This study examined nurses' assessments of pain in postoperative patients. Subjects were 119 registered nurses and 119 abdominal surgery patients at two community and two university hospitals. Patients completed a visual analogue scale (VAS) when they indicated the presence of pain. Simultaneously, the nurse with the primary responsibility for the patient's care completed a VAS to indicate an assessment of the patient's pain. The results indicated nurses' assessments were modestly, but significantly, correlated with their patients' pain. The majority of the nurses underassessed their patients' pain. Further inspection of the data, however, revealed that nurses' assessments were related to the degree of pain experienced by patients. Nurses underassessed more severe pain and overassessed mild pain. Nurses' accuracy in assessments were related neither to patients' age nor to the length of time since surgery. Implications for clinical practice and suggestions for further research are discussed.