The Ecology of Dispersal: Causes and Consequences of Phenotype-dependent dispersal

Dispersal is one of the few traits shared by all organisms. The study of dispersal has a long history in both ecology and evolutionary biology. Ecologists have long sought to understand how dispersal affects population dynamics and species coexistence, while evolutionary ecologists have sought to understand the evolution of dispersal despite significant costs to the individual. Remarkably, these two discussions have rarely informed each other — in particular, ecologists often ignore the role that dispersal costs may play in influencing population-level processes. This thesis seeks to unify these two sub-disciplines of dispersal biology by investigating how among- individual variation in phenotype (e.g., dispersal duration and larval size) influences dispersal behaviour, habitat selection, post-settlement performance, and ultimately population dynamics. I used marine bryozoans (Bugula neritina and Watersipora arcuata) as an experimental system and combined the results from laboratory and field studies with theoretical models. In chapter 2, I asked whether the population effects of phenotypic variation among individuals match or exceed the influence of spatial and temporal variation in the number of individuals. To do this, I simultaneously manipulated the phenotype (dispersal duration) and the density of colonizers and measured subsequent population structure. I found that the phenotype and abundance of individuals colonizing a patch strongly interacted to affect subsequent reproductive output of the population. In fact, populations founded by a few individuals with short dispersal durations (i.e., in relatively good ‘condition’) actually had a similar reproductive yield to populations founded by many individuals with long dispersal durations (in relatively poor ‘condition’). In chapter 3, I conducted a series of experiments in the field and the lab to estimate the relative importance of direct and indirect deferred costs of dispersal. I then used those data to parameterise a theoretical model to describe how dispersal costs interact with the spacing and quality of habitat to influence population connectivity. I found that the deferred costs of dispersal can result in the strength of connectivity to distant good quality patches being the same as that to nearby poor quality patches. Costs of dispersal therefore have a number of implications for understanding the spread of invasive species and developing spatial conservation plans. In chapter 4, I showed that dispersal duration and larval size affect larval behaviour prior to settlement, and these effects are likely to interact with the more well studied external settlement cues. I also present a habitat selection model, parameterized with the experimental results, to explore when these phenotype-dependent behaviours might be adaptive under different scenarios of habitat abundance and quality. I suggest that presettlement larval behaviour can be viewed in the context of informed dispersal where individuals use internal and external cues to assess the current settlement habitat in relation to unknown habitat elsewhere. In chapter 5, I used an experimental approach that examined the effects of maternal exposure to temperature on the phenotype and performance of offspring in different temperatures. I also analysed autocorrelation in water temperature from field measurements to estimate intergenerational predictability in environmental correlations. I found that the responses of larvae to different water temperatures depended on the temperature that their mothers experienced. Analysis of time series data on temperature in the field indicated that offspring are likely to experience similar thermal environments as their mothers, particularly during the larval and very early post- metamorphic life-history stages. In this paper, I provide estimates of absolute and relative maternal fitness and discuss how the benefits of transgenerational plasticity depend on the importance of frequency- (or density-) dependent selection. The need to understand dispersal and colonization processes has become increasingly apparent in the context of managing threatened or invasive species, as well as predicting biological responses to environmental change. The main implications of the results presented in this thesis are that costs of dispersal, phenotypic variation, and phenotypic links between life-history stages at the individual level can have consequences at a population level: consequences that would not be predicted assuming dispersers were homogeneous. Studying the costs of dispersal, and the role of phenotypic links among life-history stages, is therefore fundamental for a mechanistic understanding of population processes. It is hoped that this work will contribute to the broader understanding of the links between ecological and evolutionary dynamics that are essential to properly manage biodiversity.