Serial reflectivity measurements from paired instruments are examined during two field campaigns in order to examine the precision of the measurements. The instruments studied are two collocated Joss-Waldvogel disdrometers (JWD) at Wallops Island, VA and two collocated profilers deployed at Ji-Parana, Brazil during Tropical Rainfall Measuring Mission (TRMM) Large-Scale Biosphere-Atmosphere Experiment. Differencing the measured reflectivity from the instrument pairs eliminated most of the temporal and large-scale precipitation variability, reducing the error fluctuations to those of the instrument precision plus fluctuations due to precipitation variability over the small differences in sample volume and distances between the instruments. For both pairs of calibrated instruments we found that the observed time-series of one-minute dBZ differences were not autocorrelated and exhibited a Gaussian-like distribution. Consequently, the difference time-series could be meaningfully characterized by their standard statistics, including the rms difference or standard deviation, and the standard error about the mean. While the disdrometer pair exhibited an rms difference of 2.1 dBZ, a standard error about the mean of less than 0.1 dBZ for the 12-hour rain event was achieved. The profiler pair exhibited an rms difference of 0.4 dBZ, with a standard error of only 0.05 dBZ for the 90-minute stratiform rain event. Since it is currently difficult to routinely calibrate radars in an absolute sense to better than 1--3 dBZ, the precisions of a few tenths of a dBZ obtained here suggest the potential for substantially improving these calibrations, and open the door to examination of subtle sampling and stability effects.