385 
H. L. Duke 
As, under ordinary circumstances, the glands of flies in my experiments 
were not dissected out unless flagellates were seen in the gut, it is possible 
that a fly with positive glands and negative gut, if such occurred, would 
escape notice. This is, however, not probable. 
The possibility of natural contaminating infections has been most care¬ 
fully investigated. In addition to control monkeys kept for months undei 
daily examination, a particularly severe test was devised to settle this point. 
Two baby monkeys, brought in with their mothers and still unweaned, were 
allowed the freedom of the monkey village. They roamed about all day long 
among their friends, and were examined daily. The mother of one was in¬ 
fected with the relatively virulent buck strain. Both these babies spend much 
of their time in the boxes of infected monkeys near by, and sit hugged up close 
for long periods at a time. They have been under observation for months, and 
neither they nor the other controls have ever become infected. I have nevei 
seen an instance of a natural infection among the monkeys at this laboratory. 
I think, therefore, that it is justifiable to assume that the infective flies in 
those positive experiments, such as No. 87, in which no positive flies weie 
found on dissection, were overlooked during the removal of the daily deads 
until they had dried up; and this may well happen, unless special care is 
taken, owing to the tendency of dead flies to become caught up in the comeis 
of the boxes. 
(d) Effect of different kinds of blood on the developing flagellates 
(as revealed by Tables V, VI and VII). 
Lizard. Kleine at one time held the view that monkey’s, and more 
especially man’s blood, was more favourable than that of ruminants to the 
development of trypanosomes in the fly (Kleine 1911 b). Subsequently he 
gave up this opinion and expressed himself as convinced that the type of 
blood taken up by the fly was a wholly unimportant factor (Kleine, 1913). 
At Mpumu a single series of experiments (Duke, 1913 b) pointed to yet a 
different conclusion, namely that ruminant blood was more favouiable than 
monkey blood; the percentage of infected flies in the experiments were: on 
ruminant blood 9*1 per cent.; on monkey blood 2*1 per cent. In the expeii- 
ments recorded in the present paper an attempt was made to puisue this 
matter further and, keeping in view practical issues, to discovei wkethei the 
various strains of trypanosomes employed were capable of developing in flies 
fed for a prolonged period on reptile blood. 
Observers who have attempted to feed laboratory-bred flies on crocodiles 
have been struck with the difficulty experienced in persuading the flies to 
feed. As no crocodiles were available, water-lizards ( Varanus sp.) weie em¬ 
ployed, since they are easily obtained and live for months in captivity on a 
diet of shell-fish and crabs. 1 
At first the greatest difficulty was experienced in making the flies feed. 
One would feed, and then perhaps one or two more, but often three-quaiters 
