Forty-nine patients with chronic wrist pain were examined by radiocarpal arthrography, followed by radiocarpal arthroscopy. One patient had bilateral involvement. None of these patients showed deformities or lesions on standard and dynamic radiographs. Not all patients who underwent arthrography underwent radiocarpal arthroscopy. Sensitivity of arthrography was 52%, specificity 50%, positive predictive value 92%, and negative predictive value 8% compared with arthroscopy. In this patient group, a very low negative predictive value was found: in 92% of the negative arthrographic examinations an arthroscopic lesion was found. Arthrography of the radiocarpal joint is a good examination for confirming a clinical suspected lesion, but is of limited value when negative.