Ten commonly used physical activity questionnaires were evaluated for reliability and validity in 78 men and women aged 20-59, with varying physical activity habits. One month reliability was found to be high for all questionnaires except those pertaining only to the last week or month. Longer term test-retest reliability tended to be lower. Validity was studied in relation to treadmill exercise performance, vital capacity, body fatness, the average of 14 4-wk physical activity histories and the average of 14 2-d accelerometer readings. No questionnaire measure was correlated with the accelerometer reading, and correlations with vital capacity were generally low. Only the Minnesota Leisure Time Physical Activity Questionnaire household chores measure was correlated with habitual performance of household chores. Most questionnaires, even very simple ones, were related to performance of heavy intensity physical activity and treadmill performance; these same questionnaires tended to be related to percent body fat. Fewer questionnaires related to performance of light or moderate activity. Occupational activity was unrelated to any of the validation measures. It is concluded that there are multiple, nonoverlapping dimensions of physical activity, reflected in multiple nonoverlapping validation realms. More important than the length or attention to detail of a questionnaire seems to be the logic of its questions. Important areas of physical activity that should be addressed in future questionnaires include sleep, light, moderate and heavy intensity leisure activities, household chores, and occupational activity. Recent versus habitual activity should also be considered.