324 C. F. M. SWYNNERTON. 



(and moisture retaining) upland dolerite tlian on the sandstone-shale formation 

 at the same elevation. The Brachystegia-Burkea groves of the far drier, deeply 

 cracking, compact basalt soils of the plains, effectively drained by the low-flowing 

 Buzi and highly insolated, were losing their leaves earlier than those of the granite, 

 especially on stony outcrops ; while similar wooding on a quartzite strip on the 

 Buzi, at the same elevation and distance from the river and only half a mile away, 

 was far later than the latter, as was the Brachystegia wooding (doubtless with better 

 rainfall, however) of the sedimentary area of the higher Buzi. 



(5). Altitude, other things being equal, makes a difference. Travelling between 

 Maruma and the escarpment at Makwiana's kraal on 11th August I saw no Albizzia 

 (munjerenshe) or other secondary wooding that was not completely leafless. On 

 approaching the foot of the descent I began to find some in leaf — and at once found 

 fly, both brevipalpis and pallidipes. I have come across many other examples 

 both in our individual tree-species and in our woodland types that were thus remi- 

 niscent of the instance of the plane tree — simultaneously leafless at Innsbruck, 

 yellow-leafed at Garda, dark green at Palermo and in parts of Greece evergreen ; 

 and I have little doubt that latitude must also count for much in this connection. 



Altogether it is certain that no unqualified statement disregarding rainfall, soil- 

 moisture, frosts, elevation, latitude and exceptional seasons can be made for parti- 

 cular types of wooding as regards the period for which they will be without leaves 

 and inhospitable to tsetse. It is, I think, equally certain that an exhaustive study 

 of these factors may be of great assistance not merely in the explanation of local 

 outbreaks, but in the actual control of the fly. Certainly no investigator can at 

 present lay down the law for another area than his own as to the behaviour of 

 particular tvpes of wooding and, therefore, for the seasonal behaviour there of the 



fly- 



A point that cannot be too greatly emphasised is that, if we are to fight the 

 fly economically as well as eflectivety, we must base our measures not on the normal 

 but on the exceptional season. I do not refer to the season that is so exceptional 

 as to occur only once in 50 years — as the great droughts of 1860 (approximately) 

 and 1913 ; nor to cases in which haste is vital ; but, let us say, to the driest or 

 coldest season in ten. 



The effect of the factors that I have mentioned above may be seen even when 

 leaf -fall has been brought about artificially, as by a fire. 



V. — Effect of the Annual Grass-fires. 



These very commonly anticipate leaf-fall. Except in the loftier bush, and to 

 Bome extent there, they scorch and wither even the higher leaves and defoliate the 

 saplings and shrubs, destroying the shade. The very factors, local and seasonal, 

 that accelerate or retard the date of the normal leaf-fall hasten or postpone that 

 of the fires also. In 1918 — except on the uplands where the frosts were considerable 

 and dried the grass — ^the earlier fires were a month or more later than usual ; in 

 1913 they were exceptionally early (after a deficient rainfall) and provided some 

 excellent instances of what might be accomplished by severer annual fires towards 

 the destruction of tsetse-haunted bush. The Buzi flats with their cracking basalt 



