Figure 2. Model outputs illustrating potential epidemiological significance of altered vector competence arising from biting time.

a, Effect of altered vector competence on malaria prevalence with mosquitoes biting predominantly in the evening (run 2 and 5 in Supplementary Table 6), at midnight (run 1 and 4 in Supplementary Table 6) or in the morning (run 3 and 6 in Supplementary Table 6) in the absence of bed nets (LLINs). In these and subsequent figures the solid lines represent the means and the matching coloured bands, the 95% confidence intervals. The black line shows the control scenarios where competence is the same for all mosquitoes (run 7-12 in Supplementary Table 6). In these cases of constant vector competence, prevalence is identical regardless of biting time. If we allow competence to vary, there is little effect on prevalence if mosquitoes bite predominantly at midnight. However, variation in competence leads to increased malaria prevalence when feeding patterns are skewed towards evening biting, and reduced prevalence when skewed towards morning biting. b, Impact of LLINs on malaria prevalence when mosquitoes bite predominantly in the evening, at midnight, or in the morning either with altered (run 1 – 3 in Supplementary Table 6) or constant (run 7 – 9 in Supplementary Table 6) vector competence, assuming all mosquitoes have equal probability of contacting an LLIN. Under these assumptions, LLINs lead to reduced overall infection prevalence, but the efficacy of LLINs is less if biting is skewed towards the evening relative to midnight or morning biting, as evening biters have the greatest vector competence and hence, higher overall transmission potential. c, Impact of LLINs on malaria prevalence when mosquitoes bite predominantly in the evening, at midnight, or in the morning either with altered (run 4 – 6 in Supplementary Table 6) or constant (run 10 – 12 in Supplementary Table 6) vector competence, but assuming that mosquitoes feeding in the evening or morning have reduced contact with LLINs. Under these assumptions the relative efficacy of LLINs is reduced, but most markedly when feeding is dominated by evening biting mosquitoes with highest vector competence.