Musculoskeletal markers are frequently used to reconstruct past lifestyles and activity patterns. Yet, the reliability of muscle marker measurements has been called into question because they allegedly fail to correlate with cross-sectional properties and exercise patterns, and are confounded by body size. In this study, the principle of aggregation was used to sum muscle markers over 7 insertion sites (4 humeral, 2 radial, and 1 ulnar) and examine the effects on them of body size, age, sex, and cross-sectional properties. Analyses were made of a sample of 91 (66 males, 25 females) Native British Columbians (3500–1500 years BP) and 18th century Quebec prisoners. Muscle markers were measured using three-point observer rating scales; size was measured by standard methods; age and sex were determined through pelvic, cranial, and dental morphology; and cross-sectional properties were calculated from radiographs. Whereas any single muscle marker component failed to correlate with age, size, sex, or cross-sections, aggregate muscle marker correlated with: age, r = 0.49; size, r = 0.38; sex, r = 0.40; and, cross-sections, r = 0.38; P < 0.001. Older individuals had greater muscle markers, as did larger individuals, males, and those with more robust cross-sections. Based on partial correlations and regression analyses, age was the best overall predictor of aggregate muscle marker. Am J Phys Anthropol, 2003. © 2003 Wiley-Liss, Inc.