Recognition of sacroiliac disease has been integral to arthritis diagnostic approaches. However, intra- and inter-observer variation have been recognized as limiting factors in radiologic assessment of the sacroiliac joint. The assumption that such irregularly shaped joints can be radiologically assessed of the sacroiliac joint. The assumption that such irregularly shaped joints can be radiologically assessed has not actually been rigorously evaluated. Direct arthroscopic visualization of the joints was therefore used as a standard against which to assess the validity of routine radiologic views, stereo-radiology, tomography and computed tomography, in the analysis of a group of sacroiliac joints drawn from the Todd Collection (Cleveland Museum of Natural History). The most sensitive techniques for the recognition of erosions or fusion are 15 degree angulation and stereo, but the frequency of false positives makes the technique of questionable clinical utility. While sensitivity for joint space narrowing is high with several techniques (30 degree anterior-posterior or posterior-anterior projections and tomograms), specificity is still poor, with a high rate of false positives. Fulfillment of the New York Criteria for sacroiliac disease is especially insensitive and non-specific. Radiologic techniques therefore have major limitations for the assessment of sacroiliac disease and greater reliance on clinical acumen is required for patient categorization.