Similar patterns of cognitive and motor deficits have been widely reported from manganese exposures in welding, metallurgical and chemical industry workers. A risk assessment was performed based on studies reported in the literature, extending some earlier work, and deriving new estimates of exposure response and excess risk. Many investigations of manganese neurological effects in humans have insufficient information to derive an exposure response; however, findings from a chemical manufacturer, two smelter and two welder populations permitted application of the benchmark dose procedure for continuous end-points. Small particles and aggregates of condensation fume (condensing vaporized metal, <0.1 μm in diameter) appear to have a higher potency per unit mass than larger particles from dusts (>1.0 μm). Consideration was given to long-term effects of continuous low exposures that instead of producing increasing toxicity attain a steady-state condition. Impairment was defined as excursions beyond the 5th percentile in a normal population and the concentrations of manganese predicted to result in 1% excess prevalence of impairment over different time periods were calculated. Over five years, exposures resulting in 1% excess prevalence of impairment (for purposes of discussion) were in the vicinity of 10 μg/m3 for manganese fume and 25 μg/m3 for larger particle dusts. These levels are below current recommendations for occupational limits on manganese exposure in the United States.