Hence the expenses incurred by any company in a given; time must be divided into two parts, one being the expense incidental to insurance, and the other that incidental to investment, which parts are to be compared respectively with the insurance claims met, and interest receipts of the company for the same time; or what is equivalent in the latter case, the net rate of interest earned after deducting the incidental investment expense may be found. When this process shows that one company has earned a higher rate of interest than another, at the same time that its insurance expenses bear a lower ratio to its insurance claims paid, _there is no escape from the conclusion that during the period under observation it has served its policy-holders more economically_, and the test is therefore scientific. Though, if one company shows a higher rate of interest, while the other shows a lower ratio of insurance expense, it will still be necessary, to complete the test, to equate either the rates of interest or the ratios of insurance expense (it does not practically matter which), and note how this affects the relation of the duly corrected ratios on the other score. To be exact, if the average vitality of the members of the two companies differ (other things being equal, it is always cheapest to belong to that company which has the lowest death rate), the ratios of insurance expense to expected, as well as actual, claims of each must be found, and equated. The science of this procedure, or mode of testing expenses, and also its practical simplicity, may be more clearly perceived by reference to its practical application in the following table: