Studies in plant population ecology are concerned with explaining the patterns in spatial and temporal variation of abundance. These patterns are a consequence of the demographic patterns of birth, death, emigration and immigration. These studies are easier on short lived species but generalisations to other species, in particular long lived species, may not be valid. Xanthorrhoea johnsonii is a long lived perennial plant endemic to Australia. Seed set in this species is highly variable with individual plants producing anything from a few seeds to over 9 000 seeds in one flowering season. The aim of this study is to examine the spatial and temporal variation in seed set and to try and determine whether or not the observed variation could be correlated with levels of seed predation or various plant traits such as age, size of floral display, synchrony of flowering with other individuals in the population, or distance to nearest flowering neighbour. Spatial autocorrelation analysis from four study sites indicated that seed set was not random but patchy. Correlations of seed set with various plant traits showed some significant trends but were not consistent across all sites. It seems that, while individual sites show very different trends, the sites that had a higher density of flowering plants had higher seed set per plant. This has been suggested by others as 'masting seed set' and is one way of avoiding high levels of seed predation. Allozyme electrophoresis was used to examine genetic variability and while the numbers of readable, variable loci were limited (5), analysis of the data suggests that there is some evidence of limited gene flow in the populations examined. Progeny arrays were also examined from ten plants from four sites to determine the level of outcrossing in the population. Plants at all sites consistently showed high levels of outcrossing which is consistent with the protandrous nature of the flowering pattern. Significant deviations from total outcrossing was observed at a site that had extremely low flowering density. At this site the plants that flowered early in the flowering season were separated both temporally and spatially from nearest flowering neighbours and showed significant levels of true selfing. This situation is analogous to small fragmented populations. Germination response to temperature indicated that high germination rates could be expected under field conditions during summer seasons with good rainfall. Germination response showed no correlation with the maternal outcrossing rates but was correlated with seed weight. These results lend support to the argument that the maternal environment of the plant (both environmental and genetic) influences the germination response in seeds.