This is a particularly difficult branch of the subject to investigate. It might be thought at first sight that the death rates at different age periods might be compared, and this is often done; but when it is noted that the death rates at different age periods are organically connected, it is obvious that such a comparison is statistical or actuarial and completely neglects the biology of the subject. To illustrate the problem a diagram is given. This is constructed on a principle open to objection, but, if that is remembered, it illustrates a number of points. The healthy district life table H2 has been taken as a standard of comparison. Now a healthy district life table labours under certain disadvantages. A district may, as the Register General says, be excluded because it contains an institution drawing its inmates from a wider district. This is not a serious objection. A more serious objection is the fact that whether deaths which might be considered to belong to the district are returned to it or not, the result is equally unsatisfactory. Everyone who has had practical experience knows many instances in which it is impossible to allocate a death to the district to which it properly belongs. When all is balanced, I think that probably on the whole the healthy districts get credit for more than their share of deaths, that is, for more deaths than would occur if there was no process of intermingling of town and country. For one death the town gets credit for, the country gets credit for another, in the one case some old person dying of a cancer in a hospital, in the other, some poor young person, who having tried city life, returns broken, to die of phthisis or some similar disease. Thus a healthy district life table is open to objection as a criterion. It is, however, for the purpose at present required, the only one at our disposal.