It has been shown that G. morsitans and G. tachinoides annually shift their breeding-ground in March from the edge of a residual forest island to the centre— a distance of some 40 yards. The move is associated with an increase in mortality amongst the last puparia to be found, death usually having occurred late in pupal life.

The object of the investigation has been to discover what prompts the female to move just before pupal mortality increases.

It has been found that there is little difference in the temperature conditions of the two breeding-grounds, but that the evaporation rate is one-third less in the centre of the forest. However, the abandonment of the forest margin is not explained by attributing to the female the ability to detect differences in the evaporative power of the air, but rather by the negative reaction to light developed by the whole community under conditions of universally high temperature. The period of migration coincides with an abrupt increase in the number of hours when the shade temperature is above 90°F. (32°C), the temperature at which Jack & Williams found that the negative reaction to light begins. During March the diurnal temperature is below this figure only immediately after dawn, when under laboratory conditions parturition rarely occurs. Since the temperature remains above 90°F. until well after activity has ceased at dusk, it is probable that at any rate the non-hungry tsetse spends most of the day and all the night in the centre of the forest. In the evenings field observations suggest that the stimulus of hunger may overcome the negative reaction to light, as then flies will undoubtedly attack in the forest edge, but since pregnant females rarely feed for several days before parturition, this counter-stimulus is unlikely to be operative at the time the female extrudes her larva.

By moving into the centre of the forest the adults will gain a reduction in temperature of only 3 or 4°F., but they will be less susceptible to heat, as the much lower evaporation rate will reduce their water loss.

The climatic conditions in the pupal environment of the forest edge at 1 ½ inches depth were found to favour survival, so it was concluded that the factor which inhibited emergence must have been operative before the puparia were formed. Possibly high surface evaporation and unsaturated soil atmosphere affect the formation of the pupal integument at the time of pupation, rendering subsequent emergence impossible.

The meteorological data collected in this investigation are more detailed and localised than anything yet published by the writer, and should give a better picture of the dry season adult and pupal environments in a locality which is very close to the hot, dry limits of the range of both species.

Puparia are found in large numbers in soils whose mean monthly temperatures range from 69°–82°F. (20·6—27·8°C). Field observations suggest that a constant temperature of 72°F. (22°C.) with a R.H. of 80 per cent., or 77°F. (25°C.) at R.H. 100 per cent., are conditions which should provide very healthy puparia if reproduced in the laboratory (see Appendix).