The evolutionary ecology of dispersal Ulf Dieckmann a dieckman@iiasa.ac.at, Bob O'Hara b bob.ohara@helsinki.fi and Wolfgang Weisser c weisser@ubaclu.unibas.ch [a] International Institute for Applied Systems Analysis, Schloßplatz 1, A-2361 Laxenburg, Austria[b] Division of Population Biology, University of Helsinki, PO Box 12, FIN-00014 Helsinki, Finland[c] Zoology Institute, University of Basel, Rheinsprung 9, CH-4051 Basel, Switzerland

Dispersal is a life-history trait that has profound consequences for populations. Viewed from an ecological perspective, dispersal influences the dynamics and persistence of populations, the distribution and abundance of species, and community structure. From an evolutionary perspective, dispersal determines the level of gene flow between populations and affects processes such as local adaptation, speciation and the evolution of life-history traits. In fact, it is difficult to imagine any ecological or evolutionary problem that would not be affected by dispersal.

The various consequences of dispersal are extensively discussed in the ecological and evolutionary literature (the Science Citation Index gave more than 1000 occurrences of 'dispersal' in the abstract or title of papers for the year 1997 alone). Surprisingly, however, the question of why particular dispersal strategies evolve has received much less attention. Part of the problem is that many of the mechanisms proposed to affect the evolution of dispersal ( Box 1) are notoriously difficult to test in the field. There exists a serious gap between theory and data, and consequently our understanding of why particular organisms disperse in specific ways is still limited. A recent workshop in Finland provided an opportunity to survey the state of the field.

The workshop – Evolution of Dispersal – took place in October last year at the Tvärminne Zoological Station of the University of Helsinki (financed by the Finnish Academy of Sciences through the Spatial Ecology Program in the Division of Population Biology). The organizers, Liselotte Sundström and Mikko Heino (both at the Division of Population Biology, Helsinki University) brought together an impressive array of researchers with diverse backgrounds and diverse approaches to the evolution of dispersal.


Adaptive dynamics of dispersal strategies

To understand present states and potential changes in dispersal traits, we have to evaluate the selective pressures that underly their evolution. These pressures arise from interactions between individuals of the dispersing population and their environment. Because dispersal often occurs in spatially heterogeneous environments, the resulting population dynamics and ecological feedbacks tend to be intricate.

It is difficult to incorporate such complicated feedbacks between an evolving population and its ecological environment into models of population genetics. Consequently, models of evolutionary game theory have been used, but these tend to oversimplify strategies and feedbacks by relying on payoff matrices. An alternative approach for studying the evolution of dispersal uses adaptive dynamics ^1–3 , where selective pressures and resulting adaptive changes are derived from population dynamics ( Box 2).

Mats Gyllenberg (University of Turku, Finland) and Hans Metz (Leiden University, The Netherlands) presented a technique for predicting invasibility of metapopulations. For the first time, their method has allowed the initial growth rate of rare mutants in resident metapopulations to be obtained analytically. Ulf Dieckmann demonstrated how correlation dynamics (where spatially extended populations are described not only by densities of individuals, but also by densities of pairs of individuals) can provide insights into tradeoffs between competitive and dispersal abilities.


From methods to mechanisms

One new development facilitated by adaptive-dynamics theory is the inclusion of population dynamics into evolutionary models. Stefan Geritz (University of Turku) discussed how the evolutionary dynamics of dispersal rates in metapopulations are affected by the existence of multiple demographic attractors. Michael Doebeli (University of Basel, Switzerland) showed that complex population dynamics could lead to an 'evolutionary cycling' of dispersal rates: out-of-phase fluctuations select for increasing dispersal rates until dispersal synchronizes the dynamics. If costly, dispersal is then selected against until the dynamics are again asynchronous so that the cycle can repeat itself. Other adaptive-dynamics models for studying the effects of spatial and temporal heterogeneities (both internally generated and externally imposed) on the evolution of dispersal rates were presented by Kalle Parvinen (University of Turku) and Andrea Mathias (Eötvös University, Budapest, Hungary).

Findings from these different models all point towards a common conclusion: in spatially structured populations, interactions between ecological and evolutionary dynamics can lead to polymorphisms in dispersal rates through repeated 'evolutionary branching'.

Olof Leimar, Ulf Norberg (both at Stockholm University, Sweden) and Graeme Ruxton (University of Glasgow, UK) used lattice models to investigate causal mechanisms for the evolution of dispersal. Justin Travis (Imperial College, Silwood Park, UK) and Calvin Dytham (University of York, UK) explored the effects of habitat heterogeneity by using random fractals to create spatial and temporal fluctuations in carrying capacities of habitats. If spatial fluctuations were autocorrelated (red noise), greater dispersal rates evolved than when the fluctuations were not autocorrelated (white noise). Autocorrelated temporal fluctuations caused lower dispersal rates to evolve than nonautocorrelated temporal fluctuations. Francois Rousset (University of Montpellier, France) and Nicolas Perrin (University of Lausanne, Switzerland) demonstrated the importance of kin selection for the evolution of dispersal, and Pekka Pamilo (Uppsala University, Sweden) discussed his investigations of the effects of social structure on dispersal in ants.

The evolution of dispersal has consequences for other life-history traits, which in turn can affect dispersal rates. Éva Kisdi (University of Turku) analysed the joint evolution of dispersal and a trait determining survival in two different types of habitat with environmental stochasticity. In her adaptive-dynamics model, evolution often resulted in low dispersal rates and local adaptation (thus yielding an evolutionarily stable dimorphism of two phenotypes, each of which is a specialist for only one habitat). However, differences between habitats and the magnitude of temporal fluctuations have a strong effect on evolutionary outcomes.

Three speakers explicitly aimed to identify causes or consequences of dispersal in particular organisms. Janis Dickinson (University of California, Berkeley, USA) argued that differences between sexes in the relative success of philopatric versus dispersing individuals might be a reason for sex-biased dispersal in western bluebirds (Sialia mexicana) – although problems in following dispersers made quantitative fitness estimates very difficult.

Habitat fragmentation could lead to a decrease in dispersal rates, because genes associated with dispersal will be lost from isolated populations when individuals leave the habitats. Because the decrease in dispersal propensity can influence the persistence of a species in metapopulations, this process has implications for conservation biology. Chris Thomas (University of Leeds, UK) presented data from several butterfly species to suggest that the ability to disperse might indeed be decreasing in isolated or fragmented populations.

Jean Clobert (University of Paris VI, France) argued that several of the factors that are suggested by theoretical models to influence the evolution of dispersal might act together, even within the same population. Because many factors lead to similar predictions, identifying their relative importance is a major goal that can only be achieved experimentally. A recurrent result of Clobert's studies on the common lizard (Lacerta vivipara) is that dispersal is condition-dependent – a fact largely ignored by current models.


Measuring dispersal

Some of the practical statistical problems of measuring dispersal in the field were outlined by Walt Koenig (University of California, Berkeley, USA) when discussing his findings on acorn woodpeckers (Melanerpes formicivorous). Koenig emphasized that if the scale over which dispersal is measured is smaller than the scale over which organisms actually move, then average dispersal distances can be grossly underestimated. This is a 'right censoring' problem, familiar to those analysing medical trials (where not all patients die or relapse before the end of the trial). Unfortunately, there was no clear shape in the dispersal pattern that would have allowed extrapolation of measurements to longer distances.

Individuals moving too far was not a problem faced by Bruno Baur (University of Basel) in his tracking of snails, which can move as far as 7 m per year. Indeed, Baur suggested that catastrophes such as avalanches and floods after torrential rains were the major mechanism for long-range dispersal. Wolfgang Weisser (University of Basel) demonstrated difficulties in delineating local populations of aphids. David Jenkins (University of Illinois, USA) discussed empirical data taken from artificial ponds and argued that the movement of zooplankton between ponds is a much rarer and less predictable phenomenon than previously thought. Bruce Rannala (State University of New York, USA) assessed the utility of Wright's island model and used a Bayesian framework to develop methods for the estimation of past immigration, based on population genetical data.

By itself, the measurement of dispersal is merely a descriptive exercise. Linking measurements to mechanisms, Jens Roland (University of Alberta, USA) has estimated the effects of spatial pattern of woodland and meadow on dispersal behavior in Parnassius butterflies, which inhabit meadows that arise in gaps created by forest fires. Roland has shown that the intervening landscape between sampling sites had a predictable effect on the amount of movement between sites.


Dispersal and metapopulation viability

Are extinction risks for endangered species reduced by evolving dispersal strategies? In a process known as 'adaptive rescue' ⁴, populations exposed to environmental threats can increase their viability through the evolution of critical life-history traits. However, what is beneficial to the population as a whole is not necessarily favored by individual selection. In a given ecological setting, the evolutionary stable rate of dispersal need not be identical to the rate that would optimize population persistence (Isabelle Olivieri, University of Montpellier). Also, the response of these two rates to changing ecological conditions can be qualitatively different. Coevolution of dispersal rate and reproductive effort might enhance metapopulation persistence in highly disturbed landscapes.

Pierre-Henri Gouyon (University Paris-Sud, France) illustrated (from empirical data and theoretical analyses) the importance of dispersal for the persistence of threatened plant metapopulations. Transitions between vegetation types, brought about by environmental change, can result in extinctions if adaptation of dispersal strategies cannot occur fast enough.

Régis Férričre (Ecole Normale Supérieure, Paris) presented models of metapopulations that are driven to extinction by natural selection acting on dispersal rates. In contrast to adaptive rescue, such populations actually undergo an 'adaptive suicide'. A degrading environment can obstruct the evolutionary path of a dispersal trait towards more viable rescue states. From within such an 'adaptive trap', gradual evolution of dispersal can no longer prevent population extinction.

In this workshop, the recent rise of adaptive dynamics theory was very apparent, with many speakers using this tool to explore different aspects of dispersal evolution. In the real world, however, detailed knowledge about dispersal in many organisms remains scarce. Some contributions suggested that new techniques, such as those from molecular biology, might help to overcome this shortage. It will remain a challenge to integrate the various approaches presented, so that more theoretical predictions can be tested in the field. A forthcoming symposium * in France will provide the next opportunity to see how close we are to finding a unifying approach in the study of dispersal.


Unlinked references

5–14



References

[1] Metz J.A.J. et al. (1996) Adaptive dynamics: a geometrical study of the consequences of nearly faithful reproduction.
In: van Strien S.J. and Verduyn Lunel S.M. (Eds.) Stochastic and Spatial Structures of Dynamical Systems. (pp. 183-231) : North Holland

[2] Dieckmann U. (1997) Can adaptive dynamics invade?
Trends Ecol. Evol., 12:128-131. [ScienceDirect] [Cited by]

[3] Geritz S.A.H. et al. (1998) Evolutionarily singular strategies and the adaptive growth and branching of the evolutionary tree.
Evol. Ecol., 12:35-57. [Cited by]

[4] Gomulkiewicz R. and Holt R.D. (1995) When does evolution by natural selection prevent extinction?
Evolution, 49:201-207. [Cited by]

[5] Maynard Smith J. and Price G.R. (1973) The logic of animal conflict.
Nature, 246:15-18. [Cited by]

[6] Johnson M.L. and Gaines M.S. (1990) Evolution of dispersal: theoretical models and empirical tests using birds and mammals.
Annu. Rev. Ecol. Syst., 21:449-480. [Cited by]

[7] van Valen L. (1971) Group selection and the evolution of dispersal.
Evolution, 25:591-598. [Cited by]

[8] Comins H.N., Hamilton W.D. and May R.M. (1980) Evolutionarily stable dispersal strategies.
J. Theor. Biol., 82:205-230. [MEDLINE] [Cited by]

[9] Roff D.A. (1994) Habitat persistence and the evolution of wing dimorphism in insects.
Am. Nat., 144:772-798. [Cited by]

[10] Hamilton W.D. and May R.M. (1977) Dispersal in stable habitats.
Nature, 269:578-581. [Cited by]

[11] Hastings A. (1983) Can spatial selection alone lead to selection for dispersal?
Theor. Popul. Biol., 24:244-251. [Cited by]

[12] Levin S.A., Cohen D. and Hastings A. (1984) Dispersal strategies in patchy environments.
Theor. Popul. Biol., 26:165-191. [Cited by]

[13] Doebeli M. and Ruxton G.D. (1997) Evolution of dispersal rates in metapopulation models: branching and cyclic dynamics in phenotype space.
Evolution, 51:1730-1741. [Cited by]

[14] Shields W.M. (1987) Optimal inbreeding and evolution of philopatry.
In: Swingland I.R. and Greenwood P.J. (Eds.) The Ecology of Animal Movement. (pp. 132-159) : Clarendon Press



Footnotes

^[*] 24 April–1 May 1999 Causes, consequences and mechanisms of dispersal, Roscoff, Britanny, France. (Jean Clobert, Universite Pierre et Marie Curie, Bëtiment A, Case 237, 7 quai Saint Bernard, 75252 Paris, France. jclobert@hall.snv.jussieu.fr)