126 
A Study of Trypanosome Strains 
least suggest there is auy identity, but comparisou may bring home to the 
trypanosome worker the average sizes of the two components*. The differences 
of the variabilities are, however, much larger, and the influence of host on 
variability as well as on mean ought to be studied. 
It will be seen at once that the divergence in the individual means of T. minus 
from the general mean is very slight, at most a micron, and well within the limits 
which arise, as we have seen, from difference of host. It is a most remarkable fact 
that from six independent reductions the mean size of T. minus should come 
out so nearly 19"8 microns. In T. majus the correspondence is not so good ; the 
average of about 26 microns falls to 24 in the Mzimba strain and rises to 28'8 in 
the case of Chitulukaf. Still it does not appear to me that these changes of 
mean of the T. viajus component are absolutely beyond the variation due to differ- 
ences of host and treatment. Another more serious matter is the comparatively 
wide range found for the variabilities ; but even here it is impossible to assert that 
such differences will not occur with difference of host. For example the Mvera 
cattle strain, a fair sample of the simple T. pecorum, gives : 
Host 
Mean 
Standard 
Deviation 
Coefficients of 
Variation 
Goat . . . 
13-80 
1-462 
10-592 
Rat ... 
14-75 
-839 
5-689 
Dog ... 
13-79 
1-087 
7-885 
Here while the means are within one micron, the differences in variability are 
of the same order as those found in T. majus from different hosts. 
Again, taking a pure homogeneous strain as T. caprae with goat and sheep as 
host, which are scarcely so differentiated as man and antelope, we find : 
Host 
Mean 
Standard 
Deviation 
Coefficients of 
Variation 
Goat 
Sheep 
25-31 
25-60 
2-187 " 
1 -923 
8-642 
7-512 
Lastly, taking T. simiae for goat and monkey we have : 
Host 
Mean 
Standard 
Deviation 
1-403 
1-687 
Coefficients of 
Variation 
Monkey . . . 
Goat 
17- 26 
18- 11 
8- 127 
9- 315 
* Tlie maximum average length of T. caprae is 26'8 in the waterbuck and of T. simiae 18-1. 
t It should be noted that with tlie ichole of the human data tlie mean is 26-33 and that Chituluka's 
mean is very exceptionaL 
