36 NATURAL HISTORY OF GLOSSINA PALPALIS 



of distinguishing between avian and reptilian blood was 

 crude and open to a wide margin of error, 



I repeated this work with a greater number of flies and 

 at greater length, and using as standards such sources 

 of food as the flies might meet with, viz. cormorant, croco- 

 dile, Varanus, python, frog and lizard. 



At Jinja I had found that the proportion of mammalian 

 to non-mammalian blood in over 12,000 wild flies was 

 indicated by 



M : N :: 31-5 : 68-5. 



On Damba Island the proportions in over 6,000 flies were 



M : N : : 21 : 79. 



Now in both these places antelope's blood was readily 

 obtainable, for Situtunga abounded on Damba and bush- 

 buck were constantly seen in the fly area at Jinja ; hippo- 

 potami were of course available at both localities. In 

 1914 further work was done and blood was examined 

 from flies taken on islands where there were no Situtunga, 

 and others such as Kome, where Situtunga abounded. 



In flies from five small islands not inhabited by Situ- 

 tunga the proportions of mammalian to non-mammalian 

 blood were : 



M : N : : 4 : 96. 



On two isles inhabited by the buck the figures were : 



M : N : : 25 : 75. 



Now on all islands hippopotamus blood is obtainable ; 

 indeed, on those of the first group it was the only possible 

 mammalian blood, and accounts for the 4 per cent, of 

 cases. 



But even where Situtunga are present only about one 

 quarter of the food is obtained from them, and it appears 

 that Glossina palpalis much prefers non-mammalian 

 blood. 



