344 C. r. M. SWYNNERTON. 



any rate and under certain local conditions, be exceedingly localised. It is a fact 

 also that the two animals that I have suggested to be the fly's most important 

 carriers are amongst the three that suffered worst from the rinderpest. It is possible, 

 therefore, that the actual effect of the rinderpest in certain areas may have been 

 not directly to starve the fly, but to confine it the year through to a far greater 

 extent to its dry season centres by the destruction of its chief carriers. The narrow- 

 ing in this way of the area tapped for food in the wet season might react also on the 

 numbers of the fly. The suggestion is tentative and perhaps may not be borne 

 out in the belts actually reduced, especially if man, a carrier of morsitans, was 

 present in numbers, living or passing, about the dry season centres. 



The failure of the rinderpest to destroy the fly here to any appreciable extent 

 teUs decisively against the view that the rinderpest blood was in some way poisonous 

 to the fly — ^for all my informants agreed as to its severity everywhere, except in a 

 portion of Makwiana's highlands. The animals sometimes died " whole herds 

 together," and I myself can witness indirectly to the effect on the buffalos. 



Its failure tells also against the hope that we may exterminate the fly in the 

 Mossurise district by destroying the bigger game only. 



XI. — Factors determining the Presence of Tsetses. 



Shade and Undergrowth. 



It would appear from the present and other observations that the fly must be 

 protected from the drying effect of continuous sunlight either by a sufficient supply 

 of shade, or by a readily accessible moisture supply, or by both ; and that this, 

 with food, is its primary desideratum. 



Glossina brevipalpis, so far as 1 have seen (and I was with it for some weeks in 

 all), relies very greatly indeed on shade and is rarely found away from fairly heavy 

 shade. The requisite degree of shading is provided by wooding with leafy under- 

 growth. This may be either primary forest (in which the undergrowth is sometimes 

 such as to give sufficient shelter alone (PL xiii, fig. 2) or secondary bush in leaf 

 with thickets and sapling clumps below (PI. xii, fig. 2). Of these types the primary 

 forest (PI. ix ; PI. xiii, fig. 2) and, in many places and most seasons, heavy Brachy- 

 stegia bush are the best capable of carrying this tsetse through the dry season. It 

 is not at all dependent on the presence of vleis, and I have found it waiting in all 

 the thickets at some distance from water of any kind and in hot weather in 

 September mth the ground baked. It is true that it was in greater numbers near 

 certain little streams than in the Brachystegia thickets, but (as results from the 

 same type of forest away from streams appeared to show conclusively) this was only 

 because those streams were lined with fringing forest of primary type. 



So dependent does the fly appear to be on good shade that, except in the early 

 morning, after sunset and on dull days, it wiU leave animals it is on as soon as they 

 emerge from the shady bush into the sunlight. On dull and rainy days it will follow 

 freely into the most open country and at high noon, so that the term " crepuscular,'* 

 which has been applied to this fly, is not altogether justified by my observations. 

 Thermometer readings taken at the same time in primary forest, primary forest with 

 its undergrowth cleared, and, thirdly, in a sapling thicket in Uapaca wooding along- 



