Single-addition and sequential net acid generation (NAG) experiments were carried out on natural and prepared samples with varying contents of organic matter and pyritic S. Results show that carbonaceous matter can produce significant quantities of organic acids during peroxide oxidation in the NAG test, which leads to overestimation of acid-forming capacity and misclassification of benign samples as potentially acid forming (PAF). Organic matter effects were most pronounced in samples with low S (less than 0.5 per cent) and total organic carbon (TOC) contents of greater than seven per cent. The single-addition NAG test will tend to overestimate acid potential in samples with little or no pyrite and high organic matter content, potentially leading to sample misclassification. High pyritic sulfur samples tend to show little difference in single-addition NAG acidities with increasing organic matter, and it appears that pyrite is preferentially oxidised by peroxide when present in significant concentrations (greater than one per cent pyritic S). In sequential NAG tests overestimation was greatest in samples with little or no pyrite, but highly carbonaceous (>20 per cent TOC) pyritic samples were affected in later stages as the pyrite was depleted. A modified NAG method is proposed to account for the relative proportions of pyrite derived acidity and organic acidity in a given NAG solution.