The next and all-important stage was to devise an experiment or a series of experiments on man so as to test the efficiency of the method against cholera attacks. This part of the investigation could only be done in a cholera-stricken country, where opportunities would arise of comparing the incidence of the disease in inoculated and uninoculated. Such opportunities are limited. Except in certain parts of India and China, cholera appears in localities unexpectedly and does not last long. In the places where the disease is endemic the cases are scattered over large areas. These features rendered the demonstration of the effect of the vaccine a matter of particular difficulty. In 1893 I went to India, and in the course of a year inoculated some twenty-three thousand people in the northern parts of the country; but no cholera appeared in their midst to show whether the vaccine was of value or not. In the spring of 1894 the inoculations were introduced into Bengal, and, with the assistance and co-operation of Prof. W. J. Simpson, of King’s College, London, at that time Health Officer of Calcutta, and of his staff, efforts were made to induce the inhabitants of the _bustees_ of Calcutta to get themselves inoculated. These bustees are isolated villages consisting of groups of mud huts inhabited by the poorer class. Owing to the consumption of water from the ponds or tanks belonging to these villages, the inhabitants of the bustees are subject to periodic visitations of cholera. It was in one of these bustees that the first observation was made as to the effect of the cholera vaccines. The spring is essentially the cholera season in Calcutta. About the end of March two fatal cases of cholera and two cases of choleraic diarrhœa occurred in Katal Bagan Bustee, in a population grouped around two tanks. This outbreak led to the inoculation of one hundred and sixteen persons in the bustee out of about two hundred. After the inoculation there occurred nine more cases of cholera, seven of which proved fatal, and one case of choleraic diarrhœa. All the ten cases occurred among the uninoculated portion of the inhabitants, which formed the minority, none of the inoculated suffering. The results were more interesting when analyzed in detail. Some of the cases had occurred in families in which some of the members had been inoculated and others not, and the disease selected the non-inoculated members, sparing the inoculated. Thus, in one house six members out of eight had been inoculated. The attack, a fatal one, occurred in one of the remaining two. In another house eleven members out of eighteen were inoculated. The eleven members remained free while four out of seven not inoculated were attacked.