92 
A Study of Tri/panosome Strains 
(iv) T. caprae. This is found in waterbuck, ox, goat and sheep. The dis- 
tribution of length is apparently homogeneous and the mode at 25 microns*. 
I leave out of account several forms of trypanosome referred to by Sir David 
Bruce and colleagues, e.g. T. vivax, T. uniforme, T. ingens, etc., of which no large 
series of measurements were at my disposal. 
With the exception of T. simiae, which occurs in the warthog, the above 
trypauosomes appear to be found generally in the wild game and all of them are 
found in the Olossina viorsitans. Sir David Bruce and his colleagues suppose the 
differentiation into these classes to precede the consideration of individual strain, 
but the exact modus differ entiationis is not clear from the memoirs. 
(3) Method of Investigation. The actual formula employed in the present 
investigation is very simple and can be applied by anyone able to do ordinary 
arithmetic. If N and N' be the sizes of two samples and the corresponding 
frequencies : 
/2> fit fj, ••• fpy ••• fs! 
fl > f-z , fi , •■■fpy • • • ./s > 
where fp, f J are the frequencies falling in the p^"^ category, then if 
be calculated, the probability P that the observed or a greater divergence between 
the two series would arise from sampling the same population is obtained by 
determining P from ;)^- by my method of testing " goodness of fit." This method 
was first published in the Phil. Mag., Vol. .50, p. 157, 1900. The shortest method 
of actually determining P is by aid of Palin Elderton's tables for P with argument 
issued in Bionietrika, Vol. l. p. 155, 1902. This is the process used in the 
measurements of sameness and divergence provided below. 
(4) On the Probability of the identity of the Strains discussed by Sir David 
Bruce and others. 
{a) I take first the question of the "sameness" of the Wild-game strains of 
trypauosomes as isolated from five antelopes — reedbuck, waterbuck, oribi, and two 
hartebeeste. Sir David Bruce and others discuss these strains in a paperf of 
February, 1912, and conclude, apparently from the statistical data, that "the five 
Wild-game strains resemble each other closely and all belong to the same species." 
Now these Wild-game strains have a distinct advantage for they are all 
obtained from the trypanosomes ultimately taken from the rat as host ; they were 
passed from the infected antelope through healthy goat, monkey or dog, which 
* R. S. Proc. Vol. 86, B, p. 278. 
t R. S. Proc. Vol. 8G, B, p. 407, 1913. In the Table p. 405 for 2500 trypanosomes under the 
heading 31 microns read a frequency of 33 not 53. 
