1. Shinyanga is divisible into:

(a) Thorn-bush country containing plenty of tsetse but no population (Pl. x, figs. 1 and 2). This now covers far more country than it did fifty years ago, owing, first, to depopulation caused by natives wars, secondly, to depopulation caused in the past fifteen years by (chiefly) the advance of tsetse; and its distribution has also changed greatly.

(b) Open grass-land, produced in the past from the thorn bush by the settlement in it of people in sufficient numbers to clear it and keep it clear through their cutting for firewood and building and the browsing of their stock (Pl. xi, fig. 2; Pl xv, fig. 1).

2. Except in one respect the position in the open country is ideal. There is complete segregation of tsetse and people, and a situation exists in consequence amounting, for practical purposes, to control of the tsetse. All the people are able to keep cattle and all the children to get milk, tsetse are excluded, sleeping-sickness is excluded, and, owing to the compactness of settlement, mammalian pests are excluded except from the margins, and the people generally are accessible for purposes of administration, medical treatment, education, and agricultural development.

3. The exception is this. The tree roots still survive in the open country and, whenever population thins for a time, shoot up into bushes and trees. Where these are adjacent to tsetse bush, the fly is carried into them, while they are still small, on persons passing to and fro, and cattle begin to die. The people then fall back, allowing new bush to spring up and the process to be repeated. In Shinyanga this retreat and the reconquest by fly has been for some time at the rate of a mile a year, and new “fronts” have been formed, so that the cattle-areas are now nearly surrounded.

4. The position is serious and urgent. If it had not been taken in hand the cattle areas were doomed, as were the present schemes of development, and the railway now planned on the strength of the presence of population might have led ultimately to a depopulated waste.

5. The remedy is in the hands of the people themselves. They have merely each year to combine to cut back the young shoots and the bush and the tsetse cannot advance. And persons must not allow the flies to travel on them into cleared areas.

6. But the retreat, and natural increase of stock, have caused congestion of the cattle in the areas still populated, with severe shortage of grazing, and deaths; and the destruction of grass has led to erosion and serious loss of soil (Pl. ix, figs. 1 and 2). The remedy is again easy. When the people come out each year to cut back the young shoots, they should take the opportunity to clear for some little distance into the bush and so get enough grazing. Also, any new agriculture likely to demand much ground, and so still further to reduce the grazing, should be located so far as possible in the fly-bush. All this was explained to the people, and they asked us to organise an effort on their part. This we did, in May and June 1924, when about ten thousand men went into camp under their respective Sultans and headmen and cleared in the most energetic manner many square miles of tsetse bush on a great part of the Shinyanga tsetse “front.” It only remains to organise this effort as a popular annual custom and to locate the clearings knowledgeably and in conformity with a thoughtout plan in order gradually to break up and destroy the fly-belt. The lessons as to management that were learned from this year's clearing are detailed on p. 331.

7. A small, typical, infested block of the fly-bush has been cut off by clearing and an experiment initiated: (a) in-inducing natives to settle in the barrier, so as to widen it automatically and keep it clear; (b) in clearing the block itself of fly; and (c) in testing the width of barrier which, combined with such precautions as may prove of use, will prevent the block from becoming reinfested from the main bush. The initial response to the inducements offered (p. 000) has been good.

8. The first step towards breaking up the fly-belt on a greater scale has been taken by clearing, to a width of 400 yards on either side, a road traversing a point (Samuye to Kizumbi) at which the fly-belt narrows to seven miles. It is planned to make (and greatly widen) further breaks and to attack in detail the blocks broken off, should it prove difficult otherwise to prevent reinfestation of country that may be cleared of fly by grass-burning, etc.

9. Felled bush continued to harbour fly for some time, at any rate if people and game continued to wander between it and the bush as yet unfelled. Complete destruction of the timber by twice piling and burning it made it uninhabitable by fly, but if time is of no account, the bush can lie till the twigs and small branches of it are warped downwards into a closer mass by the next rains and so are destroyed by the grass fires of the following year.

10. A direct prohibition, enforced by a patrolling guard here and there, reduced nearly to nothing the carrying of flies into cleared or cattle country by man. Game wandered in and out almost entirely at night. The stationing of a pair of fly-boys on a road with nets, combined ultimately with dark shelters alongside for the attraction of the flies, was most useful during an entire year in preventing the carrying in of flies by cars and wayfarers. The attachment of brown paper smeared with “rat varnish” to the hoods of cars and the backs of natives was useful in destroying flies frequenting a road.

11. Mere cutting of tsetse bush is useless, for it springs again, so that it is futile to try to gain ground by this means faster than we can consolidate it. The following measures for the consolidation of clearing are being tested now in Shinyanga: The introduction of native settlement to keep the ground clear automatically; the special cutting down once or twice a year by each village of the new shoots for a few hundred yards round; the encouragement of the keeping of goats; the diversion of large scale agriculture by ploughing, which entails stumping, to the clearings it is wished to consolidate; the digging up of stumps by the natives for firewood; and the killing of trees by poisons and particular forms of ring-barking.

12. The annual postponement of the grass fires till October, and then lighting them on an organised plan, was advocated some years ago (Bull. Ent. Res. xi, 1921, p. 382), and was encouraged by circular in Tanganyika Territory in 1921, when also I carried out a successful experiment near Kilosa (Pl. xv, figs. 1, 2). Father Cirvegna took up the organisation of the measure over a small piece of fly-belt at Madibira, in the Iringa district, and there seems to be little doubt that he has cleared the area of fly by means of it in three burnings.

A large scale experiment by myself this year in the Nzega and Shinyanga subdistricts seemed to show: (a) that by means of native guards, each with an area, and some aid from the headmen, it is perfectly possible to preserve a great piece of country nearly unburned; (b) that the late grass fire, lighted on a broad front with a good wind and followed up by the burners, produces, when the grass growth is sufficient, a “drive” of the flies on the wing; these then become congregated in enormous numbers beyond the limits of the fire and in any appreciable patches of ground which escape it; (c) that in these places, by the prompt use of nets and bird-lime, it is possible nearly to exterminate the congregated flies. The fewer these congregations the more practicable is this measure, so that the destruction of the more extensive fire-excluding thickets by hand will be a useful preliminary measure; (d) that far more numerous breeding thickets and breeding logs are destroyed by such a fire than by an earlier or ill-organised fire, and that with the expenditure of some labour in the littering with dry grass of thickets and logs unlikely to be burned otherwise all may be burned, or many may be cleared by hand; (e) that a much larger percentage of dead pupae (52 per cent. in this case) may be found afterwards in breeding places thus burned through than in those that are not burned (9 per cent.).

13. Some of these indications merely confirm results previously obtained by me in Rhodesia and Tanganyika; others need to be confirmed by means of further experimentation. Meantime, combined with Cirvegna's result, they suggest that late organised grass burning is likely to be of value over the very considerable areas in Africa in which the grass and the dry season are long and organisation is possible. But even if this should definitely prove true, means of preventing reinfestation must be studied, for these may turn out to be a serious difficulty.

14. The relations of the fly and the game are not being ignored. They are being studied and will be the subject of special experimentation and, if necessary, of special localised measures fitted into our general scheme of control.

15. This scheme amounts to an attempt to destroy and control tsetse by means of agencies and resources already in existence and merely requiring to be diverted specially to that purpose and organised, and by enlisting the co-operation of the people themselves who are to be aided and protected. This is both the most practical way of dealing with the problem and the way that best accords with the sound policy of educating the natives to understand and attack their own problems and advance themselves through their own efforts.