The past recommendations of the ICRP about dose limits to the skin are reviewed. Recently, an ICRP Task Group has been revisiting the old arguments and setting them against new data. With the exception of the function of cells in the skin associated with immunocompetence, non-stochastic effects have been well characterized and threshold doses are known with a precision appropriate for setting radiation protection standards. The current dose limitations of 0.5 Sv per year and a working lifetime dose limit of 20 Sv should protect the worker population against deterministic effects. When the ICRP made its recommendations in 1977 for dose limits there was no appreciation of the importance of the interaction of ultraviolet radiation (UVR) and X-rays. Both clinical and experimental data show that the risk of ionizing-radiation-induced cancer is significantly increased by subsequent exposures to UVR. Therefore, risks for sun-exposed areas of skin differ from those that are shielded. The risk estimate for skin cancer is very dependent on the selection of the projection model and on the mortality rate assumed. Based on the relative risk model a mortality rate of 0.2 per cent and summing risks for both UVR exposed and shielded skin the risk is about twice (1.94 x 10(-4) Sv-1) that which ICRP derived in 1977. With the absolute model the risk is considerably less, about 0.5 x 10(-4) Sv-1. There is still insufficient understanding of the effects of multiple or protracted exposures on the risk of skin cancer induction. Experimental results suggest that exposures, at least to relatively high total doses, that are protracted over a long period are more carcinogenic than a small number of exposures over a short period.