The aim of the study was to compare the effects of ischaemic and non-ischaemic training on aerobic performance. In 10 subjects, peak oxygen uptake (peak VO2) and time to fatigue (TTF) for one-legged exercise were measured before and after 4 weeks (4 times week-1) of one-legged training. Each training session started with one leg training for 45 min with 20% blood-flow reduction induced by local application of a supra-atmospheric external pressure of 50 mmHg (ischaemic leg; I-leg). We have previously shown that this decreases leg blood flow by about 20%. The contralateral leg (non-restricted-flow leg; N-leg), serving as a control, then trained with an identical power-output profile for 45 min but without flow restriction. In the I-leg the average training-induced increments in TTF and peak VO2 were 27 and 24%, respectively. In the N-trained leg TTF and peak-VO2 increased 10 and 14%, respectively. Both increments were significantly greater (P < 0.05) in the I-trained leg. Moreover, the performance increase in the I-trained leg was exaggerated (P < 0.05) in the ischaemic test condition, i.e. there was a specificity in the training response. In conclusion, ischaemia acts as an additive stimulus to training leading to an exaggerated increase in endurance and peak-VO2 compared to identical training without blood-flow restriction. The main explanation is probably an enhanced local adaptation in the I-trained leg.