This study was conducted to examine the carcinogenic effects of inhaled beta-particle-emitting radionuclides, particularly in lower dose regions in which there were substantial uncertainties associated with available information. A total of 2751 F344/N rats (1358 males and 1393 females) approximately 12 weeks of age at exposure were used. Of these, 1059 rats were exposed to aerosols of 144CeO2 to achieve mean desired initial lung burdens (ILBs) of 18 kBq (low level), 247 rats to achieve mean ILBs of 60 kBq (medium level) and 381 rats to achieve mean ILBs of 180 kBq (high level). Control rats (total of 1064) were exposed to aerosols of stable CeO2. Based on the 95% confidence intervals of the median survival times and the cumulative survival curves, there were no significant differences in the survival of groups of female and male exposed rats relative to controls. The mean lifetime beta-particle doses to the lungs of the rats in the four groups were: low level, 3.6 +/- 1.3 (+/-SD) Gy; medium level, 12 +/- 4.5 Gy; and high level, 37 +/- 5.9 Gy. The crude incidence of lung neoplasms increased linearly with increasing doses to the lungs (controls, 0.57%; low level, 2.0%; medium level, 6.1%; and high level, 19%). The estimated linear risk coefficients for lung neoplasms per unit of dose to the lung were not significantly different for the three dose levels studied. The risk coefficient at the lower level was 39 +/- 14 (+/-SE) excess lung neoplasms per 10(4) rat Gy; at the medium level the risk was 47 +/- 12; and at the higher level the risk was 50 +/- 9.0. The relationship of beta-particle dose to the lung and the crude incidence of lung neoplasms was described adequately by a linear function. We concluded that the risk of lung neoplasms in rats per unit of radiation dose did not increase with decreasing mean beta-particle dose to the lung over the range of 3.6 to 37 Gy. The weighted average of these three values was 47 +/- 6.4 (+/-SE) excess lung neoplasms per 10(4) rat Gy. To extend the risk coefficients for lung neoplasms to lower doses by experimentation will require much larger numbers of rats than used in this study.