1. From a block of country of four square miles, infested by Glossina swynner-toni and a few G. pallidipes, grass-fires have been excluded for three years in succession. The tsetse-population has decreased in that time by nearly nine-tenths (from 9.4 to 1.0 flies per 1,000 yards of transect). In a block of two square miles it has decreased in two years without burning to the same average of one fly per thousand yards of transect from a high previous figure unmeasured. Two “barrier” strips following roads that previously were highly infested have also been unburned for three and four years respectively and have become similarly inhospitable to the flies.
2. In a control block of only 1¾ square miles, the grass has been burned in the ordinary native fashion annually during the same period. The average fly-catch per thousand yards of transect has increased since this was done from 4.2 to 13.1.
3. In each of these blocks game is about as plentiful as it was at the first. The 4-square-mile block shows a slight gain.
4. In the four experimental blocks and strips, but not in the controls, extensive thicket is growing up everywhere in the grass as a result of the exclusion of the fires. Thus the previous marked difference, both spatial and seasonal, between the respective densities of the different vegetational communities is commencing to disappear. Much grass-country remains but its average condition is one of long grass rising from a mulch, thick or thin and sometimes absent, of the previous year's grass that keeps the soil damp after rain—the more so since run-off is checked. Longer and denser grass is encroaching into the previously very open concentration sites of G. swynnertoni, which are also being invaded by thicket.
5. The contrasts between burned and unburned country are especially strong in the short wet season. At this time in burned country the flies have just gone through a period of intense desiccation, short hunger cycle and hardship, and the population is at its annual minimum. The ground, however, is very open and visibility of the other sex and of food animals and breeding-places is unimpeded. Conditions are suddenly made ideal by the advent of the rains. Humidity is added to temperature; the hunger cycle is lengthened; the openness and insolation of the ground now probably play a further beneficent part by preventing an excess of humidity. The flies are active, they breed freely, the shade conditions enable them to spread, and the density curve soars upward.
On the unburned ground the water-content of the soil is increased in the early rains, dampness is retained between showers, the temperature is lowered, evaporation is reduced, and light intensity in the grass becomes less (Scott, 1934). The activity and breeding of the flies is quite likely affected. The striking increase in the population that characterised unburned country in the early rams and short dry season is absent, and the density, having failed to rise at this time, continues low or falls further.
6. In two cases (G. morsitans, Nash 1933, at Kandaga; G. swynnertoni in Block 11 at Shinyanga) the second annual impetus to breeding, associated normally with the advent of the long dry season, duly produced its usual effect on the tsetse-populations of country that was in its first year of no burning. In the case of the block that has been unburned for three years this effect has been slight.
7. That the conditions, though deteriorating, are as yet not lethal through hunger, due to poor visibility or otherwise, is suggested by the fact that the very small fly populations referred to maintain a nearly stationary balance in at least the larger experimental block and do not show special hunger or other sign of distress. It is possible that most of the diminution hitherto has been caused merely by emigration to better conditions, for the blocks are not yet fully isolated. On the other hand, were the main effect on the breeding, distress would not be detected and the result might take place independently of any emigration. Reduced visibility of food animals may become operative later.
8. It is suggested by the results so far on the tsetse and by the present appearance of the country in these four experiments that with the further advance of the vegetational succession towards its natural climax the remaining fly population will probably dwindle further and disappear—at least if the areas are isolated; the growing homogeneity in the density of the whole vegetation will reach a point at which it is unlikely to satisfy the daily and seasonal needs of an insect which for food, breeding and regulation of moisture and temperature, has a separate and vital use for each of the three densities furnished by thicket, savannah wooding and nearly open hard-pan and vlei.
9. The concentration sites of G. swynnertoni combine in a small space the types of vegetation the flies need, both the denser and the more open. They are important at all times but especially as refuges from unfavourable conditions elsewhere and as meeting-grounds of the sexes when the population is reduced. They are operative all the year round, and an effective change in these refuges alone in either the wet season or dry should be a shrewd blow at the fly. In experiments elsewhere in Shinyanga the clearing of the thickets at these places has quickly brought down the fly-population through its dry-season effect. In Blocks 4A and 10A the cessation of grass-burning is apparently already affecting these sites by hemming them in with areas presenting unfavourable conditions—probably mainly through excessive shade and over-humidity in the rains. The cessation of burning might even suffice if confined to these concentration sites and their neighbourhood.
10. As regards the probable effect of not burning the grass on other species of tsetse, it is considered that the method may prove useful against G. morsitans under moderately extensive vegetational conditions, though it may in this case take much longer and may not apply to large areas. Berlinia-Brachystegia wooding went to thicket in the writer's early experiments in a specially high rainfall area in Rhodesia, and (under lower rainfall) the exclusion of fire for five years at Itundwe (Central Province, Tanganyika) has produced strong densification in the Berlinia. In the Kikore plots (5 years protected) savannah type seedlings are being suppressed by the new conditions, but the flourishing state of the few shade-tolerating thicket plants suggests that all that is wanted is the introduction of a suitable seed-supply. This will be necessary in any case in all places without natural regeneration.
11. In an observation by Nash (Kandaga, Central Province) G. morsitans reacted initially in much the same way as G. swynnertoni, but the experiment was interrupted. In Itundwe G. morsitans and its pupae are much less numerous in the unburned Berlinia than in the burned.
12. The reaction of G. pallidipes, as distinguished from that of its habitat, is uncertain but so far has been promising under Shinyanga conditions; but the new conditions at Itundwe look as though they might favour this fly. G. brevipalpis is definitely likely to be favoured by the effect of not burning the grass.
13. The increase in humidity and lowering of the temperature produced by not burning the grass may be expected to be specially effective towards the wet side of a tsetse-fly's range. Towards the dry side it might even at first assist the fly.
14. Experiments to test the possibility of success against morsitans are being taken in hand, as are experiments on larger pieces of country infested by swynnertoni and pallidipes.
15. The control of grass-fires offers difficulties, but our experience over a wide area and many years shows that with determination and organisation it is quite possible, though an occasional minor set-back must be expected. The advantages of this measure from other points of view than that of tsetse extirpation, and the means of restoring the pasture, are referred to on p. 425.
