OBJECTIVES: The aim of the study was, on the basis of new information on nickel species and exposure levels, to generate a specific exposure matrix for epidemiologic analyses in a cohort of Norwegian nickel-refinery workers with a known excess of respiratory cancer. METHODS: A department-time-exposure matrix was constructed with average exposure to total nickel estimated as the arithmetic mean of personal measurements for periods between 1973 and 1994. From 1972 back to the start of production in 1910, exposure concentrations were estimated through retrograde calculation with multiplication factors developed on the basis of reported changes in the metallurgical process and work environment. The relative distribution of water-soluble nickel salts (sulfates and chlorides), metallic nickel, and particulates with limited solubility (sulfides and oxides) was mainly derived from speciation analyses conducted in the 1990s. RESULTS: The average concentration of nickel in the breathing zone was