OBJECTIVES: To establish whether choice and availability of takeaway and restaurant food consumption are associated with increased obesity. DESIGN: Crossroads Undiagnosed Disease Study: a cross-sectional study undertaken between June 2001 and March 2003. SETTING: A regional centre and six shire capitals of variables size in rural Victoria, Australia. PARTICIPANTS: In total, 1454 residents of randomly selected households. MEASUREMENTS: Obesity (by body mass index (BMI) or waist circumference), weekly recreational activity, self-reported frequency of takeaway consumption, number of local takeaway and restaurant food outlets in the area. RESULTS: The prevalence of obesity ranged from 25.5-30.8% and was higher than the general Australian population among both men and women. Those in the regional centre were less likely than those in large and small shire capitals to participate in recreational activity of 150 min or more (39.7 vs 48.4%, 46.0% respectively, P=0.023) and yet reported better access to facilities and amenities for physical activity. Recreational activity of > or =150 min/week was associated with 0.75 (0.58-0.97) fold less risk of obesity. BMI was unrelated to takeaway consumption. Waist circumference was significantly lower among those eating no takeaways, but similar whether takeaways were consumed <1/month or > or =1/week. Increased takeaway consumption was associated with increased consumption of higher fat preparations of dairy and meat products. Availability of takeaway outlets and restaurants was unrelated to obesity. CONCLUSION: The obesity epidemic exists among those without significant consumption of or availability to takeaway foods. In a setting of easy availability of food, the obesity epidemic relates strongly to reduced physical activity, but not to consumption of takeaway food.