Two experiments were conducted in the semi-arid (400 millimeter annual rainfall) Macquarie region of New South Wales, Australia, at the Trangie Agricultural Research Station (31° 59'S; 147° S7'E), to examine (a) the way botanical parameters can be used to separate grazing and climatic impacts on range vegetation, and (b) how this delineation affects application of range science tenets (site, condition and trend) to different pasture types. Two range sites were studied: Site 1, a light soil type, was dominated by annual grasses, legumes and forbs, whereas perennial grasses (mainly Chloris acieularis and Chloris truncata) dominated the heavy soils of Site 2. Although designed to determine carrying capacity of these sites grazed at 2.5, 3.7 and 4.9 sheep per hectare, botanical data (plant cover by species, and density, diameter and basal area of Chlorisgrasses) collected during a seven year (1967 to 1974) grazing trial provided botanical inputs for Experiment 1 which was designed to: (1) determine range condition in 1967 and 1974 using two methods (Quantitative Climax Method - Method 1, and Christie's Method - Method 2); (2) determine if range condition and sheep production are positively correlated; (3) distinguish the roles of grazing and climate in community change; and (4) determine the value of demographic parameters in defining processes of community change. For Objective 1, the efficacy of Method 1 was hindered by inadequate reference areas, climate-induced variability in plant cover, and uncertainty in classifying perennial grasses as "increasers" or "invaders". Method 2 was superior because slower response of basal cover to weather enabled it to detect grazing-induced changes. Rated by Method 2, grazing treatments improved condition over the seven years, although moderate grazing was most successful, and Site 2 was in better range condition than Site 1. Contrary to the expected pattern for Objective 2, "poor" condition Site 1 produced more wool and higher sheep liveweight than Site 2. This occurred because invading annuals are more palatable, nutritious and productive than native perennial grasses. For Objective 3, simple climatic models showed that quantity and composition of pastures on both sites are determined first by timing and effectiveness of autumn rainfall, and subsequently modified by grazing intensity and plant competition. Two important demographic results were observed for Objective 4. First, population size of important Chloris grasses was regulated by density-dependent mechanisms when average density exceeded 10 plants per meter square for Chloris aciaularis (the dominant grass) and 1.6 for Chloris truncata, a weak perennial. Spatial distribution explains the differential between species. Second, constant death risk in cohorts and long life-span produce mixed-age Chloris acicularis populations which are stable to changing climate and grazing. For Chloris truncata, short life-span and exponential death risk in cohorts renders it unstable to climate and grazing. The management implications of these results are discussed. Community stability was examined further in Experiment 2 by observing regeneration of perennial grasses on both sites at the three stocking rates following five months of grazing at 25 sheep per hectare. The results confirmed Experiment 1: on neither site did Chloris aciaularis density decline significantly under abusive grazing. The reasons for the tenacity of the species are discussed in terms of physiological adaptations. In contrast, Chloris truncata populations. were decimated by abusive grazing, but produced enormous cohorts in the post-perturbation period suggesting that the species exercises a r-strategy for survival. It is concluded that while plant cover and basal area describe functional stability of vegetation to grazing and climate, they cannot detect structural changes in communities. Using plant size, age and density, demography can fulfill this function. Suggested applications of demography to practical situations are given and the implications discussed.