The influence of variation in recruitment relative to variation in mortality or emigration in determining variation in population size and growth has been analyzed in several taxa living across a range of ecosystems. There is no consensus on the relative importance of recruitment. In most terrestrial vertebrates, adult survival typically varies much less than recruitment through time, space, or with population density. In marine ecosystems, the 'recruitment limitation hypothesis' states that recruitment patterns account for most of the temporal or spatial variation in population abundance of most species. However, some studies of tropical reef fishes and of marine invertebrates have reported that changes in postrecruitment mortality rate and postsettlement processes can be more influential than changes in recruitment rate. Several methodological issues might account for such inconsistencies. Before we can safely identify the ecological and evolutionary causes of variation in the relative importance of recruitment on population dynamics, such problems have to be solved.
First, recruitment is not defined consistently among studies. Obviously different measures of recruitment will provide different answers. Early recruitment stages (i.e., number of propagules produced, early survival) are generally much more variable than late recruitment stages (i.e., late survival and settlement). The relative contribution of the different stages to recruitment variation, however, is rarely known and can vary from site to site for a given species. Moreover, the parameters describing the outcome of the different recruitment stages have a multiplicative, rather than an additive, influence on the whole recruitment process. Thus, under some conditions, the overall variation in recruitment will be much larger than the observed variation in any recruitment stage, whereas under other conditions, some compensation between recruitment stages could dampen the observed variability of recruitment. The contribution ofthe interactions among recruitment stages to changes in population abundance could be much higher than the contribution of the main effects of these stages. Serial autocorrelation among recruitment stages, although rarely assessed, is thus expected to play a major role in population dynamics.
Second, the analysis of the relative importance of recruitment on population dynamics strongly depends on the scale of analysis. Analyses at different scales answer different questions. The relative contribution ofrecruitment can be calculated from demographic analysis by performing a sensitivity or elasticity analysis. The question asked is then ''By how much 'would' population growth 'change if' recruitment was increased or decreased by a given amount (sensitivity analysis) or a given proportion (elasticity analysis)?'' Asking the same question for other processes (e.g., postrecruitment survival) will provide a measure of the relative contribution of recruitment to population dynamics. Note that this approach ignores the variation in recruitment observed in the field, because it only estimates a potential influence. In most cases the expected relative contribution ofrecruitment assessed from elasticity analyses reflects the position of the study species on the slow-fast continuum of life histories. Recruitment tends to exert a stronger influence on population dynamics in species in which individuals produce a large number of propagules early in their life and have short life span compared to species where individuals produce few propagules late in their life and have long life span. For example, the relative importance of seedling recruitment varies substantially between annual plants, perennial herbs, and shrubs and trees, being much larger in herbs than in woody plants.
The relative contribution ofrecruitment can also be used to answer a different question: ''By how much does the 'observed' variation in population growth 'change in response to observed' changes in recruitment?'' Asking the same question for other processes (e.g., postrecruitment survival) will also provide a measure of the relative contribution of recruitment to population dynamics. That approach accounts for the variation in recruitment observed in the field, so that the influence of both the observed variation (the larger the variation in recruitment, the higher its relative contribution, holding other demographic processes constant) and the potential impact (the larger the elasticity or sensitivity, the higher the relative contribution of recruitment) of demographic processes are included. Using these approaches the relative contribution of recruitment is context specific and thereby less easy to predict from a species' life history than the expected contribution. It is quite straightforward to understand why predicting the potential impact of changes of a demographic rate on the population growth rate is easier than predicting the relative contribution of the same demographic rate to observed changes in population growth. In the first case, we perturb all demographic rates by the same proportional amount, while in the second the observed perturbations vary between demographic rates. However, some general patterns could exist in relation to the taxa studied or to the demographic status of the populations studied. For instance, most populations of passerine birds in the UK have been reported to be limited mainly through recruitment to the breeding population. In both marine invertebrates and large herbivores, the relative importance of recruitment versus adult survival could decrease with increasing densities. More studies are required to assess the general relevance of such findings. Depending on changes in environmental conditions, there can be much variation in the relative contribution of recruitment among, but also within, populations of a given species. Moreover, depending on whether we assume a perturbation (potential impact) or measure observed perturbations (relative contribution) of demographic rates, the importance of recruitment relative to other demographic processes can strongly differ for a given study.
Was this article helpful?