New Zealand's South Island tussock grasslands have been highly modified by human activities, including burning, grazing and introductions of exotic plants for pastoralism. Studies suggest that tussock grasslands are degraded, in that native species have declined, and exotic species have increased in both diversity and abundance. These trends are primarily thought to be related to the impacts of grazing and subsequent grazing removal. Few studies have assessed long-term changes that have occurred in tussock grasslands, and those that have are generally limited to one particular location. This thesis aimed to investigate temporal changes in community structure in tussock grasslands, and relate these changes to environmental variables and land tenure. Data were used from 90 permanently-marked vegetation transects, which were set up on 19 geographically widespread properties in areas of tussock grassland across Canterbury and Otago in the South Island of New Zealand. The transects were on land in both conservation and pastoral tenure. Each transect was 100 m, and consisted of 50 0.25 m² quadrats. The transects were measured between 1982 and 1986 (first measurement), were re-measured between 1993 and 1999 (second measurement) and again between 2005 and 2006 (third measurement). A total of 347 vascular species were observed over the 90 transects and three measurement times. Species richness declined between the first and second measurements (first time interval), and increased between the second and third measurements (second time interval), at both the small (quadrat) and large (transect) scales. Both native and exotic species declined in mean quadrat species richness during the first time interval, and then increased during the second time interval. Changes in mean quadrat species richness were similar on transects in both conservation and pastoral tenure. Multivariate analysis of species' occurrences in quadrats identified a long gradient in species composition for these 90 transects. Four key plant communities were identifed along this gradient and differed in their mean elevation: (1) Highly-modified pastoral community, (2) Short-tussock grassland community, (3) Tall-tussock grassland community, (4) Alpine mat-forming species community. A detailed investigation into temporal changes that occurred on 53 transects that occurred in short- and tall-tussock grassland communities showed that changes in species composition were not consistent over time. Transects on different properties changed in species composition by different amounts. Specifically, in ordination space, transects on two properties changed in composition significantly more than transects on one other property. The property that a transect was on also affected the way that it changed in composition, i.e. native species were more likely to have increased on transects on some properties. Transects in conservation tenure did not change in species richness or composition differently from those in pastoral tenure. Considering that many native plants in tussock grasslands are relatively slow-growing, and that these areas have been grazed and burned for more than a century, we may expect it to be some time before we can detect differences in vegetation dynamics on conservation land from that on pastoral land. The changes in the community structure of these tussock grasslands were related to a combination of environmental factors, such as soil chemistry, climate, and management factors. This study has allowed greater understanding of vegetation change in tussock grasslands, and demonstrates the importance of long-term ecological monitoring in making reliable and accurate predictions about landscape-scale changes in tussock grassland community structure.