A modified form of the Community Atmosphere Model, ver. 3 (CAM3) developed at the National Center for Atmospheric Research (NCAR) is validated in its warm-season mean precipitation diurnal cycle for two regions of the United States. For all grid boxes of each region, simulated and observed precipitation records over four four-month periods are separated into discrete precipitation events. These events are binned into mutually-exclusive categories, and the diurnal harmonic for each category is estimated. In this way, the model is validated over the spectrum of precipitation episodes, and biases in the overall seasonal-mean diurnal cycle can be attributed to particular kinds of events. The results of the study indicate that the model's total seasonal precipitation is overwhelmingly weighted in the extremely long events and that these events contain the source of any biases in the seasonal mean.