BACKGROUND Given the continuing epidemic of obesity, policymakers are increasingly looking for levers within the local retail food environment as a means of promoting healthy weights. PURPOSE To examine the independent and joint associations of absolute and relative densities of restaurants near home with weight status in a large, urban, population-based sample of adults. METHODS We studied 10,199 adults living in one of four cities in southern Ontario, Canada, who participated in the Canadian Community Health Survey (cycles 2005, 2007/08, 2009/10). Multivariate models assessed the association of weight status (obesity and body mass index) with absolute densities (numbers) of fast-food, full-service and other restaurants, and the relative density (proportion) of fast-food restaurants (FFR) relative to all restaurants within ~10-minute walk of residential areas. RESULTS Higher numbers of restaurants of any type were inversely related to excess weight, even in models adjusting for a range of individual covariates and area deprivation. However, these associations were no longer significant after accounting for higher walkability of areas with high volumes of restaurants. In contrast, there was a direct relationship between the proportion of FFR relative to all restaurants and excess weight, particularly in areas with high volumes of FFR (e.g., odds ratio for obesity=2.55 in areas with 5+ FFR, 95% confidence interval: 1.55-4.17, across the interquartile range). CONCLUSIONS Policies aiming to promote healthy weights that target the volume of certain retail food outlets in residential settings may be more effective if they also consider the relative share of outlets serving more and less healthful foods.