We examined changes in muscle buffer capacity (beta m(in) (vitro)), VO2peak and the lactate threshold (LT) after 5 weeks of high- intensity interval training (INT) above the LT or moderate-intensity continuous training (CON) just below the LT. Prior to and immediately after training, 16 female subjects performed a graded exercise test to determine VO2peak and the LT, followed 2 days later by a resting muscle biopsy from the vastus lateralis muscle to determine bmin vitro. Following baseline testing, the subjects were randomly placed into the INT (n = 8) or CON training group (n = 8). Subjects then performed 5 weeks of cycle training (3 days per week), performing either high-intensity INT (6-10 x 2 min at 120-140% LT with 1 min rest) or moderate-intensity CON (80-95% LT) training. Total training volume was matched between the two groups. After the training period, both groups had significant improvements in VO2peak (12-14%; P < 0.05) and the LT (7-10%; P < 0.05), with no significant differences between groups. The INT group, however, had significantly greater improvements in bmin vitro (25%; 123 +/- 5-153 +/- 7 lmol H(+center dot)g muscle dm(-1 center dot)pH(-1); P <= 0.05) than the CON group (2%; 130 +/- 12-133 +/- 7 mu mol H(+center dot)g muscle dm(-1 center dot)pH(-1), P > 0.05). Our results show that when matched for training volume, high- intensity interval training above the LT results in similar improvements in VO2peak and the LT, but greater improvements in bmin vitro than moderate- intensity continuous training below the LT. This suggests that training intensity is an important determinant of changes to beta m(in) (vitro).