Thirty-five college athletes with lower leg pain underwent radiography and radionuclide studies to rule out a stress fracture. Their asymptomatic extremities and 13 pain-free athletes served as controls. Four main patterns were observed: (a) sharply marginated scintigraphic abnormalities and positive radiographs; (b) sharply marginated scintigraphic abnormalities and negative radiographs; (c) ill-defined scintigraphic abnormalities and negative radiographs; and (d) negative radionuclide images and negative radiographs. Since the patients with the first two patterns were otherwise identical medically, the authors feel that this scintigraphic appearance is characteristic of bone stress in the appropriate clinical setting, regardless of the radiographic findings. A schema is proposed to explain the occurrence of positive radionuclide images and negative radiographs in the same patient, using a broad conceptual approach to the problem of bone stress.