Only certain types of physical activity are osteogenic. Most methods to quantify historical physical activity fail to account for bone specific loading. Physical activity questionnaires (PAQ), activity records, and pedometers are amongst the most common methods of physical activity measurement, however, their limited ability to predict bone mass at clinically relevant sites likely reflects their lack of specificity for the skeletal system. PURPOSE: To compare the ability of several common methods of physical activity assessment with a bone-specific PAQ, to predict bone mass and DXA-derived mechanical characteristics at the hip, spine, whole body and heel in healthy, young adults. METHODS: We recruited 40 healthy men and women (20 M; 20 F) aged 18 to 30 years inclusive. Participants completed the Modified Activity Questionnaire (MAQ) and Bouchard 3-day Activity Record (3DR), and wore a pedometer for 14 days to record number of steps taken during that time. We also administered our purpose-developed bone-specific physical activity questionnaire (BPAQ), which takes into account recent and historical physical activity with an emphasis on load magnitude of recorded activities. Calcaneal broadband ultrasound attenuation (BUA) (QUS-2, Quidel, USA) and bone mineral density (BMD) (XR-36, Norland, USA) were measured from all subjects. Mechanical characteristics of the hip and spine were computed according to the formulae described by Seivanen and colleagues (1996). RESULTS: The BPAQ was a significant predictor of variance in dominant (R2 = 0.78, p < 0.001) and non-dominant (R2 = 0.38, p = 0.001) BUA, index of bone strength (IBS) (R2 = 0.30, p = 0.012), and dominant (R2 = 0.26, p = 0.001) and non-dominant (R2 = 0.08, p = 0.045) FN bone mineral apparent density (BMAD). MAQ, pedometer steps, and energy expenditure based on 3DR were only weakly associated with some bone measures. CONCLUSION: Our BPAQ and algorithm predicted indices of bone strength at skeletal sites used to predict osteoporotic fracture risk better than generic instruments normally employed to determine levels of physical activity. The finding likely reflects the importance of current and historical "bone friendly" physical activity on the skeletal system and emphasises the imperative to utilise bone specific PAQs in skeletal research.