The measurement of sound‐power injection into a reverberant room is subject to statistical uncertainties owing to large fluctutations of sound‐pressure level with variations in source frequency and source and/or receiver position. Especially for low‐frequency pure‐tone sources, point‐source theory indicates a requirement of many microphone and source positions if radiated power is to be accurately measured. An investigation has been conducted to determine the effect of finite source size on the power statistics in a reverberant room for pure‐tone excitation. Theoretical results indicate that the standard deviation of low‐frequency pure‐tone finite sources is always less than that predicted by point‐source theory and measurably less when the source dimension approaches one‐half of an acoustic wavelength. A supporting experimental study was conducted utilizing an 8‐in. loudspeaker and a 30‐in. loudspeaker at eleven source positions. The resulting standard deviation of sound‐power output of the smaller speaker with source position is an excellent agreement with both the derived finite‐source theory and existing point‐source theory if the theoretical data is adjusted to account for experimental incomplete spatial averaging. However, the standard deviation of the sound‐power output of the larger speaker with source position is considerably lower than point‐source theory indicates, but in good agreement with the derived finite‐source theory.