with contact discontinuities that separate species of different thermodynamic properties. Such problems as numeri-Existing shock-capturing schemes have difficulties with multispecies computations, creating nonphysical glitches at species inter-cal oscillations and computational inaccuracies around faces. We attribute these glitches to inconsistencies in the equation contact discontinuities have been recognized and discussed of state in cells containing several species. Our remedy is to define in [1, 14, 17, and 25]. To circumvent the problem, Karni mixtures within a grid cell as a collection of species which can [14] proposes to solve the equations in terms of primitive possess distinct temperatures. This formulation requires solving an variables. With this approach, pressure is solved directly additional set of species energy equations. Computational results show that the glitches have been eliminated. For chemically reacting instead of computed from the conservative variables; flow simulations, existing splitting methods often generate nonhence, oscillations in the pressure field can be eliminated. physical waves at stiff reaction fronts. We show that this numerical The drawback is that the resulting scheme is nonconservaphenomenon is due to the mixture model that overestimates the tive; it can have problems in predicting shock speeds correaction temperature. This is avoided by introducing an enforcerectly. Using the high-order conservative scheme of , ment on the reaction temperature that depends on the temperatures of each species. We demonstrate that the method computes detona-Ton et al. show that the contact discontinuity is better tion waves with time steps and grid sizes much larger than would resolved as the order of accuracy is increased, but the be allowed to resolve reaction zones.