There is increasing anecdotal evidence that simple occupational tests of aerobic fitness impose a systematic bias against heavier personnel when predicting fitness for load-carrying tasks. This study tested the hypothesis that simple field tests of aerobic fitness are not good predictors of load-carrying performance and that personnel with greater body mass are more able to perform occupationally relevant load-carrying tasks. Twelve healthy male volunteers ran on a level treadmill at 9.5 km/h for 4 min, with (T18) and without (T0) an external backpack load of 18 kg. During each exercise period, steady-state oxygen uptake (VO(2)) was assessed. On a subsequent occasion (at least 7 days later), 11 of the subjects ran to exhaustion at 9.5 km/h whilst carrying the 18 kg external load (ETT18). There was a strong inverse linear relationship between relative VO(2) and body mass (r = -0.87, P < 0.01) and between VO(2) and lean body mass (r = -0.74, P < 0.01) during the T18 trials. Furthermore, there was a moderately strong relationship between exercise time (ETT18) and body mass (r = 0.69, P < 0.05) and between exercise time and lean body mass (r = 0.71, P < 0.05). There was no relationship between exercise tolerance time and VO(2) (r = 0.12). The results show that fitness tests that determine aerobic power in units relative to body mass (e.g. timed distance run) incur a systematic bias against heavier personnel. Such tests are therefore inappropriate when predicting the ability of personnel to work in occupations that encompass load-carrying tasks.