The effects upon shrubby and herbaceous vegetation of two low-intensity fires and/or grazing by feral rabbits and wombats were studied in a subalpine environment between 1977 and 1984. Fire reduced shrub cover and biomass to low levels and their recovery was slow. When unburnt, tall shrubs increased, then decreased as a result of extreme frost and drought and subsequently continued to decrease. The recovery of the herb cover and biomass after fire was complete within a year except when delayed by drought andlor disturbance by rabbits. Bare ground was less than 2% in the absence of fire and rabbit grazing but increased to nearly 20% after fire in the presence of rabbits. Rabbits reduced the cover and biomass of 39 species of forb, in some cases to zero. Feral rabbit populations survived and multiplied on burnt areas but decreased on areas left unburnt. The implications of this research for the management of similar environments are discussed.

Measurements of overlapping cover of plant species on permanent transects between 1959 and 1978 in the alpine tract at two sites near Kosciusko and Gungartan generally confirm an earlier hypothesis that the cover at Kosciusko, after 15 years' protection from grazing and burning (i.e, by 1959), had attained a degree of equilibrium. A long-term cycle of first decrease and then increase was measured in the cover of the dominant snowgrasses (Poa) and a few other species. Small colonizing species (notably Neopaxia australasica of the short alpine herbfield) increased and then decreased; temporarily spreading onto the dead snowgrasses. Very little bare soil was exposed at any time at Kosciusko; however, the cover of most species fluctuated in the short term in response to dry summers and late-lying snow. At Gungartan, where grazing continued until 1958, the cover of species underwent similar fluctuations but, whilst not reaching the levels measured at Kosciusko, it showed a strong upward trend. Neopaxia australasica, which had occupied many slopes below snowdrift areas, was largely replaced by snowgrasses and tall hygrophilous sedges. The naturalized sorrel (Rumex acetosella), after spreading onto some of the Neopaxia sites and bare soil, was also largely replaced by taller species. Despite a marked decrease in the amount of bare soil, some of the most severely eroded areas at Gungartan were still unvegetated after 20 years' protection. Observation of nearby ground-water areas confirmed-ds indicated by a limited number of measurements at Gungartan, i.e. that hygrophilous sedges recovering frorn the effects of grazing were blocking small streams and that Sphagnum moss and associated shrubs were beginning to spread. The above-mentioned changes in vegetation were evaluated in terms of catchment protection, water yield and nature conservation and recreation. Near-optimal conditions for these uses were maintained at Kosciusko. At Gungartan recovery to the conditions of the Kosciusko vegetation was restricted to the least-disturbed sites.

This paper examines vegetation changes in the absence of fires at Kosciusko during the 14 years following the general withdrawal of livestock grazing at Kosciusko in 1958 on two pairs of plots on subalpine grassland at Dainers Gap (elev. 1700 m) and in the Hotel Kosciusko Water Reserve (elev. 1660 m). One plot in each pair was experimentally grazed. and the other plot was ungrazed. The vegetation of the Kosciusko Reserve was in a relatively natural condition, characterized by continuous forb-rich snowgrass associated with regenerating snowgum woodland. At Dainers Gap the snowgrass was less continuous, and there was no snowgum regeneration. In the Kosciusko Reserve the main effect of grazing was to reduce the cover of the large palatable forbs. This change appeared to be largely reversible within 6 years after the 8 years of grazing treatment was discontinued. At Dainers Gap the tussocks spread laterally in the absence of fire. on both the grazed and ungrazed plots. Under grazing (5 sheepjha/season), small grazing-adapted herbs maintained themselves on the intertussock spaces, but in the absence of grazing they were largely replaced by taller-growing species, many of which appeared only after this treatment had ceased. Shrubs, although uncommon, increased rapidly in size, especially on the grazed plot where competition from herbaceous species was reduced. On the grazed plot the area of bare soil did not decrease appreciably but on the ungrazed plot it decreased on all but the largest bare patches where the lateral spread of tussocks and shrubs was not sufficient to cover them. Under grazing the standing crop on the intertussock spaces at Dainers Gap remained at low levels of about 0.3 kg/m2. but when not grazed for 10 years it increased to about 1 kg/m2, with a resultant increase in surface soil organic matter and decrease in bulk density. During the one season when sheep were weighed at monthly intervals. they gained weight at both sites at the start of the grazing season until the supply of the relatively palatable and nutritious herbs and grass inflorescences was exhausted, then lost weight as mature snowgrass leaves became the main part of their diet. Seasonal weight losses of sheep increased over the 15 years of the experiment as most of the major forbs were eliminated by grazing from the Kosciusko Water Reserve plot, and the area of minor herbs was reduced at Dainers Gap by the encroachment of snowgrass tussocks onto the intertussock spaces. At Dainers Gap in the absence of grazing and fires; water catchment and nature conservation values improved but. without fire, unpalatable snowgrasses and shrubs increased and the value of the area for grazing was reduced. With grazing, the previously unsatisfactory conditions for water catchment and nature conservation were maintained. At the Kosciusko Water Reserve nature conservation and grazing values were reduced by grazing. Introduction Transhumant grazing increasingly became a major issue in the use and management of the Snowy Mountains, following completion of the Hume Dam in the early 1920s and the report by Byles (1932) documenting the deterioration of parts of the Hume

The effects of two low-intensity fires andlor grazing by rabbits and wombats upon trees of Eucalyptus pauciflora and E. stellulata. in a subalpine environment were studied between 1977 and 1984. The experimental plot design permitted comparisons to be made between unburnt and burnt areas, each of which was subjected to four grazing regimes. Tall shrubs, mainly Bossiaea foliosa, which were present under the trees during the first fire but not the second increased fire intensity and flame height and hence damage to trees. The first fire in March 1978 did not completely kill any trees, but killed all the stems 2 m or more in height on 34% of trees and stimulated basal regrowth on a further 17% without killing all the stems. The first fire also reduced the basal area of live stems by about 17%. Small saplings less than 2 m high and seedlings were generally top-killed but resprouted from the base. In the second fire only lignotub- erous regrowth and seedlings were affected as they were in the first fire; the older trees showed no sign of damage. Rabbits killed a significant number of trees by first ring-barking the stems and then grazing the regrowth. Both the fires facilitated the establishment of tree seedlings in the frost-hollow by exposing bare soil, but few of these seedlings survived to the end of the 7-year study period. Frost was probably the main cause of tree death, but drought and competition from other species may also have contributed to losses.