High incidences of lung cancers have been observed in a number of studies in which rats were chronically exposed by inhalation to high concentrations of diesel engine exhaust and carbon black particles. These particles have previously been viewed as being relatively innocuous compared with other particles such as benzo[a]pyrene that are carcinogenic because of specific chemical properties. Studies with mice and Syrian hamsters exposed to similar concentrations of diesel exhaust did not produce an excess of lung cancer or yielded equivocal outcomes. Diesel exhaust soot and carbon black particles are readily inhaled and deposited in the pulmonary region, where they are retained with a long half-life because of their low solubility. Substantial evidence indicates that the increased incidence of at lung cancers results from the accumulation of large burdens of particles in the lungs, altered clearance of particles from the lungs, persistent inflammation, increased cell turnover, and induction of mutations in lung epithelial cells. The mutations and subsequent rumors are hypothesized to occur as a result of persistent inflammation and increased cell turnover rather than as a result of direct interaction of chemical constituents of the particles with DNA of lung cells. The observed effects in rats appear to be threshold phenomena that occur only with prolonged exposure to high concentrations of particles. Thus the rat lung cancer findings at high concentrations should not be extrapolated to low concentrations using the linearized multistage model typically used as a default assumption for assessing the cancer risk of chemicals. This article reviews past approaches to evaluating the carcinogenic risk of diesel exhaust and carbon black and suggests alternative approaches to characterizing their human cancer risk.