We review information from the past 5 years on changes in diesel exhaust (DE) emissions and developments in the study of DE toxicity. New DE technologies have changed the composition of DE considerably, reducing emissions of many of the components of health concern. The increasing similarity of modern diesel and compressed natural-gas engine emissions needs to be reflected in any regulatory analysis. Even for historical DE emissions, considerable study of DE exposure in animals or humans has not produced data useful for risk assessment. DE inhalation exposure in most species (hamsters, guinea pigs, mice) does not produce lung tumors. Inhalation studies of DE in rats found lung tumors only with lung overload, and tumors also occurred when inert dusts were inhaled at overload concentrations. The animal data are reassuring and show that DE is not a concern at ambient exposure levels. Re-analyses of occupational epidemiology have been shown to have serious shortcomings that do not allow the derivation of a quantitative cancer risk. Studies of railroad workers found no dose response; rather cancer risk appeared to decrease for individuals with greater DE exposure. Studies of truck drivers also suffer from serious flaws because of misconceptions about the year that diesel was introduced, lack of an adequate latency period, and the realization that drivers exhibited increased lung cancer risk even prior to the diesel era. Recent industrial hygiene studies of drivers show that DE was not likely to be a primary source of particles or polycyclic aromatic hydrocarbons. Further, there is no dose response across occupations. In fact, the occupation (underground miners) with the highest exposure to DE does not exhibit increased cancer risks. This new information seriously weakens earlier risk characterizations of DE by various regulatory groups. New research and better exposure measurements are needed before a reliable risk assessment of DE can be produced.