Integrated empirical approaches to better understand species’ range limits

Plants encompass tremendous evolutionary diversity associated with their radiation into almost all habitats on Earth. Striking variation in growth form, life history, and reproductive system is apparent in many plant groups, even among closely related species. Yet almost all species exhibit evolutionary stasis in the limits to their geographic distributions (Sexton et al., 2009). Why don't species continually expand their distributions through incremental local adaptation to conditions at the range edge? What evolutionary constraints account for the striking variation in range size and location within clades? These questions have been brought into sharper focus by concern over how species' distributions will be altered by climate change, habitat fragmentation, and other human impacts. Plants have played a dominant role in recent progress understanding range limits because their distributions are often well known from herbarium records, fitness components can be measured in the field, and they can be experimentally manipulated. In his seminal book, The Structure and Dynamics of Geographic Ranges, Gaston (2003, p. 4) noted that interest in range limits "derives from many quarters" involving diverse perspectives and tools often applied in isolation. Since then, it is the integration of multiple approaches that is producing the most exciting work exploring the question "Why do species have stable range limits?" (Table 1). Fitness increased toward N limit and beyond Negligible local adaptation across N portion of range, and edge populations not better adapted to habitat beyond limit Little gene flow among populations towards limit Neutral diversity declined close to limit Genetic diversity declined toward limits Gene flow not high enough to limit adaptation of edge populations Genetic variation declined but bottlenecks not more frequent toward limit Asymmetric gene flow from center to edge Habitat suitability declined toward and beyond limits But unoccupied suitable habitat at N limit suggests role for dispersal limitation λ of transplanted population > 1 beyond N limit ENMs did not predict transplanted population performance Niche-limited S edge but evolutionary constraints on expansion unclear Dispersal-limited N edge but constraints on dispersal unclear No decline in genetic diversity toward limits, except right at range edges No evidence of asymmetric gene flow from center to edge A species is widely expected to occur where environmental conditions fall within its realized niche, that is, where individual fitness is high enough that populations are self-replacing. Range limits may also arise when a constraint on dispersal prevents a species from colonizing or maintaining itself in suitable habitat. While niche and dispersal limitation are not mutually exclusive, their relative importance has been a major focus of research investigating distributional limits (Sexton et al., 2009; Hargreaves et al., 2014). If a species' distribution is at least partially niche-limited, there are two main explanations for why that species might not adapt to conditions beyond the range. First, it may lack genetic variation for traits or trait combinations that enable persistence beyond the range (genetic constraint hypothesis), which is likely if the range limit coincided with steep changes in abiotic or biotic factors that affect fitness. Additional constraints could arise if range-edge populations are small and isolated because new advantageous mutations will likely be lost by drift before they are favored by selection and because drift may fix deleterious alleles more frequently, further reducing fitness (Willi and Van Buskirk, 2019). Second, adaptation at the range edge may be stymied by gene flow from the range core. If core populations are more productive than peripheral populations and if individual movement is random with respect to habitat quality (as in most plants), then genes will tend to flow from core to edge. In this case, selection in edge populations may favor alleles that enhance persistence at, and perhaps beyond, the range edge, but these populations are held away from local optima by the influx of maladapted alleles from core populations (gene swamping hypothesis). Dispersal can limit ranges in three ways. First, dispersal beyond the range may be prevented by some physical barrier. However, for many species, there is no obvious barrier, and suitable habitat seems available and accessible beyond the range. Second, range limits can occur on heterogeneous landscapes if dispersal into unoccupied habitat patches fails to keep pace with extinction from occupied patches (Samis and Eckert, 2009). Finally, recent abiotic or biotic environmental change may create suitable habitat beyond the range that a species has not yet had time to colonize. Such range-edge disequilibrium might become more frequent owing to anthropogenic climate change. Both theoretical and empirical work has advanced our understanding of range limits, but here we focus on four main empirical tools that have been profitably used, largely independently, to explore the ecological and evolutionary limits to species' ranges. If ranges occur on environmental gradients and are niche-limited, individual fitness and therefore demographic vital rates and abundance may be highest in the range core and decline toward the edges (abundant center hypothesis). This prediction has been tested for many plant species with mixed results (Pironon et al., 2017), possibly because most tests do not use comprehensive measures of population viability (e.g., population growth rate, λ). In addition, range limits might coincide with abrupt environmental changes, so the absence of a gradual decline in demographic parameters toward the edge would not rule out niche-limitation. Ecological niche models correlate species occurrences with environmental data to define suitable habitat, which can then be used to test whether there is suitable habitat beyond the range (Angert et al., 2018). If so, the range may be dispersal-limited; if not, niche limitation is likely. However, ENMs will overestimate the potential range if they do not account for important biotic interactions or if edge populations are sinks, requiring integration with demographic studies (Eckhart et al., 2011) or transplant experiments (Bayly and Angert, 2019). Comparing populations transplanted at the range edge with those transplanted beyond is the most direct test of niche vs. dispersal limitation, especially if experiments are well replicated over space and time and measure lifetime fitness in unmanipulated habitat. Transplants have yielded key insights; for example, elevational ranges seem commonly niche-limited while geographic ranges much less so (Hargreaves et al., 2014). Including multiple "source" populations can reveal the degree of local adaptation across the range, though this is rarely done. Moreover, the processes underlying fitness variation are difficult to identify without knowing how genetic variation and gene flow vary across the range (Samis et al., 2016; Willi and Van Buskirk, 2019). Genetic analyses can reveal variation in population demography over much longer time scales than observational or experimental studies and can test whether genetic variation or gene flow vary toward the range edge. Genetic diversity within populations often declines toward range edges (Pironon et al., 2017). Combined with demographic surveys, these studies can distinguish contemporary population dynamics that might limit ranges from historical processes such as postglacial colonization. However, almost all genetic studies assay putatively neutral polymorphisms rather than quantitative genetic variation, which would more directly reflect evolutionary constraints (Paccard et al., 2016) Three case studies illustrate the importance of integrating empirical approaches (Table 1: details, references, and additional case studies). Combining tools has identified genetic constraints on range expansion in the outcrossing annual Clarkia xantiana subsp. xantiana (Onagraceae). Ecological niche models have suggested that habitat suitability declines toward and beyond the eastern range limit, and niche limitation was further supported by a parallel decline in population growth rate and very poor performance of transplants beyond the eastern limit. Putatively neutral genetic variation also declined toward the range edge and estimated gene flow was asymmetric, suggesting a role for gene swamping. Direct analyses of quantitative genetic variation in ecologically relevant traits, however, presented a different picture. Genetic variation for individual traits did not decline toward the limit, and populations were phenotypically differentiated from each other, indicating that gene flow has not prevented adaptation. Analysis of selection agents at the range edge suggested that range expansion would probably require an unlikely multi-trait response to selection exerted by abiotic and biotic challenges including insufficient pollination and heavy herbivory. Despite examples like C. xantiana where distributional limits seem to reflect niche limits, populations planted beyond geographic limits usually enjoy fitness high enough to ensure self-replacement, suggesting dispersal limitation (Hargreaves et al., 2014). Such is the case with the dune endemic Camissoniopsis cheiranthifolia (Onagraceae). Abundance, plant size, and reproductive output did not decline toward the northern range edge, and mean fitness of transplanted populations actually increased toward and beyond the limit. Single-generation transplant experiments may not reveal sporadic episodes of low fitness that knock a species range back to its observed limit, yet C. cheiranthifolia planted beyond the limit has persisted for at least a dozen generations (R. L. Cross and C. G. Eckert, unpublished data). If the range is dispersal-limited, gene flow into edge populations may be low, allowing local adaptation. Gene flow among northern populations of C. cheiranthifolia is indeed low, but there is little evidence for local adaptation across what appears to be substantial ecological variation toward the northern range edge. Although dispersal limitation seems likely, the underlying mechanism remains unclear. Individual species may exhibit range limits along multiple geographic and elevational gradients, yet studies investigating more than one limit in a single species are very rare. Ecological niche models combined with large-scale surveys of occurrence suggested that the perennial herb Erythranthe (Mimulus) cardinalis (Phrymaceae) is niche-limited at its southern limit but may be dispersal-limited at the northern edge. Accordingly, population growth rate increased toward the northern limit, and populations transplanted beyond exhibited positive growth. The genetic constraints on expansion at either range margin are less obvious. Combining common garden experiments and population genetics suggested that gene flow prevents populations from reaching their local optimal phenotype, but not more frequently at geographic range edges. Moreover, artificial selection unexpectedly revealed a stronger evolutionary response of reproductive phenology at the southern, apparently niche-limited, edge than at the range center or northern limit, emphasizing that different range edges may involve different constraints. What constrains species distributions has long been a core question in ecology and evolution and is increasingly relevant to conserving biodiversity in the face of rapid anthropogenic change. Therefore, a growing body of research has creatively integrated theoretical perspectives and empirical tools (Table 1). A better understanding of range limits will involve research on many fronts, including large-scale biogeographical and macroecological analyses (Gaston, 2003). However, the underlying ecological and evolutionary mechanisms will be revealed by multifaceted species-level case studies and their eventual synthesis. We thank Pamela Diggle for encouraging us to write this essay; Karen Samis, Sara Dart, Adriana Lopéz-Villalobos, Anna Hargreaves, and Dave Ensing for ideas and inspiration; Pamela Diggle, Alyson Van Natto, Anna Hargreaves, Jannice Friedman, and two anonymous reviewers for comments on the manuscript; and the Natural Sciences and Engineering Research Council of Canada for funding our work on range limits through a Canadian Graduate Scholarship to R.L.C. and Discovery Grants to C.G.E. R.L.C. and C.G.E. developed the concept for this paper and wrote and revised the manuscript.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.