THE TSETSE PROBLEM IN NORTH :M0SSURISE. 321 



foliage, creviced stem and branches and furrowed, protectively coloured bark, is 

 a particularly useful tree to the tsetse and is abundant in Brachysteyia bush at 

 low elevations. It is the foreground tree in PL xvi, fig. 1. 



Tondo-bush is the tsetse bush par excellence. Brachystegia-Uapaca wooding 

 in our present connection is divisible into three varieties, two of which carry each 

 its own tsetse. It is divisible into (a) high "itondo," characterized by the greater 

 size of the trees and generally denser wooding and forming fine " savannah " forest 

 (PI. xvi, fig. 1) ; (6) lower, more open *' itondo," still mostly savannah forest, 

 but poorer and intermixed with more definite savannah (PI. xv, fig. 1). It is 

 (with other conditions correct) the special haunt of G. niorsitans. The former, 

 finer variety is divisible further into (1) clean itondo, devoid of woody undergrowth 

 and of the large fly (PI. xiii, fig. 1), and (2) itondo with woody undergrowth, 

 inhabited by G. brevipalpis (PL xii, fig. 2). 



This undergrowth, usually patchy, consists for the most part of saplings of the 

 trees themselves, often in close-packed clumps (PL xii, fig. 2 ; PL xvi, fig, 1), 

 and of large scattered shrubs such as the umtunduluku (Xhnenia americana) and 

 the chigwenderes (Bauhinia galpini and B. petersiana) ; but in parts of the Iiigh 

 itondo may be found numerous clumps of " ravine " type primary forest 

 undergrowth, extensive or small, of bird and baboon-carried genera that have made 

 a start and maintained themselves where the ill-regulated fires have enabled thein 

 to do so. PL xiv, fig. 1, gives some idea of their nature. 



Two highly important points about itondo in relation to tsetses are these : — 

 (a) The grass-gro\\i;h in general below high tondo-bush is sparser than in some of our 

 other secondary sub-types and does not burn really efiectively till late in the season ; 

 (6) itondo as a whole, under our local conditions of rainfall, tends to retain its 

 leaves through the winter better than any other of our secondary formations. The 

 mutsatsas {Brachysteyia bragaei, etc.) have often regained their leaves before the 

 associated mutondo trees {B. glohiflora) have lost theirs, and the Uapacas only 

 lose theirs (and that not completely) when nearly all the other species are busy 

 regaining their leaves. This tendency to keep in leaf without some slight 

 interregnum is less marked at high elevations, in particularly cold and dry 

 winters, and in the di'ier parts of the lowland areas; but on the Buzi, and in the 

 " Oblong " (see Map) generally, it was sufficient in 1918 to carry G. brevipalpis 

 in numbers everywhere through the winter, and it is probable that in most if not 

 all years it would do so in most parts of this piece of wooding. 



(4). Dense secondary forest includes especially (a) the very fine t3rpe of wooding 

 dominated" by musara and musunganyemba (Milletia stuhhnanni and Pteleopsis 

 myrtifolia) — definite " monsoon forest " ; and (6) a more mixed type. In places, 

 owing often to the coppicing efiect of unthorough native cultivation with the good 

 seed-bed offered by old gardens to seeds from the surrounding bush, and to the 

 relative exclusion of fire consequent on the absence of grass, the ordinary trees of the 

 savannah form more or less dense thickets (PL xvii, fig. 1). Eventually trees and 

 shrubs that are normally found on the outskirts of primary forest introduce them- 

 gelves and finally dominate. Such are the munjerenshe (Albizzia fastigiata chirin- 

 densis), the large umkadhlo {Rauivolfia inebrians) and the fine shrubs Vernmiia 



