We examined mean effects and variance moderating effects of measures of physical activity and fitness on six measures of adiposity and their reciprocal effects in a subsample of the population-representative Danish Twin Registry. Consistent with prior studies, higher levels of physical activity suppressed variance in adiposity, but this study provided further insight. Variance suppression appeared to have both genetic and environmental pathways. Some mean effects appeared due to reciprocal influences of environmental circumstances differing among families but not between co-twins, suggesting these reciprocal effects are uniform. Some variance moderating effects also appeared due to biases in individual measures of adiposity, as well as to differences and inaccuracies in measures of physical activity. This suggests a need to avoid reliance on single measures of both physical activity and adiposity in attempting to understand the pathways involved in their linkages, and constraint in interpreting results if only single measures are available. Future research indications include identifying which physical activity-related environmental circumstances have relatively uniform effects on adiposity in everyone, and which should be individually tailored to maximize motivation to continue involvement.