Festuca novae-zelandiae (Hack.) Cockayne is a long-lived, caesptiose grass indigenous to New Zealand. It has expanded dramatically in range in 1000 years of human settlement. This study was conducted to determine how generalist life-history attributes and 'adaptive' variation have enabled this species to adjust to environmental change. A range of aspects of the biology of F. novae-zelandiae were investigated. Reproduction in 155 individuals over four seasons was monitored. Only 17.4% of individuals flowered every year and 36.1 % did not flower at all. However, the reproductive output of the population was still relatively consistent between years. Reproduction of individuals and the pre-dispersal fate of seeds was compared among eight sites. Differences existed among sites for most attributes measured. Site factors appeared to mask any effects of tussock density on culm production. Total seed predation was positively related to density and negatively to altitude. A flightless fly, Diplotoxa moorei (Diptera, Chloropidae) was found to be the most common identifiable seed predator, accounting for up to 8% of seeds produced in some populations. Seed germination was compared for eight populations in laboratory trials and most seeds germinated easily. Mean final % germination across all populations was 86%. Seedlings were monitored in the field for 30 months. They commonly established on mat-vegetation, were very slow-growing (mean of 0.6 leaves yr-1) and had a half-life of 12 months. Tillers in 26 tussocks were tagged and monitored for 30 months. Tillering was concentrated in spring, tillers lived on average 15.2 months and the majority (64%) did not produce daughter tillers. Tillering rate was not affected by either position within a tussock or size of the tussock. A stage-based approach was used to investigate population structure. Stages were defined using discriminant analysis. The discriminant function was then used to assign stages to 255 tussocks which had been measured at the beginning and end of a two year period. Transition probabilities were calculated between stages. Transitions were distinctly non-linear and tussock size and condition could vary widely in the space of two years in response to environmental fluctuations. The genetic identity of mapped tussocks was investigated using isozyme electrophoresis and it was found that clonal fragmentation did not contribute significantly to the maintenance of population densities. Patterns of variation at different taxonomic levels were compared using three taxa: F. novae-zelandiae s.s., a distinct high altitude form of F. novae-zelandiae and a closely related species, F. matthewsii. Vegetation composition was investigated using ordination and classification techniques. Environmental gradients in altitude, temperature and rainfall were important in explaining observed patterns. The vegetation composition at "high altitude" F. novae-zelandiae sites was more distinct from that of F. novae-zelandiae s.s. sites than the latter was from F. matthewsii sites. Morphological variation in the same taxa was investigated using Principal Components Analysis. The same environmental factors were important to observed patterns of variation. Populations within F. novae-zelandiae possessed some genetically-determined differentiation that related to habitat. A narrow zone of intermediacy was found between F. novae-zelandiae s.s. and F. matthewsii. Plants from populations of the three taxa were grown in cultivation. Culms of F. matthewsii emerged four weeks prior to F. novae-zelandiae s.s. "High altitude" F. novae-zelandiae was more similar to F. matthewsii in flowering phenology. Levels of biochemical variation in populations of all three taxa were investigated using isozyme electrophoresis. All populations were characterised by high within-population variation and relatively low between-population variation. The differences between the three taxa were small. Variation among populations related to environmental factors but not to the proximity of populations. The adaptiveness of populations to their own environment was tested using reciprocal transplants. No 'home-site' advantage was found. All populations proved to be highly plastic in growth responses. The findings of this study are discussed in terms of generalist versus specialist strategies in long-lived, polyploid, perennial grasses. I conclude that in environments characterised by unpredictable, short-term fluctuations, long-lived species will show adaptation to large-scale, long-term environmental trends only, and adopt a generalist strategy in the face of short-term fluctuations.