The purpose of this study was to calculate optimal pedaling rates based upon external work (EW) rate and mechanical work (MW) rate criteria that respectively exclude and include the internal work (IW) rate of the lower limbs. Metabolic and kinematic data were collected as 12 males pedaled an ergometer at rates of 40, 60, 80, and 98 rpm while producing external power outputs of 49, 98, and 146 W. Energy expenditure (EE) was calculated from steady rate oxygen uptake and respiratory exchange ratio values. The IW rate was determined from digitized kinematic data by modeling the thigh, shank, and foot as a three-segment linked system and calculating their changes in potential, translational kinetic, and rotational kinetic energy. The EW rate was calculated from the observed pedaling rate and the ergometer resistance. The MW rate was defined as the sum of the EW rate and IW rate. At each level of external power output, the MW rate increased linearly with pedaling rate increments while the EE displayed a curvilinear relationship. Both gross efficiency (GE = EW rate/EE) and mechanical efficiency (ME = MW rate/EE) responded quadratically to pedaling rate treatments but a repeated measures ANOVA revealed significant differences in their beta 0, beta 1, and beta 2 regression coefficients. Optimal pedaling rates calculated from ME were consistently higher (82 to 101 rpm) than those determined from GE (35 to 57 rpm). The pedaling rates that optimized ME, but not GE, are similar to the rates reported to be biomechanically optimal and preferred by trained cyclists. This study demonstrates that the choice of a work rate criterion can alter the meaning and interpretation of metabolic data.