chance that each votes rightly be the ratio of a to b, then the 

 ratio o£ a* to b* will be the odds that the affirmative voters arc 

 right ; and the ratio of a* to b^ will be the odds that the negative 

 voters are right; and the ratio of a* b^ to b* a', or a to b, will be 

 odds refulting that the affirmative voters are right. 



If there were eight voters, the lowefl majority muft be five and 

 three; and the odds that the affirmative voters were right would 

 be a^ to ^> ; and the odds that the negative voters were right 

 would he a^ to ^^ ; and the refulting odds that the queflion was 

 juftly decided would be a^ b^ to a^ b' or <z* to b'. 



In general therefore it appears, that the odds for the truth of 

 the decifion, will be that power of the odds that each perfon votes 

 juftly, whofe index is the difference between the number of affir- 

 mative and negative voters. 



And hence we may corred the error of thofe who imagine, that 

 the probability, cateris paribus, is the fame, if the proportion of the 

 number of affirmative witnefles to the number of negative witnefles 

 be the fame ; whereas the probability is to be cflimated by the dif- 

 ference of thefe numbers. 



We have already remarked, that in the enading of a new law, 

 we ought to have at leaft that probability for the expediency of 

 the law, below which a perfon cannot a£l without imprudence. 

 As the manner of determining this degree of probability is ex- 

 tremely 



