influence of volume, weight, specific 

 gravity, elasticity, width, length, depth, thickness, shape, 

 &c., of the substances employed. Try substances of dif- 

 ferent degrees of transparency to light, conductivity for 

 heat and electricity, magnetic capacity, solubility in acids, 

 &c. Considerable assistance in forming the list may be 

 obtained by consulting the index of a suitable book on 

 science. 



In these processes it is entirely by means of a know- 

 ledge of their effects that causes are discovered. The 

 discovery of causes by such means is called induction, and 

 consists in imagining hypothetical explanations of pheno- 

 mena already known, including new results and observa- 

 tions, deducing consequences which must occur if those 

 explanations are true, and testing by suitable experiments 

 or observations whether those consequences really take 

 place. In proving a supposed cause we must be able to 

 deduce new effects, which must follow if the assumed 

 cause be the true one. 



One phenomenon may be invariably connected with 

 another, either as cause or effect, as a necessary condition, 

 or as a mere coincidence. In the discovery of causes we 

 must remember that a phenomenon A may be related to 

 another one, B, in several different ways in different cases. 

 1 . B may be a direct result of A ; thus the fall of a body 

 to the earth is a direct result of the universal force of 

 gravity. 2. B may be an indirect result of A, some other 

 phenomenon or intermediate cause (perhaps of an obscure 

 kind) intervening between them ; thus I found that if a 

 metallic ball was placed upon two horizontal metal rails, 

 insulated from each other, and an electric current passed 



