Karl Pearson 
109 
Percentage of Posterior -Nuclear Forms among 
Sliort and Stuinpy Forms. 
From first Extraction 
7 and under 
S 7,, and over 
Totals 
6 days and under 
12 
18 
30 
7 days and over 
18 
12 
30 
Totals 
30 
30 
60 
leading to r = — 'SOD. 
In other words using tsetse fly strains and not wild-game strains, but the same 
host, we find that now the correlation is negative or the longer the infection the 
smaller the percentage. Actually the five G. viorsitans strains show remarkably 
irregular results compared with the results for the wild-game strains; the ex- 
tractions were spread over much the same period, 13 to 14 days on the average, 
but were somewhat more numerous for the G. morsitans. Thus even the same 
method of extraction may give widely vai'ying results according to the nature 
of the strain producing the infection, although the host be the same. 
To the statistician who examines the frequency distributions provided by 
Sir David Bruce and his colleagues for both wild-game strains and Glossina 
morsitans strains, there can hardly remain a doubt about the heterogeneity of 
the material in each case. We have already demonstrated this statistically for 
the wild-game strains. These strains not only differ by immense differences 
inter se, but intra se they are clearly heterogeneous. Whether this heterogeneity 
is due to the mixture of separate strains, to dimorphism within the strain, or to 
the combination of material drawn from the rat at various stages of infection, it is 
not possible on the material at present available to determine finally. The same 
remarks apply with even greater certitude to the wild G. morsitans strains than to 
the wild-game strains. But we shall return to this point in the last section of this 
paper. We have already noted that Sir David Bruce and his colleagues identify — 
against the weight of the statistical evidence — the Mvera cattle strain, the wild-game 
strain and the wild G. morsitans strain as belonging to the same species T. pecorum*. 
They had previously identified other strains in wild game, G. morsitans and human 
beingsf with T. rhodesiense which they elsewhere describe as vel hracei\. This is 
again, I hold, against the weight of statistical evidence. But it is not clear from 
the memoirs themselves what is the exact process by which an individual fly, an 
individual human being, or the blood from a specimen of wild game is credited 
with carrying a liomogeneous strain. The sizes are so different in the cases of 
T. pecortim and T. simiae that there may be no difficulty in distinction, but the 
range is so great and to the statistician the material seems so heterogeneous in the 
case of T. hrucei vel rhodesiense that, perhaps, a fuller description by the authors 
* R. S. Proc. Vol. 87, B, p. 26. 
t R. S. Proc. Vol. 86, B, p. 42. 
X R. S. Proc. Vol. 86, B, p. 426. 
