conservation, populations, conservation biology, dispersal, population, David W. Macdonald, consequences, Cambridge University Press, extinction, Journal of Animal Ecology, endangered species, selection, Hanski, behaviour, Trends in Ecology and Evolution, population densities, dispersal mortality, dispersal patterns, environment, habitat disturbance, stochastic factors, predation pressure, genetic structure, Macdonald, Barreto, G. R., Animal Behaviour, animal populations, selection pressure, fragmented habitats, metapopulation
25. Dispersal in theory and practice: consequences for Conservation Biology
David W. Macdonald and Dominic D.P. Johnson ABSTRACT Dispersal is a fundamental parameter of population processes. In small and isolated populations, linked only when individuals are able to disperse between them, it becomes a particularly critical one which ultimately determines whether a species will become extinct. Predictions for these populations are limited because estimates of dispersal variables from the field are difficult to obtain and therefore scarce. Models are therefore also limited because they cannot be parameterized accurately, and have until recently ignored behavioural and spatial variation in dispersal. This problem is compounded for endangered species
which most desperately need accurate viability analyses and management plans, but about which little is usually known and opportunities for research may be few. Threatened populations are not the only conservation application for dispersal. Predators and pests are a significant conservation concern, and advances in the management of these populations also require detailed understanding of dispersal behaviour to forecast their expansions. More detailed behavioural research into dispersal is clearly needed. Aside from their academic merit, the chapters within this book provide an excellent basis for furthering the application of research on dispersal to conservation biology. Keywords: conservation, viability analyses, models, parameters, extinction, management, fragmentation, habitat. Introduction Dispersal is one of the crucial parameters in even the most simplistic population models, providing (along with the birth/death ratio) one of two principle modes of population change (Hanski 1999a). It is automatically, therefore, an influential variable in population dynamics, and especially so in sub-populations (Hanski 1999a; Levins 1970b) prone to extinction within a metapopulation. This is because, among sub-populations increasingly isolated by habitat change and fragmentation, a trickle of dispersal, the `rescue effect', may stave off extinction. Indeed, if local extinction does occur, the only prospect for reversing it lies in the arrival of dispersers (Brown and Kodrick-Brown 1977; Hanski 1999a). This will only be successful if the proximate cause of the extinction can be identified and eradicated, which may require significant conservation efforts or large-scale habitat restoration. The ability to model, predict, and ultimately to manage these eventualities thus depends on knowledge of the processes and parameters of dispersal. Not only does dispersal generate considerable theoretical interest, but its direct consequences and repercussions for population dynamics are of vital importance to the conservation of threatened species
world-wide (Simberloff 1988). This is widely relevant across the globe, from the small and isolated habitats and populations in logged tropical rainforests areas in the third world (Whitmore 1997), to those of the fragmented agricultural landscapes of Western Europe (Macdonald and Smith 1990, 1991). Global climate change
David W. Macdonald and Dominic D.P. Johnson
threatens species with specific habitat requirements, where their dispersal is not sufficient to adjust their distributions accordingly (Davis 1989). The many examples of species for which dispersal is a critical matter would fill a book by themselves. However, examples which illustrate the central importance of dispersal to a diversity of issues include the viability of fragmented populations of Water Voles, Arvicola terrestris, and the expansion of their introduced predator, the American Mink, Mustela vison, in European waterways (Macdonald and Strachan, 1999; Barreto et al. 1998a); the mechanisms of inbreeding avoidance amongst Ethiopian Wolves, Canis simensis, (SilleroZubiri et al. 1996) and African Hunting Dogs, Lycaon pictus, (McNutt 1996); limited dispersal ability of plants in the face of climate change (Primack and Miao 1992), and the epizootiology and control of rabies in red foxes, Vulpes vulpes
, (Macdonald 1995). Dispersal is, in the words of John Wiens (chapter 7), `the glue that binds populations together'. This glue is an indispensable item in the toolbox for maintaining dwindling populations and managing pests over the many damaged landscapes across the globe. We first discuss the range of problems that dispersal poses for conservation. Then, we discuss the importance of, and some caveats with, population viability models, which are particularly important as tools for conservation since empirical field tests are rare. We then explore the problems that the process of environmental change itself is likely to cause for dispersal, followed by how knowledge about dispersal may be applied in conservation biology. We conclude with some remarks emphasizing the urgency of the conservation of species, and how much remains to be discovered about dispersal.
The problems of dispersal for conservation
Lack of data and reliance on models A decade ago, Johnson and Gaines (1990) noted a `prodigious' number of models of dispersal evolution. Unfortunately, as three recent reviews identify, there remains an almost complete absence of empirical data on dispersal and other relevant behavioural parameters in recent population extinction studies (Caro 1999; Doak and Mills 1994; Reed and Dobson 1993), and there are difficulties in modelling and understanding evolutionary mechanisms of dispersal (Dobson 1985). Much of the recent work on extinction processes has focussed on the metapopulation paradigm (Hanski and Gilpin 1991; Hanski and Gilpin 1997), upon which, according to Doak and Mills (1994: 619), "essentially all of the most influential papers on the subject, have been modelling efforts". In fact, evidence has not always been forthcoming to support the population turnovers and extinctions predicted by metapopulation dynamics in patchy habitats (Harrison 1991; Harrison et al. 1988; Taylor 1990; Taylor 1991). Yet, empirical studies are constrained by the difficulties of obtaining information about dispersing individuals (North 1988; Verhulst et al. 1997) and, as a consequence, many models for population viability analysis (PVA) remain largely untested with field data. Previous models (including `packaged' PVA models) have treated dispersal as a fixed trait, whereas it is becoming increasingly clear that there is considerable phenotypic plasticity in dispersal behaviour (van Vuren and Armitage 1994; Reed, 1999; Murren et al. chapter 18; O'Riain, chapter 10) and that it may be condition- or situation-dependent (Dufty and Belthoff, chapter 15; Ims and Hjermann, chapter 14; Stamps, chapter 16).
Dispersal in theory and practice: consequences for conservation biology
Lack of behavioural detail in dispersal studies Reed (1999) specifically targets the lack of inclusion of behavioural mechanisms in population modelling, a damaging omission because behaviour is an important influence on population persistence for many taxonomic groups spanning birds, mammals, insects, and marine animals. This is especially pertinent given the renaissance in recognizing the importance of `Allee' effects, which result in decreased reproductive rate or survival at low densities (Courchamp et al. 1999). This negative feedback
obviously has significant implications for the conservation of small populations (Stephens and Sutherland 1999). The mechanisms by which the Allee effect occurs are many, but social species are likely to be particularly disadvantaged as many of them rely heavily on conspecifics for cooperative behaviours, and this adds to other compounding problems (Woodroffe and Ginsberg 1998). Worse still, Reed argues, by omitting behavioural aspects of dispersal behaviour, predictions made for surviving population numbers may be over optimistic. This delusion could arise where models assume equal dispersal tendency, whereas the reality might encompass behaviour patterns that result in resistance to dispersing. Variation in dispersal behaviour For example, dispersal is generally incorporated into population models using a random diffusion/random walk paradigm, where animals disperse in random directions and stay in the first suitable habitat
(Wiens, chapter 7). In reality, however, behaviour has been shown to operate as a selective process in dispersal (references in Reed 1999; Stamps, chapter 16; Danchin et al., chapter 17). For example, some bird species with strong habitat specificity do not venture from their patch into the intervening landscape and therefore are constrained in dispersing whereas others readily fly from one fragment of habitat to another (Wiens, 1994, chapter 7; Johnson and Mighell 1999). Thus, dispersal estimates from natural habitats may be markedly different from those when the animal is faced with a matrix of unfamiliar habitat types. Landscape structure influences not only the sites between which animals may disperse, but also the permeability of a given route. Understanding the constraints upon dispersal demands knowledge of the species' natural history: expansive water may be a barrier to colonization of islands by American Mink (Herteinsson 1992), whereas Dormice use thin hedgerows in agricultural landscapes to disperse through expanses of open field which would otherwise be impenetrable (Bright and Morris 1996). Frank and Woodroffe (in press) draw attention to how species which disperse readily (like lions Panthera leo and coyotes Canis latrans) are relatively resilient in the face of population control, while some species (badgers Meles meles and Spotted Hyaenas Crocuta crocuta) are remarkably reluctant to disperse and therefore recolonize cleared areas extremely slowly. Tuyttens and Macdonald (2000) emphasize the likelihood that populations perturbed by attempted management are likely to behave differently than their undisturbed counterparts: One of the behaviour patterns likely to be perturbed by management is dispersal. Clearly, versatile species with generalist habitat requirements are likely to be able to disperse through a wide array of habitats, and those that are most mobile may do so over great distances; for example, red foxes may disperse over tens, or even hundreds, of kilometres (Lloyd 1980; Macdonald 1984). Nonetheless, behaviour holds many surprises: In the UK, the home ranges of Water Voles are generally confined to the water's edge, but recent evidence reveals that individuals may disperse overland across watersheds (X. Lambin, pers. comm.). The complications outlined above, which reveal the shortcomings of progress in dispersal studies for practical conservation planning, are widely and profitably discussed in the chapters
David W. Macdonald and Dominic D.P. Johnson
of this book. Several authors note significant failings: inconclusive tests of hypotheses (Lambin et al., chapter 8), lack of data and appropriate tests (Ronce et al., chapter 24; Wiens, chapter 7), problems with analyses (Peacock and Ray, chapter 4), and lack of knowledge of proximate factors (Dufty and Belthoff, chapter 15). Happily, however, solutions are emerging, with: proposals for how better to measure dispersal (Bennetts et al., chapter 1; Ross, chapter 3); methods of separating interactive effects (Gandon and Michalakis, chapter 11; Perrin and Goudet, chapter 9; Ronce et al., chapter 24); and reviews of possible complexities which need to be controlled for in future studies, such as variation in dispersal due to developmental conditions, environmental cues (Stamps, chapter 16), behaviour, or physiology (O'Riain, chapter 10; Murren et al., chapter 18). These latter topics have further consequences for metapopulations and the concept of effective population size, because of differences in genotype and fitness in those individuals that tend to disperse (Whitlock, chapter 19).
The urgency of conservation and dependency on models of dispersal Clearly, we are in an extinction crisis, with species disappearing at one hundred-fold, or even one thousand-fold, background rates of extinction. The practical goal of the science of conservation biology is to illuminate ways of slowing this trend. Dispersal is a crucial element of population dynamics and, consequently, a central issue in conservation biology. Our ability to forestall this crisis will depend, in some measure, on better understanding dispersal. There are, however, significant limitations to this end. The principal problem appears to be lack of empirical data (Reed 1999), but also a lack of experiments which have been able to disentangle the multiple causes of dispersal and the relevant spatial scale (Ims and Hjermann, chapter 14; Ronce et al., chapter 24) as well as problems associated with assumptions of hypotheses and interpretations of data (Peacock and Ray, chapter 4; Lambin et al., chapter 8). Ims and Hjermann (chapter 14) suggest that failure to account for the spatial scale of environmental and related variables, relative to a species' dispersal capacity, may explain inconsistent results concerning condition-dependent dispersal in the literature. Biological Conservation
ists do not enjoy the luxury of waiting for academic developments to arise that will eventually resolve these difficulties. They must be urgently stimulated to produce, at the least, approximate expectations and remedies on which to base conservation actions. Models have therefore become very important in targeting the correct variables and identifying the principle processes. Of course, in some circumstances an extreme paucity of information means that Realistic models
cannot be constructed. Nevertheless, models may very usefully arise in two general types. The first are theoretical models which formalize or examine, qualitatively or quantitatively, particular processes and variables in order to understand dispersal behaviour and its population consequences. The second type are those which specifically aim to forecast a species' population dynamics given certain specific parameters and starting conditions, and, in some cases, the details of the landscape, population viability models. It is the latter class that has come to be particularly associated with conservation projects, and on which we shall concentrate here. As a consequence of the above, many conservation management schemes have only model projections on which to base decisions. Such an example comes from the much-debated population viability models developed for the Northern Spotted Owl, Strix occidentalis caurina, in diminishing old-growth coniferous forests (e.g. Lande 1988; Lamberson et al. 1992). Empirical studies for each species are desperately needed (ideally case by case) as many of these models go largely unverified, or, at best, they rely on substitution of data
Dispersal in theory and practice: consequences for conservation biology
collected from similar species, or `off-the-shelf' PVA models that make unrealistic assumptions about dispersal behaviour. Bennetts et al. (chapter 1) set out some practical guidelines to achieve dispersal estimates, and were surprised by the lack of empirical tests of models despite the advent of new methods and new technologies for tracking and modelling animal movements using GPS or satellites and GIS (see also Wiens, chapter 7). That it is substantially easier to develop a model than to test it is a fact that should be noted by those prioritizing research funding. It is a poignant inevitability that the lack of data is particularly apparent for endangered species. They, by definition, are few and are limited in distribution, and disturbing them to obtain such life-history data may jeopardize the remaining individuals (Doak and Mills 1994; Reed 1999).
Consequences of dispersal on genetic structure of populations The genetic structure of the population also depends, among other things, on dispersal rates and dispersal distances (Endler 1977; Hanski and Gilpin 1997); these parameters, therefore, ultimately impose constraints within which local populations can evolve (see also Roff and Fairbairn, chapter 13). Most theoretical studies of dispersal have focussed on the consequences for population genetic structure and the mechanisms of its evolution (Johnson and Gaines 1990). Authors such as Peacock and Ray (chapter 4) and Ross (chapter 3) discuss advances in studying dispersal through genetic data, although Rousset (chapter 2) concludes that molecular approaches have provided only few success stories, against an undeveloped theoretical background and using methods largely untested in the field (see also Peacock and Ray). Barton (chapter 23) also draws attention to problems associated with using genetic markers as estimates of dispersal and associated gene movements. He suggests greater emphasis on measuring fitness directly in the field; clearly, this is difficult, but critical in order to clarify the consequences of dispersal. Ross (chapter 3) warns that different genetic methods can give different results, but encourages the development of detailed understanding of dispersal and gene flow from genetic approaches within a "predictive framework" developed from the natural history of the species, which can only be done with long-term and large-sample studies. Peacock and Ray (chapter 4) and Bennetts et al. (chapter 1) also suggest integration of genetic and demographic approaches in dispersal studies, not only during data collection
but also throughout the process of model development, right through to their testing. Peacock and Ray (chapter 4) reassert the point that hypotheses for dispersal (or anything else) cannot be tested properly in the absence of adequate data, and in this case that includes data on mating patterns and parentage.
Some further complications of dispersal for conservation In addition to noting the problem of lack of data on dispersal, Lambin et al. (chapter 8) draws our attention to the fact that convincing evidence is scarce for certain beliefs about the functional significance of dispersal, for example kin-competition. As another example, evidence is mounting for negative density-dependent dispersal rates (Lambin et al., chapter 8; see also Ims and Hjermann, chapter 14). Amongst carnivores, for instance, there is no evidence that dispersal is less common in controlled populations because there are more vacancies for youngsters at, or close to, home. If anything, control seems only to increase
David W. Macdonald and Dominic D.P. Johnson
dispersal (see Woodroffe, in press) and risk social perturbation, with potentially counterproductive effects amongst survivors (Swinton et al. 1997; Tuyttens and Macdonald 1998). Another current, but poorly understood issue vital for conservation planning concerns feedback mechanisms in dispersal behaviour (Ronce et al., chapter 24; Van Baalen and Hochberg, chapter 21). While the multiple causes of dispersal are difficult enough to disentangle (Gandon and Michalakis, chapter 11, Ronce et al., chapter 24), feedback from dispersal consequences on the population level may further cloud interpretation of Empirical results
of mechanisms and measurements of dispersal. This is especially pertinent at present, as many animal populations are undergoing declines (or increases) in densities following anthropogenic change and therefore exhibit dynamic dispersal parameters as the perturbed populations experience shifting selective pressures. A final word of caution before proceeding further is that we have thus far heralded dispersal as a sort of panacea for species conservation, especially in fragmented habitats and metapopulations. There are, however, several facets of dispersal that, conversely, may help to drive a species to its own extinction, especially in fast-changing and altered landscapes. For example, dispersing individuals are vectors for pathogens to move between otherwise isolated populations of susceptible conspecifics, the effects of which may become more risky to small populations (Caro 1999). Alternatively, a high dispersal rate can, in theory, cause the establishment of synchronous local population fluctuations, exposing the whole metapopulation to similar dynamics where stochastic effects are correlated (Hanski 1989; see also Lande et al. 1999 for a recent treatment). Thus, the probability of extinction of the whole metapopulation at a high dispersal rate may exceed the product of the independent probabilities of all local extinctions when the dispersal rate is low. Because of such adverse consequences, it is vital to keep to the fore an attention to the multitude of ramifications that dispersal may have for populations. Management of damaging side effects
may be as important as administering the cure.
Progress in models of dispersal in conservation Empirical knowledge can often usefully be complemented by modelling. As has already been emphasized, this opportunity becomes a somewhat frustrating necessity in many investigations of dispersal, due to the scarcity of empirical data. Indeed, in assessing the relative importance of potential variables to population viability, models may themselves be useful in identifying which parameters most merit expending further effort in the field to obtain accurately. A synergy between reality and simulation is nevertheless important in conservation biology for at least two other reasons: First, conservation planning generally involves evaluation of alternative management options; it may be feasible to make a preliminary evaluation by harmlessly manipulating the virtual world of a model before risking expensive or irreversible action on the ground (if, and only if, the model is realistic). Second, fieldwork is almost always sufficiently labour-intensive as to be confined to a small scale, whereas policies are often implemented on a wide scale, sometimes regionally or nationally. Therefore, models can be useful if they facilitate extrapolation to wider geographical scales. However, models that make `predictions' of population viability can be dangerous because they may entertain an illusion of broad application across a range of conservation management programmes before they have undergone rigorous testing or validation. As an example, a predictive model (developed with empirical data) for Mountain Sheep populations (Ovis canadensis) (Berger 1990) which had specific management implications, sparked a vehement
Dispersal in theory and practice: consequences for conservation biology
debate about whether its predictions were valid (Berger 1999; Wehausen 1999), possibly partly because of assumptions made about dispersal. As another indication of the perils of generalizing with models, an experimental study of two closely related species of Drosophila showed how the division of single populations into small sub-populations had very different effects on persistence depending on the species, even though they were otherwise very similar (Ims and Stenseth 1989). What is observed to happen in one species is not necessarily, therefore, a good model for another, even if they are closely related. Modelling population persistence for species of particular conservation concern has now become a routine process, and there are now several widely available population viability analysis (PVA) programs (Boyce 1992). These provide a useful framework within which to identify priority causes of decline for management. However, although they generally rank the influence of input variables consistently, different versions may give different predictions with identical problems and data (Macdonald et al. 1998). Since they contain prescribed algorithms (Lima and Zollner 1996) they are essentially a `black-box', and application to a novel species may be biased from the outset. Of course, they cannot reasonably be designed to incorporate all situation-specific details across the great diversity of novel problems. Given that they are often the only solution, an ideal compromise would be for programmers and conservationists to work together towards a customized adaptation of the original. Apart from these problems, PVAs have hitherto also been limited by the general absence of spatial information
. Along with the development of GIS technologies and sophisticated computer-generated prediction models based on two- or three-dimensional outputs on digitized landscapes, the scope for creating ever more `realistic' (and complicated) models is increasing. Several chapters in this book also provide new extensions to older models by specifically incorporating dispersal parameters describing variation between individuals or habitats or interactions (e.g. Murren et al., chapter 18; Hanski, chapter 20; Mouquet et al., chapter 22; Perrin and Goudet, chapter 9), while van Baalen and Hochberg (chapter 21) introduce the complexity of inter-trophic interactions, which may have been a source of bias in former models. However, as with any statistical routine, and process-based spatially explicit population modelling in particular, the increase in complexity goes hand in hand with more parameters and a decrease in interpretability and confidence in results. However, to the extent to which the models are reliable, they allow us to explore various options for each species. For example, work might be undertaken to improve the habitat for Water Voles by restoring riparian banks, or for Dormice by planting and coppicing Hazel. These activities might be approached in various ways, each with different cost implications: How many Hazel copses should there be, and distributed in which pattern? We have illustrated how our models, or their refined successors, could inform just such decisions elsewhere (Macdonald et al. 1998, Macdonald and Rushton unpublished manuscript, Rushton et al. in press). Of course, it would be folly to base action on such models without some reassurance that there are grounds for confidence in them. For example, in simulations of spatial spread, Macdonald and Strachan (1999) present tentative evidence of similarities between simulated and observed Water Vole populations (see also Rushton et al. in press). However, these are correlations only, and so again we urge caution. At the least, however, such explorations highlight the topics on which Field Research
might usefully be focused. Despite the absence of parameter estimates
for even simple models, spatially explicit models of populations used in conservation biology are becoming something of a cottage industry, so caution in their applicability is crucial, particularly regarding dispersal. However, a recent paper reports that some of the errors claimed to be associated with such models were grossly exaggerated (Mooij and DeAngelis 1998 and references therein). Indeed, at least as a
David W. Macdonald and Dominic D.P. Johnson
didactic, exploratory tool, the power added to individual, process-based models by advances in Geographic information systems
(GIS) appears to offer a useful practical guide for conservation planning and a good means of identifying gaps in field data (e.g. Macdonald et al. 1998, Rushton et al. in press, Macdonald et al. 2000b, South et al. 2000). However, while the apparent realism of this GIS-based approach strikes us as useful, we realize it can also be beguiling, and so we echo Doak and Mills's (1994) observation of a magnetism between conservationists, who are often keen (and sometimes desperate) to apply new theoretical techniques to urgent problems, and modellers, who are equally keen to find applications for their models.
Influence of environmental change on dispersal There are many direct effects of human influence on dispersal, for example the extinction of natural seed dispersers or destruction of habitat through which individuals are willing to disperse. The many indirect effects though, can be separated into two broad categories. Firstly, global climate change means that many sedentary species, and especially plants, will need to disperse latitudinally (or vertically, as in altitude) in accordance with shifting suitable habitat, perhaps 10 times faster than during the last ice-age retreat (Davis 1989; Primack and Miao 1992; Wilson 1992). Primack and Miao (1992) report that this may be too fast for certain plant species to persist and warn that a similar situation may exist for aquatic species of isolated lakes as well as, particularly, for vertebrate species of montane peaks. This is of particular concern, given that such areas contain endemic species
especially prone to extinction (Balmford and Long 1994; Stattersfield et al. 1998) and whose distributions appear to be closely linked to environmental variables (Johnson et al. 1998). Such isolated populations of restricted-range species may not have a suitable habitat to move into even if they are able to disperse. Large-scale environmental stochasticity is likely to be a particularly problematic influence because, of course, correlated extinctions of populations increase the risk of metapopulation extinction (Hanski 1999a). Indeed, it may be that today's rare species have become rare as a result of low dispersal and high correlation of between-population extinctions (Hanski 1989). The synchrony in dynamics between populations of a species may become especially critical. In addition to these indirect global climate change effects, man's deterministic influences are increasingly on a large enough scale to be labelled as `global' with regard to those species with limited geographic distributions. This is enough to warrant concern that entire metapopulations will be annihilated in one fell swoop, denying an opportunity for even some fragments to persist. The second broad category is that of the local knock-on effects of habitat destruction and fragmentation and the implications for the connectivity of the populations. Wiens (chapter 7) warns that there is not necessarily any reason to suppose that movement pathways bear a `close relationship to theoretical optima'. This introduces a considerable problem that may have to be faced by the study of dispersal in the future. While this volume has documented many likely evolutionary and mechanistic influences on dispersal tendency and characteristics, these may not operate as they formerly did, because of human-induced change resulting in species under considerable stress from habitat destruction and contamination. For example, migratory orientation has been shown to be influenced by organophosphorus pesticides (Vyas et al. 1995). There is a danger, therefore, that species' dispersal decisions may now be maladaptive if based upon cues of the environment which have been changed (Stamps, chapter 7), or that they may select habitats according to conspecific attraction (Stamps 1988; Reed
Dispersal in theory and practice: consequences for conservation biology
1999), which may be inappropriate if population densities reflect habitat disturbance rather than an otherwise approximately ideal free distribution (Reed and Dobson 1993). Furthermore, new studies indicate that animals use conspecific reproductive success, rather than just presence, as a cue (Boulinier and Danchin 1997; Danchin and Wagner 1997), in which case depressed reproduction due to environmental degradation
could directly alter the normal patterns of dispersal. Ims and Hjermann (chapter 14) recognize a need for studies to test whether it remains optimal for animals to employ condition-dependent dispersal when environmental change may directly or indirectly alter the cues and methods which animals use to disperse. Such effects could have broad influences, as there is evidence that individual variation in dispersal may arise due to a multitude of condition- or situation-dependent factors which human activities may have caused to change, including prior experience (Stamps, chapter 16), degree of parasitism (Boulinier et al., chapter 12a), predation pressure (Weisser, chapter 12b), hormones, body condition (Dufty and Belthoff, chapter 15), conspecific attraction, and patch choice criteria (Danchin et al., chapter 17). Another complication of environmental changes may be their interaction with other pressures to make dispersal more critical than usual. For example, Macdonald et al. (1999) explored the effects of predation (either naturally or in the form of human activities) on small populations of prey. A simulation model illustrated that, in small populations, stochastic factors always caused significant impacts while, in general, predation could only limit populations rather than extinguishing them. However, these simulations suggested that, in theory, predation pressure on already-small, isolated populations by generalist predators can cause local extinctions, particularly if these populations are also threatened by other factors. Thus, an interactive effect of depressed populations (for environmental reasons) can expose isolated populations of species, unable to disperse between habitat fragments, to extinction from factors that are not usually lethal to a large connected population (Barreto et al. 1998b; Macdonald and Strachan 1999). Similarly, movement and dispersal patterns relative to the size and configuration of protected areas emerge as especially important in the planning of nature reserves for the protection of wide-ranging carnivores (Woodroffe and Ginsberg 1998, Woodroffe, in press), because anthropogenic causes of mortality are intensified at reserve edges. Indeed, Woodroffe (in press) found a positive correlation between dispersal distance and proneness to extinction among large carnivores. Among four species of conservation concern, recent spatially explicit models (Macdonald and Rushton, unpublished manuscript) showed that interactions between life histories and the environment resulted in dispersal having very different influences on persistence depending on the species. For two rodent species, Water Voles and Hazel Dormice Muscardinus avellanarius, maximum dispersal distance (among other factors) was a significant partial correlation coefficient of the number of populations which persisted, as well as of dispersal mortality in the dormouse. However, for two mustelid species, American Mink and Pine Marten Martes martes, neither dispersal distance nor dispersal mortality emerged as a significant correlate of the total population size. Adult mortality was a significant correlate for both predators, as well as home range size for the mink and litter size and juvenile mortality for the pine marten. These species-specific differences do not necessarily hint at a generalization for their taxanomic group, but they do emphasize that the role of dispersal in population dynamics may vary substantially between species. Of course, these are dependent on the characteristics of the environment and the parameters used, but it serves here as a brief example of how dispersal appears to differentially affect populations depending on the details of the environment, its relative spatial scale and the particular life-history parameters of the species.
David W. Macdonald and Dominic D.P. Johnson
As a concluding caveat, for species facing extinction, which will often stem from HUMAN IMPACTS
on their environment (Balmford and Long 1994), dispersal may have unexpected interactive influences on persistence. The behaviour of populations is an emergent property of the reactions of individuals to their circumstances, which are now changing rapidly. Dispersal is likely to be a crucial element of this behavioural response. It is such threatened species that most urgently require models (albeit less urgently than they require data) to determine management plans, as they are also the ones for which mistakes would bear the greatest cost.
Conservation applications Because dispersal favours metapopulation survival, it is sometimes a specific goal of conservation management to boost a species' dispersal ability through habitat recreation, stepping-stones, or corridors. The observation that dispersal is more complicated than previously thought is almost a truism, but one which nevertheless may account for difficulties in reaching consensus in conservation biology. For example, under-estimating the complexity of dispersal may explain why, despite evidence that linear links improve population viability in some studies (e.g. chipmunk populations in hedgerows, Henderson et al. 1985), the value of corridors linking protected areas remains controversial (Gonzalez et al. 1998; Simberloff et al. 1992). It may also explain failures when attempting to model population dynamics with field data from another population of the same species (Hanski and Thomas 1994). Conservation programmes increasingly involve the translocation of animals to reinforce failing populations (once, of course, the cause of initial decline has been identified and stopped). Indeed, re-introduction could be necessary (and hopefully would still be possible) if all of these other measures have failed. This requires reliable information about dispersal. Previous re-introductions have often been poorly managed, and there is a lack of scientific follow-up studies to learn from mistakes, improve methods, and detail dispersal behaviour (Sarrazin and Barbault 1996). Settlement behaviour in particular can be critical to success and can be directly assisted if understood (Sarrazin et al. 1996). Introduction experiments offer a controlled environment in which to identify key features of settlement behaviour (Massot et al. 1994b); however, the conditions of departure from the original territory and of the transient stage during dispersal are far harder to investigate. Nevertheless, conservation programmes could be doing more to establish specific hypothesis-testing research as part of their reintroduction schemes in order to answer questions about, among other things, dispersal (Sarrazin and Barbault 1996). There are of course many ways in which knowledge of dispersal might potentially be applied to practical action in conservation. Two examples strike us as particularly relevant to the discussion within this book, and are discussed below. Models of metapopulation extinctions regard populations as essentially approximating to some variant of the Ideal Free Distribution (IFD) (see Holt and Barfield, chapter 6). However, it has been suggested, also, that individuals may aggregate in relatively large numbers within a metapopulation, using conspecifics as a cue of habitat quality (Smith and Peacock 1990). This asymmetry in distribution, with a component independent of the IFD, may increase the likelihood of extinction of some populations insofar as "eggs" are distributed between fewer `baskets', representing higher losses and population extinctions as well as fewer sources for dispersal (e.g. Ray et al. 1991). However, Danchin et al. (chapter 17) suggest that, where settlement is problematic, conservationists may be able to employ this phenomenon, perhaps with strategically placed dummies or captives, in order to encourage
Dispersal in theory and practice: consequences for conservation biology
aggregation in suitable habitats which can then be targeted for protection (Reed and Dobson 1993; Sarrazin et al. 1996). Dispersal has a critical influence on social species, and, reciprocally, sociality may influence dispersal and resulting gene flow. A mechanism of group formation common amongst social species is offspring philopatry (Ross, chapter 3). There is a possibility that social species with particularly strong philopatric tendencies might be extinction-prone under modern circumstances. For example, Krutchenkova, Goltsman and Macdonald (unpublished manuscript) have described an island population of Arctic Foxes amongst which dispersal is a rarity, with the dangerous consequence that young adults remain in their natal range, exposed to social suppression of reproduction, while neighbouring territories lie vacant. The African Hunting Dog has a declining distribution across Africa, and efforts by conservation biologists to reintroduce them have been hindered by, in addition to many logistical factors, complex social factors apparently stemming from rather specific social prerequisites (Woodroffe, Ginsberg and Macdonald, 1998). An understanding of dispersal in this context may be crucial for their conservation. Indeed, in southern Africa, `metapopulation management' in the form of translocating wild dogs between reserves may be a fruitful conservation tool (Mills et al. 1998), and such interventive management mimicking dispersal may increasingly, if tragically, be necessary as a last resort in human-dominated landscapes (Macdonald et al. 2000a, Woodroffe and Ginsberg, 2000).
Some caveats for future interpretation of dispersal The significance of various parameters to a population's viability is contingent upon the interaction between its life history, its competitors, and the landscape. In particular, dispersal plays a different role in the population processes of small restricted-range species from that in large wide-ranging ones. In the modern landscape, long journeys are highly likely to involve encounters with Human Development
, and such encounters generally involve risk. Thus, it is notable, perhaps, that larger species at the top of the food chain are typically in greatest danger of extinction and often disappear first (Woodroffe in press). This may provide some clues to the importance of spatial scale as a factor in dispersal success and survival probability (Wiens, chapter 7). While habitat loss is continuous, fragmentation is a `threshold phenomenon', occurring specifically at certain critical percentages of loss (Wiens, chapter 7); this will differentially affect species according to the fractal dimension relevant to their movements. Thus, we must be cautious in interpreting at what stage differently mobile species are in danger of isolation and at what point dispersal really becomes the limiting factor for species persistence. For example, it has been difficult to predict with accuracy the values of certain threshold population sizes, below which populations have been shown in models to rapidly decline (Lamberson et al. 1992), but also within which a small number of dispersers can drastically change the genetic structure of populations (Whitlock, chapter 19). As well as identifying conservation priorities, it may sadly become necessary to avoid squandering effort in attempting the conservation of lost causes. Hanski (chapter 20) reminds us that some populations exist in unsuitable habitats as sinks, simply because immigration rates exceed those of death and emigration (Harrison et al. 1988; Paradis 1995). Are such populations worth conserving if they could not survive once in isolation? They could be sources for re-colonization of temporarily vacated source habitats (Thomas and Jones 1993), and could be deliberately used as such if maintained as zoo populations for captive breeding. Even if reintroduction were not possible, however, when all else had failed, there would seem
David W. Macdonald and Dominic D.P. Johnson
to be great virtue in a living museum, who would not wish now for a bank of dodos or thylacines in captivity? Nevertheless, it remains important to discover which wild populations should be prioritized for investment. More generally, three further considerations may be of note in the context of dispersal and conservation. Hamilton and May (1977) found that the ESS dispersal rate was 0.5 (due to kin competition effects) even when dispersal mortality was 100%. A wide range of further and extended models also reduced to the same underlying rate (Johnson and Gaines 1990). Thus, if these models are correct, high dispersal mortality observed in nature may not necessarily be construed as a cause for concern. Secondly, although the literature on inbreeding and gene flow and the consequences for conservation is vast, there is little evidence from wild populations of demographic decline due to inbreeding (Caro and Laurenson 1994). This may, however, be partly due to lack of detailed data from populations with which to test for this; such evidence that has emerged for increasing extinction risk with inbreeding are from detailed, long-term study populations (Saccheri et al. 1998; Westemeier et al. 1998). Thirdly, while there is a clear emphasis on habitat destruction and fragmentation, some species may have distributions at least partly defined by competitive exclusion (Hassell et al. 1994; see also Mouquet et al., chapter 22). A thorough knowledge of the whole community, therefore, may be necessary to predict species dynamics where other species, and especially predators, have already become extinct.
What is the likely course of dispersal evolution in altered environments? Very generally speaking, according to models, spatial variation selects against dispersal, while temporal variation predicts selection for dispersal (Johnson and Gaines 1990 and references therein). Thus, habitats undergoing change with man's activities should encourage dispersal, while the resulting static mosaic should decrease it. However, while this may be true for flexible dispersers, habitat fragmentation presumably halts the dispersal of `cautious' species, alluded to before (see Wiens, chapter 7). Woodroffe (in press) similarly points out that, amongst carnivores, the Nee-May model predicts that good dispersers should do well in fragmented habitats. A high rate of dispersal is not always a panacea for fragmented metapopulation persistence. Too high a rate of dispersal might result in the source itself being deprived of a viable population. For this reason, and due to the increased mortality of dispersers in more fragmented habitats and the consequently increased metapopulation extinction probability (Hanski, chapter 20), one may predict the evolution of short dispersal in endangered populations. Indeed, Thomas and Jones (1993) showed that more butterflies disperse out of small patches (with long edges), with the result that populations living on small patches tended to go extinct purely through dispersal behaviour. Conversely, one may also expect the opposite possibility of selection for greater dispersal to areas where fitness is higher and thus for selection of longer dispersal distances as suitable habitat patches become ever more distant (Hanski, chapter 20). In the latter case, those who disperse furthest will have a disproportionate influence on the survival of metapopulations and gene flow (Wiens, chapter 7). Roff and Fairbairn (chapter 13) review evidence that many physiological and behavioural traits defining migratory propensity are at a level of heritability enabling rapid response to selection pressure. They also report that different migratory and fitness traits are genetically correlated; thus, selection on one may have considerable pleiotropic effects on the species' life history and behaviour (see the experimental references in Roff and Fairbairn, chapter 13).
Dispersal in theory and practice: consequences for conservation biology
Thus, the evolution of dispersal in altered habitats seems both to be feasible and to have important potential consequences (is it also possible that selection might simply favour behavioural flexibility?). Nevertheless, Hanski (chapter 20) tells us that empirical studies on the evolution of dispersal rates in areas experiencing increasing fragmentation are "practically non-existent", and theoretical work also remains to be done in this area. One can only, therefore, make a plea to fill this most important of gaps in our knowledge. What if we were to succeed in developing models and empirical results for `model' populations (Stenseth and Lidicker 1992), only to find that, when applied to endangered animals, their long exposure to new selection pressures in an anthropogenic landscape has altered their dispersal behaviour?
Conclusions: crisis management
? Several authors have observed that a comprehensive review on dispersal is no longer a practical possibility because of the enormity of relevant literature. Yet, as has been repeatedly pointed out in this book, there are still large gaps in our knowledge, and disappointingly few available estimates of dispersal variables from natural populations. The contributions to this book have both pointed out these gaps and made significant efforts to fill them, or at least pointed the way for future studies to do so. Academic and conservation-based interest in the subject, because of the increasingly critical state of the latter, should be combined. `The sixth great extinction spasm of geological time is upon us' (Wilson 1992: 327); furthering our understanding of dispersal is by no means an idle concern. Future work for conservation can focus on three principle angles of attack: detailed field studies of metapopulations (Hanski et al. 1995a), experimental manipulations which can control for the biases of field studies and allow detailed hypothesis-testing (Massot et al. 1994b; Sarrazin and Barbault 1996), and, finally, detailed individual process-based modelling (Macdonald and Rushton unpublished manuscript). Conservation has two principle problems: firstly, reducing deterministic habitat destruction by humans and contamination leading to reduced species populations and, secondly, minimizing the effects of stochastic processes on these resulting small and isolated populations. Dispersal is particularly crucial to determining a species' ability to cope with both of these processes. Dispersal may be the last saving grace sustaining dwindling populations above a threshold under which demographic or genetic stochastic events become lethal. Conservation often involves active management, and in particular habitat restoration and damage minimization. These types of management activities often tend to be on a large scale, partly because of national environmental policy, and therefore are undertaken without time or opportunities for experimentation, yet with considerable costs attached to making mistakes (Doak and Mills 1994). So, despite the current paucity of parameter estimates, once equipped with reliable models, in the future it might be possible to confine some of these mistakes to the virtual reality of the computer, thereby better informing decisions that are required in the field. Because conservation decisions ultimately have to be enacted in real landscapes, process-based models conducted in geographic information systems
provide an arena for potentially powerful predictive analyses. Obviously, models cannot precisely simulate the behaviour of the model species. Indeed, models are only models, and provide projections rather than predictions, so we do not expect even the best of them to capture every nuance of behaviour that could be observed in the field. Rather, we would judge them useful if they approximated reality sufficiently to help identify the main parameters that decisionmakers must bear in mind, and on which conservationists might focus attention (Verboom et
David W. Macdonald and Dominic D.P. Johnson
al. 1993). One can cautiously (Mills et al. 1999) examine, through SENSITIVITY ANALYSIS
, specifically how dispersal ranks in importance as a predictor of population size or persistence and thus implement management of the key variable responsible for decline. Accurate parameterization remains a vital prerequisite for all these models, so reliable results are still limited by poor field data. The plight of endangered species poses a thrilling impetus for dispersal modelling, offering the chance for researchers to make a significant contribution. With that opportunity comes responsibility: in the practical arena, models face a daunting test, insofar as the cost of misjudgements may be measured in extinctions.
Acknowledgements. We are grateful to Drs. W. Cresswell, J. McNutt, and R. Woodroffe for helpful comments on an earlier draft, and to Jean Clobert for his invitation to attend the Roscoff workshop which led to this review.
Dispersal in theory and practice: consequences for conservation biology References Balmford, A. & Long, A. (1994) Avian endemism and forest loss. Nature 372: 623-624. Barreto, G. R., Macdonald, D. W. & Strachan, R. (1998b) The tightrope hypothesis: an explanation for plummeting water vole numbers in the Thames catchment. United Kingdom Floodplains. Westbury, pp. 312-327. Barreto, G. R., Rushton, S. P., Strachan, R. & Macdonald, D. W. (1998a) The role of habitat and mink predation in determining the status and distribution of declining populations of water voles in England. Animal Conservation 1: 129-137. Berger, J. (1990) Persistence of different-sized populations: an empirical assessment of rapid extinctions in bighorn sheep. Conservation Biology 4: 91-98. Berger, J. (1999) Intervention and persistence in small populations of bighorn sheep. Conservation Biology 13: 432-435. Boulinier, T. & Danchin, E. (1997) The use of conspecific reproductive success for breeding patch selection in terrestrial migratory species. Evolutionary Ecology 11: 505-517. Boyce, M. S. (1992) Population viability analysis. Annual Review of Ecology & Systematics 23: 481-506. Bright, P. W. & Morris, P. A. (1996) Why are dormice rare? A case study
in conservation biology. Mammal Review 26: 157-187. Brown, J. H. & Kodrick-Brown, A. (1977) turnover rates
in insular biogeography: effect of immigration on extinction. Ecology 58: 445-449. Caro, T. (1999) The behaviour-conservation interface. Trends in Ecology and Evolution 14: 366-369. Caro, T. M. & Laurenson, M. K. (1994) Ecological and genetic factors in conservation: a cautionary tale. Science 263: 485-486. Courchamp, F., Clutton-brock, T. H. & Grenfell, B. (1999) Inverse density dependence and the Allee effect. Trends in Ecology and Evolution 14: 405-410. Danchin, E. & Wagner, R. H. (1997) The evolution of coloniality: the emergence of new perspectives. Trends in Ecology and Evolution 12: 342-347. Davis, M. B. (1989) Lags in vegetation response to greenhouse warming. climatic change
15: 75-82. Doak, D. F. & Mills, L. S. (1994) A useful role for theory in conservation. Ecology 75: 615626. Dobson, F. S. (1985) Mulitple causes of dispersal. American Naturalist 126: 855-858. Endler, J. (1977) Geographic variation, speciation, and clines. Princeton University
Press, Princeton, N.J. Frank, L. & Woodroffe, R. B. (in press) Behaviour of carnivores in exploited and controlled populations. In: (J.L.Gittleman., R. K. Wayne, D. W. Macdonald & S. Funk) (ed) Carnivore conservation. Cambridge University Press, Cambridge. Gonzalez, A., Lawton, J. H., Gilbert, F. S., Blackburn, T. M. & Evans-Freke, I. (1998) Metapopulation dynamics, abundance and distribution in a microecosystem. Science 281: 2045-2047. Hamilton, W. D. & May, R. M. (1977) Dispersal in stable habitats. Nature 269: 578-581. Hanksi, I. (1989) Metapopulation dynamics: does it help to have more of the same. Trends in Ecology and Evolution 4: 113-114.
David W. Macdonald and Dominic D.P. Johnson Hanksi, I., Pakkala, T., Kuussaari, M. & Lei, G. (1995) Metapopulation persistence of an endangered butterfly in a fragmented landscape. Oikos 72: 21-28. Hanksi, I. & Thomas, C. D. (1994) Metapopulation dynamics and conservation: a spatially explicit model applied to butterflies. Biological Conservation 68: 167-180. Hanski, I. (1999) Metapopulation Ecology. Oxford University Press, Oxford. Hanski, I. & Gilpin, M. (1991) Metapopulation dynamics: a brief history and conceptual domain. Biological Journal of the Linnean Society 42: 3-16. Hanski, I. & Gilpin, M. E. (1997) Metapopulation Biology: Ecology, Genetics and Evolution. Academic Press, San Diego. Harrison, S. (1991) Local extinction in a metapopulation context: an empirical evaluation. Biological Journal of the Linnean Society 42: 73-88. Harrison, S., Murphy, D. D. & Ehrlich, P. R. (1988) Distribution of the bay checkerspot butterfly Eukphydryas editha bayensis: evidence for a metapopulation model. American Naturalist 132: 360-382. Hassell, M. P., Comins, H. N. & May, R. M. (1994) Species coexistence and self-organizing spatial dynamics. Nature 370: 290-292. Henderson, M. T., Merriam, G. & Wegner, J. (1985) Patchy environments and species survival: chipmunks in an agricultural mosaic. Biological Conservation 31: 95-105. Hersteinsson, P. (1992) Mammals of the Thingvallavatn area. Oikos 64: 396-404. Ims, R. A. & Stenseth, N. C. (1989) Divided the fruitflies fall. Nature 342: 21-22. Johnson, D. D. P., Hay, S. I. & Rogers, D. J. (1998) Contemporary environmental correlates of endemic bird areas derived from meteorological satellite sensors. Proceedings of the Royal Society
, Series B 265: 951-959. Johnson, D. D. P. & Mighell, J. S. (1999) Dry-season bird diversity in tropical rainforest and surrounding habitats in North-east Australia. Emu 99: 108-120. Johnson, M. L. & Gaines, M. S. (1990) Evolution of dispersal: theoretical models and empirical tests using birds and mammals. Annual Review of Ecology and Systematics 21: 449-480. Lamberson, R. H., McKelvey, R., Noon, B. R. & Voss, C. (1992) A dynamic analysis of northern spotted owl viability in a fragmented forest landscape. Conservation Biology 6: 505-512. Lande, R. (1988) Demographic models of the northern spotted owl (Strix occidentalis caurina). Oecologia 75: 601-607. Lande, R., Engen, S. & Sжther, B. (1999) Spatial scale of population synchrony: environmental correlation versus dispersal and density regulation. American Naturalist 154: 271-281. Levins, R. (1970) Extinction. In: (M. Gustenhaver) (ed) Some Mathematical Questions in Biology. American Mathematical Society
, Providence, Rhode Island. pp. 77-107. Lima, S. L. & Zollner, P. A. (1996) Towards a behavioural ecology of ecological landscapes. Trends in Ecology and Evolution 11: 131-135. Lloyd, H. G. (1980) The red fox. Batsford, London. Macdonald, D. W. (1984) Running with the fox. George Allen & Unwin, London. Macdonald, D. W. (1995) Unresolved questions for the control of wildlife rabies: social perturbation and interspecific interactionsRabies in a Changing World. Proceedings of the British Small Animal Veterinary Association, Cheltenham, UK. pp. 33-48. Macdonald, D. W., Mace, G. & Rushton, S. (1998) Proposals for the future monitoring British mammals: Department of the Environment, Transport and the Regions.
Dispersal in theory and practice: consequences for conservation biology Macdonald, D. W., Mace, G. & Rushton, S. P. (2000) Conserving British mammals: is there a radical future? In: (A. Entwhistle & N. Dunstone) (ed) Priorities for conservation of mammalian biodiversity: has the panda had its day? Cambridge University Press, Cambridge. Macdonald, D. W., Mace, G. M. & Barreto, G. R. (1999) The effects of predators on fragmented prey populations: a case study for the conservation of endangered prey. Journal of Zoology (London) 247: 487-506. Macdonald, D. W. & Smith, H. (1990) Dispersal, dispersion and conservation in the agricultural ecosystem. In: (R. G. H. Bunce & D. C. Howard) (ed) Species dispersal in agricultural habitats. Belhaven Press, pp. 18-64. Macdonald, D. W. & Smith, H. (1991) New perspectives on agro-ecology: between theory and practice in the agricultural ecosystem. In: (L. G. Firbank, N. Carter, J. F. Darbyshire & G. R. Potts) (ed) The Ecology of Temperate Cereal Fields: 32nd symposium of the British Ecological Society. Blackwell Scientific Publications, pp. 413-448. Macdonald, D. W. & Strachan, R. (1999) The mink and the water vole: analyses for conservation. Oxford: Wildlife Conservation Research Unit, University of Oxford. Macdonald, D. W., Tattersall, F. H., Rushton, S. P., South, A. B., Rao, S., Maitland, P. & Strachan, R. (In press) Reintroducing the beaver (Castor fiber) to Scotland: a protocol for identifying and assessing suitable release sites. Animal Conservation . Massot, M., Clobert, J., Lecomte, J. & Barbault, R. (1994) Incumbent advantage in common lizards and their colonizing ability. Journal of Animal Ecology 63: 431-440. McNutt, J. W. (1996) Sex-biased dispersal in African wild dogs (Lycaon pictus). Animal Behaviour 52: 1067-1077. Mills, M. G. L., Ellis, S., Woodroffe, R., Maddock, A., Stander, P., Rasmussen, G., Pole, A., Fletcher, P., Bruford, M., Wildt, D., Macdonald, D. W. & Seal, U. (1998) Population and habitat viability assessment for the African wild dog (Lycaon pictus) in Southern Africa. Final Workshop Report. IUCN/SSC Conservation Breeding Specialist Group, Apple Valley, MN, USA. Mills, S., Doak, D. F. & Wisdom, M. J. (1999) Reliability of conservation actions based on elasticity analysis of matrix models. Conservation Biology 13: 815-829. Mooij, W. M. & DeAngelis, D. L. (1998) Error propogation in spatially explicit population models: a reassessment. Conservation Biology 13: 930-933. North, P. M. (1988) A brief review of the (lack of) statistics of bird dispersal. Acta Ornithologica 24: 63-74. Paradis, E. (1995) Survival, immigration and habitat quality in the Mediterranean pine vole. Journal of Animal Ecology 64: 579-591. Primack, R. B. & Miao, S. L. (1992) Dispersal can limit local plant distribution. Conservation Biology 6: 513-519. Ray, C., Gilpin, M. & Smith, A. T. (1991) The effect of conspecific attraction on metapopulation dynamics. Biological Journal of the Linnean Society 42: 123-134. Reed, J. M. (1999) The role of behaviour in recent avian extinctions and endangerments. Conservation biology 13: 232-241. Reed, M. J. & Dobson, A. P. (1993) Behavioural constraints and conservation biology: conspecific attraction and recruitment. Trends in Ecology and Evolution 8: 253-256. Rushton, S. P., Barreto, G. W., Cormack, R. M., Macdonald, D. W. & Fuller, R. (In press) Modelling the effects of mink and habitat fragmentation on the water vole. Journal of Applied Ecology .
David W. Macdonald and Dominic D.P. Johnson Saccheri, I., Kuussaari, M., Kankare, M., Vikman, P., Fortelius, W. & Hanksi, I. (1998) Inbreeding and extinction in a butterfly metapopulation. Nature 392: 491-494. Sarrazin, F., Bagnolini, C., Pinna, J. L. & Danchin, E. (1996) Breeding biology during establishment of a reintroduced Griffon vulture Gyps fulvus population. Ibis 138: 315325. Sarrazin, F. & Barbault, R. (1996) Reintroduction: challenges and lessons for basic ecology. Trends in Ecology and Evolution 11: 474-478. Sillero-Zubiri, C., Gottelli, D. & Macdonald, D. W. (1996) Male philopatry, extra-pack copulations and inbreeding avoidance in Ethiopian Wolves (Canis simensis). Behavioural Ecology & Sociobiology 38: 331-340. Simberloff, D. (1988) The contribution of population and community biology to conservation science. Annual Review of Ecology and Systematics 19: 473-511. Simberloff, D., Farr, J. A., Cox, J. & Moehlman, D. W. (1992) Movement corridors: conservation bargains or poor investments. Conservation Biology 6: 493-504. Smith, A. T. & Peacock, M. M. (1990) Conspecific attraction and the determination of metapopulation colonization rates. Biological conservation 4: 320-323. South, A. B., Rushton, S. P. & Macdonald, D. W. (2000) Simulating the proposed reintroduction of the European beaver (Castor fiber) to Scotland. Biological Conservation 93: 103-116. Stamps, J. A. (1988) Conspecific attraction and aggregation in territorial species. American Naturalist 131: 329-347. Stattersfield, A. J., Crosby, M. J., Long, A. J. & Wege, D. C. (1998) Endemic bird areas of the world: priorities for biodiversity conservation. Birdlife International, Cambridge. Stenseth, N. C. & Lidicker, W. Z. (1992) animal dispersal
: Small Mammals as a Model. Chapman & Hall, London. Stephens, P. A. & Sutherland, W. J. (1999) Consequences of the Allee effect for behaviour, ecology and conservation. Trends in Ecology and Evolution 14: 401-405. Swinton, J., Tuyttens, F. A. M., Macdonald, D. W. & Cheeseman, C. L. (1997) Social perturbation and bovine tuberculosis in badgers: fertility control and lethal control compared. Philosophical Transactions of the Royal Society of London 352: 619-631. Taylor, A. D. (1990) Metapopulations, dispersal and predator-prey dynamics: an overview. Ecology 71: 429-436. Taylor, A. D. (1991) Studying metapopulation effects in predator-prey systems. Biological Journal of the Linnean Society 42: 305-323. Thomas, C. D. & Jones, T. M. (1993) Partial recovery of a skipper butterfly (Herspera comma) from population refuges: lessons for conservation in a fragmented landscape. Journal of Animal Ecology 62: 472-481. Tuyttens, F. A. M. & Macdonald, D. W. (1998) Sterilization as an alternative strategy to control wildlife diseases: bovine tuberculosis in European badgers as a case study. Biodiversity & Conservation 7: 705-723. Tuyttens, F. A. M. & Macdonald, D. W. (2000) Consequences of social perturbation for wildlife management and conservation. In: (L. M. Gosling & W. J. Sutherland) (ed) Behaviour and Conservation. Cambridge University Press, Cambridge. pp. 315-329. van Vuren, D. & Armitage, K. B. (1994) Survival of dispersing and philopatric yellow-bellied marmots - what is the cost of dispersal. Oikos 69: 179-181. Verboom, J., Metz, J. A. J. & Meelis, E. (1993) Metapopulation models for impact assessment of fragmentation. IALE Studies in Landscape Ecology 1: 172-191.
Dispersal in theory and practice: consequences for conservation biology Verhulst, S., Perrins, C. M. & Riddington, R. (1997) Natal dispersal of great tits in a patchy environment. Ecology 78: 864-872. Vyas, N. B., Kuenzel, W. J. & Hill, E. F. (1995) Acephate affects migratory orientation of the White-throated Sparrow
(Zonotrichia albicollis). Environmental Toxicology and Chemistry 14: 1961-1965. Wehausen, J. D. (1999) Rapid extinction of mountain sheep populations revisited. Conservation Biology 13: 378-384. Westemeier, R. L., Brawn, J. D., Simpson, S. A., Esker, T. L., Jansen, R. W., Walk, J. W., Kershner, E. L., Bouzat, J. L. & Paige, K. N. (1998) Tracking the long-term decline and recovery of an isolated population. Science 282: 1695-1698. Whitmore, T. C. (1997) Tropical rainforest disturbance, disappearance and species loss. In: (W. F. Laurence & R. O. Bierregaard) (ed) Tropical Forest Remnants: Ecology, Management and Conservation of Fragmented Communities. Chicago University Press, Chicago. pp. 3-12. Wiens, J. A. (1994) Habitat fragmentation: island v landscape perspectives on bird conservation. The Ibis 137: 97-104. Wilson, E. O. (1992) The Diversity of Life. Penguin, London. Woodroffe, R. & Ginsberg, J. (1998) Edge effects and the extinction of populations inside protected areas. Science 280: 2126-2128. Woodroffe, R. & Ginsberg, J. (2000) Ranging behaviour and vulnerability to extinction in carnivores. In: (L. M. Gosling & W. J. Sutherland) (ed) Behaviour and Conservation. Cambridge University Press, Camrbidge. pp. 125-140. Woodroffe, R. B. (in press) Strategies for carnivore conservation: lessons from contemporary extinctions. In: (J.L.Gittleman., R. K. Wayne, D. W. Macdonald & S. Funk) (ed) Carnivore conservation. Cambridge University Press, Cambridge. Woodroffe, R. B., Ginsberg, J. R. & Macdonald, D. W. (1997) The African wild dog: status, survey and conservation action plan. IUCN/SSC Canid Specialist Group, Gland, Switzerland.