This thesis reports the results of ecological modelling research based on the use of absolute fossil pollen data as a proxy for actual measurements of historic vegetation change. The theme of this project is community organisation while the topic of special interest is community behaviour. The theory of community organisation includes some of the most fascinating and controversial problems in ecology. As defined in this thesis, community organisation is the study of the forces or mechanisms responsible for shaping the structure and behaviour of a community through time. An unresolved and persistent problem in community ecology has been the question of whether or not natural communities are capable of equilibrium behaviour. During the 1970s, key theoretical developments led community ecologists to question this traditional assumption. Today, the non-equilibrium paradigm is considered to be a more likely model for community organisation. The history of these major theoretical changes is reviewed in this thesis with the objective of deriving and testing hypotheses of plant community organisation. The hypothesis testing reported in this thesis is based on a new modelling paradigm. This framework required a flexibility that would permit its application to a wide range of ecological models and a sensitivity to the limitations of fossil pollen accumulation rate (PAR) data. Field data has been obtained from two plant communities: a Mangrove community located on the coastal plains of the Fijian island of Totoya and a lowland mixed beech forest, located within the ring plain of one of New Zealand's active volcanoes (Mt. Ruapehu). This combination of study sites permits the research hypotheses that are advanced in this thesis to be tested in plant communities of varying environmental, temporal, spatial and structural characteristics. The results of this research indicate that the plant communities studied exist in a low to intermediate density region, well below any theoretical region of density-dependency. The density vague behaviour of these communities appears to the result from stochastic domination. In conclusion, these results support the non-equilibrium model of community organisation. The implications of this research for palynologists involved in human impact studies and empirical modelling research are evaluated. Future objectives for theoretical modelling research in this area are suggested.