For some years no,% investigators have studied the changes in numbers of tsetse flies by means of regular catches conducted over fixed routes, and the changes in mean weight and its constitution in samples of flies. For the most part these investigations have failed to afford any explanation of why different species are more or less segregated in distinct habitats, a phenomenon inexplicable by a priori reasoning, and one which therefore offers hope of unpredictable discoveries, and perhaps ultimately of practical advantage. A notable exception is the work of Nash (1936, 1940) in Nigeria, who has observed that Glossina tachinoides Westwood is somewhat more susceptible than is G. submorsitans Newstead to very high temperatures, and he has been able to relate the differences in late dry-season habits of concentration of these two species to the difference in viability. It seemed that in East Africa it might be profitable to compare the two closely related species G. morsitans Westwood and G. swynnertoni Austen, in order to obtain some clue to the reason for their vegetational or climatic preferences. It should be recalled that G. morsitans is generally associated in East Africa with what is known as the miombo formation, dominated by associations of Isoberlinia-Brachystegia broad-leaved deciduous woodland; while G. swynnertoni seems to avoid this type and is found in the thornbush dominated by Acacia and Commiphora communities. There are, however, many known exceptions in the distribution of G. morsitans even in Tanganyika. It was considered that the best way to explore this subject was to introduce each of the two species into the habitat of the other, and to throw in also G. palpalis Robineau-Desvoidy, normally a waterside species, as likely to afford a more striking indication of how a tsetse community might be expected to appear when introduced into an environment to which it was presumably quite unsuited. Some preliminary field observations have now been made, and have been supported by a small amount of laboratory work. So far it has been possible only to introduce G. morsitans (twice) and G. palpalis (once) into the habitat of G. swynnertoni (living with G. pallidipes Austen); the converse experiment of putting G. suynnertoni and G. palpalis into the habitat of G. morsitans has not yet been attempted, but it is hoped to do this in the future. Observations of interest bearing on tsetse population studies, and on the accumulation of fat, have been made in the course of this investigation, but discussion of them will be reserved to a more appropriate occasion. Flies which have lost weight rapidly, or which have not recently fed, are deficient in water, fat and residual dry weight, especially the first two. Under favourable conditions, also, fat and residual dry weight increase with age, at least for some weeks. The weight of fat is therefore an indication of the well-being of the fly community at any one time of year. However, the weight of flies at emergence varies with the seasons, smaller individuals emerging at the hotter times of year; and in any batch of flies emerging on the same day there is always considerable variation in weight, which is related to variation in size. To correct weight for size, I have made use of a wing-vein measurement (see below). This method is not satisfactory for individual flies, but works fairly well for batches.