The first intensive observation period (IOP) to simultaneously measure diffuse horizontal shortwave irradiance (scattered solar radiation that falls on a horizontal surface) with a wide array of shaded pyranometers suggested that a consensus might be reached that would permit the establishment of a standard with a smaller uncertainty than previously achieved. A second IOP has been held to refine the first IOP measurements using a uniform calibration protocol, offset corrections for all instruments and validation of those corrections, improvements in some of the instruments, and better data acquisition. The venue for both IOPs was the Department of Energy's Atmospheric Radiation Measurement central facility in northern Oklahoma. The 9 days of measurements in October 2003 included a better mixture of clear and overcast conditions than during the first IOP and revealed considerable differences among the instruments' responses for different cloud conditions. Four of the 15 instruments were eliminated as candidates to be included in the standard because of noisy signals, inadequate offset correction, or instability with respect to the majority of the measurements. Eight pyranometers agreed to within ¿2% for clear-sky conditions. Three others have a high bias on clear days relative to these eight, but all 11 agree within ¿2% on overcast days. The differences and causes of this behavior under clear and cloudy skies are examined.