Animal studies were used in addition to epidemiological studies to investigate the effects of exposure, exposure rate and other factors in predicting risks resulting from human exposures. The advantage of animal data is that animal experiments are generally conducted under well- controlled conditions and that exposure and exposure rate can be estimated more accurately. A trend towards increasing tumour risk with decreased exposure rate was observed in Sprague-Dawley rats exposed at cumulative exposure varying from about 0.72 J h m(-3) (200 WLM) up to 10.8 J h m(-3) (3,000 WLM) and high exposure rates varying from 0.09 J h m(-3) (25 WLM per week) to 1.8 J m(-3) (500 WLM per week). In contrast, the results obtained at low cumulative exposure, comparable to domestic indoor exposures showed no evidence of an inverse exposure-rate effect. Chronic radon exposure at 0.09 J h m(-3) (25 WLM), protracted over a 18 months period, at a potential alpha energy concentration (PAEC) of 0.042 mJ m(-3) (2 WL), resulted in fewer lung carcinomas in rats than a similar cumulative exposure protracted over 4 to 6 months at a PAEC of 2.1 mJ m(-3) (100 WL). Moreover, the lung cancer incidence in rats exposed at low exposure rate (0.60%) was slightly lower than that in control animals (0.63%). The preliminary results of a new series of experiments carried out to investigate the influence of exposure-rate on lung cancer induction in rats at relatively low cumulative exposures of 0.36 J h m(-3) (100 WLM), and PAEC varying from 0.22 mJ m(-3) (10-12 WL) to 3.15 mi m(-3) (150 WL) indicate that at relatively low cumulative exposures comparable to lifetime exposures in high-radon houses or current underground mining exposures, the risk of lung cancer in rats decreases with decreasing PAEC, i.e., exposure rates. These data suggest that in terms of risk of induction of lung cancer, there is a complex interplay between cumulative exposure and exposure rate, resulting in an optimal exposure rate at a given exposure level. The significance of exposure rates in assessing the hazards of domestic radon exposure was addressed on biophysical grounds by Brenner, who concluded that, when cumulative exposures are sufficiently low that multiple traversals of target cells by alpha particles are rare - that is the case for typical domestic radon exposures -, all exposure-rate enhancement effects disappear. Our recent data in rats appear to follow the same trend.