NOTES ON THE BIONOMICS OF GLOSSINA MORSITANS. 131 



a most necessary distinction between apparent and real abundance of fly, and shows 

 how the percentage of females caught in the ordinary way varies in inverse ratio, 

 and the number of pupae secured in direct ratio, with the abundance of game. Here 

 we have direct observations indicating that scarcity of game induces hunger and 

 hunger inhibits breeding. Both these observations have received support from 

 other investigators, including the comparatively scanty observations in this territory, 

 and appear to be in a fair way to become estabhshed. If this proves to be the case 

 the effect of hunting in the early days in the Transvaal and the apparent efiect of 

 the rinderpest would appear to be quite comprehensible, in spite of the fact that the 

 game was not completely ehminated in either case. 



The result of a heavy reduction of game would seem to follow logically, quite apart 

 from complete ehmination, namely irregular meals at long intervals — perhaps some 

 individual starvation — ^in any case a greatly reduced birth rate. If the birth rate 

 were checked at all suddenly, especially during the season of concentration when 

 the larvae are deposited within a Umited area, the fly's parasites might be expected 

 to destroy a much higher percentage than before. The writer is very much in agree- 

 ment with Duke's contention as to the probable effect of a sudden or " cataclysmic " 

 change in the distribution and numbers of game animals. The local distribution of 

 G. morsitans in relation to the habits of game animals, even during the time of year 

 when the fly is dispersed, is frequently very marked. In this way it appears that the 

 effect of shooting and of the rinderpest might be explained, but it is only com- 

 prehensible on the supposition, which the writer beheves to be justified, that the fly's 

 abihty to perceive its hosts is strictly Hmited, that it is incapable of following up a 

 trail for any great distance by scent and of attaching itself to and keeping in touch 

 day after day with such game as may be present in the infested area. In the latter 

 part of the dry season its hmitations in locating game, except such as may visit its 

 haunts, seem to be obvious, and if, as is probable, this is one of the most important 

 breeding periods, the effect of both shooting and rinderpest seems clear enough. 



It may seem improbable that a species of insect should be so iU-provided for taking 

 care of itseK under varying conditions as the views outhned in this paper suggest, 

 but if there is one thing more striking than another about the tsetse it is just this 

 disabihty to adapt itself to a change in its environment, otherwise it would not have 

 retreated before the advance of civihsation, in the absence of any direct effort to 

 drive it back. 



Summary. 



In the foregoing pages the following tentative views have been put forward in the 

 hope that they may receive either confirmation or criticism from other investigators : 



(1). The latter part of the dry season is probably, in general, the most favourable 

 portion of the year to the fly. 



(2). The later the onset of the rains the greater the capacity for increase. 



(3). Seasons of unusually heavy rainfall are inimical, at least in certain areas. 



(4). A series of years of low rainfall is favourable. 



(5). Regions where the fly commonly attains the greatest concentration of numbers 

 are the regions where it is most hable to be reduced or to die out, namely, parts of the 

 infested areas where first-grade foci predominate. 



