The population dynamics of plants

Abstract Long-term studies of plant populations are reviewed, and their dynamics summarized in three categories. Many short-lived plants have ephemeral, pulsed dynamics lasting only a single generation, with recruitment determined almost entirely by germination biology and by the frequency and intensity of disturbance. Such populations are not amenable to traditional population models. At the other extreme, some long-lived plants have such protracted tenancy of their microsites that it is impossible to establish what pattern of dynamics (if any) their populations exhibit. A relatively small number of species show what we would traditionally regard as population dynamics at a given point in space (i.e. reasonably predictable trajectories that can be modelled by Nt+1= f(Nt)). A major difficulty in generalizing about plant dynamics is that the majority of species are successional; their recruitment depends upon the death, through senescence or disturbance, of the dominant plants. Where we do have data spanning several generations, it is clear that: (i) the populations are regulated by density dependent processes; (ii) in contrast to some animal populations, numbers appear to vary less from year to year in places where mean density is higher, and less from place to place in years when mean density is high than when density is low; (iii) few, if any, plant populations show persistent cyclic or chaotic dynamics, but (iv) there are several robust generalizations that stem from the immobility and phenotypic plasticity of plants (the law of constant yield; self-thinning rules, etc.). These generalizations are analysed in the context of simple theoretical models of plant dynamics, and the patterns observed in long-term studies are compared with similar data from animal populations. Two important shortcomings of traditional plant demography are emphasized; (i) the dearth of simple manipulative experiments on such issues as seed limitation, and (ii) the tendency to locate study plots around existing mature individuals (the omission of ‘empty quadrats’ may introduce serious bias into the estimation of plant recruitment rates).