Nuclear magnetic resonance measurements have been made in metallic lithium, sodium, and rubidium, using pulsed radio-frequency power. The experimental data are values of ${T}_{1}$, the spin-lattice relaxation time, and of ${T}_{2}$, the inverse line width or spin-spin relaxation time. Measurements were made at several Larmor frequencies and at temperatures between -65\ifmmode^\circ\else\textdegree\fi{}C and 250\ifmmode^\circ\else\textdegree\fi{}C. Over a considerable temperature range, ${T}_{1}$ is found to be primarily determined by interaction with the conduction electrons. The magnitudes of ${T}_{1}$ agree fairly well with the Korringa theory in all three metals. The lithium data in particular indicate that $\frac{{P}_{f}}{{P}_{a}}\ensuremath{\cong}0.6$. In lithium and sodium a dependence of ${T}_{1}$ on the resonance frequency is observed, which cannot be explained on the basis of the Korringa theory. Information about the atomic self-diffusion process is also obtained. Portions of the ${T}_{1}$ and ${T}_{2}$ data are interpreted using the theory of Bloembergen, Purcell, and Pound. In addition, some of the ${T}_{1}$ data are interpreted in terms of the lattice diffusion theory of Torrey. These analyses yield values for $D$, the coefficient of self-diffusion, of ${{0.24}_{\ensuremath{-}0.10}}^{+0.17}\ifmmode\times\else\texttimes\fi{}\mathrm{exp}(\ensuremath{-}13200\ifmmode\pm\else\textpm\fi{}\frac{400}{\mathrm{RT}})$ ${\mathrm{cm}}^{2}$/sec in lithium and ${{0.20}_{\ensuremath{-}0.15}}^{+0.56}\ifmmode\times\else\texttimes\fi{}\mathrm{exp}(\ensuremath{-}10000\ifmmode\pm\else\textpm\fi{}\frac{600}{\mathrm{RT}})$ ${\mathrm{cm}}^{2}$/sec in sodium. Although there is some ambiguity in the interpretation of the rubidium data, they indicate $D\ensuremath{\cong}0.23\mathrm{exp}(\ensuremath{-}\frac{9400}{\mathrm{RT}})$ ${\mathrm{cm}}^{2}$/sec. Unusual broadenings of the resonance lines are observed at the melting points in all three metals. These broadenings are not presently understood, but some features of the data can be correlated with a mechanism involving magnetic local fields which arise from lattice imperfections.