Source-sorted municipal organic waste from different dwelling types in five Danish cities was sampled during one year. The samples were from permanent, full-scale systems or temporary, experimental systems for collection of source-sorted municipal organic waste. Pre-treatment of the organic waste prior to biological treatment was used in all cities to remove foreign objects and provide size reduction. All sampling was performed after pre-treatment in order to obtain more homogeneous and representative samples. The sampling included both the pre-treated waste and the reject from the pre-treatment allowing for estimation of the composition of the original waste. A total of 40 waste samples were chemically characterised with respect to 15 parameters. The waste generally consisted of around 88% VS of which an average of 80% was easily degradable. The average content of N, P and K in the dry matter of the organic waste was 2.5%, 0.4% and 0.9%, respectively. A general analysis of variance was applied to show the influence of the collection system, dwelling type and annual season on the waste composition. The content of plastic and crude fibres in the waste differed the most among the samples, probably due to use of different bag types (plastic and paper) in the different collection systems. Variations in the ash content and the calorific value might be explained by differences in the sorting instructions (whether soil and cat litter are allowed in the organic fraction). Significant seasonal variations were seen for ash, S and Cl. Dwelling type showed no statistically significant influence on any waste components. A test for uniform distribution of the p-values from the analysis of variance (Kolmogorov-Smirnov test) showed that the overall composition of the collected waste was strongly affected by the collection system (city) and season, while dwelling type had no significant influence. (c) 2006 Elsevier Ltd. All rights reserved.