Karl Pearson 
143 
forms, but I am inclined to think that the index really passes through the value 
100. Undoubtedly this range of index, or possible athwartness of the nucleus is 
not conspicuous in the simple strains like T. pecoram, T. simiae and T. caprae. 
Conclusions, (i) If appeal be made to statistical measurements, judgment 
between identity and diversity of strain must be formed by means of accepted 
statistical processes and not by mere comparison of graphs. 
(ii) Statistical processes show that the conclusions already formed as to the 
identity of trypanosome strains from mere inspection of the graphs cannot be 
confirmed. 
(iii) There must be some standardised process of treatment both in regard to 
host, and to method of and stage of infectivity at extraction. 
(iv) Even making allowance for differences due to host and treatment, wc 
find remarkable divergences in the very strains asserted to be identical. 
(v) It would appear that some order would be brought into the chaos, if we 
could consider the strains described as T. brucei, T. rhodesiense, T. gamhiense, the 
wild-game, the Mzimba, and very probably the tsetse fly and the human strains 
as really consisting of two components, which for the time I have termed 
T. minus and T.majus. It is highly desirable that additional measurements should 
be made (? a nuclear index ascertained) to determine whether these lead also 
to similar components. 
I do not assume that this is a final solution of the problem, nor do I assert that 
T. minus and T.majus represent necessarily, although probably, distinct strains; 
they may be dimorphic forms of one and the same strain occuiring in different pro- 
portions. But, I believe, that the suggestion of their existence may help to explain 
some anomalies of the present chaos. I ought also to state quite frankly that this 
paper is not written in a merely critical spirit. I believe that the trypanosome 
v^orkers have undertaken in their elaborate systems of measurements most laborious 
and most valuable work, but, I think, the time has now come when without 
trained statistical aid, but little further progress will be made in a very important 
and urgent matter. 
The very large amount of arithmetical work in this paper would never have 
got carried through had I not had the ever ready assistance of my colleague 
Miss Julia Bell ; to Mr H. E. Soper also I owe help in the arithmetical work, but 
I have to thank him in particular for the careful preparation of the diagrams, and 
the planimetric determination of their frequencies by aid of which the for all 
but two of the compound curves was found. In the case of T. brucei and 
T. rhodesiense actual calculation of the areas of the normal curves was used. 
