The accurate estimation of field metabolic rates (FMR) in wild animals is a key component of bioenergetic models, and is important for understanding the routine limitations for survival as well as individual responses to disturbances or environmental changes. Several methods have been used to estimate FMR, including accelerometer-derived activity budgets, isotope dilution techniques, and proxies from heart rate. Counting the number of breaths is another method used to assess FMR in cetaceans, which is attractive in its simplicity and the ability to measure respiration frequency from visual cues or data loggers. This method hinges on the assumption that over time a constant tidal volume (VT) and O2exchange fraction (?O2) can be used to predict FMR. To test whether this method of estimating FMR is valid, we measured breath-by-breath tidal volumes and expired O2levels of bottlenose dolphins, and computed the O2consumption rate (V?O2 ) before and after a pre-determined duration of exercise. The measuredV?O2 was compared with three methods to estimate FMR. Each method to estimateV?O2 included variable VT and/or ?O2 Two assumption-based methods overestimatedV?O2 by 216-501%. Once the temporal changes in cardio-respiratory physiology, such as variation in VT and ?O2, were taken into account, pre-exercise restingV?O2 was predicted to within 2%, and post-exerciseV?O2 was overestimated by 12%. Our data show that a better understanding of cardiorespiratory physiology significantly improves the ability to estimate metabolic rate from respiratory frequency, and further emphasizes the importance of eco-physiology for conservation management efforts.