132 



upon man where wild hosts are very few and the female percentage 

 very high than when they are very many and the female percentage low; 

 the female percentage may thus be a very valuable index to the 

 chances favouring the transmission of human trypanosomiasis. 

 Shelter is an important factor in determining the sex ratio, high 

 percentage of females being associated with a type of vegetation 

 Imown to be particularly repugnant as shelter and low percentage 

 with vegetation known to be attractive. 



While Glossina pcdjjcdis is a riparian species, it is found that water 

 or humid conditions are not of direct benefit to the fly, nor required 

 by it. The proper combination of food, shelter and breeding-places 

 requisite to its existence, however, occur so infrequently away from 

 the shores of lakes or banks of larger streams that it is perforce riparian 

 in habit, its dense occurrence usually being limited to about 50 yards 

 inland from the shore ; beyond 100 to 200 yards flies appear only as 

 stragglers ; at 300 to 500 yards they will practically disappear, unless 

 some special cause for their occurrence exists, such as the presence 

 of unusually large numbers of game animals. The two favourite hosts 

 of G. palpalis, namely, the crocodile and Varanus (monitor lizard) 

 are amphibious in habit. They are not only the most attractive to 

 the fly, but the most favoured breeding-grounds of the fly are frequently 

 identical with spots selected by the crocodile for breeding places or by 

 Varanus as a basking ground. G. palpalis therefore becomes almost 

 a specific parasite of these reptiles. The habits and habitats of the 

 various hosts of the fly in the region of Victoria Nyanza are discussed, 

 with notes and observations indicating the practicability of controlling 

 the fly through extermination of its hosts. It is considered that 

 complete extermination of the four principal hosts (crocodile, Varanus, 

 situtunga and hippopotamus) would cause a reduction in density of 

 infestation amounting to 95 to 99 per cent. Food, however, even 

 when abundant and of a preferred variety, is valueless to the fly 

 imless adequate protection in the form of both shelter and breeding 

 grounds also occurs within reach of the food. If protection is adequate, 

 the range and density of G. palpalis are controlled by the quantity 

 of preferred food occurring within reach, and the distance separating 

 food from protection is a factor of the first importance in determining 

 the prevailing degree of infestation. 



Breeding-places are frequently chosen in sand or gravel, but fine, 

 dry vegetable debris serves as weU. Typical breeding-grounds in 

 places distant from the shore are described and the correlation between 

 these and density of the fly is discussed. An important conclusion 

 reached is that massive shelter is requisite to the life of G. palpalis, 

 and that, while lightly sheltered areas may provide the best and most 

 attractive hunting, breeding and assembling grounds for the species, 

 unless they he within easy reach of the fly from massive shelter they 

 will not be infested. 



The necessary constituents of a proper combination of the requisites 

 of life for the fly are discussed. While no specific observations on the 

 subject were made, the tentative conclusion is reached that unless 

 all three requisites, namely, good shelter, breeding-ground and the 

 blood of vertebrates, occur within a radius of less than 100 yards from 

 a central point, the conditions of life are so unfavourable that the 

 species ca'nnot exist. Within such favourable zones, natural increase 



