NOTES ON THE BIONOMICS OF GLOSSINA MORSITANS. 117 



is shown by the fact that the latest onset of the rains recorded in recent years in the 

 territory, namely in 1912, when no appreciable rain fell till towards the middle of 

 December, was followed next year by the largest extension of fly recorded in connec- 

 tion mth the Sebungwe fly area. 



The Effect of Grass Fires. 



Another point is the stress laid by certain investigators on the efficacy of grass 

 fires in reducing the fly. The present writer has been unable to obtain the shghtest 

 evidence that grass fires in this territory have any material effect on the number of 

 fly, indeed the evidence to the contrary is almost conclusive. One of the most 

 striking instances was recorded on the Gaori River below the escarpment in the 

 Lomagundi district. Fly in November 1910 was still confined to a dry season haunt, 

 wliich consisted in this case of shady trees on the bank of the dry water-course 

 mentioned above. For some reason the grass had not been burnt until three days 

 (according to the natives) before the wTiter's visit, and the presence of still smoul- 

 dering trees and the general appearance of the ashes of the grass showed it to have 

 been burnt extremely recently. The fire had swept through the country over a very 

 wide area. Fly was present in considerable numbers, as akeady mentioned, attacking 

 the party very persistently. When a second visit was made to the spot in April of 

 the next year, it was found that the grass throughout the " focus " stood as high as a 

 man's chin and was so thick that it was far too laborious to attempt to make headway 

 through it. The fire the previous November must therefore have thrown up flames 

 twelve to fifteen feet into the air. These fires are an annual occurrence, here as 

 elsewhere, and yet this little water-course is well knowTi to all using the path which 

 crosses it, as a fly centre, where the pest is always very much in evidence in the dry 

 season. This was, as a matter of fact, the reason why that particular foot-path was 

 followed. Many other less striking instances have come to the writer's notice dm'ing 

 the past ten years, and the accumulated evidence has induced an entire loss of faith 

 in the efiicacy of grass fires in reducing the fly. As a matter of fact if tsetse-flies 

 could not avoid grass fires they would surely long ago have been exterminated in 

 parts of the country inhabited by natives. Of course even in Southern Rhodesia 

 much patchy burning of the grass occurs, but the long, intensely dry winter pre- 

 sumably favours burning over wide areas more than conditions in Uganda, for instance. 

 It is difficult to see how the flies could avoid a roaring furnace like that along the 

 Gorai River by mounting in the air unless they went to a very great height (for a fly), 

 but the contiguous mopani would afiord a refuge, as in this type of forest the grass is 

 always very short, and is sometimes almost altogether wanting. If large areas of 

 country infested with fly were covered with long thick grass better results might be 

 anticipated. Possibly this is the reason that greater effect on the fly has been 

 noticed elsewhere. 



The Value of Mopani Country to the Fly. 



Both Duke and Fiske, referring to Uganda, speak of the fly showing hunger in the 

 long grass season, which is of course the latter part of the wet season and earlier part 

 of the dry. This is attributed to difficulty in locating game. The point is extremely 

 interesting, because from Duke's description of the Masindi area comparatively long 



