Summary form only given. In the University of Nevada Reno Megagauss Experiment small rods with diameters near 1 mm were driven by the Zebra pulser resulting in a current pulse that rises to a peak just over 930 kA in 160 ns. A green filtered photodiode in an experiment with a 1 mm diameter aluminum rod shows a strong signal rising at about 100 ns on this same time scale when the current is near 450 kA1. Lagrangian magnetohydrodynamic (MHD) simulations of the experiment with MACH2 using three different equations of state - two with van der Waals loops and one with Maxwell constructions - and two different resistivity models all show plasma formation in the 96 to 102 ns range. Plasma forms in these simulations when a thin layer of the outer edge expands to 5% or less of solid density, cooling as it expands. The resistivity of this warm vapor layer is a few thousand times that of solid aluminum and is increasing rapidly as the density and temperature fall. The effect in our simulations that reverses the cooling of this vapor and creates plasma is, paradoxically, Ohmic heating since that should decrease with increasing resistivity in a constant electric field. However, the numerical error in the field increases substantially with increasing resistivity introducing erroneous Ohmic heating. In fact, we were able to suppress the plasma formation by decreasing the simulation time step and tightening up on the field diffusion error tolerance. The vapor then continued to expand and cool, a more reasonable behavior for the MHD model. Nevertheless, we believe that our MHD simulations' correct prediction of the time of plasma formation in the experiment is more than a coincidence. Rather, we believe that plasma formation in the experiment is associated with non-MHD electrical breakdown of the vapor expanding from the surface, and that it is the timing of this expansion that is correctly predicted by MHD. In this presentation, we will show these simulation results and will offer a simple physical explanation for the agreement between the timing of the vapor expansion across all three simulations and the experiment.