Fig. 6

Number of Anopheles arabiensis females recaptured per night indoors (a) and outdoors (b) at the hut of the non-user (i.e. the sentinel hut). Both median (black line crossing the box plot) and estimated means (red square inside the boxplot) are shown. Each black dot represents actual number of mosquitoes recaptured in different experimental nights. All the huts had intact untreated bed nets for basic protection. Observations were made at the sentinel non-user hut before fitting the neighboring huts with traps or transfluthrin-treated eave ribbons (i.e. controls), and then after the neighbors were fitted with either eave ribbons, traps or both traps and eave ribbons fitted to same hut. There was also observations based on push–pull mosaic system, where half of the neighbors huts had eave ribbons, and the other half had traps, i.e. Neighbors have EITHER traps OR eave ribbons (*)