All physical measuring devices have finite response times determined by the characteristics of their optical, electrical, and thermal components. In a system which is not time-critical, one merely waits for the system to stabilize, and then reads the measurement. In a system which is in motion, such as the CERES scanners, the spatial location of any measurement is dependent upon the response time of the detector. In order for the remotely sensed data to be accurately geolocated, the systematic delays between the time-of-observation of the scene and the time-of-measurement must be determined. The first CERES instrument was launched aboard the Tropical Rainfall Measuring Mission (TRMM) platform from Japans Tanegashima Island Space Center on 28 November 1997. The next two CERES instruments are scheduled for dual launch on the EOS/AM-1 platform in late 1998. In this paper we specifically address the delay between the time of observation of a scene, and its actual measurement as determined from Point Response Function Source data taken in the CERES calibration chamber at TRW in Redondo Beach; and we compare the theoretically generated Point Spread Functions to these measurements. The agreement between the measured and theoretical contours in excellent, in the total and shortwave channels. For all 3 instruments, the calculated time-delay of 0.023ms to the centroid,and the measured peak delay of 0.022ms are very consistent. The result of on-orbit coastline detection algorithms are currently being analyzed in comparison with the PSF delays used in the geolocation algorithms in order to further validate the proper geolocation of the measured data.