Cadence is one of the only variables cyclists can adjust to manage their performance and fatigue during an event. Not surprisingly, cadence has received a great deal of attention from the scientific community in an effort to identify the cadence that optimizes power output while minimizing the fatigue that is incurred. The literature appears to present conflicting results with little consensus as regards the optimal pedalling cadence. This is in large part due to the inconsistent definition of the term “optimal” cadence, which has been used to describe energetic cost, muscular stress, and perception of effort. The issue is further confounded by the workload-dependent nature of the “optimal” cadence - that is, at higher power outputs, the optimized cadence is different from that at lower power outputs. Although the optimal cadence is different for energetic, muscular, and perceptual definitions, the curves that describe the effect of changes in cadence on these variables consistently exhibit a J-shaped response. This suggests that there is an underlying principle that is common to each of the definitions. Indeed, it would appear that the response of both the cardio-respiratory system (energetic cost) and the muscular system (muscular stress) is determined by the types of muscle motor units that are recruited during the exercise. Furthermore, although part of the response may be due to the inherent differences in the characteristics between the different motor units, the absolute contraction velocity relative to fibre type optimum may be of greater significance. Even when the power output is increased, the shape of the response curves to changes in cadence remains constant, although the nadir of the curve does shift to the right for increasing power outputs. We propose that the point at which the energetic vs. power and the muscular stress vs. power curves intercept is defined by the cadence at which the perceived effort is minimized (i.e. the preferred cadence). However, cadence fluctuations occur under field conditions that are unrelated to physiological factors and, therefore, the ability to identify an “optimal” cadence is limited to the laboratory environment and specific field conditions.