For open end cavities commonly used in gyrotrons, the lack of a well-defined geometrical boundary gives rise to distinctive features as well as physics issues that are basically different from the closed cavity, There is a question concerning the consistency between the time- and frequency-domain interpretations of the resonant frequency, There are also inherent uncertainties in cold testing an open-end cavity by the trans mission method in which one probe is located inside the cavity, Further, properties of the cavity can be significantly altered by a slightly mismatched load. Here, we address these issues by developing a spectral model based on an integrated quantity, the total field energy in the cavity, which properly accounts for effects due to the open boundary. Resonant frequencies determined from the field-energy spectrum are shown to be fully consistent with those of the temporal model, The field-energy spectrum can be constructed from the power of the reflected waves, thus allowing cold test by the reflection method. In the cold test, calibration modification and phase averaging have been employed to minimize the effects of external coupling. Results of the cold test are compared with single-mode code and high-frequency structure simulator (HFSS) calculations. Measured data are found to be in good agreement with the HFSS predictions, while the single-mode code considerably overestimates the Q value at large output taper angles.