The realism of models in plant demography

This paper discusses the question of why we need to validate the realism of models used in plant demography. A model is a simplified representation of an existing or supposed reality which assists in making calculations and predictions and in formulating theories. In other words, a model mimics, as closely as possible, the relevant features of a system. The term demography is employed in various ways by various authors to cover many uses depending on the context. Essentially, this term originates from the human sciences. Demographers are social scientists who study the dynamics of human populations (Ehrlich et al. 1977). They are primarily concerned with the questions of how and when birth rates change in response to social or environmental pressures. Some ecologists consider demography simply as a method to describe how it is, not why it is, and thus use demographic techniques to quantify the varying densities of plants in their areas of distribution. Others include demography within the broader field of population biology and even consider concepts of natural selection (e.g. Harper & White 1974; Harper 1977; Solbrig & Solbrig 1979; Solbrig et al. 1979; Blom 1987). With this approach, the principal objectives of plant demography are natality, death, immigration, emigration, and the behaviour of plants during all stages of their life cycle. These various aspects of the life cycle determine the life history of a plant. Through studies relating life-history characteristics to environmental factors, insight can be gained into the underlying processes that cause both changes in population sizes and variation in the behaviour of the members of that population. The underlying processes are linked to those environmental circumstances that may force species to adaptations brought about by natural selection. Discussions on the phenotypic determination of environmental and genetic influences on demography can be found in Bradshaw (1965), Levin (1976), Jain (1979), Blom (1983), Gottlieb (1984), Haeck & Woldendorp (1985), Lotz & Blom (1986), Schlichting (1986) and Ernst (1987). In order to understand processes as they occur in nature, a large number of measurements are needed, and many experiments have to be carried out that test the effects of environmental factors, either separately or in combinations, on the behaviour of organisms. The development of theories and the predictions of future events are important goals for ecologists. Therefore, the construction and use of models have to be considered both as an exciting challenge and as a tool for scientists in this field of research. Although the field of ecology is only a relatively small biological discipline, there is an abundance of literature available on models. It appears that in the past each researcher constructed his own model, forgetting that model making should not be an aim in itself, but a tool to a better understanding of reality. Ecology has, as have various other branches of biology, grown top-heavy with models. Categorization of models greatly depends upon the viewpoint of the classifier and, as a consequence, many different classifications are found in the literature.