The partitioning of iron during pulverized coal combustion was investigated both experimentally and theoretically. Emphasis was on determining how coal variables and combustion conditions influenced the formation of slagging precursors. Experimental work consisted of burning a suite of six well-characterized coals in an aerodynamically well-defined 17 kW downflow combustor. Speciation of iron in collected ash samples was determined by M/Sssbauer Spectroscopy. A model was developed to predict the partitioning of iron for the coals and experimental conditions examined. The model was based on a competition between pyrite oxidation and iron capture by silicates and required as input, CCSEM data on the initial distribution of mineral matter in the coal. It used literature-derived mechanisms and kinetics to calculate pyrite oxidation and char combustion rates, together with a new glass formation mechanism, coupled with current knowledge of sintering. Requiring only two unknown parameters to be fitted, the model agreed well with 26 sets of experimental data, which covered a wide range of iron partitioning in the fly ash. Both experimental and theoretical results suggest that removal from the parent coal, of extraneous iron alone, may not be very effective in eliminating slag formation in boilers.