130 EUPERT W. JACK. 



inroads into the number of buffalo. As a matter of fact, events in this region seem 

 to have contributed largely to the sohdifying of Selous' opinion that the fly was 

 dependent upon buffalo. 



The rainfall in the Zambesi valley frequently differs very considerably from that 

 in the remainder of the territory, and it is unfortunate that no records in this area 

 for the period concerned appear to be available. 



The records, are however, instructive in regard to the influence of a series of seasons 

 of low rainfall. At Kariyangwe in the Sebungwe district the rainfaU recorded for 

 the three seasons 1909-10, 1910-11 and 1911-12 was extremely low. This station 

 was then closed down, but referring to the Wankie records, somewhat further from 

 the scene, we find that the next two years were also very low indeed. During this 

 period the extension of the Sebungwe fly area proceeded at its maximum speed, as 

 may be seen by a glance at the author's map (Bull. Ent. Kes. x, map 1). The 

 records of advance in this area are moderately accurate as regards the three year 

 intervals marked on the map, with the exception of the north, where the discovery 

 of tsetse-fly on the lower Sengwe River in 1913 indicated that the previous hmits in 

 this area, visited only at long intervals by officials who traversed the country rapidly,, 

 were probably faulty. It is not possible, however, to obtain accurate yearly records. 

 The dry season of 1913 was remarkable for advance in almost every direction, ex- 

 cluding a portion of the western hmit where the fly appears to have reached the 

 division between favourable and unfavourable country. The rains were very late 

 in 1912, no serious showers falling until near the middle of December. Dry season 

 conditions were thus unusually prolonged, and it is beheved that such conditions 

 favour increase of fly. 



To recapitulate, whilst the writer is far from attaching any exaggerated importance 

 to these hmited records, they are certainly suggestive of abnormally heavy wet 

 seasons having a deleterious effect on the fly in certain areas, and of a series of seasons, 

 of low rainfall, especially if late, favouring increase. Intense and prolonged dry 

 season conditions, such as occur during a cycle of years of low rainfaU, are judged to 

 favour the fly at centres which resist the action of drought best, owing to the con- 

 sequent concentration of game. A rise in the water-table after an exceptionally 

 heavy season or two causes a disturbance in the usual conditions by inducing a more 

 general distribution of good grazing and water throughout the country, and a pro- 

 portionately reduced attraction of the " focus " to game. 



It is much to be regretted that the writer's itinerary in 1918 did not include Sipane 

 Vlei. The tremendous rains of the 1917-18 season had a very marked effect on the 

 forest lying between the railway hne and the Gwaai River on the western side of the 

 territory, and it is reasonable to suppose that conditions at Sipane Vlei may have 

 been similar to those in 1919, seeing that at Gokwe, the nearest meteorological station, 

 43'03 inches feU in the 1917-18 season and 33*10 inches the 1918-19 season, although 

 even the latter is well above the normal for that district. If the writer's theory is 

 correct, fly should have been numerous and hungry in the dry season of 1918, pupae 

 being, however, scarce, resulting in a great diminution of fly by 1919. 



This theory is very much in accordance with observations pubHshed by Lloyd 

 (Bull. Ent. Res. vii, pp. 67-79) with reference to Northern Rhodesia. Lloyd draws 



