360 RUPEKT W. JACK — OBSEr.VATIONS ON THE 



tree was away from the influence of the river, so far as providing shade in 

 August was concerned, and there were no Glnssina congregating near it. Pro- 

 bably the larvae were deposited when the surrounding trees were in leaf 

 (PI. XXI, fig. 2). 



(10) Ababa tree (botanical name unknown) on level ground near the river. 

 The grass had been burnt off on the previous day : soil sandy, but dead leaves 

 and leaf mould absent. One live pupa was found sheltered under a hollow made 

 by previous grass-fires in the trunk. The soil sloped down towards the trunk, 

 but the pupa was in a ridge of soil against the trunk itself (PL XXI, fig. 1). 



(11) Mvumila tree {\)Vohiih\j Kirkia acicminata) an & bank sloping sharply to 

 the liver. Three puparia found in a protected position close to the trunk on 

 the upper side ; soil loose and sandy, vegetable debris abundant. 



Searching in the leaf mould, etc., beneath bushes on the banks of the 

 rivei', near the trees where the pupae were found, again proved fruitless at 

 Sinombi, and this, coupled with the results at Manzituba, and the very prolonged 

 searches made in such situations near the Gorai River, Lomagundi, in November 

 and April, leads me to the opinion that such spots are rarely if ever selected for 

 the deposition of the larvae. The tsetse-fly is such a comparatively slow breeder 

 that it can scarcely aflTord to expose its pupae to the scratchings of the game 

 birds which frequent exactly the same haimts as the fly during the dry weather, 

 and often in amazing numbers. Along the Gorai River in November the 

 numbers of guinea fowl, " pheasant " (PferKM;<?« ) and " redwing " (i^ra?ieo/Mii<.?) 

 were astonishing. They rose at almost every step along the banks, and all the 

 ground imder the bushes had been scratched over and over again. In April 

 they were, however, not more than ordinarily abundant. 



It has already been stated that G. morsitans congregates on the shady banks 

 of rivers and borders of vleis in the dry weather. In the wet season, when the 

 forest is shady, the fly is much more generally distributed through the veld. 

 This statement is based on the following observations : — («.) In the Hartley 

 district there are one or two farms on the railway line, and within three or four 

 miles of a fly-belt on the Surisuri River. It was found that for three consecutive 

 years some cattle on these farms became fly-struck in February and March, but 

 no cases occurred during the dry weather. The wet season opens in December 

 there. The writer took almost monthly observations on the fly-belt in 1909-10, 

 and the fly was taken only near two tributaries of the Surisuri during the dry 

 weather. In February, 1910, however, fly was found in the bush at several points 

 away from these tributaries, and one specimen on the borders of one of the farms 

 where cattle had been struck, within a mile and a half of the homestead, and 

 this (an important factor) on the outroard journey, {b.) BeloAv the Escarpment 

 ill Lomagundi, in November, 1910, G. morsitans, was found congregated on 

 the Gorai River, and on the Dandi and Ambi Rivers near their junction, and 

 also in the shady borders of a vlei not far from the Dandi. In April the fly was 

 not congregated at any of these places. Near the Gorai River, it occurred 

 perhaps mostly in the mopani belts, a mile or two from the river, but was to be 

 found all over the veld, and the same state of things was observed near 

 the Ambi and Dandi Rivers, and in the vlei mentioned above, (r.) Near 



