THE TSETSE PROBLEM IN NORTH MOSSURISE. 319 



The last figure, though obtained in British territory, gives an idea of the higher 

 altitudes in the lower part of Gwenzi's country, about the Chinvika River south of 

 Spungabera. 



III. — Wooding of the District Analysed.* 



A classification of woodland types in relation to tsetse would properly be founded 

 on the degree to which and the length of period during which they lose their leaves 

 in winter. It so happens that the division between evergreen (or with high, dense, 

 heavy undergrowth and lianas) and deciduous coincides here in the main with 

 the natural division of the wooding into two principal types. 



These are : — (1) primary forest, pyrophobic or fire-fearing, dominated either 

 by evergreen or deciduous trees ; (2) secondary, pyrophytic or fixe -supporting 

 forest of various sub-types. The latter replaces the former where that is burnt 

 out. Bark adaptations enable it to withstand fires, and it is burnt through annually. 

 Where it so thickens up as to reduce the grass growth, and consequently the severity 

 of the fires, it tends to be re-invaded by the primary forest. A large proportion 

 of the finer rubber forest in the area investigated consisted of secondary forest 

 in course of re-invasion, and this harboured the large tsetse {Glossina brevipalpis). 



Primary forest (PI. ix ; PL xiii, fig. 2) consists of lofty, densely growing trees 

 that support many woody lianas and shade lower tiers of evergreen shrubs and 

 small trees, with a carpet and fringe that do not readily bum. Primary forest of 

 the '* rain-forest " type exists in the highlands, mostly in small patches at 

 Spungabera and elsewhere, and is characterised by such splendid trees as 

 Khaya nyasica (East African mahogany or mubaba), Chrysophyllum fulvum (large 

 muchanja) and Piptadenia huchanani (umfomoti). The last-named species 

 dominates most forest of this type seen in the Lusitu-Sitatonga rubber country. 

 It gives it almost the character of " monsoon " forest, for it differs from the others 

 in being regularly deciduous ; but the period of leaflessness is not very long, and 

 the lianas and high close-packed evergreen shrubs (PL xiii, fig. 2) continue to 

 afford the requisite shade for the forest fly, G. brevipalpis, and to favour its 

 activities at all hours of the day. 



There is much striking evidence in favour of the view that primary forest once 

 covered this country, and that during very many thousands of years past it has 

 gradually been reduced to its present small dimensions through the agency of burning 

 and cultivating man. From this and from the analogy of the distribution of our 

 other forest animals, quite apart from the results of the movements of game, would 

 follow the comforting deduction that primary forests that to-day lack the large 

 tsetse possess some advantage, elevational or otherwise, in regard to it, and are 

 likely therefore never to harbour it for long. Such forests are those of Spungabera, 

 ]\Iount Silinda and the border generally. 



Apparent primary forest here has three sub-types, " mountain," " heavy " and 

 " ravine." The first does not concern us; the second (PL ix) is characterised by 

 the shrubs, Conopharyngia usamharensis and Dracaena fragrans, and the liana, 



♦These descriptions 'of woodland types are necessarily greatly condensed, but their 

 recognition will be facilitated by reference to the accompanying illustrations. 



