118 A Study of Tri/2)anosome Strains 
Microns. 
From Dog 
11 
lt2 
13 
15 
16 
17 
Totals 
Wild-Game Strain 
12 
31 
57 
50 
24 
6 
180 
Mvera Cattle Strain 
3 
11 
27 
30 
21 
8 
100 
This leads to = 15-992, P = -0138. 
Previously (p. 98) on the total series of different hosts we had found 
P = •000,243. Thus by referring our material to individual hosts, we have reduced 
the degree of divergency between the wild-game and Mvera cattle strains, but it 
would be still hazardous to state that these strains are identical. 
Lastly, we turn to the Mvera cattle strain and tlie wild 0. morsitans strain 
dealing with dog and goat as hosts separately* : 
Microns. 
From Goat 
9 
10 
11 
13 
13 
u 
15 
16 
17 
Totals 
Wild G. morsitans Strain ... 
1 
3 
12 
■21 
55 
60 
32 
12 
4 
200 
Mvera Cattle Strain 
1 
1 
3 
14 
22 
26 
19 
13 
1 
100 
This gives x' = 7-968, P = -4368. 
And again : 
Microns. 
From Dog 
9 
10 
11 
i 
12 1 13 
I 
U 
15 
16 
17 
Totals 
Wild G. morsitans Strain 
3 
14 34 
41 
40 
19 
9 
160 
Mvera Cattle Strain 
3 
11 27 
30 
21 
8 
100 
resulting in = 11-120, P = -0852. 
The Mvera cattle strain and the Glossina morsitans strain had for all hosts a 
divergence measured by P = -000,008. Thus the great bulk of this divergence 
is due to multiplicity of hosts -f-. 
To sum up the results obtained for T. ■pecorum in Mvera cattle, wild G. morsitans 
and wild-game strains, the identification of these strains was quite illegitimate on 
the basis of the compound host frequencies. It is reasonable on the basis of 
* R. S. Proc. Vol. 87, B, pp. 3, 10—11. 
t It is worthy of note that in comparisons with the cattle strain the goat appears to give closer 
results than the dog, hut the dog appears the better in the comparison of the G. morsitans and wild- 
game strains. 
