300 
Mammalian Trypanosomes of A frica 
gation of the sick, and in their territory also, both east and west, the disease 
died down to a practically negligible factor in the death-rate of the Colony. 
The direct-transmission hypothesis accounts both for the origin and for the 
dying down of the virulent form of trypanosomiasis in these at one time 
populous areas. 
A possible factor in the reduction of that broad contact between fly and 
man which this hypothesis invokes as the essential factor in the epidemic, 
is afforded by the high lake level of 1906. The effect of the high water of 1917 
on the density of the fly was most conspicuous, and led to a definite reduction 
in their numbers on most of the well-known fly-shores visited by me early 
in 1918. At the time of my visit the water had again begun to fall, and when 
Carpenter visited these parts some months later he found but little if any 
diminution in the fly density. That a considerable reduction did occur, how¬ 
ever, there is not the slightest doubt. 
Now in Koch’s reports (Koch, 1906-07) there occur two interesting ob¬ 
servations. Speaking of the distribution of the fly round Mwanza in 1906 
he says “The harbour proper of Mwanza is free from Tsetse. This is not in 
agreement with Feldmann, who claims to have found them especially in the 
harbour. This contradiction is possibly due to the lake being 1 \ metres higher 
than in previous years.” Again, referring to certain women at Kisiba who, 
in contrast with all the other infected people at that place, had never visited 
the Uganda fly areas, he says that all attempts to find G. palpalis in the 
places visited by these women failed, the search being made more difficult 
on account of the high water. These two references are supported by the 
chart of the lake levels since these were first recorded (Duke, 1919 a). The 
levels of 1906 and 1917 were just about the same; and from analogy alone, 
we can presume that there was a definite diminution in the numbers of the 
fly at the time of Koch’s visit—a diminution varying in degree in different 
localities according to the steepness of the shore-line. 
This factor, however, can hardly have played any essential part in the 
diminution in virulence of the epidemic in Buvuma; for, as the death returns 
indicate, this diminution began in 1903, when the lake level was not high, 
while in Sesse the change was already noticeable in 1905. 
(d) Role of cyclically-in fected flies in the spread of the epidemic. 
The part played by cyclically-infected flies in the spread of the epidemic 
is not at all clear. Only a very small number of G. palpalis seem capable of 
sustaining full cyclical development of the parasite, even under the most 
favourable conditions for picking up trypanosomes. The only available in¬ 
formation on the infectivity-rate under the conditions prevailing during the 
epidemic, are the Entebbe figures given below (cf. Part II, sec. (a)). This 
point is of great importance, and, unfortunately, the number of flies used is 
small. 
