This study was conducted to assess whether respiratory frequency can be used as a valid parameter for estimating ventilatory threshold and for examining differences in exercise modes such as a cycle ergometer and a treadmill. 24 men and 12 women performed an incremental exercise test to exhaustion on a cycle ergometer and on a treadmill. Oxygen uptake, carbon dioxide output, pulmonary ventilation, ventilatory frequency, and heart rate were measured continuously every 30 sec. during the test. Three different and independent reviewers detected the ventilatory threshold point and break point of respiratory rate, which were then compared. Analysis indicated that (1) ventilatory threshold was well correlated with break point of respiratory rate for both cycle ( r = .88, p<.001) and treadmill exercise ( r = .96, p<.001). However, on the average, ventilatory threshold was only 71% (cycle) or 88% (treadmill) of break point of respiratory rare. (2) The regression equation for treadmill exercise was more accurate than that for cycling, but the detected data samples were smaller. The break point of respiratory rate was more easily detected for the cycle ergometer test (33 of 36 subjects) than for the treadmill test (only 15 of 36). The cycle ergometer test identified the break point of respiratory rate more easily than did the treadmill test. (3) There was an association between physical fitness and whether the break point of respiratory rate was detectable, and the more fit the subject (above average), the more likely the break point was to be undetected. Our study demonstrates that the break point of respiratory rate is closely associated with ventilatory threshold and that the cycle ergometer test is more conducive than the treadmill test to the detectability of break point of respiratory rate.