It is generally believed that more muscle mass in the lower body reflects more muscular strength and power. If true, greater ground reaction forces may be facilitated from increasing muscle mass, which may result in higher maximal running speeds. One method to assess the potential contributions of body composition (i.e. muscle mass and fat mass) to sprint performance is to examine the associations between those variables within each sprinter across time. The present study examined the relationships between longitudinal changes in body composition and sprint performance. Twelve elite female sprinters (9 of whom participated in the Olympics and/or the World Championships) had total body muscle mass, leg muscle mass, and body fat estimated using ultrasound. For each participant, these measurements lasted at least 5 consecutive years and the longest was 10 consecutive years (average 7 ± 1.5 years). The best sprint time in each season was used for evaluating sprint performance. Our unadjusted model indicated that muscle mass, and body fat correlated favourably with sprint performance over time. After adjusting for the error variance attributable to body fat, the relationship muscle mass with sprint performance is largely eliminated. In the first adjusted model, the partial correlations were r = -0.20 for leg muscle mass and r = 0.46 for body fat. In the second, the partial correlations were r = -0.20 for total muscle mass and r = 0.50 for body fat. In conclusion, these findings may suggest that reducing fat mass may have a greater impact on sprint performance than increasing muscle mass.