The drought, frequent lightning strikes, and resultant large fires of 1988 in Yellowstone National Park were considered a several-century event for the area. They presented an unparalleled opportunity to document the effects of large fires on forage production, forage quality, and herbivory for the largest elk (Cervus elaphus) population in a natural area in North America. We documented elk-grassland dynamics on the Blacktail Plateau on Yellowstone's northern elk winter range following the burning of 25% of the study area in 1988. Contrary to predictions of earlier scientists, grazing of the grasslands by elk prior to the fires did not result in warmer, drier, or less productive bunchgrass communities. Soil moisture was equivalent, soil temperatures were cooler, aboveground grass biomass was equivalent in two of three years, and N, macronutrients (Ca, Mg, P, K), and digestibility concentrations were higher on grazed than ungrazed grasslands. Forb biomass, Poa sandbergii biomass, and litter accumulations, however, were less on grazed sites, and more bare ground occurred on grazed (35%) than on ungrazed (24%) grasslands (P < 0.05). Elk herbivory resulted in more documented ecosystem effects for Blacktail Plateau elk than did burning. Two years postfire, burning had increased aboveground biomass of grasslands by 20%. However, digestibility of only one (Festuca idahoensis) of three grass species (F. idahoensis, Pseudoroegneria spicata, and Koeleria macrantha) was enhanced. N, cellulose, and macronutrient concentrations in grasses, and digestibility in P. spicata, K. macrantha were unaffected by the firest. In grazed vs. ungrazed areas F. idahoensis, P. spicata, and K. macrantha averaged a 21% higher N concentration, and a 7% higher digestibility. Aboveground herbaceous biomass was greater in grazed than ungrazed areas in one of three years, and total aboveground N yield was greater in two of three years. Forage biomass did increase postfire and could have benefited elk foraging efficiency. Elk use of burned grasslands increased following the fires (P < 0.05). Elk avoided burned forests during the first three winters postfire (P < 0.05), but elk obtain few of their forages from forest communities on the northern Yellowstone winter range (<10% of feeding observations prefire). The possibility exists that shrub and herbaceous biomass in forest understories will increase after &#8805; 3 yr postfire. Other studies suggest slower recovery, or increases to preburn levels in forests after approximately 6-8 yr. Eventually, the elk might benefit from increased quality and biomass of forages in burned forests. Grazing, by reducing fuels, can alter the extent of burning and create a more patchy fire pattern than occurs on ungrazed areas, thus conserving N that would otherwise be volatilized by burning. We observed no such interaction between elk herbivory and burning, in spite of a sixfold reduction in litter on grazed sites. Burning did not affect N concentrations on either grazed or ungrazed study sites, apparently because accumulations of dead aboveground plant material were still very light on both treatments in comparison to, e.g., tallgrass or mixed-grass prairies in the Great Plains. The relatively minor effects of the fires of 1988 on the grassland study sites were probably a result of the relatively fast front fires, with little residual burning, and the relatively small dead plant accumulations in bunchgrass communities.