64 NATURAL HISTORY OF GLOSSINA PALPALIS 



Ipomaea and others which grow abundantly in these 

 localities are trained so as to make a thick mass of young 

 greenery on the roof, and the soil is lightly strewn with 

 bark, dead sticks or leaves. At regular intervals of a 

 week or ten days these shelters can be visited and all 

 pupae easily collected by boys searching in the sand. 



Should the method prove feasible as a means of destruc- 

 tion on a large scale, it might be possible to put sieves 

 or f)erforated trays in the sand under the shelter of a 

 kind that would allow the pupae to be quickly sifted 

 out. Or another method might be tried of destroying 

 the pupae by the sun's heat. The roof could be made in 

 two halves, and at regular periods one half could be lifted 

 off so that the pupae beneath would be killed by prolonged 

 exposure to the sun's heat, while next time that half 

 would remain covered and the other half be taken off. 

 The rainy season, however, would introduce complica- 

 tions into this method, for often there are days with very 

 little sun, and if the sand was allowed to get wet, the 

 flies would no longer deposit pupae therein. The method 

 is a simple and inexpensive one, and should at least 

 greatly reduce the numbers of the fly. There is no doubt 

 whatever that these shelters are very attractive, and 

 approved of by the fly as suitable for their young. Be- 

 fore I came away from the islands in 1919 for leave, a 

 number of artificial breeding grounds were prepared 

 on Bulago and Kimmi and are now being tested. The 

 pupae are collected weekly from each, and also from an 

 ideal natural breeding ground on each island for com- 

 parison, that on Kimmi being the upturned tree roots as 

 described. From figures that have reached me so ' far, 

 it is clear that my artificial shelters are just as attractive 

 as, or sometimes more than, the natural shelters, for on 

 some occasions not so many pupae are found under the 

 tree as under one of the shelters close by. During July 

 over two thousand pupae were destroyed on Kimmi by 

 this method, Kimmi being barely a mile in diameter. 



