The general expression of the transcription factor gene Pax-6 homologues and the overall presence of the photopigment opsin have cast doubt on the polyphyletic evolution of photoreceptors. Herein it is proposed that the evolutionary pathway of photoreceptors reflects two different, successive processes: (i) the (monophyletic?) differentiation of photoreception as such, mediated by a specific transcription factor gene (such as Pax-6 or sine oculis) and (ii) the genetic information of that induction factor (normative unit for photoreception). The latter stimulates the (polyphyletic) differentiation of the photoreceptors themselves through its multiply convergent co-options with variable network-modifications (intercalation of different genes). The expression of transcription factor genes does not per se imply homology of the differentiated photoreceptors (but at most some pattern of homoiology). The differentiation of both receptor types, ciliary versus rhabdomeric, in one and the same cell during development (of veliger larvae; Blumer 1996) shows them to be interchangeable structures (mere morphs). Apparently dependent of functional requirements, the structural type of the receptive organelle has no direct bearing upon the homology identification of the photoreceptors. This obviates the need to propose separate (ciliary and rhabdomeric) precursor cells in metazoans. A possible primitive “dermal” receptive cell which was polyphyletically adapted for metazoan photoreceptors is discussed. The rich morphological diversity of photoreceptors (including the larval organs) in Mollusca appears to represent in-group differentiations. Their polyphyletic lines are surveyed and the fine structure of eyes of three pteriomorph Bivalvia species—Lima lima (Linnaeus, 1758) (with a subdivided retina), Chlamys varia (Linnaeus, 1758), and Pseudamussium peslutre (Linnaeus, 1771) (with incomplete proximal retina)—is reported.