EFFECTS OF GRAZING BY UNGULATES ON UPLAND BUNCHGRASS COMMUNITIES OF THE NORTHERN WINTER RANGE OF YELLOWSTONE-NATIONAL-PARK

Herbivory by native ungulates, primarily elk (Cervus elaphus), was stu died on xeric, bunchgrass-dominated slopes on the Northern Yellowstone ungulate winter range of Yellowstone National Park. Plant cover, stan ding green biomass, and nutrient concentrations were compared between grassland sites protected from grazing for 24-27 years and adjacent pl ots grazed mostly during winter by elk, and to some extent by bison (B ison bison). Accumulated litter and standing dead vegetation were four times more abundant on the ungrazed plots. However, few consistent di fferences were observed in grass or forb biomass from elk herbivory. T here was no overall effect of grazing treatment on forbs or other nong rasses (P < 0.05). Total grass biomass wits less on grazed than ungraz ed sites in 1986, after a dry, warm spring, but there was no differenc e in 1987 (P < 0.05). Biomass of junegrass (Koeleria cristata) and thi ck-spiked wheatgrass (Agropyron dasystachyum) was greater on grazed th an ungrazed sites. Three nongrasses (Artemisia frigida, Phlox hoodii, Antennaria, microphylla) were more abundant on some ungrazed sites alt hough the effect on grazing treatment was not significant (P > 0.05). The numbers of grasses (n = 4.5), forb (n = 17.5) and shrub (n = 3.5) species did not differ between grazed and ungrazed plots (P < 0.05). V egetative culms of bluebunch wheatgrass (Pseudoroegneria spicata), Ida ho fescue (Festuca idahoensis), bluegrasses (Pea spp.), needle-and-thr ead grass (Stipa cornara),junegrass, and thick-spiked wheatgrass were shorter on winter grazed plots (P < 0.05), but grazing did not affect the numbers of vegetative culms or the height or numbers of reproducti ve culms of grasses (P > 0.05), with the single exception mat more veg etative culms of junegrass occurred on grazed sites. Protein content a veraged 16% higher in bluebunch wheatgrass, 36% higher in Idaho fescue , and 10% higher in junegrass (P < 0.05) on grazed sites. Bare ground averaged 38% more on grazed sites than on ungrazed sites (P < 0.05), b ut pebble cover was 50% less on grazed sites (probably because of hoof compaction); as a result all bare surfaces (bare ground plus pebble c over) averaged only 18% greater on grazed sites. Dead clumps of the gr azing sensitive bluebunch wheatgrass and Idaho fescue can be expected on an overgrazed range, but dead bunchgrass clumps did not vary betwee n grazed and ungrazed sites (P < 0.05). The effects of herbivory by na tive ungulates was observed in swale sites that had been disturbed by agricultural activities until the 1930s. Three grasses (Stipa viridula , Poa pratensis, P. compressa) possessed greater cover on grazed swale s, but one forb, field pennycress (Thlaspi arvensi, was less abundant on grazed sites. Herbivory by native ungulates greatly stimulated gras s but not forb cover on these productive swale sites (grass cover was double that of ungrazed sites) (P < 0.05), and grazing mostly by elk c ontributed to the continued dominance of these sites by the exotic, gr azing resistant Poa pratensis even 54 years following cessation of agr icultural activities.