•70 LL. LLOYD. 



36 days, so that, the period during which the pupae might have been deposited was 

 from 5th August to 10th. September. The days on which the flies emerged were 

 noted and the dates of deposition thus calculated. It was found that six pupae had 

 been deposited before 24:tli August, and 83 between this date and 10th September. 

 (Twenty-two were broken in collecting and 21 died. The mortality was not in this 

 case apparently due to the fires, since in series collected in unburnt country the 

 mortality was about the same, and was due to some cause at present unknown ; 

 possibly slight damage in collecting, or some unhealthy condition in the fly tubes, 

 e.g., dessication). Free breeding in this area therefore began about ten days after 

 the fire had passed and the new grass in the swamp commenced to grow. In all the 

 other series at Ngoa the deposition of the pupae had been fairly even over the possible 

 periods between 28th July and 27th September. Fly in this area was very trouble- 

 some, but the game had been absent some days when the collection was made. 



(4). Consists of the edge of a wide vlei along the Kanchibia River, and some old 

 native gardens very near. The area had been burnt and had a good new growth of 

 grass when examined. At all seasons this is a popular place with the game animals, 

 which feed on the vlei and retire to the old gardens. In six collections 718 pupae and 

 521 cases were taken. Fly was seen in very small numbers and the workers were 

 rarely bitten. 



(5). This consists of a piece of flat country bordering a wide vlei called the Pembe. 

 The vegetation consists of long grass and bracken with scattered trees, and in one or 

 two places small patches of ordinary forest. The whole of the country had been 

 burnt off a month previously and the new growth was well developed. Game is 

 always to be found here and the whole vlei was pitted with the spoor, while deeply 

 cut game paths ran between the more favoured spots. In four collections 989 pupae 

 and 1,325 cases were taken. As in the previous case, very few tsetse were seen and 

 these were not hungry. It is curious that in this and the previous case no swarms 

 of male flies were seen. A possible explanation is that movement when collecting 

 is very slow, only a few yards at a time with long pauses. 



(6). Along the edges of an unburnt vlei which is often attractive to game, 

 especially zebra and hartebeeste. No recent spoor was seen and the animals had 

 doubtless left for the more attractive burnt country near. One search yielded 10 

 pupae (only two of which produced flies), and 71 empty cases. A great deal of fly 

 was seen and all the workers were badly bitten. 



The above facts are reproduced in Table III. The subject has been discussed in 

 some detail because by comparing the numbers and behaviour of the fly in areas 

 where game is plentiful and those in which it is scanty, some hints may be obtained 

 as to what would be the result of the much desired experiment in game destruction. 

 Half measures appear to be worse than useless. On the other hand, if the game were 

 reduced to a minimum so that the fly is much attracted to man there might be some 

 hope of destroying it by means of nets in very hmited areas. 



The reduction of game in a fly area apparently leads to the following results. The 

 casual observer would notice no reduction at first, but might even see an increase 

 owing to the increased hunger of the fly ; the number of female flies in an indis- 

 criminate catch would rise towards fifty per cent.— an important point if netting were 

 used as a means of exterminating them ; and pupae would be found in fewer numbers, 

 showing that an actual reduction had taken place. 



