The McGill Pain Questionnaire (MPQ) is a recent empirically derived instrument designed to provide quantitative information on major dimensions of pain. Although widely used as an outcome measure in clinical research, little attention has been directed specifically at the instrument itself. The present study addressed this need. Detailed findings were obtained for both single and multiple administrations of the MPQ in two subject samples, each composed of 18 cancer outpatients in pain. These data were compared to similar, but less extensive, data reported by Melzack [6]. MPQ indices proved highly replicable over the two subject samples tested and were remarkably similar to the findings reported by Melzack for a different cancer pain patient sample. No differences were found between the written form of MPQ administration used in the present study and the oral procedure followed by Melzack. The consistency of pain descriptor subclass choice in the present samples was high, ranging from 66% to 80.4% over 4 administrations, and these values compare well with the value of 70.3% reported earlier by Melzack. However, the present subjects selected a larger set of pain descriptor words compared to the word set reported to be characteristic of cancer pain by Dubuisson and Melzack [2]. Both individual and group analyses indicated the MPQ is best used as a measure of immediate pain, and not as a summary measure of past pain over a defined period of time. These findings support the use of the MPQ as a reliable, multi-dimensional measure of immediate pain, and suggest the potential value of future research aimed at refining the psychometric properties of the instrument.