1. Evidence is produced to show that T. rhodesiense may retain its cyclical transmissibility by Glossina for at least 600 days in an antelope.

2. Prolonged residence in these animals tends to impair the power of a strain to infect man, when infection is attempted by the cyclical route. Thus of six volunteers exposed to cyclically infected tsetse carrying T. rhodesiense that had been for many months continuously in antelope, only one became infected. All these flies derived their infections from the antelope itself. As the cyclical method is the one operating in nature, these observations suggest that although certain species of antelope are admirable hosts for T. rhodesiense, yet as a reservoir from which tsetse can become infected with trypanosomes pathogenic to man these animals do not constitute so great a menace as has hitherto been supposed.

3. In contrast to (2), in every instance where the trypanosomes from the antelope, before being tested on man, were inoculated by the syringe into a monkey, every volunteer exposed became infected.

In the single experiment where the monkey was infected by tsetse instead of by direct inoculation from the antelope, one volunteer became infected and another did not. There is thus a suggestion that the passage through the monkey prepared the trypanosome for survival in man. On the other hand, the behaviour of strain Tinde I shows that T. rhodesiense may lose its power to infect man in spite of repeated passage through monkeys.

4. Some further indirect evidence is produced to show that T. rhodesiense may owe its origin to T. brucei, and that pathogenicity to man may be a property possessed in different degrees by different strains of T. brucei in nature.

5. These researches suggest, also, that human beings differ in their susceptibility to T. rhodesiense, but that these differences only operate within a narrow range of variation in the power of strains to infect man. In other words, strains strongly pathogenic to man will infect anyone, but strains whose power has been weakened will only be able to use individuals of subnormal resistance.

6. Two different strains of T. rhodesiense, have been shown to behave differently in the same man. One did and the other did not infect him.

7. There may be a difference in the suitability of the various species of antelope to act as hosts to T. rhodesiense. Thus the bushbuck seems to be a better host to the trypanosome than the oribi.

8. A young hyaena infected with T. rhodesiense for 180 days remained in excellent health. Flies infected from a monkey sub-inoculated with the blood of this hyaena infected a volunteer: the trypanosome had then been 80 days in the hyaena.

9. Clean flies that had taken their first two meals off monkeys infected with a strain of T. rhodesiense non-pathogenic to man, were nourished entirely on human blood during the first 3 weeks of the cycle of development of the trypanosomes in their interior. These flies, on the completion of this cycle, were still unable to infect man.

10. A strain of T. rhodesiense, after a series of cyclical passages through a reedbuck and six monkeys, was found to be non-pathogenic to man. The possibility has to be borne in mind that this strain owed its original association with man to meeting an abnormally susceptible individual. This strain since its arrival at the Institute has been tested on nine different volunteers, and none of them became infected. There is therefore nothing to distinguish it from T. brucei, save the fact of its isolation from man.

11. It is recorded that a single cyclically infected fly infected a volunteer at a single feed.

12. It is shown that the appearance of a tender indurated swelling at the site of the bite of an infective fly is a not uncommon symptom of an infection with T. rhodesiense. Local tenderness and swelling were also noted where the glands of an infected fly had been inoculated subcutaneously, even in cases where no infection ensued. Where inoculations of blood were employed on man the local disturbance disappeared rapidly when no infection resulted, but when infection took place the local symptoms steadily increased during the last few days of the incubation period, and before trypanosomes were recognised in the peripheral blood.

13. The only infection of an antelope with T. gambiense carried out during this research was that of oribi III with strain LI. This strain at its first isolation was not very transmissible by G. palpalis, and after a few months in the antelope showed signs of failure to adjust itself of this particular host. Another strain, mentioned in a previous paper (Duke, 1935), failed to infect an adult female situtunga, although two sheep exposed to the same strain at the same time duly became infected.