A series of sensitivity measurements on a highly sensitive broadcast receiver over a period of one month indicates that large variations in sensitivity may occur. The importance of these variations in sensitivity ts pointed out with particular reference to intercomparison of receiver measuring equipment and production testing of radio receivers. In this particular study, a high degree of correlation between relative humidity and sensitivity was found; i.e., days of high relative humidity were also days of low sensitivity and vice versa. This effect is probably attributable to increased losses in the radio-frequency transformers during periods of high relative humidity. It is shown that changes in sensitivity may be delayed as much as four days after the corresponding changes in relative humidity. IMPORTANCE OF THE PROBLEM FOR THE intercomparison of radio receiver measuring equipment, it is customary to have a receiver measured under standardized conditions by the various laboratories concerned. The results are then compared, usually on the assumption that characteristics have remained unchanged over the total period of time involved. In a previous paper' by Van Dyck and Dickey it was stated that a 10 per cent variation in characteristics of a particular set was due to changes within the set itself as it was moved from one laboratory to another. However, no data have been furnished to show how this conclusion was reached. It is obviously important to recognize the magnitude of changes in characteristics and the conditions affecting them. It is only in this way that the true magnitude of experimental and systematic errors can be determined and the results of such intercomparison properly interpreted.Some systems of production testing of radio receivers are based on comparison, either direct or indirect, of the product with a master set of approved characteristics. The reliability of such a system is dependent to a large extent on the retention of calibration by the master set.In both of the above instances it is difficult, if not impossible, to attribute changes in sensitivity to any specific cause. In the case of the * Dewey decimal classification: R161. Presented at organization meeting of the Cincinnati Section, September 16, 1929.