106 
A Stff(h/ of Tri/imnosome Strains 
Clearly then two samples as divergent as those found would occur on the 
average once in three trials. It follows that two individual rats are really inter- 
changeable and we note that the extent to which ox is interchangeable with dog 
or goat for the cattle strain is very much the degree in which two rats are inter- 
changeable. To judge from this single instance, individuality within the same 
species of host is not very important, and when we find two hartbeeste differing 
as those considered on p. 93, it seems much more likely, with the information we 
have at present got, that the hartebeeste were infected with different strains of 
trypano.some than that their individuality produced the enormous divergence 
noted. Again the sensible divergence between Mzimba strain in dog and rat on 
p. 104 is probably due to difference of host, but the enormous difference in the 
wild-game strain between a single rat and dog and goat on p. 103 is probably due 
to differences in the strains of trypanosomes in the various types of wild game 
dealt with. We may consider whether the dog and goat data for the wild-game 
strain differ sensibly. We have* 
Microns. 
11 
12 
IS 
u 
15 
16 
17 
18 
Totals 
Wild-Game Strain, Goat ... 
1 
16 
37 
73 
38 
26 
8 
1 
200 
» „ >, Dog ... 
12 
31 
57 
50 
24 
6 
180 
Here t^'^ = 6'04 and P = ■•5378. Thus in more than half the trials we should 
obtain from homogeneous material pairs of samples more divergent than those for 
dog and goat. This confirms the view formerly expressed that as far as trypano- 
somes are concerned dog and goat are interchangeable. W^e cannot yet say that 
they are not interchangeable with the rat, as the mixture of strains in dog and 
goat and the uniqueness of strain in the rat may account for the marked 
divergence of the latter. Sir David Bruce and his colleagues do not appear to 
have noticed the wide divergence of the distribution of the rat from the dog and 
goat either as indicating the heterogeneity of the wild-game and the cattle strains 
of trypanosomes, or as suggesting such wide differentiation of strain by the host, that 
rat-material cannot be mixed with that from dog and goat. They do, however, 
remark of the wild-game strain : " In this the rat is not a suitable animal, since 
many strains of T. pecorum have no effect on itf." This suggests that T. pecorum 
is not homogeneous and that the rat exercises a selective influence on its strains. 
The suggested rejection of the rat data seems, however, to be based upon the in- 
convenience of its non-infectivity, and not on what might turn out to be of great 
importance a selective influence on wild-game or cattle strains. It is not possible 
to test this selective power in the present instance, as we do not actually know 
how heterogeneous either the cattle or wild-game material used really was. 
* R. S. Proc. Vol. 87, B, p. 7. 
t R. S. Proc. Vol. 87, B, p. 7. 
