Glossina pallidipes was sampled using biconical traps with stationary and revolving attachments in the form of blue and white flat screens. A modified biconical trap with attachment of four black strips captured the highest number of both female and male G. pallidipes Austen. However, all the stationary traps under-sampled teneral and young tsetse in post-teneral category. Biconical trap attached to a revolving component making 20 rev/min gave the best representative sample with regard to age category of females. Screens making faster revolutions resulted in larger proportions of males and hardly any old females. Flies that visited the trap with revolving screens were non-hungry, and probably mate-seeking. Those in stationary traps were mostly hungry and probably seeking blood meal.

Biconical traps were studied for their efficiency in sampling for Glossina morsitans morsitans Westw., G. pallidipes Aust., G. austeni Newst. and G. brevipalpis Newst. in three localities in Mozambique where these species occur sympatrically or singly. Studies were also carried out to determine whether acetone vapour was effective in enhancing the efficiency of the biconical trap. Various comparisons were also made between biconical traps and Langridge traps, and with fly-round patrols, with and without bait oxen.

It was demonstrated initially that the blue biconical trap (BBT) was more efficient for G. morsitans and G. pallidipes than red or white models or Langridge traps. In all subsequent studies, therefore, the BBT was used. It was concluded that: the BBT was ineffective for sampling G. morsitans, with or without acetone; the BBT was markedly more effective than fly rounds for sampling G. pallidipes, and its efficiency was significantly enhanced by acetone; and G. austeni was attracted to the BBT, but not all flies would enter the non-return cage—G. brevipalpis was readily caught in the BBT. Acetone with the BBT gave inconclusive results for these species.

Although biconical traps are now widely used for the sampling of tsetse populations, they provide only relative population estimates determined by the fly density, activity and trap response. Studies were therefore undertaken to convert such biconical trap indices into unbiased absolute population density estimates. The first step was to determine the range of attraction of such traps for the main trypanosome vector in Kenya, Glossina pallidipes, although some data were also obtained for G. brevipalpis. Two different techniques were used. Firstly traps were placed at varying distances apart, resulting in competition and consequent reduction of catch sizes per trap when ranges of attraction overlapped. Secondly, the catches in grassland at varying distances from a forest edge were compared with two predicted functions relating catch size to distance. Results from the competition experiments showed a marked decline in catch per trap when the intertrap distance was less than 30–40 m. The resultant estimate of maximum range of attraction of 15–20 m for G. pallidipes was compatible with the observed pattern of decline of catches at increasing distances from the forest edge. The range for G. brevipalpis was tentatively estimated as up to 10–15 m.