Capsule Empirical support is provided for the design and field methods associated with a replicated supplementary feeding experiment.Aims We tested assumptions associated with a supplementary feeding experiment in which counts of House Sparrows were used to infer impacts of feeding mealworms and seed on reproductive output and population size. Methods Supplementary food was provided at 33 House Sparrow colonies spread across suburban London, with 33 unfed colonies serving as controls. We compared the characteristics of fed and unfed colonies and tested whether competition for mealworms increased with colony size. We tested whether counts of young sparrows reflected variation in reproductive output and whether supplementary feeding affected the distribution of young sparrows. Results Fed and unfed colonies were similar with respect to habitat, cat density, background feeding and initial colony size. Intensity of feeder usage increased with colony size. Counts of fledglings and juveniles were positively correlated with counts of active nests. Supplementary feeding influenced the distribution of juveniles but not that of recent fledglings. Conclusions Fed and unfed colonies were similar in character. Competition for mealworms was greater at larger colonies and fledgling counts provided a useful measure of variation in reproductive output across colonies.Supplementary feeding is a commonly used research technique to assess the effects of food limitation on birds and other animals (Boutin 1990, Robb et al. 2008. A large number and variety of studies have recorded mainly positive impacts of supplementary feeding on various components of avian breeding performance, including earlier laying, larger clutches, increased chick growth rates and higher fledging success (Newton 1998, Robb et al. 2008) although a few negative impacts have also been reported (Harrison et al. 2010). Most studies have focused entirely on impacts on reproductive success and have been conducted over small spatial and temporal scales. Fewer studies have measured impacts of supplementary feeding outside of the breeding season or on population density (van Balen 1980, Jansson et al. 1981, Brittingham & Temple 1988, Siriwardena et al. 2007. Furthermore, remarkably little is known about the impacts on birds of the ubiquitous global phenomenon of supplementary feeding of garden birds despite concerns about potential negative impacts such as increased disease transmission and 'dependency' on anthropogenic food sources (Jones & Reynolds 2008).While it is relatively straightforward to provide accessible and palatable food in supplementary feeding experiments, interpretation of apparent impacts requires knowledge of the intensity and geographical scale of the utilization of that food (Boutin 1990, Newton 1998. If the distances over which target species travel in order to exploit supplementary food exceed the distances between experimental units (fed and unfed areas), there is a risk of treatment 'contamination' and compromise of study design.