The highly ordered neuronal projections from the retina to the tectum mesencephali (optic tectum) in several vertebrate groups have been intensively studied. Several hypotheses so far have been proposed, suggesting mechanisms to explain the topographical and biochemical specificity of the retinotectal projections during ontogeny. In the present paper we compare the main hypotheses of retinotectal development with respect to the nature of specificity envisaged, the activity-dependence versus inheritance criterium and the strategy of argument, in casu the descriptive versus interferential type (...) of argument. Matching of the current developmental hypotheses and mechanisms for elimination or persistance of contralateral, respectively ipsilateral connections, is attempted with respect to the known neuroanatomical connectivity data in the amphibian optic tectum and with respect to the symmetry relations between ipsilateral and contralateral connection patterns described in amphibians and other vertebrate groups. Local mechanisms influencing the survival potential of synaptic contacts between retinal afferents and tectal neurons are probably essential for generating global symmetry patterns. Finally, a global topological approach is discussed with respect to its applicability in the amphibian retinotectal projection system. Basic assumptions of topological modeling appear to rely on anatomical and functional properties of the visual system like left-right symmetry, dichotomy, (absence of) convergence and segregation of fibers and neurons into columnar information units. It is concluded however, that more neuroanatomical, physiological and ultrastructural data are needed to establish a formalized operation model of the amphibian retinotectal system. (shrink)