73 



" This experiment was carefully conducted, and careful notes were 

 kept upon the behaviour of the flies on the islet towards man and also 

 towards certain domestic animals which were tethered there .... 

 They were so strikingly affected by the banishment of their reptilian 

 hosts as to leave no doubt that they had been principally dependent 

 on them, and that the increase in the percentage of females from 15-5 

 to 57-2 was the direct result of food shortage." 



After giving precise statistics as to the attacks of the flies upon the 

 domestic animals (two goats and two pigs) and upon the human 

 personnel employed in the foregoing test, Fiske writes : 



" This experiment of hunting the wild hosts of tsetse was accidentally 

 repeated on the peninsula of Neozi on the island of Bugalla (Sesse) 

 where camp was pitched in November 1914. There was a not heavy 

 or noticeable infestation by tsetse, and the principal host of it was 

 situtunga. But no sooner was the camp occupied than these animals 

 evacuated the peninsula and were seen crossing the isthmus connecting 

 it with the mainland one or two miles distant only a few hours after 

 the men began work on the tent and huts. On the following day 

 (Sunday) the behaviour of the flies was not notably changed, but 

 on the third day they became so unbearably persistent in their attack 

 that (in view of the possibility of human infection from their bites) the 

 camp was abandoned." 



A high female percentage in catches of G. palpalis, which, as has 

 already been shown, indicates a dearth of favoured hosts, and con- 

 sequent hunger and extreme pertinacity in attack on the part of the 

 females, is, in fact, as is explained by Fiske, an important index of 

 danger of transmission of human trypanosomiasis, and therefore a 

 phenomenon of far more than mere academic importance. 



" Making allowances for error at every point," writes the author, 

 "it is clear that density of infestation being equal, the fly is several 

 hundred times more likely to feed upon man where wild hosts are very 

 few and female percentage very high than when they are very many 

 and female percentage low. 



" Perhaps the most pertinent point in this connection is that the 

 chances favouring transmission of the virus of sleeping sickness from 

 man to man are vastly less proportionately when few flies feed on man 

 than when many do so. The same fly must feed on or bite the human 

 host twice in order to transmit disease from an infected to a healthy 

 man. If only one fly in 500 or 1,000 actually bites man, the chances 

 that that same fly will attack man a second time are absurdly small ; 

 if every second or third fly feeds upon or bites man the chances that 

 that same fly will attack man a second time are stupendous in com- 

 parison. 



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

 chances favouring transmission of human disease." 



In order that sex percentage data obtained in different localities 

 and on different dates shall be inter-comparable, it is important that 

 the flies shall be caught in a definite manner. On this subject Fiske 

 remarks : 



" The most convenient method for measuring variations in the 

 density of Glossina palpalis is that of employing expert ' fly boys ' and 

 of counting the number of flies which can be caught per boy per hour 

 under standardised conditions. This method had already been used 

 by Dr. G. D. H. Carpenter and others, and when care was taken to 

 eliminate sources of error very reliable figures were secured." 



