444 Trypanosoma leimsi 
We are therefore doubtful about the act of biting conveying ‘ late ’ 
infection. 
As it seemed to us that late infection was not conveyed by the 
bite of the flea, we thought that it might ensue from the act of the rat 
biting and swallowing the flea. We therefore performed the following 
experiment. 
Experiment 19. A ‘clean’ rat D xii was fed with bread in which was incor¬ 
porated about 12 fleas which were thought to be infective, as 96% of their fellows 
from the same flea-box exhibited small trypanosomes in their hindguts. However 
no infection resulted. 
It is certainly possible that the mechanism of late infection is a 
contaminative one, due to the voiding of the infective agents on to the 
skin of the rat. It has been stated that during the act of feeding 
certain fleas squirt excrement from their recta, but we can say that 
this is not so in the case of G. fasciatus or G. agyrtes for we have 
observed many hundreds in the act of feeding and it has never then 
occurred. Moreover we have never observed that the coat of a rat 
infested with fleas becomes soiled with their excrement, although we 
would note here that Brumpt has found the coats of heavily infested 
mice to be badly soiled by flea excrement. Again, it is very difficult 
to find any trypanosome forms in the voided excreta of fleas which are 
even heavily infected with the small trypanosomes in the hindgut and 
rectum. If infection were due to contamination, one would expect 
rather a large number of the parasites to be excreted, especially if 
there were many in the rectum. On the other hand the fact of the 
presence of the parasites in the hindguts and recta of fleas seems in 
itself to lend colour to the idea of the infection of the rat being due to 
contamination. We would also mention here that Manteufel (1909) 
had no difficulty in infecting rats by placing blood infected with 
T. lewisi on their skins, while Hindle (1910)' has lately obtained 
infection of rats in a similar way by T. gamhiense. We cannot say 
what the mechanism of infection really is, but our experiments would 
appear to indicate that the bite of the flea inoculating the parasite 
into the rat is not the usual mode of transference. 
On the other hand the following experiment seems to indicate that 
‘ mechanical ’ transmission is effected by the bite. 
Experiment 18. Ten fleas from an infected rat were fed within 24 hours on 
a clean rat through gauze, with resultant infection. After this the fleas were 
' In presg, to appear shortly in the Brit. Med. Journ. 
