122 A Stndji of Trypimosome Strains 
We conclude that the Gaussiaii adequately describes the distributioa of 
T. simiae. In more than half the trials we should get a worse sample. See for 
graphical fit, Diagram III. 
(c) T. caprae (see p. 110). 
Mean = 25'50S microns. S.D. = 2'1011 microns. 
Microns 
Observed 
Calculated 
Values 
Values 
30 and under 
4 
4-28 
21 
8 
9-82 
22 
23 
23-95 
23 
49 
46-74 
2}t 
79 
73-05 
25 
95 
91 -38 
26 
80 
91-54 
27 
68 
73-45 
28 
57 
47-16 
29 
24 
24-26 
30 
9 
9-98 
31 and over 
4 
4-38 
x2 = 5-175 
P= -921 
This is a still more excellent fit ; if the Gaussian represented the population, 
in 92 "7o samples we should get a more divergent sample than that observed. 
The curve is given in Diagram IV. 
150r 
140- 
Microns. 
Diagram IV. Gaussian fitted to T. caprae Frequency. 
It will be clear from the above three illustrations of what we may term 
homogeneous trypanosome strains that the Gaussian curve of frequency suffices 
to describe adequately such matei'ial. It is equally clear that no Gaussian can 
