THE TSETSE PROBLEM IN NORTH MOSSURISE. 323 



visit on this trip, but which (from what I saw of them in 1906-7) I should quite 

 suspect of being capable of harbouring G. jjallidipes and, in many places, 

 G. br^vipalpis. 



IV. — Factors Influencing Leaf-fall. 



I have described above the relative deciduousness of our various types of bush. 

 We have seen that primary forest (PI. ix), and bush which it is seriously invading, 

 is never leafless enough here to discourage its special fly — hrevipalpis ; and that 

 of the secondary types, Brackystegia wooding (PI. xvi, fig. 1) and particular varieties 

 of the " dense secondary " formation tend to retain in this district sufficient 

 leaf to carry fly through the year. 



However, within any particular type of wooding actual leaf-fall is hastened 

 locally, and the gaining of the leaf correspondingly deferred, by several factors. It 

 is true that some of the differences are apparently individual, though unseen root 

 disabilities might account for them. It is at^nv rate of frequent occurrence to 

 see two trees of the same species and size (e.g. of Pterocarpus) standing together, 

 one of them leafless the other not. Ficus subcalcarata (the mutowetowe fig, an 

 isolated large specimen of which will sometimes harbour fly) is an extraordinarily 

 erratic leaf-shedder. But general rules are apparent also. 



(1). Season affects leaf-fall. After the abnormally heavy rains of 1917-1918 

 (2,514*5 mm. at Spungabera) leaf -fall generally took place, I should say, quite 

 three weeks later than usual, where frost was absent. A low rainfall (such as that 

 of 1913), or early and severe cold, hastens the phenomenon. It must be remembered 

 further that rainfall, hence leaf-fall, varies with locality, particularly in relation 

 to the mountain ranges. The rainfall appears to be distinctly lower east of the 

 Sitatongas than on them or even west of them, and is lower here again than on 

 the hills of Spungabera, the '' Jihu " (Gwenzi's) and the British border. 



(2). Proximity of moisture in the soil affects leaf-fall in two opposite ways. 

 On damp low-lying ground liable to frost the leaf-fall is hastened, and even where 

 there is no actual frost, trees bordering on ill-drained, cold, swampy ground tend 

 annually to lose their leaves earlier then those of the same species that stand back 

 on well-drained soil. But trees standing on warm, damp soil lose their leaves 

 later than the latter. These differences are to be clearly seen in fly-harbouring 

 Brackystegia woods, but the particular illustration I shall quote is concerned with 

 some scattered wooding of Pterocarpus angolensis (mubvangazi — PL x, fig. 1) that 

 is under my continuous observation. Here not only is there, in May, an increased 

 leaflessness as one approaches the cold, damp lower valleys, but of the trees within 

 the latter, those with their roots raised on ant-heaps are practically in full leaf 

 when those on the general ground-level are completely leafless. 



(3). Diflerences in the amount of humus present has appeared to account for 

 a difference in leaf-fall between woods of the same species — e.g. of Pterocarpus 

 sericeus (mumbhungu) on the dolerite — and it is to be noted that ant-heaps exert 

 an influence in this direction also. It is probably again mainly a matter of relative 

 moisture retention. 



(4). Geological formation affects leaf-fall once more on the same lines. I have 

 noticed that Pterocarjms sericeus loses its leaves a good deal later on the fertile 



