The purpose of this paper is to determine whether melt inclusions from three well studied, compositionally zoned ash flow sheets from the Southwest Nevada Volcanic Field can be used to determine the evolution of the parent magma bodies. All samples are from the quenched, glassy pumice fragments that occur in the upper and lower portions of the Topopah Spring Tuff (68.7--78.7% SiO2), the Rainier Mesa Tuff (55.0--76.3% SiO2), and the Ammonia Tanks Tuff (57.5--76.3% SiO2) (all SiO2 values anhydrous). Most of the melt inclusions in this study are in feldspar; a few are in quartz. There are large variations of volatile concentrations in melt inclusions within an individual phenocryst and large variations in melt inclusions in different phenocrysts within a pumice fragment. This finding demonstrates that the volatiles did not reequilibrate after the formation of the ash flow sheet. The volatile contents of most of these melt inclusions represent decompression and loss of some volatiles and therefore do not represent conditions at the time of entrapment. The presence of CO2 in many inclusions demonstrates that the magma evolved a gas phase in its preeruption evolution. The nonvolatile chemical variations of the melt inclusions from the Rainier Mesa and Ammonia Tanks Tuffs are similar. Melt inclusions from the lower-silica pumice fragments show a wide range of chemical variation and monitor the liquid as it evolved. In both tuffs some melt inclusions are significantly more silica-rich than the matrix glass, and indicate mixing with a more evolved magma. Most melt inclusions from the high-silica pumice fragments from each of these tuffs show little chemical variation and are similar to the matrix glass. The fact that no lower-silica melt inclusions occur in the high-silica pumice fragments supports the model that the high-silica magma did not evolve from the low-silica magma. The maximum H2O concentrations of inclusions in each pumice fragment from the Rainier Mesa and Ammonia Tanks Tuffs show a general positive correlation with SiO2 concentrations of the pumice fragment. No such relationship is observed with the Topopah Spring Tuff. ¿ American Geophysical Union 1996