Computed tomography (CT) is an established diagnostic modality for evaluation of canine elbow dysplasia. However, diagnostic sensitivity for elbow abnormalities may be affected by variations in CT image quality. Currently there are no data-based recommendations for elbow CT image display parameters. The purpose of this study was to test the effect of CT display parameters on observers' diagnostic certainty for CT abnormalities reportedly associated with canine elbow dysplasia. Fifty dysplastic elbows and ten clinically normal elbows were selected for imaging. Computed tomographic image data from selected elbows were filmed in transverse, sagittal and dorsal planes, with each plane filmed in window widths of 1500, 2500, and 3500 HU. Two veterinary radiologists independently evaluated each set of images for the presence or absence of hypoattenuating MCP subchondral defects, in situ MCP fissures, discrete MCP fragments, irregularity of the radial incisure of the ulna, subchondral sclerosis of the trochlea humeri, osteochondrosis or kissing lesions, and joint incongruity. Level of diagnostic certainty for each CT abnormality was recorded for each observer using a visual analog scale system. The effect of elbow status, plane, and window on the degree of observer certainty was tested. Overall, observers demonstrated higher diagnostic certainty for normal elbows than dysplastic elbows. Observer certainty for the presence of altered subchondral density (presumed osteomalacia or sclerosis) was primarily affected by window width, whereas certainty for structural defects of the MCP, radial incisure, trochlea humeri, and joint incongruity was primarily affected by image plane. / Master of Science