Using sibship reconstructions to understand the relationship between larval habitat productivity and oviposition behaviour in Kenyan Anopheles arabiensis

BACKGROUND: Strategies for combatting residual malaria by targeting vectors outdoors are gaining importance as the limitations of primary indoor interventions are reached. Strategies to target ovipositing females or her offspring are broadly applicable because all mosquitoes require aquatic habitats for immature development irrespective of their biting or resting preferences. Oviposition site selection by gravid females is frequently studied by counting early instar larvae in habitats; an approach which is valid only if the number of larvae correlates with the number of females laying eggs. This hypothesis was tested against the alternative, that a higher abundance of larvae results from improved survival of a similar or fewer number of families. METHODS: In a controlled experiment, 20 outdoor artificial ponds were left uncovered for 4 days to allow oviposition by wild mosquitoes, then covered with netting and first and second instar larvae sampled daily. Natural Anopheles habitats of two different types were also identified, and all visible larvae sampled. All larvae were identified to species, and most samples of the predominant species, Anopheles arabiensis, were genotyped using microsatellites for sibling group reconstructions using two contrasting softwares, BAPS and COLONY. RESULTS: In the ponds, the number of families reconstructed by each software significantly predicted larval abundance (BAPS R(2) = 0.318, p = 0.01; COLONY R(2) = 0.476, p = 0.001), and suggested that around 50% of females spread larvae across multiple ponds (skip oviposition). From natural habitats, the mean family size again predicted larval abundance using BAPS (R(2) = 0.829, p = 0.017) though not using COLONY (R(2) = 0.218, p = 0.68), but both softwares once more suggested high rates of skip oviposition (in excess of 50%). CONCLUSION: This study shows that, whether in closely-located artificial habitats or natural breeding sites, higher early instar larval densities result from more females laying eggs in these sites. These results provide empirical support for use of early instar larval abundance as an index for oviposition site preference. Furthermore, the sharing of habitats by multiple females and the high skip-oviposition rate in An. arabiensis suggest that larviciding by auto-dissemination of insecticide may be successful.