Here the endemic index was 12, remarkably low compared with 70 per cent, to 80 per cent., a common figure in West African children; and it was only on reaching the foot of the Himalayas that we found the condition of things parallel to what we had previously found in Africa. Here the endemic index ranged from 55 per cent, to 72 per cent, in two villages examined by us. How, then, can we explain this change from 0 to 72 in a region where, so far as we could judge, the climate and other conditions were the same? There was, however, one important point in which these districts differed. Whereas M. rossii was the prevailing anopheline at Calcutta, it was not so at the foot of the Himalayas: here the prevailing anopheline was M. listoni, and, moreover, dissection showed us that this species was the carrier of malaria here. It seems, then, that the difference in malarial en demicity had to do with a difference in the species of Anophelince, and, in fact, we were able to prove this later in quite a different way. We argued if this idea were correct, that when we found the two species together, one would be infected and the other not. When, then, the opportunity arose, we proceeded to put the matter to the test by dissection. The two species (in this case M. culicifacies and not M. listoni) were captured under identical conditions-that is, they were caught in the same house, in the same village, at the same time. The following table gives the results of our dissections made in two widely distant districts:

So that here, under conditions where M. culicifacies was transmitting malaria , M. rossii was not. The result of the experiment was thus of the greatest importance, showing that all Anophelince do not transmit malaria, or at least in these two instances M. rossii did not. Further dissections of M. rossii have all proved so far negative, and although this species abounds in tropical towns, such as parts of Madras, Calcutta, and Bombay, there is no reason to suppose that it can convey malaria in nature. In experiments it is not the same, for by feeding M. rossii on infected cases, it was found to carry malaria, though perhaps not so readily as other species; and before we consider here the bearing of this question of species on malaria endemicity, it will be well to mention here that similar results to ours in India have recently been gotten in the United States. Thus

Hirshberg made a series of comparative experiments by feeding A. maculipennis and A. punctipennis respectively on the same cases of malignant tertian malaria , placing each species under the same conditions. The two species were kept in the incubator at 30° C. Under these conditions sporozoites developed in eight days. Out of 58 A. punctipennis, none developed sporozoites, while out of 48 A. maculipennis, 8 were positive. It is very evident then that the species of anopheline is a most important factor, and in all probability it will be found that in nature also A. punctipennis does not carry malaria, but we would again observe that this is a point that must not be assumed, but decided by actual dissections of this species caught in houses where malaria prevails. If we may venture to draw conclusions from a single series of experiments, we should be inclined to think that here also A. maculipennis was not a particularly good carrier as compared with tropical species, but the number of times that the mosquitos were infected by feeding may be the explanation of this. Of the actual cause of this difference in carrying power of two species of the same genus we have at present no explanation. This is one aspect of the importance of species, but we must consider another phenomenon which is equally peculiar. We have already noted the disappearance of malaria from districts where it formerly prevailed, and have noted the explanations generally given for this fact, but the following instances will show that we can have all (?) the conditions necessary for the appearance of malaria in a tropical country, and yet a complete absence of malaria. Thus in a district of Madras investigated by Christophers and myself in India we found in two villages, about a mile apart, with the usual teeming native population, myriads of anophelines and numerous breeding places,-and, moreover, here the anopheline present was M. culicifacies, one which we knew to be an efficient carrier of malaria,- that an examination of the blood of the native children showed that they were completely free from parasites. The possibility of the introduction of malaria into these villages, if we must suppose this necessary, was a daily occurring one, for the villages were the halting stages of the large traffic passing up and down the country. Malaria was rife in the district, moreover, and 15 miles away, at an elevation of 1000 feet, the endemic index was 50. These are instances of what Celli has termed Paludismus sine malaria. A minute examination of these villages and the conditions prevailing there failed to give us any clue as to the explanation of this remarkable condition. In the light of so exceptional a condition in a malarial district we think that the explanations usually given for the disappearance of malaria from a given spot are inadequate. While giving their due importance to these exceptional cases, we must not overlook the very close relationship between the distribution of the Anophelince and endemic malaria, and yet, while we can establish this, we shall find that in considering even this normal relationship we encounter peculiarities in the endemicity or malarial index of localities that we cannot at present explain. In endeavoring, then, to elucidate the factors responsible for the varying endemic index in different parts of India, Christophers and myself selected a number of villages, isolated, but all within a circle of about 10 miles in diameter. Our method was to determine the endemic index of each village, and, at the same time, the species of Anophelince caught in the native huts, to map out accurately every breeding place, actual and potential, and the larvae found there, to note what breeding places had dried up, and, in fact, to determine, as exhaustively as possible, every factor that could possibly influence the endemic index. We examined in this way 10 villages. The differences in the endemic index were striking, varying from 45 per cent, to 5 per cent. The differences depended on one factor, the presence or absence of Anophelince in the native huts, and this depended directly on another factor, the presence or absence of breeding grounds. Thus, to take two examples: in the first case, where the endemic index was 45, anophelines were caught with ease in the huts (and among them M. culicifacies), and their breeding grounds were only a few yards away, whereas where the endemic index was only 5, no anophelines were found by careful search in the huts, and there were no breeding grounds nearer than half a mile. It is, perhaps, necessary to explain how the endemic index was 5 and not zero where no Anophelince existed, but this is due to a residual infection persisting after the rainy season, when throughout the country it would be possible for Anophelince to breed in many places.