382 Estimate of the Risk of Life [July, 



(1),in the first year, being together 6 ; also the appointments of the year 

 1831 (29), which could not be included in the computation, unless the 

 year 1832 were brought in. The number, therefore, of Civil Servants ap- 

 pointed between 1790 and 1831, who were in the condition to abide the 

 chance of life for a first year, commencing from the 1st January after their 

 appointment, was 905. Our table assumes 904, because there was one 

 retirement in the year we are reviewing ; and, as those who retire give 

 the risk of their lives only for the broken period before their departure, 

 the number of persons on the register on the 1st January requires cor- 

 rection on their account. It is assumed, that retiring servants give, 

 on the average, the risk of their lives for half the year in which they 

 retire; accordingly deduction is made, from the registered numbers on 

 the 1st January, of half the number of the retirements between that 

 date and the 31st December following. Where the number is uneven, as 

 in this case being one, the deduction is made alternately to include or 

 reject the unit. We thus begin with 904 lives, of which the proportion 

 of deaths in the first year was 17, being in the ratio of 188 out of 10,000. 

 For the computation of the ratio in the second year, we have the 905 men 

 of the first year, minus the whole number of retirements (1), and deaths 

 (17), in the course of that year, and minus also the appointments of 1830, 

 which were 25, that is, 905 — 43 or 862. Of this number of registered 

 servants on the 1st January of the second year, four retired in the course 

 of the year : deducting two, therefore, for the diminished risk upon their 

 lives, the number 862 stands corrected to 860, of which number 21 died 

 in the year, being in the ratio of 244 out of 10,000. A separate value 

 has thus been computed for each year of life in India, from the first to 

 the thirtieth inclusive; after which the number of registered servants be- 

 comes too small to afford results at all trust-worthy. Before we proceed to 

 the tables exhibiting the estimate of the value of life for each year, form- 

 ed on this principle, it will be proper to exhibit separately the precise 

 number of survivors in India, on the first of January of each year of 

 their service; it being my purpose, as above explained, to compare these 

 numbers, not with the original appointments as in the table of the 

 Gleanings, but with the deaths of the year then commencing. 1 shall 

 presently explain why no correct result could be extracted, from a 

 comparison of the survivors with the number of original appointments, 

 or even with that of arrivals in India. See Table, No. II. 



Taking the table of survivors as the material for calculation, the 

 ratio of deaths is exhibited for each year in the table which follows, 

 !No. III. The sameratiois reduced to a decimal proportion upon 10,000, 

 for more ready deduction of the average of each five years, and for 

 comparison with the tables of Europe. The number of deaths assum- 



