Although in vivo P-31 NMR permits unique measurements of the body's stores of high energy phosphate metabolites, their low concentrations place critical limits and demands on spatial resolution, localization techniques, and reproducibility. We argue that optimum reliability and reproducibility of in vivo measurements of metabolite ratios and concentrations is best achieved by preselecting the minimal requirements for spatial localization suited to, and varying with, the particular organ and disease type. Strategies for measuring metabolite ratios and concentrations in the human heart and brain are discussed and demonstrated in the context of the expeditious choice of operating parameters, including localized volume size, localization technique, NMR coils, radiofrequency-field uniformity, patient positioning, and corrections for spectral distortion caused by NMR relaxation, to optimize the return of useful metabolic information in practical clinical research exams of about an hour duration.