In an attempt to differentiate between loose and infected total joint replacements, seventy-nine patients were studied with both technetium-99 pyrophosphate and gallium-67 citrate bone scans. Seventy of these patients had positive technetium scans. Nineteen had positive gallium scans and all nineteen had positive cultures at the time of operation, while one patient had a false-negative gallium scan. Ten other patients who had negative technetium-99 scans and a painful total joint were followed without operative intervention. Of these, only two required operation after their technetium-99 scans became positive. Based on these studies, gallium-67 citrate scans combined with technetium-99 pyrophosphate scans are believed to be useful in differentiating between loose and infected prostheses. A negative technetium scan should weigh heavily against surgical exploration in an attempt to find the source of pain.