Global images of far-ultraviolet dayglow with the Earth Camera on board the Polar spacecraft were used for a comprehensive study of transient, localized decreases of intensities, also known as atmospheric holes. An automated determination for detection of an atmospheric hole was developed and employed in this survey of the characteristics of this phenomenon, which avoided the possibility of aliasing by visual inspection of the images. An extensive investigation of possible instrumental effects included (1) evaluation of random rates for a period in mid-January for which there were no atmospheric holes, thus providing an excellent inflight calibration of the instrument's noise, (2) the effects in the images due to energetic electrons in the outer radiation zone, (3) the nonuniform sensitivities for the pixels of the sensor, or hot spots, and (4) the contributions of longer wavelength radiation from the atmosphere to the camera's responses. For the straightforward selection criteria for the identification of atmospheric holes the instrumental effects were not major contributors to the occurrence frequencies of this phenomenon. The subsequent search for geophysical effects revealed (1) a strong altitude dependence of the frequency of atmospheric holes, (2) a substantial local-time variation of rates which favored locations in the local-morning sectors of the atmosphere relative to those in the evening, (3) increasing sizes of the atmospheric holes as the spacecraft altitude decreases, and (4) large seasonal variations in the hole rates during the period November 1997 through late January 1998 which were remarkably similar to those observed during the same months but 16 years earlier with Dynamics Explorer 1. ¿ 1999 American Geophysical Union