[ 92 ] 



and leveral black papers, in this cafe if it be required to deter- 

 mine the probability of drawing a white paper, we muft proceed 

 by trials, that is, we are to draw out a paper and obferve its 

 colour, then replacing it, we are to draw out a fecond, which is 

 to be replaced likewife, then a third, a fourth, and fo on. It is 

 clear, fays Diderot, that the firft drawn paper, being white» 

 gives but a very low degree of probability that the number of 

 the white exceeds that of the black ; a fecond white one being 

 drawn would encreafc this probability, a third wo\ild augment 

 it. At length, if a great number of white papers ftiould be 

 drawn out, without interruption, we would conclude that they 

 were all white j and that with fo much the more probability as 

 we fhould have drawn out more papers. But if, of the three firft 

 flips of paper, two fhould appear to be white and one black, we 

 would infer, that there was fome low degree of probability, that 

 there was twice as many white papers as black. If of the fix firft 

 drawn papers, four fhould be white and two black, this proba- 

 bility would encreafe ; and it would encreafe fo much the more 

 in proportion, as the number of trials fhould continually confirm^ 

 the fame proportion of the white papers to the black. 



This manner of determining, probably, the ratio of the chances 

 for the happening of an event to thofe of its failing, is applicable 

 to every thing which is contingent in nature. 



It may be afked, fays Diderot, whether this probability, admit- 

 ting of infinite increafe by a feries of repeated experiments, can 



arrive 



