The volatility of mercury can cause large errors during quantification when methods based on acid digestion are applied. In this study, mercury levels in coals have been determined by an atomic-absorption-based instrument hitherto not used with coals or coal-derived materials. The results have been compared with ＇certified＇ values of reference materials. The instrument is relatively easy to use: solid and liquid samples may be introduced directly, without pretreatment. The range of samples studied included the Argonne Premium Coal Samples and other coals, coal-derived products, biomass materials, sand and kaolin. The instrument gave correct mercury concentrations for certified reference materials. Samples of silver birch and forest residue contained similar concentrations of mercury as observed in coals; other biomass samples contained far smaller quantities of mercury. Thus, the use of biomass in power generation would not necessarily lead to any great reduction of mercury in emissions from power plant: careful selection appears necessary. Only 3% of the mercury of the original coal was detected in the filter cake (corresponding to undissolved coal and mineral matter) from a coal liquefaction pilot plant. Mercury in the original coal appears to have passed either to (i) gas formed during liquefaction (between 2 and 4% daf) or to (ii) the dissolved coal extract, possibly in the form of organometallic complexes. The presence of 0.15 ppm mercury in a coal tar pitch also suggests the organometallic retention of mercury in this fraction. All solid residue streams from a gasification pilot plant, operating at about 930-960 degrees C, have shown nearly negligible mercury, compared to the original coal, indicating that the bulk of the mercury in the feedstock was released into the product gas.