expressed by a New Formula. 7 



observations with the year 1813. Both series of observations, 

 with insignificant interruptions, have been continued to the pre- 

 sent year (1866). 



The method of observation in both series has been substan- 

 tially the same. In Sweden and in England there have been 

 made periodical enumerations of the living and dying, in quin- 

 quennial intervals of age, from the age of 5 years upwards, and 

 in annual intervals for ages below five years. The Swedish ob- 

 servations have been satisfactorily made, judging from the 

 internal evidence of correctness afforded by uniformity of in- 

 crease of ratios of dying to living for consecutive quinquennial 

 intervals of age. Similar success has not attended the English 

 observations for quinquennial intervals of age ; for the several 

 indicated rates of mortality for even quinquennial intervals of 

 age ending at the ages 30, 40, 50, 60, and 70 years are nearly 

 as great as the several indicated rates of mortality for odd quin- 

 quennial intervals of age, five years greater, ending at the ages 

 35, 45, 55, 65, and 75 years. This will be seen on reference 

 to column 5 in Table VI. hereunto annexed. The amount of 

 error is considerable, and remains unaccounted for. The English 

 Life Tables Nos. 2 and 3 are founded on these observations, cor- 

 rected by the assumption that the errors are in opposite direc- 

 tions and neutralize one another, the true rate of mortality 

 being assumed to be halfway between consecutive contradictory 

 quinquennial rates. The publication of the numbers dying in 

 quinquennial intervals of age ceased with the year 1850, the 

 intervals of age now adopted for publication being decennial for 

 ages exceeding 25 years. 



On comparison of the rates of mortality for quinquennial inter- 

 vals of age, of the male population of Sweden and England (the 

 former for 61 years, and the latter for 7 years), it will be found 

 that, according to the general average for all ages exceeding 25 

 years, the Swedish rates exceed the English rates at the same 

 ages in the proportion of 120 to 100. When, however, the 

 comparison is separately made for even and for odd quinquen- 

 nial intervals of age, it is found that the Swedish rates exceed 

 the English rates in the proportion of 109 to 100 for even quin- 

 quennial intervals of age, and in the proportion of 131 to 100 

 for odd quinquennial intervals of age. 



According to all observations on human mortality, there exists 

 a short period near the age of 12 years at which the rate of 

 mortality is at a minimum and apparently stationary. In the 

 English observations this period extends from age 10 to age 15 

 years, in the Swedish observations from age 10 to age 20 years. 

 If the English observations at the quinquennial interval of age 

 15 to 20 are to be relied upon, it would ensue that the propor- 



