Comparative attraction of <i>Anopheles quadriannulatus</i> and <i>Anopheles arabiensis</i> to humans estimated by comparing their relative abundance in samples of mosquito larvae and adults collected across an ecologically heterogeneous landscape in

Abstract As most malaria parasites of humans are strict anthroponoses, mosquito preference for human blood strongly influences transmission intensity and intervention strategy. Here, retrospective analyses of observational entomological data assessed the attraction of Anopheles arabiensis and Anopheles quadriannulatus (Diptera: Culicidae) to humans by comparing their abundance in larval and adult samples collected across an ecologically heterogeneous landscape in southern Tanzania. Surveys of mosquito larvae and adults were conducted across a landscape mosaic of different habitat types, with a gradient of land use practices ranging from comprehensive conversion to agriculture and human settlement through to essentially intact natural ecosystems inside well‐protected conservation areas. Larvae were collected from all water bodies within a 2 km radius of each of 40 mobile camping locations, while adults were surveyed using four light traps and one interception netting barrier trap at each transient camp. Light traps were placed at defined locations, specifically beside a human‐occupied tent, near the camp, in a nearby streambed, and in an open natural glade, while the barrier trap was also placed in an open natural glade. Almost all adult Anopheles gambiae complex mosquitoes caught were unfed and presumably host‐seeking. Breaking this complex down by sibling species, two to four times more An. arabiensis were caught in the light traps placed beside human‐occupied tents, although barrier traps achieved somewhat higher capture rates again. In contrast, An. quadriannulatus catches were consistently low across all traps, even in wild areas where it dominated larval populations, but were highest in open glades away from the camp location and activities. The overall proportion of An. arabiensis in adult collections was higher than in larval samples (98.7% vs. 78.3%, p &lt; 0.0001) and adults caught beside human‐occupied tents had 20 times higher odds of being An. arabiensis , rather than An. quadriannulatus . Similarly, the barrier trap placed away from the camp, but frequently visited by human researchers, exhibited 22‐fold enrichment of An. arabiensis . These results confirm strong attraction of An. arabiensis to humans, contrasting with complete non‐responsiveness of the non‐vector An. quadriannulatus . Light traps beside human‐occupied tents efficiently capture anthropophagic mosquitoes outdoors, suggesting those occupants act as de facto bait hosts. In contrast, traps farthest from people give apparently unbiased representations of larval population composition, albeit with very low efficiency. However, frequent collector visits to netting barriers appear to attract anthropophagic mosquitoes, turning them into human‐baited traps in practical terms.