A detailed tsetse-fly survey of 41,000 sq. miles in the inland savannah zone of the Gold Coast (Ghana) was completed in five years using an entirely African staff and working only six months each year.

This undertaking was made possible because studies of the species of Glossina concerned, G. palpalis (R.-D.), G. tachinoides Westw. and G. morsitans sub-morsitans Newst., had provided knowledge of their habitat, food-hosts and activity rhythm, on which could be based a standard yet simple survey technique, within the compass of independent native teams.

The restricted activity rhythm of the flies, subject to both daily and seasonal variations, meant that nothing less than a ‘day’ of eight hours of observation could be accepted as a valid sampling unit.

With tsetse flies at low densities, a day's search was inadequate, giving only 40 per cent, reliability compared with 6-day observations. Four days' search gave 90 per cent, reliability and was adopted as the minimum period of search.

Each survey team was under the charge of a Kecorder or Field Assistant, with fly-boys devoting the whole of their time to catching. Comparability of survey data was ensured by keeping teams to the same size, two fly-boys being the optimum, since it was found that the size of the catch was not always proportionate to the number of boys, the relationship varying according to a variety of factors.

In surveying G. palpalis and G. tachinoides the ‘picket’ system of more or less stationary teams, measuring their catches by the ‘day ’, was found overwhelmingly superior to the ‘fly-round’ technique.

The survey procedure described for savannah woodland country was not effective in sampling G. nigrofusca Newst. in the Ashanti forest or G. longipalpis Wied. in the coastal savannah, but Morris's ‘animal’ traps were found to take both species readily in places where they had not been detected or only occasionally caught by fly-boys. Experiments showed that trap catches were superior, both numerically and in giving a more representative sample of the tsetse population present. It was also shown that trapping could be a valuable means of studying the biology of these two ‘man-shy’ species of Glossina. Traps were also effective in taking G. pallicera Big. and G. medicorum Aust.

The greater value of traps compared with fly-boys as a means of sampling various species of tsetse fly is discussed, and suggestions are put forward for developing trapping as a method of survey.

The type of trap employed is regularly catching insects of several other families, including a number of species among the Tabanidae, and of several other orders.