xiv.] FORBES' SCIENTIFIC WORK. 479 



have said, that you must take a great number of starry 

 heavens, and take the average of all. 



'The number of persons on the different parts of 



London Bridge at any moment, when there is no spec- 



of any kind, is a pure matter of chance (i.'c. depends 



on laws which we* are not considering) ; yet, as I know 



well, they are most oddly grouped, without any sort of 



uniformity. But, taking the average of 100 different 



states in the same day, I apprehend that the mean would 



very uniform treading of the pavement. . . .' 



From R. LESLIE ELLIS. 



4 AXGLESEA HOUSE, GREAT MALVERN, 



Sept. 3rd, 1850. 



' MY DEAR SIR, 



' . . . I am glad to see by what you have sent 

 me, for which pray accept my thanks, that you have 

 ni to say something touching one of the popular 

 applications of the theory of probabilities. 



k The simple view of the matter seems to me to be, 

 tliut whenever we perceive a relation among phenomena 

 re disposed to infer the existence of a relation among 

 their causes, and that the strength of such an inference 

 varies with an infinity of circumstances, and in particular 

 with what may be called the d pnori conceivability of 

 the , e of the causal relation. If in the midst of 



the stream of people whom you meet in the Strand you 

 roup gathered together at a particular spot, you 

 ihfer that they are interested about the same object, 

 wh,-i iat may be: the strength of the infer. -nee 



with the size of the group, the similarity of their 

 appearance, &C. ; and in this case you know that a 

 v very well have produced what you 

 . T<> set this last condition aside, and to attempt 

 to estimate the force of the inference 1,\ the theory of 

 probabilities, is to leave out something wnicb is essential, 

 and to introduce something wliich is irrelevant to tin- 

 matter in hand. Mv.-rythin^ which exists is r) />< 



infinitely improba] that this sh< > per should 



