A key issue facing regulators is the interpretation of exposure-response data at low levels of exposure. The evidence for chrysotile shows that for lung cancer and mesothelioma there exist levels of exposure below which risks are for practical purposes zero. In all industry sectors except textiles there is remarkable consistency in the absence of statistically valid relationships between exposure and risk of lung cancer or mesothelioma at the lower levels of exposure, even in large populations and long follow-up. A review of epidemiological data taken with new experimental and pathological evidence gives a framework for rational decisions on risk.