Plate impact experiments in the 100-250 GPa pressure range were done on a 100 single-crystal MgO preheated before compression to 1850 K. Hot Mo(driver)-MgO targets were impacted with Mo or Ta flyers launched by the Caltech two-stage light-gas gun up to 7.5 km/s. Radiative temperatures and shock velocities were measured with 3%-4% and 1%-2% uncertainty, respectively, by a six-channel pyrometer with 3-ns time resolution, over a 500-900-nm spectral range. MgO shock front reflectivity was determined in additional experiments at 220 and 248 GPa using ≈50/50 high-temperature sapphire beam splitters. Our measurements yield accurate experimental data on the mechanical, optical, and thermodynamic properties of B1 phase MgO from 102 GPa and 3900 K to 248 GPa and 9100 K, a region not sampled by previous studies. Reported Hugoniot data for MgO initially at ambient temperature, T = 298 K, and the results of our current Hugoniot measurements on samples preheated to 1850 K were analyzed using the most general methods of least-squares fitting to constrain the Grüneisen model. This equation of state (EOS) was then used to construct maximum likelihood linear Hugoniots of MgO with initial temperatures from 298 to 2400 K. A parametrization of all EOS values and best-fit coefficients was done over the entire range of relevant particle velocities. Total uncertainties of all the EOS parameters and correlation coefficients for these uncertainties are also given. The predictive capabilities of our updated Mie-Grüneisen EOS were confirmed by (1) the good agreement between our Grüneisen data and five semiempirical γ (V ) models derived from porous shock data only or from combined static and shock data sets, (2) the very good agreement between our 1-bar Grüneisen values and γ (T ) at ambient pressure recalculated from reported experimental data on the adiabatic bulk modulus K s (T ), and (3) the good agreement of the brightness temperatures, corrected for shock reflectivity, with the corresponding values calculated using the current EOS or predicted by other groups via first-principles molecular dynamics simulations. Our experiments showed no evidence of MgO melting up to 250 GPa and 9100 K. The highest shock temperatures exceed the extrapolated melting curve of Zerr and Boehler by >3300 K and the upper limit for the melting boundary predictions of Aguado and Madden by >2600 K and those of Strachan et al. by >2100 K. We show that the potential for superheating in our shock experiments is negligible and therefore out data put a lower limit on the melting curve of B1 phase MgO in P -T space close to the set of consistent independent predictions by Sun et al., Liu et al., and de Koker and Stixrude.