During development of the retinotectal axonal projection, which connects the retina to the optic tectum in the midbrain, the axons of neighbouring retinal ganglion cells project to neighbouring positions in the optic tectum (topographical mapping). However, although retinal fibres rigidly target their destinations in some experimental circumstances, in others they adapt to grossly diverse targets. Here (see p. 335), Franco Weth and colleagues present a surprisingly simple model that explains these hitherto puzzling discrepancies. In this model, topographical axonal mapping relies solely on the balance of forward and reverse signalling by the ephrin/Eph family of guidance molecules. To test their model, the researchers develop a novel ephrin/Eph (double-cue) stripe assay and show experimentally that the simultaneous presence of forward and reverse ephrin/Eph signalling is indeed sufficient for appropriate topographic growth decisions in chick embryonic nerve fibres. Moreover, using computer simulations, they show that their new model is capable of reproducing the discrepant data collected over the years on topographic mapping by the retinotectal axonal projection.