Body mass prediction from skeletal frame size in elite athletes

Body mass can be estimated from measures of skeletal frame size (stature an d bi-iliac (maximum pelvic) breadth) fairly accurately in modern human popu lations. However, it is not clear whether such a technique will lead to sys tematic biases in body mass estimation when applied to earlier hominins. He re the stature/bi-iliac method is tested, using data available for modern O lympic and Olympic-caliber athletes, with the rationale that these individu als may be more representative of the general physique and degree of physic al conditioning characteristic of earlier populations. The average percent prediction error of body mass among both male and female athletes is less t han 3%, with males slightly underestimated and females slightly overestimat ed. Among males, the ratio of shoulder to hip (biacromial/ bi-iliac) breadt h is correlated wi-th prediction error, while lower limb/trunk length has o nly a weak inconsistent effect. In both sexes, athletes in "weight" events (e.g., shot put, weight-lifting), which emphasize strength, are underestima ted, while those in more endurance-related events (e.g., long distance runn ing) are overestimated. It is likely that the environmental pressures facin g earlier hominins would have favored more generalized physiques adapted fo r a combination of strength, speed, agility, and endurance. The events most closely approximating these requirements in Olympic athletes are the decat hlon, pentathlon, and wrestling, all of which have average percent predicti on errors of body mass of 5% or less. Thus, "morphometric" estimation of bo dy mass from skeletal frame size appears to work reasonably well in both "n ormal" and highly athletic modern humans, increasing confidence that the te chnique will also be applicable to earlier hominins. (C) 2000 Wiley-Liss, I nc.