In connection with studies on the control of tsetse flies (Glossina spp.) by insecticides, an investigation was made in Tanganyika Territory of the influence of a number of factors on the distribution of insecticide in woodland and open country. A solution of DDT in oil was disseminated as a coarse aerosol from an aircraft and its behaviour traced by the mortality suffered by caged flies. The test insects were wild-caught Musca (Eumusca) lusoria Wied., which proved more susceptible to the insecticide used than did Glossina. palpalis fuscipes Newst.

In open country, complete kills were obtained with the standard application rate (defined as one run using 10 gals, per minute of a 10 per cent, solution of DDT) for at least three hundred yds. downwind of the line of emission, in all the atmospheric conditions encountered.

Tests of the possibility of treating very thin woodland in comparatively high winds (for aerial dispersals of aerosols) and slightly unstable conditions were spoilt by variations in the emission rate of the insecticide, but it was shown that quite small numbers of twigs upwind provided a considerable degree of protection to the test insects and that the kills in cages completely exposed on the upwind side of trees were considerably less than at corresponding distances downwind in completely open country. There was evidence that in high winds and slightly unstable conditions, penetration through the trees was better than with moderate winds and small inversions, but in both conditions there was mortality behind obstacles.

Preliminary comparisons between the kill in the open and in continuous woodland composed of Acacia xanthophloea showed that in the latter mortality was greatly reduced and did not approach that frequently obtained in practical large-scale experiments in tsetse control. A search was made for factors which would increase the level of mortality.

Increased atmospheric stability caused greater average mortality at the expense of evenness of kill. This was not due solely to decreasing effective swathe width due to lighter winds.

Large natural openings in the canopy assisted the penetration of insecticide in unstable conditions, but in stable air the kill around and downwind of clearings was reduced compared with kills in unbroken woodland.

The principal factor in raising mortality to a generally high level was the summation of sublethal doses due to drifting downwind of the fringes of successive parallel swathes. The highest kill obtained in any of these trials was from summation of three swathes emitted 75 yds. apart, which produced a mean mortality of 84 per cent, for 75 yds. downwind of the third run.

An attempt to cover the whole area of woodland by emitting insecticide in a series of parallel runs, as in actual control procedure, showed that in conditions of low turbulence there yet may be great unevenness in the kill obtained. Mortality varied between 21 and 100 per cent., and this irregularity was almost certainly due to the fact that during each of the runs the wind was almost directly along the aircraft track. Such contingencies are unavoidable, for unsteadiness of wind direction is a constant feature of the stable, non-turbulent conditions in the free air that are needed to permit any substantial kill within continuous woodland.

The relation between these results and those obtained in practical control experiments is discussed. It is concluded that the use of caged insects and a particular type of woodland probably accounts for the comparatively low kills obtained in this investigation, from which, nevertheless, valid conclusions can be drawn concerning the variation in mortality from place to place.