This study attempted to quantify the difference in heart rate and exercise stage at which blood lactate threshold (T(bla)) occurs using 3 different modes of exercise: running, double poling (DP) on roller skis, and skating (SK) on roller skis. Nine elite collegiate cross-country ski racers (4 men, 5 women) served as test subjects. Testing was conducted on a motorized FitNex treadmill, specially designed for roller skiing. Heart rate was monitored via telemetry with values averaged over the last 30 seconds of each stage. A 40-micro l blood sample was obtained at the fingertip at the end of each 4-minute stage, and 25 micro l was analyzed for whole blood lactate concentration. The T(bla) was determined by the first exercise stage that elicited a concentration over 4.0 mmol.L(-1). The same test protocol was used for all 3 exercise modes. Mean heart rate, in beats per minute (b.min(-1)), at T(bla) was not significantly different (P < or = 0.05) for SK (mean 187 +/- 14 b.min(-1) SD) vs. running (mean 187 +/- 12 b.min(-1) SD); however, heart rate was significantly lower at T(bla) for DP (mean 161 +/- 17 b.min(-1) SD) vs. running and DP vs. SK. The mean exercise protocol stage that induced a blood lactate value which exceeded T(bla) was significantly different (P < or = 0.05) for running (5.22 +/- 1.20 mmol.L(-1) SD) vs. DP (1.89 +/- 0.78 mmol.L(-1) SD), running vs. SK (3.67 +/- 0.71 mmol.L(-1) SD), and SK vs. DP. It was concluded that T(bla) occurs at a lower heart rate and exercise stage during DP as compared with SK or running. Therefore, it stands to reason that the heart rate at T(bla) may vary based on mode of exercise, and when using heart rate to estimate blood lactate concentration, coaches and athletes should be aware that different modes of exercise elicit a different blood lactate concentration at a given heart rate depending on exercise mode used.