250 ESSAY ON PROBABILITIES. 



to render themselves safe against fluctuation. The state 

 of opinion upon this matter is somewhat unsettled ; one 

 party advocating the practice of approaching near to the 

 line which separates security from insecurity ; another 

 insisting upon what appears to the first a most super- 

 fluous degree of caution. Without expressing an opinion,, 

 I will describe the various risks, and the method of 

 avoiding them which has usually prevailed. 



1. The insecurity of data, that is,, of existing tables 

 of mortality. This divides itself into two parts ; that 

 relating to the young and middle aged,, and that relating 

 to old lives. With regard to the first, the data might 

 probably be obtained in sufficient numbers to justify a 

 considerable degree of confidence in them as to the 

 chances of a single life, or even of a considerable number ; 

 but when the number of lives is to be as great as the 

 number of persons who may choose to offer themselves, the 

 considerations in Chapter IV., again adverted to in page 

 242, present themselves in force. I am not aware that 

 any writer on the subject in this country has formally 

 taken into consideration the uncertainty of tables, arising 

 from their limited numbers, except Mr. Lubbock, who 

 has made use of (Cambridge Phil. Trans., and Treatise 

 on Probability Lib. Usef. Know.) the correction which 

 the probability of living a given number of years should 

 receive on that account. But, considering the probable 

 errors of the data, this correction is small, and the 

 question how far an office proceeding upon such data 

 can deal with the public to any amount is yet in its 

 infancy, though the necessity for its consideration is 

 approaching, and it is one of vital importance to the 

 interests of the middle and lower classes. 



The constructor of tables of mortality draws a number 

 of balls from an urn which contains an infinite number, 

 and, having sorted them into red, blue, black, &c., 

 presents them to the world as a necessary representation 

 (or very nearly so) of the proportions in which those 

 colours are scattered throughout the whole urn. He 

 commits an error which is in all probability very small, 



