Skin 31P MRS measurements might detect metabolic damage from irradiation, chemotherapy, or ischemia. Although rat and cadaver data have demonstrated this potential (C.D. Cuono, et al., Plast. Reconstr. Surg. 81, 1-11 (1988), H.W. Klein, et al., Ann. Plast. Surg. 20, 547-551 (1988)), few studies of in vivo phosphorus human skin spectra have been published (A. Zemtsov, et al., J. Dermatol. Surg. Oncol. 15, 1207-1211 (1989), A Zemtsov, et al., J. Am. Acad. Dermatol. 30, 959-965 (1994)), and those likely reflect underlying muscle as much as skin. To separate 31P skin and muscle spectra, we have developed a unique two-layer "flotation" phantom for mapping coil sensitivity and an associated semiempirical two-power RF depth-resolved technique. Phantom and method have been applied in a study of 17 normal volunteers to obtain human in vivo 31P skin spectra uncompromised by muscle contamination and to quantitate ratios of major phosphometabolites. Skin results consistently showed low ratios of phosphocreatine (PCr) to adenosine triphosphate (ATP), high levels of phosphomonoester (PME), P(i), and phosphodiester (PDE) relative to PCr, and demonstrated a shift in pH toward greater alkalinity, compared to that with simultaneous muscle results.