(With two Graphs.)IN the last memoir of this series (Greenwood, Topley and Wilson, 1931) we discussed at some length the modifications of the course of herd sickness effected by immunisation of entrants to the herds. Briefly, these were our conclusions: efficient artificial immunisation (the criteria of efficiency being the usual standardising methods of a laboratory) confers a considerable temporary advantage upon mice entering an infected herd. The risk of death during the period of herd life when the rate of mortality is especially high is greatly reduced, so that when one has the case of a herd or group temporarily exposed to special risks, the value of the expedient is very great. But no method of artificial immunisation we have tried will render the immunised animals impervious to the risks which their membership of the herd entails.The experiments now to be described were designed to test two points. First of all we wished to study the course of herd mortality when recruitment was confined to immunised animals. Secondly we wished to compare the effects of immunisation by different routes.Exp. 1. Taking the former problem first, if the reasoning of our last paper were just, we should expect that the rate of mortality in a herd recruited from immunised animals, or any function of the rate of mortality, would over a fair range of herd age be much more favourable than that of a herd recruited in the usual way from unsalted stock, but that mortality from the specific cause would not be extinguished. The experiment was inaugurated on 14. iii. 29 by bringing together 25 mice, each inoculated with 1000 Bact. aertrycke, and 100 normal mice. Three normal mice were added daily for 59 days. On 13. v. 29 the normal immigrants were replaced by 3 mice daily which had been vaccinated.As in our previous experiments the mice were vaccinated by intraperitoneal injection of a killed formolised suspension of Bact. aertrycke, which contained both the "H" and the "0" antigens. Two doses of 500 x 106 bacilli were given, with an interval of 1 week between the two doses. The interval between the administration of the second dose of vaccine and addition to the infected herd varied from 7 to 13 days. A certain number of mice were vaccinated each week, over and above those required for the daily additions; these were anaesthetised, bled to death from the jugular vein, and their sera tested for the presence of agglutinins. As it was known from previous experience (Topley 1929) that "H" agglutinins would almost invariably be