Figure 5

Per cent differences in episodes averted by a mass LLIN distribution, depending on host-searching behaviour. The per cent difference was calculated as 100×(A-B)/B, where A and B are host-searching behaviour assumptions. a) Fully probabilistic host-searching behaviour, with A is behaviour with a π_i value of 0.6 and B is behaviour with a π_i value of 0.9. b) Fully determined host-searching behaviour, with A is behaviour with a π_i value of 0.6 and B is behaviour with a π_i value of 0.9. c) Host-searching behaviour with a π_i value of 0.6, with A fully determined behaviour and B fully probabilistic behaviour. d) Host searching behaviour with a π_i value of 0.9, with A fully determined behaviour and B fully probabilistic behaviour. Horizontal dotted lines are at zero difference in episodes averted. Legend further as in Figure 3.