The validation of the radiometric calibration of virtually all infrared radiometers has previously been carried out under carefully selected, generally spatially uniform conditions, with the assumption that the radiometric accuracy of the data may be dependent on scene brightness temperature, but is independent of other scene unique conditions, such as scene spatial uniformity. The availability of AIRS and CrIS observations from polar orbits with the identical ascending node presents an opportunity to evaluate the validity of this assumption. For each day between May 2012 and January 2014 we collected 22,000 Random Nadir Spectra (RNS). We then analyzed the time series of the daily differences between AIRS and CrIS Probability Density Function in the 900 cm-1 atmospheric window channel. Under polar conditions the PDF differences between AIRS and CrIS are typically less than 50 mK for the 10%tile, the mean and the 90%tiles values of the PDF. Under area representative global conditions day and night CrIS is about 0.2K colder than AIRS at the 10%tile and mean values. These differences are well within the limits of the instrument requirements specification. However, the difference between AIRS and CrIS have a complicated zonal distribution, particular for the tropical zone. For day tropical land CrIS is 0.3 K warmer in the mean, 1K warmer in the 10%tile value (cold tails of the PDF) than AIRS. The reasons for these differences are still under investigation. A number of modifications to the CrIS radiometric calibration algorithms have been proposed.