This paper describes how the AMBR model explains multiple analogies and more specifically how the use of a superficially similar analogical base, that turns out to be inappropriate (we call it a bridge analogy), may actually lead to the re-representation of the target and the activation of a more appropriate remote analogical source. A simulation is described that demonstrates this capability of the model. A specific prediction of the model about the re-representation that the presence of the bridge analogical source is causing is tested in a psychological experiment.