Background: To contribute to the current debate as to the relative influences of dietary intake and physical activity on the development of adiposity in community-based children. Methods: Participants were 734 boys and girls measured at age 8, 10 and 12 years for percent body fat (dual emission x-ray absorptiometry), physical activity (pedometers, accelerometers); and dietary intake (1 and 2-day records), with assessments of pubertal development and socioeconomic status. Results: Cross-sectional relationships revealed that boys and girls with higher percent body fat were less physically active, both in terms of steps per day and moderate and vigorous physical activity (both sexes p<0.001 for both measures). However, fatter children did not consume more energy, fat, carbohydrate or sugar; boys with higher percent body fat actually consumed less carbohydrate (p = 0.01) and energy (p = 0.05). Longitudinal analysis (combined data from both sexes) was weaker, but supported the cross-sectional findings, showing that children who reduced their PA over the four years increased their percent body fat (p = 0.04). Relationships in the 8 year-olds and also in the leanest quartile of all children, where adiposity-related underreporting was unlikely, were consistent with those of the whole group, indicating that underreporting did not influence our findings. Conclusions: These data provide support for the premise that physical activity is the main source of variation in the percent body fat of healthy community-based Australian children. General community strategies involving dietary intake and physical activity to combat childhood obesity may benefit by making physical activity the foremost focus of attention.