We describe an experiment that uses the grouping tendencies and navigational abilities of the homing pigeon (Columba livia) to investigate the possibility of socially mediated information transfer in a field setting. By varying the composition of paired–release types, we allowed some naive birds to receive an accurate demonstration of the home route whilst others were paired with similarly naive conspecifics. After this ‘paired phase’, we predicted that if any learning of spatial information occurred then naive members of the former pairs would outperform their untutored conspecifics when re–released individually during the subsequent‘single phas’ of the experiment. This prediction was not confirmed. Neither homing speed nor initial orientation was superior in individually released tutored versus untutored birds, despite the fact that both performance measures were better in the earlier‘paired phas’ with experienced demonstrators. Our results suggest that although naive homing pigeons clearly interact with their experienced partners, they are unable to transfer any individually useful spatial information to subsequent homing flights.