We solve the equations of radiation hydrodynamics in the two-temperature fluid approximation on an adaptive grid. The temperature structure depends upon the electron-ion energy exchange length, $l_{ei}$, and the electron conduction length, $l_{ec}$. Three types of radiating shock structure are observed: subcritical, where preheating of the unshocked gas is negligible; electron supercritical, where radiation preheating raises the temperature of the unshocked electron fluid to be equal to the final electron temperature; supercritical, where preheating and electron-ion energy exchange raise the preshock $T_{e,i}$ to their final post shock values. No supercritical shock develops when $l_{ei}$ is larger than the photospheric depth of the shocked gas because a negligible amount of the ion energy is transferred to the electrons and the shock is weakly radiating. Electron conduction smooths the $T_e$ profile on a length scale $l_{ec}$, reducing the radiation flux.

We calculate the quasi-stationary structure of a radiating shock wave propagating through a spherically symmetric shell of cold gas by solving the time-dependent equations of radiation hydrodynamics on an adaptive grid. We show that this code successfully resolves the shock wave in both the subcritical and supercritical cases and, for the first time, we have reproduced all the expected features -- including the optically thin temperature spike at a supercritical shock front -- without invoking analytic jump conditions at the discontinuity. We solve the full moment equations for the radiation flux and energy density, but the shock wave structure can also be reproduced if the radiation flux is assumed to be proportional to the gradient of the energy density (the diffusion approximation), as long as the radiation energy density is determined by the appropriate radiative transfer moment equation. We find that Zel'dovich and Raizer's analytic solution for the shock wave structure accurately describes a subcritical shock but it underestimates the gas temperature, pressure, and the radiation flux in the gas ahead of a supercritical shock. We argue that this discrepancy is a consequence of neglecting terms which are second order in the minimum shock compression ratio [$\eta_1 = (\gamma-1)/(\gamma+1)$, where $\gamma$ is the adiabatic index] and the inaccurate treatment of radiative transfer near the discontinuity. In addition, we verify that the maximum temperature of the gas immediately behind the shock is given by $T_{+} = 4 T_1/(\gamma+1)$, where $T_1$ is the gas temperature far behind the shock.