Abstract
: The swarming behaviour of natural populations of Anopheles arabiensis was investigated by conducting transect surveys on 10 consecutive days, around dusk, from March to April and from September to October 2012 in Dioulassoba, a district of Bobo-Dioulasso city in Burkina Faso (West Africa). Swarms were observed outside, around identified larval breeding sites on the banks of the Houet River, as well as in the open-air courtyards found at the centre of many homes in the region. Swarms were found to occur in open sunlit spaces, mostly located above physical or visual cues somehow visually distinct from the surrounding area. Overall 67 and 78 swarms were observed, respectively, during the dry season (March-April) and the rainy season (September-October) of 2012, between 1.5 and 4.5m above the ground at the centre. 964 mosquitoes were collected and analysed from dry season swarms, of which most were male, and all were An. arabiensis, as were the few resting mosquitoes collected indoors. Larvae collected from breeding sites found on the banks of the Houet River mostly consisted of An. arabiensis and only a minority of Anopheles coluzzii (formerly identified as An. gambiae M form). Of 1694 mosquitoes analysed from 78 swarms in the rainy season collections, a few An. gambiae (formerly known as An. gambiae S form) males were identified, and the remainders were An. arabiensis. The majority of larvae collected during the wet season from the same breeding sites were identified as An. arabiensis followed by An. coluzzii and An. gambiae. The same pattern of species composition was observed in resting mosquitoes, though the proportion of An. arabiensis was less overwhelming. These data support the conclusion that An. arabiensis is the most prevalent species in this area, though the difference in species composition when using different population sampling techniques is noteworthy. Further studies are required for more detailed investigations of male dispersal, feeding behaviour and mating patterns in this urban setting.