Based on studies of wolffian lens regeneration in the newt, in which the lens can be regenerated from the iris pigmented epithelium, we have shown by cell culture studies that the capacity of lens transdifferentiation is not limited to the newt cells, but widely conserved in pigmented epithelial cells (PECs) of chick and quail embryos and even of human fetuses. Recently, we have established a unique in vitro model system of chick embryonic PECs. In this culture system we are able to control each step of transdifferentiation from PECs into lens cells by regulating culture conditions and to produce a homogeneous cell population with potential for synchronous differentiation into either lens or pigment cell phenotype. These multipotent (at least bipotent) cells showed cellular characteristics resembling neoplastic cells in many ways. They did not express both lens and pigment cell specific genes analyzed so far, except delta-crystallin gene, which is expressed in developing lens of chick embryos. It has been proved by application of cell culture procedures of the system that PECs dissociated from fully-grown human eyes readily transdifferentiated into lens phenotypes in the manner observed in chick embryo PECs. In addition, we could predict that molecules detected in either cell surface or intercellular space stabilized the differentiated state of PECs in the newt and that the loss of these molecules might be one of the key steps of lens regeneration from the iris epithelium.