Culicoides were captured at a BTV-infected dairy near Gulpen in the province of Limburg (south-east Netherlands) between 14 September and 4 October 2006. Onderstepoort-type blacklight traps were used to sample Culicoides both inside and outside a partially open shed housing 11 cattle. A total of 28 light trap collections were made at the shed and yielded: - 9371 Culicoides representing 11 species; >90% comprised five potential vectors of BTV and in order of abundance were Culicoides obsoletus and Culicoides scoticus (of the Obsoletus Complex), Culicoides dewulfi, Culicoides pulicaris and Culicoides chiopterus; Culicoides imicola, the principal Mediterranean (and African) vector of BTV, was absent. - 2339 Culicoides representing seven species were captured inside (endophily) the cattle shed; >95% comprised the Obsoletus Complex and C. dewulfi. Conversely, the Pulicaris Complex, represented by five species and including C. pulicaris, showed strong exophily with >97% captured outside the shed. - 7032 Culicoides were captured outside the shed, approximately threefold more than inside. This trend was reversed on an overcast day, when eightfold more Culicoides were captured inside; this indicates that when the light intensity outdoors is low Culicoides will attack (i) earlier in the day while cattle are still at pasture, and (ii) might follow cattle into the sheds in the late afternoon leading to elevated numbers of biting midges being trapped inside the shed during the subsequent hours of darkness. - Culicoides were captured inside the shed on all 14 sampling nights. On occasion up to 33% were freshly blood fed indicating they had avidly attacked the cattle inside (endophagy); because half the cattle had seroconverted to BTV, and because no cattle were left outdoors at night, the data indicate that (i) the housing of animals in partially open buildings does not interrupt the transmission of BTV, and/or (ii) BTV is being transmitted while cattle are grazing outdo