RELATIONS OF PHENOMENA TO EACH OTHER. 429 



stance, we know that cholera arises during warm weather, 

 in localities where pestilential vapours from decaying 

 animal matter abound ; but these conditions alone are not 

 sufficient, otherwise that disease would occur every sum- 

 mer in such places ; other conditions, the nature of which 

 is at present but little known, must also be present. Good 

 health can only result from many concurring conditions. 

 And 8. Finally the two phenomena, A and B, may be 

 merely coincidences, and in no way related to each other 

 as cause and effect. Imaginary effects I have left out 

 of consideration. 



To ascertain if the assumed or suspected cause is the 

 real one, we must test it by means of such experiments as 

 either wholly or partly exclude only one condition (the 

 suspected one) at a time ; thus if B is a direct result of A 

 only (and whether it is or is not accompanied by other 

 effects), it will not appear when A is absent. If B is an 

 indirect result of A, some other phenomenon, A! (i.e. the 

 immediate cause) intervening, it will be produced when- 

 ever A' is present, whether A is present or not ; thus I 

 found in the case of the rotating ball, already mentioned, 

 that the ball rotated without the use of an electric cur- 

 rent, if it was placed on a massive pair of red-hot hori- 

 zontal rails of copper, and therefore the immediate cause 

 of the motion in both experiments was the expansion pro- 

 duced by the heat. If A is merely a phenomenon coinci- 

 dent with B, and both are simultaneous effects of one or 

 more causes, an experiment must be devised in which B 

 alone is wholly or partly excluded or varied in amount ; if 

 the two are only coincident, A will still be produced, and 

 to the same extent. If B is composed of two portions, 

 one being due to A and the other to A', it will only partly 

 appear when A alone or A' alone is absent, and not at all 

 when both are excluded. 



