The optic tectum differentiates from the alar plate of the mesencephalon and receives retinal fibres in a precise retinotopic manner. Here, mechanisms of tectum polarisation and regionalisation are reviewed. Misexpression of Pax2, Pax5 or En can change the fate of the presumptive diencephalon to that of the tectum. Ephrin A2 and A5 are expressed in a gradient in the tectum, caudal high and rostral low, and may play important roles in the formation of a precise retinotectal projection map. Retinal fibres that express receptors for these ligands, and which come from the temporal retina, are repulsed by the ligands and do not invade the caudal tectum. Both En1 and En2 can regulate posterior characteristics in the tectum by inducing ephrin A2 and A5. Transplantation experiments in chick have indicated that the mes/metencephalic boundary works as an organiser for the tectum and the cerebellum. Fgf8 is a candidate signalling molecule in the organiser. Pax2/5, En, and Fgf8 are in a positive feedback loop for their expression such that misexpression of one of these genes in the diencephalon turns on the feedback loop and can result in induction of an optic tectum. Otx2 and Gbx2 appear to repress each other's expression and contribute to defining the posterior border of the tectum. Misexpression of Otx2 in the metencephalon can change the fate of its alar plate to a tectum, and misexpression of Gbx2 in the mesencephalon can cause anterior shifting of the caudal limit of the tectum. The anterior border of the tectum may be determined as a result of repressive interactions between Pax6 and En1/Pax2. Along the dorsoventral axis of the mesencephalon, Shh contributes to ventralize the tissue; that is, Shh can change the fate of the presumptive tectum to that of the tegmentum that is the ventral structure. It is proposed that the brain vesicle that expresses Otx2, Pax2, and En1 may differentiate into the tectum.