1. The following strains produced no infection in any of the laboratorybred Glossina used in their examination; the total number of flies dissected is given for each strain:

T. gambiense, strain V, Uganda, isolated in 1926 … … … 548

T. gambiense, strain Adero, Uganda, isolated in Jan. 1933 … 1642

T. rhodesiense, Liverpool, isolated in Jan. 1933 …. … … 750

T. rhodesiense, Liverpool, arsenic-fast variant … … … … 1166

T. brucei, “Hamburg alt” from Berlin; isolated over 30 years ago 2452

Some of the flies used in testing each of these strains were kept at 95–97° F. during their infecting feeds.

2. T. gambiense, strain “Braun,” isolated in February, 1920, gave two “gut only” infections in 1137 flies. T. gambiense, strain “McA,” isolated in 1921, produced one very light infection, of the intestinal tract only, in 1410 flies employed. This infected fly died on the 27th day after its infecting feed.

T. brucei, strain “Hornby mild,” isolated at the end of 1930, gave three infections of the intestinal tract only in 1443 flies used. Some of the flies employed on these three strains were kept during their infecting feeds at 95–97° F., but the infected flies came from boxes kept at room temperature throughout.

It will be seen that all the strains hitherto summarised are, as far as these tests are concerned, non-transmissible by Glossina, and the majority are no longer capable of infecting even the intestine of the fly.

3. T. brucei, strain “Hornby virulent,” isolated in April, 1927, from a heifer, and found by Corson in 1931 to be readily transmissible by G. morsitans or G. pallidipes (or both), 4½ years after its first isolation (Corson, 1932). A month or so later this strain was found at Entebbe to be still feebly transmissible by G. palpalis and somewhat more readily by G. morsitans, although much less so than in Corson's experiments.

4. Strain “Br,” considered by Prof. J. G. Thomson to be a T. gambiense showing some resemblances to T. rhodesiense, when examined at Entebbe some three years after its isolation from a European in West Africa, proved to be still infective to both G. palpalis and G. morsitans through only very feebly transmissible. 4272 laboratory-bred flies were used in the examination of this strain: 74 of these developed infection of the intestine, and only one a gland infection—a G. palpalis dying on the 40th day after its infecting feed. Flagellates were numerous in the glands of this fly.

5. The behaviour of the “Hornby virulent” strain of T. brucei and of strain “Br” suggests that completion of the cycle in the fly may be delayed beyond the 25–30 days usually sufficient for East African strains, and it is possible that this delay may be a feature characteristic of strains whose transmissibility by tsetse is undergoing reduction. On the other hand, the solitary infective fly obtained with strain “Br” had a heavy gland infection which had in all probability been present at least for several days before the death of this insect.

6. A strain freshly isolated from a native who was infected on or near the northern shores of Lake Victoria failed to infect any of 1642 G. palpalis used in its examination, although a number of these flies were kept at 95–97° F. during their infecting feeds.

7. The results of the investigations described in this paper lend some support to the opinion already formed as the result of numerous experiments with the polymorphic group of trypanosomes, namely that T. brucei (and, as far as can be seen, T. rhodesiense) is less prone than T. gambiense to lose touch with Glossina.

It may be that the stability of this character in T. brucei is an expression of a more perfect adjustment to environment than is possessed by T. gambiense; the latter trypanosome, which is essentially dependent on man, having not yet attained biological equilibrium in this its principal mammalian host.