Karl Pearson 
111 
At any rate the exact method of differentiation adopted would be of interest 
to the statistician. The result of the paper is that the four species of trypanosomes 
occur in quite comparable permilles of tsetse flies caught in the sleeping sickness 
ai-ea of Nyasaland, and there is no evidence to show that they or other strains also 
may not occur side by side in the same fly or in the same specimen of wild game. 
Further, these compound strains would then appear in different proportions in the 
host. Some such hypothesis seems very needful to account for the extreme 
heterogeneity of the wild game, wild 0. morsitans, and human strains as recorded 
by Sir David Bruce and his colleagues. The following table gives a comparison of 
what appear to be homogeneous strains — T. pecorum, T. simiae and T. caprae — 
with what appear statistically to be heterogeneous strains, i.e. T. briicei, 
T. rhodesier^se, T. gavihiense, the Mzimba strain, the wild-game and wild 0. 
morsitans strains of human type, and the human strains themselves. The table 
Means, Standard Deviations and Goefficieyits of Variation of eleven Strains 
of Trypanosomes. 
Series 
T. pecorum 
T. simiae 
T. caprae 
(i) T. rhodesiense . 
(ii) T. brucei 
(iii) T. gambiense . 
(iv) Mzimba Strain , 
(v) O. moi-sitans 
(vi) Wild Game . 
(vii) Human Strain , 
(viii) Chituluka 
Mean 
]3-992±-019 
17-870 + -0.50 
25 -508+ -063 
23 
23 
22 
21 
22 
22 
23 
26 
•577 ±-100 
■529 ± -094 
■113 ±-081 
■413 ± ■063 
•695 ± ^058 
•622 ± ^047 
•796 ±-035 
•172 + ^084 
Standard 
Deviation 
1^2816f014 
1 -6558 ± ■OSS 
£•1011 + -045 
6764 
3938 
7867 
9586 
3002 
4541 
1262 
8414 
071 
066 
057 
■045 
041 
033 
025 
060 
Coefficient 
of Variation 
9^16f099 
9^27 ±199 
8-58+ ^184 
83 ± 
67 ± 
12± 
82 ± 
95 ± 
27 ± 
34 ± 
50 + 
■311 
■291 
■266 
•212 
•187 
•174 
•108 
•235 
above, gives the means, standard deviations and coefficients of variation of these 
strains. It will be seen that the first three are of a very different character to the 
last five. The variation of the latter is about double that of the admittedly pure 
strains, and throughout the whole course of our further work this possibility of 
heterogeneity, and the differential selection of the components by the host must 
be borne carefully in mind. Great divergences do not discourage the use of 
biometiic methods, and we get occasionally identities of strains which are quite 
beyond the limits of chance coincidence and which point to definite possibilities if 
only host, environment, and treatment are once effectively standardised. I propose 
to try to throw some light on these points in the remaining sections of this paper. 
(6) On the Probability that Strains are alike after alloiuance for the Host. 
(a) Luckily in certain cases the treatment has been more or less alike. Thus 
in the wild Glossina morsitans strain, the tsetse flies brought to the Laboratory 
