The polarizations, $P_{\mathrm{av}}\equiv \frac{〈\overrightarrow{\mathrm{J}}·{\overrightarrow{\sigma }}_{\mu }〉}{J}$ and $P_L\equiv \frac{〈\overrightarrow{\mathrm{J}}·\hat{v}〉}{J}$, of the recoils ${}_{}{}^{12}{}_{}{}^{}\mathrm{C}$(${\mu }^{-}$,$\nu$)${}_{}{}^{12}{}_{}{}^{}\mathrm{B}$(g.s.) were measured simultaneously by selective recoil implantation. Their ratio $R^{\prime }$ is largely immune to the systematics of $P_{\mathrm{av}}$ and $P_L$ and more dependent on the dynamics than either. Our result (normalized to unit ${\overrightarrow{\mathrm{P}}}_{\mu }$), $R(\mathrm{g}.\mathrm{s}.\mathrm{})=-0.506\mathrm{}\left(41\right)\mathrm{}$, yields $\frac{g_P}{g_A}=9.0\mathrm{}\left(1.7\right)\mathrm{}$ (impulse approximation) and $\frac{F_P}{F_A\left({q_m}^2\right)}=-1.03\mathrm{}\left(14\right)\mathrm{}$ (elementary-particle treatment), to be compared to partial conservation of axial-vector current (PCAC) predictions of 7 and - 0.99, respectively. Thus PCAC is quantitatively verified.

• Received 9 March 1981

DOI:https://doi.org/10.1103/PhysRevLett.46.1507