緒論：探討以跑步成績所推估的臨界速度 (critical velocity, CV) 分別與換氣閾值速度(velocity at ventilatory threshold, vVT)、呼吸代償點速度 (velocity at respiratory compensationpoint, vRCP) 的相關與差異情形，以及提升跑步成績推估的CV 在跑步競賽的實際應用。方法：以10 名經常進行跑步訓練的健康男性 (年齡22.70 ± 7.50 years、身高170.70 ± 7.06cm、體重60.20 ± 6.34 kg)，進行跑步機漸增負荷測驗，獲得VT 與RCP，依平衡次序原則進行800 m 與5000 m 跑步測驗之後，利用Riegel 預測成績模式，推估受試者3000 m、10000 m、半程馬拉松與全程馬拉松的成績。利用800 m、3000 m、5000 m、10000 m、半程馬拉松與全程馬拉松六個成績，結合三參數非線性數學模式，計算出CV1；利用800m、3000 m、5000 m、10000 m 四個成績，計算出CV2。資料分析以皮爾森積差相關分析探討vVT 與CV1、vRCP 與CV2 的相關情形；以相依樣本t 檢定探討vVT 與CV1、vRCP與CV2 的差異情形。結果：vVT (3.07 ± 0.41 m/s) 與CV1 (3.18 ± 0.52 m/s) 達顯著正相關(r = .77, p < .05) 且沒有顯著差異 (t = -1.12, p >.05) ; vRCP (3.69 ± 0.36 m/s) 與CV2 (3.54± 0.46 m/s) 達顯著正相關 (r = .77, p < .05) 且沒有顯著差異 (t = 1.64, p > .05)。結論：CV1可以有效評量vVT，CV2 可以有效評量vRCP。CV1 與CV2 可以有效作為長距離跑步比賽配速範圍的參考。

Introduction: The purpose of this study was to examine the correlation and difference between critical velocity (CV) predicted from running performance, the velocity at the ventilatory threshold (vVT), and the velocity at respiratory compensation point (vRCP) derived from maximal incremental tests. Method: Ten males with running regularly were recruited as the participants in the study (age: 22.70 ± 7.50 years, height: 170.70 ± 7.06 cm, weight: 60.20 ± 6.34 kg). In order to obtain vVT and vRCP, each subject performed maximal incremental tests on the treadmill. Subjects also completed 800 m and 5000 m running tests at their maximal effort based on counter-balanced design. The performance of various running distances (800 m, 3000 m, 5000 m, 10000 m, half-marathon, and full-marathon) was applied to estimate CV1 by a 3-parameter nonlinear model; besides, except for half-marathon and full-marathon, the other running performance was also applied to estimate CV2 in the same model. Pearson product-moment correlation was used to analyze the correlation of CV1 and vVT, CV2 and vRCP ; The dependent t-test was used to analyze the difference of CV1 and vVT, CV2 and vRCP. Results: vVT (3.07 ± 0.41 m/s) and CV1 (3.18 ± 0.52 m/s) were significantly positively correlated (r = .77, p < .05). The two parameters were not significantly different (t = -1.12, p > .05). In addition, the vRCP (3.69 ± 0.36m/s) and CV2 (3.54 ± 0.46 m/s) were significantly positive correlated (r = .77, p < .05). The two parameters were not significantly different (t = 1.64, p > .05). Conclusion: CV1 could assess vVT effectively; CV2 could assess vRCP effectively; CV1 and CV2 could be utilized effectively as pacing during long-distance running race.