Tsetse flies (Glossina morsitans Westw.) were marked and released at Daga-Iloi, in Tanganyika on four or, usually, six days weekly for over 15 months, at 75 catching points situated 100 yards apart on an “octagonal spiral” whose successive rings were separated by 250 yards. Outer “spirals” were also traversed occasionally.

The country was hilly and dominated by woodland of Brachystegia tamarindoides, broken by more open grassland with Combretum on the one hand and by thicketed ravines on the other. The ungulate animals included zebra, buffalo, bushbuck, roan antelope, hartebeest, reedbuck, duiker, wart-hog and bushpig, with very occasional elephant, rhinoceros, waterbuck and klipspringer. Baboons and Cercopithecus monkeys were also present. Spoors of the ungulates were recorded on the daily rounds: bushbuck were most common about the ravines, but in general the ungulates tended to be most common in the open grassland with Combretum.

G. morsitans was most numerous in tall Brachystegia woodland where the slope was moderate, and where open grasslands were not too far away. Thicketed ravines were avoided. The highest apparent densities were observed on the outer “spirals”, perhaps because of disturbance on the inner one traversed almost every day.

During the rains, well-marked concentrations of flies occurred in certain parts of the Brachystegia woodland, and these re-formed in the second year. The hungriest flies were always found outside the Brachystegia, and hunger was higher in the dry season. Hunger was measured objectively by the proportion of non-teneral males alighting head-upwards on members of the catching party, which hungry flies mostly do.

There was a negative correlation between numbers of flies and numbers of ungulates, and a positive correlation between the numbers of ungulates and hunger of the flies. This is in accordance with the feeding-ground theory, and is very generally observed with G. morsitans in East Africa. Certain movements of individual flies also accorded with the theory, against which, however, the suggestion has now been made that flies caught in the putative feeding grounds have not ventured out into them until attracted there by the sight of the catching party. This suggestion is to some extent countered by quotations from previous work, but it deserves further thought, and carefully planned experiments.

Passage of large herds of buffalo through portions of the “spiral” were not apparently accompanied by mass movements of flies, even when the animals passed by day and many flies fed on them.