I ^'2 ] 



trcmrly ingenious, I cannot avoid mentioning it. The objed to 

 be attained is equivalent to this, that in the enading of a law, 

 the ri£k of error fhould not be greater than what we difregard, 

 even where our own life is in queftion. Buffon and Bernouilli 

 have endeavoured to eftimate the value of this ri£k, but the fol- 

 lowing; method adopted by Condorcet, feems to be the beft. It is 

 obfervcd, thit from the age of thirty-feven years to forty-feven, 

 and from eighteen (o thirty-three, the riil< men run of dying by 

 accident or difeafes of fhortcr duration than a week, encreafes con- 

 tinually in nearly a regular manner ; and it is alfo obferved, that 

 a man of thirty-three years is not more apprehenfive of fuch kind 

 of death than a man of eighteen, nor a man of forty-feven than 

 a man of thirty-feven years; the difference of rifk therefore in 

 thefe cafes is difrcgarded : now, from the tables of mortality, it 

 appears, that, in the firft period, the difference of rifk is = 

 ^__"_--_., and in the fccond = tt 4VrT '■> ^^^ "^ t'""^" ^^^^ ^^^ latter, 

 which is the greater, as the limit of that rifk which may be dif- 

 rcgarded, and confequently r^^jt-i ^''^' be the limit of the affu- 

 rance, which we ought to have in the enading of a new law. 



If we fuppofe that the odds that each legiflator votes juflly, is 

 ten to one, then will a majority of fix be requifite to give the 

 afTurance required ; which in an afTembly of three hundred is only 

 a majoTity of one in fifty. 



Thesi 



