232 



Drs. J. W. W. Stephens and H. B. Fantham. [Mar. 28, 



We also give a chart of six hundred T. rhoclesiense taken from the same 

 of host, namely, rats : — 



Chart 2. — Curve representing distribution, by percentages in respect to length, of 

 600 specimens of T. rhodesiense from rats. (See Addendum.) 



If we now consider the graphic representation of our measurements of 

 T. rhodesiense, as seen in Charts 1 and 2, and compare them with Bruce's 

 curves of T. gambiense and T. brucei, we note the following points : — 



While T. gambiense presents a curve with a single marked peak at 20 ll, 

 T. 'rhodesiense presents a series of small irregular peaks extending from 18 ll 

 to 28 ll, with the highest peaks at 20 ll and 26 ll. In the case of T. brucei 

 there is a slightly irregular curve extending from 18 ll to 26 ll, with a well- 

 marked peak at 24 ll. 



the three curves together, we note again that T. rhodesiense 

 different from T. gambiense, but that the difference from T.brucei 



appears to be 

 is slight. 



Discussion of Results. 



1. We consider that a sample of 20 trypanosomes, at least in the case of 

 dimorphic species like T. rhodesiense, from a particular slide on a particular 

 day is too small, as the average length obtained in this way may vary in 

 extreme cases between 24'4/x and 291 /a (see Table I, Eat B 42). 



2. The day of infection on which the measurement is taken is very 

 important, for, as we have seen in Table VII, on one day 10 per cent, of 

 stumpy forms may be found, on another day 95 per cent. This must, we 

 think, be due to an actual change in the number of trypanosomes of any 

 particular length present, and not to an error of measurement. 



3. It is probable also that the host from which the trypanosome is 

 taken is an important factor. It is difficult to be quite certain of this, 



