In a spatial simulation we show that formation of a stable, narrow hybrid zone requires no selection, only that individuals can recognize their own population type since this is a prerequisite for assortative mating. The European crow occurs in two species/subspecies that meet in a long hybrid zone, the black carrion crow Corvus corone and the grey and black hooded crow C. cornix. In a previous study (Brodin A, Haas F (2006) Anim Behav 72:139) we mimicked sexual imprinting in nestling crows with artificial neural networks, using a learning process that simulates proximate retina perception. The networks were trained on 700 images to recognize either carrion, hooded or hybrid phenotypes as their own. After training the networks were exposed to 300 new, unfamiliar pictures of crows. The networks that had been trained on pure subspecies then showed strong preference for their own type. Networks trained on hybrid crows showed weaker preference for their own type. Assuming that these preferences will determine mate choice preferences we have combined them with empirical data on dispersal of young crows to investigate what kind of geographical distribution pattern this would create. We then assume that the two subspecies meet and hybridize along a straight border. In only 60 generations a stable, narrow hybrid zone that resembles the real hybrid zone was formed. This zone remained stable over time. With minor adjustments of parameters the results would not only fit the width but also the shapes of the clines in the field.
International Symposium on Speciation - From Diversification to Reproductive Isolation