XVl] BREEDING LOCALITIES 269 



Nile. The pupae were all found within lo to 25 yards of 

 water, and were in the shelter of banana plants, or of shrubs, 

 or of trees and tangled undergrowth, etc., where the soil con- 

 sisted of crumbhng vegetation. The pupae lay at a depth of 

 from half to a httle more than an inch, and the site was always 

 moderately dry. Zupitza in the Cameroons found the pupse 

 lying somewhat superficially in the humus and moss in the 

 forks of branches, and cracks in the bark of trees, especiaUy 

 in the angles of the leaf-sheaths of palms, at a height of a few 

 inches up to about 12 feet above the ground. 



On the Sesse Islands, Fraser and Marshall found that the 

 most favourable site for the pupae was the loose dry sand on 

 the shore of the water, close to the undergrowth that edged 

 the forest and sheltered to some extent by the trees. The 

 distance from high water mark was usually five yards and 

 never more than 15. Later Fraser found that the pupa 

 were very numerous in sandy parts of the lake shore, as many 

 as 800 being collected in one day. 



Carpenter, in Uganda, has confirmed these observations 

 and found that the locality where most pupae were obtained 

 (2000 to 3000 monthly) is formed of small pebbles mixed 

 with coarse sand, four or five feet above the present level of 

 the lake and about four to five yards from the water's edge. 

 The pupae are found at the edge of the belt of vegetation, which 

 faces south-east and is shaded after mid-day. 



Carpenter's conclusions are as follows : 



" I. Favourite sites for depositing pupae are those which 

 are in shade, but where there is free air circulation. 



"2. The soil, commonly gravel or coarse sand, must be 

 dry and loose. 



"3. The fly does not extrude its larva until the middle of 

 the day, always selecting a shady spot." 



Natural enemies. The peculiar mode of reproduction of 

 the tsetse-flies renders the immature stages to a great extent 

 im.mune from the attacks of many insectivorous animals. 



Nevertheless, the pupse must be devoured by many species 

 of birds, although up to the present there are no observations 

 on this point. Bagshawe observed pupal cases with a small 



