Karl Pearson 
123 
possibly describe such skew distributions as we get in the wild-game strain or wild 
tsetse fly strain of the trypanosome species identified by Sir David Bruce and 
colleagues as T. rliodesiense*. It is equally impossible in the case of the human 
strains figured in the paper of February IQl-Sf. I illustrate this on the frequency 
distribution for 6220 trypanosomes of human strains.|. 
Observed 
Calculated 
Observed 
Calculated 
i4 and under 
1 
75-45 
M 
425 
520-55 
15 
10 
62-51 
27 
372 
444-42 
16 
41 
101-51 
28 
347 
357-96 
17 
154 
155-50 
29 
307 
271-88 
IH 
325 
224-73 
30 
198 
194-81 
10 
494 
306-27 
31 
167 
131-68 
20 
528 
393-73 
32 
123 
83-91 
21 
577 
477-39 
33 
77 
50-44 
22 
512 
545-93 
3Jf 
36 
28-61 
3S 
525 
588-91 
35 
12 
15-30 
24 
511 
599-17 
36 and over 
14 
14-18 
25 
464 
575-04 
Here = 501 and P < '000,000,001. In other words description by a Gaussian 
is absolutely impossible. The histogram of observations and the curve are shewn 
on Diagram V. 
Now the suggestion that flowed at once from these results was the compound 
nature of all the material classed under the headings : 
(i) 
2\ rliodesiense. 
(ii) 
T. hrucei. 
(iii) 
T. gambiense. 
(iv) 
Mzimba Strain. 
(V) 
Wild 0. morsitans Strain. 
(vi) 
Wild-Gaine Strain. 
(vii) 
Human Strain. 
With the experience of the Gaussian fitting the homogeneous strains, the direct 
step was to investigate whether the above material could be analysed into two 
Gaussian components and to determine how nearly these components were in 
agreement. The method of carrying out this analysis was provided in the first 
of my series of Contributions to the Mathematical Theory of Evolution^. There 
was nothing to prevent the process being applied to every individual frequency 
given by the trypanosome workers, except the very laborious arithmetic. The 
method was applied to the above seven cases, and also (viii) for the purposes of 
illustration to a single human case, that of Chituluka, a native of Nyasaland, who 
* See Ii. S. Proc. Vol. 86, B, pp. 407 and 419. 
+ See R. S. Proc. Vol. 86, B, pp. 285 et seq. 
* See R. S. Proc. Vol. 86, B, p. 300. 
§ Phil. Trans. Vol. 185, A, pp. 71—110, 1894. 
16—2 
