Responses of tsetse flies (Diptera : Glossinidae) to vegetation in Zimbabwe: implications for population distribution and bait siting

In Zimbabwe, Glossina morsitans morsitans Westwood and G. pallidipes Austen were caught from traps and targets baited with ketones, phenols and 1-octe n-3-ol. Averaged over a year, catches in mopani woodland, thicket and scrub land were much the same, but the catches in the mopani relative to the thic ket changed several-fold from one season to another. Within vegetation type s, catches varied up to ten-fold between separate sites 200 m apart and sho wed site X day interactions. Continuous use of a site for three weeks showe d no catching-out of the flies in the immediate vicinity Catches were depre ssed by about 40% when traps were beside fallen trees, by up to 82% when un der leafy bushes, and by 39% when the trap entrance faced upwind. Catches doubled when traps were in clearings. Oblique shade from distant tr ees reduced catches by 30%, due to a reduced dose of odour from cooler disp ensers. Artificial canopies of leaves above traps reduced catches by depres sing the light-orientated movement inside the trap. The canopies stopped fl ies from showing their normal preference for initially alighting on the bla ck half of blue-black targets. Representations of fallen trees and leafy bu shes reduced catches by obstructing visual stimuli and denying access. Tset se flew readily through gaps in hedged enclosures provided the gap was at l east 1 m wide, on the downwind side, and odour was present. Catches from si tes in dense vegetation were enhanced up to four times by reducing the cano py and clearing bushes. Tsetse did not persist near odour sources. A computerized model of tsetse moving randomly in space restricted by dense bushes suggested that: (i) the speed of dispersal within any one vegetatio n type is not affected greatly by the abundance of bushes, but can be chang ed ten-fold where vegetation types mix; (ii) active flies tend to concentra te in areas with few dense bushes; (iii) the most reliable indices of tsets e abundance are produced by tar gets, at sites which maximize catches and w hich may be so open as ito be atypical of the vegetation type; and (iv) the response to bushes cannot alone account for tsetse distribution between ve getation types. Siting rules for catch maximization are offered, allowing catches to be inc reased up to ten times. Bait sampling alone is unsatisfactory for elucidati ng tsetse distribution. The model of movement offers a new starting point f or interpreting catches and predicting the distributions and invasion rates of G. m. morsitans and G. pallidipes but needs refinement before applicati on to other tsetse species, stomoxyines, non-biting muscoids and tabanids. Experimental aids for developing better models are suggested.