Six successful members of the British Women's Lightweight Rowing Team were assessed before and after two-month (1990) and four-month (1991) periods of weight-reduction controlled by reduced caloric intake, while engaged in their normal physical training. Fat free mass (FFM) was calculated from body weight (BW) by utilising total body potassium measurements. Maximal oxygen intake (VO2max), respiratory anaerobic threshold (Tvent), upper body anaerobic peak power (PP) and mean power (MP) outputs, and knee flexor (KF) and extensor (KE) isokinetic peak torques were among the physiological parameters assessed. No statistical differences were noted between the data obtained prior to the two weight-reduction periods, and both periods resulted in lower BW (p < 0.001) and FFM (p < 0.05); approximately 50% of the lost BW was FFM. At the end of the two-month weight-reduction period Tvent (p < 0.02) and KF (p < 0.02) decreased. In contrast, a similar BW loss during the four-month period was associated with higher VO2max (p < 0.01) and PP (p < 0.05) compared with values prior to weight reduction. Comparisons between the percentage changes pre to post BW loss showed that the longer weight-reduction period was associated with significantly improved VO2max (p < 0.01), Tvent (p < 0.005), PP (p < 0.05) and KF (p < 0.05). We conclude: a) the proportion (50%) of weight lost as FFM in the present elite rowers is higher than the suggested optimal figure of 22%, and b) compared to four months, 6-7% of BW loss over two months may adversely influence fitness-related parameters in international lightweight oarswomen.