and severity in vaccinated and unvaccinated respectively. The =age incidence of deaths from small-pox= has, since 1847, when returns classified according to age became available, undergone a remarkable alteration. Prior to 1870 the small-pox deaths in infants nearly always formed 20 per cent. or more of the total mortality from this disease, between 1870 and 1890 they did not greatly exceed 10 per cent. of the total, while since 1890 they have again begun to form an increasing proportion of the total small-pox mortality. At ages 1-5 the change is even more remarkable. Before 1870 deaths at these ages nearly always exceeded 30 per cent. of the total; since 1870 they have varied between 5 and 14 per cent. of the total; and since 1890 they have, like the proportion of deaths under one, again increased. At the higher ages the proportion of deaths has correspondingly increased, so that the curves of age incidence have become curiously inverted. The lowered birth-rate can only account for a small portion of this transference of the chief mortality due to small-pox from childhood to adult life. Furthermore it must not be supposed that the only change which has occurred is that the deaths which formerly occurred in childhood now occur in adult life. The death-rate at all ages has greatly declined. The only explanation which in my judgment satisfactorily explains this remarkable change in age-incidence of small-pox mortality is the fact that vaccination protects children from small-pox and that the protection diminishes, though it never entirely disappears, with advancing years. This conclusion is confirmed by the evidence obtained as to the proportion of vaccinated and unvaccinated attacked, and as to the severity of the attacks occurring when a community is invaded by small-pox. =Attack-rate among Vaccinated.=—If the protective effect of vaccination, like that of a preceding attack of small-pox, wears off, it will not be expected that no attacks of small-pox will occur among the vaccinated. For evidence of immunity from attacks we must examine the records as to =revaccinated= persons exposed to infection. During the six years 1890-95, out of a staff in the London small-pox hospitals varying from 64 to 320, the percentage attacked by small-pox was _nil_, except in 1892 when it was 1·4, and in 1893 when it was 1·9. Taking the experience of towns in which during recent years epidemics of small-pox have occurred, the following attack-rates have occurred. By attack-rate is meant the percentage number of attacks occurring among persons living in infected houses. By fatality is meant the number dying out of 100 attacked.