To explain the activation of transcription by a remote enhancer, two models are most often considered, namely looping and scanning. A scanning model, also referred to as 'polymerase entry site' model predicts that for two adjacent promoters the one proximal to an enhancer would be preferentially activated. Preferential activation of the proximal promoter in a tandem promoter arrangement has been found before in several laboratories, including our own, but for technical reasons the data were inconclusive with regards to the enhancer mechanism. In the work presented here, we readdress the question of preferential promoter activation by an enhancer using a more clearly defined system. Two identical promoters were kept closeby in a divergent, or directly repeated orientation. The SV40 enhancer was placed at a great distance on one or the other side of the two promoters, to see whether the enhancer position influenced the relative efficiency of the two promoters in transfected cells. Our finding that the promoter usage is virtually unaffected by the enhancer position does not favor a scanning model, but is compatible with a looping model of enhancer action.