Effects of Habitat Fragmentation on Effective Dispersal of Florida Scrub‐Jays

Abstract: Studies comparing dispersal in fragmented versus unfragmented landscapes show that habitat fragmentation alters the dispersal behavior of many species. We used two complementary approaches to explore Florida Scrub‐Jay (Aphelocoma c?rulescens) dispersal in relation to landscape fragmentation. First, we compared dispersal distances of color‐marked individuals in intensively monitored continuous and fragmented landscapes. Second, we estimated effective dispersal relative to the degree of fragmentation (as inferred from two landscape indexes: proportion of study site covered with Florida Scrub‐Jay habitat and mean distance to nearest habitat patch within each study site) by comparing genetic isolation‐by‐distance regressions among 13 study sites having a range of landscape structures. Among color‐banded individuals, dispersal distances were greater in fragmented versus continuous landscapes, a result consistent with other studies. Nevertheless, genetic analyses revealed that effective dispersal decreases as the proportion of habitat in the landscape decreases. These results suggest that although individual Florida Scrub‐Jays may disperse farther as fragmentation increases, those that do so are less successful as breeders than those that disperse short distances. Our study highlights the importance of combining observational data with genetic inferences when evaluating the complex biological and life‐history implications of dispersal.     

Consequences of ignoring dispersal variation in network models for landscape connectivity

Habitat loss and fragmentation can negatively influence population persistence and biodiversity, but the effects can be mitigated if species successfully disperse between isolated habitat patches. Network models are the primary tool for quantifying landscape connectivity, yet in practice, an overly simplistic view of species dispersal is applied. These models often ignore individual variation in dispersal ability under the assumption that all individuals move the same fixed distance with equal probability. We developed a modeling approach to address this problem. We incorporated dispersal kernels into network models to determine how individual variation in dispersal alters understanding of landscape-level connectivity and implemented our approach on a fragmented grassland landscape in Minnesota. Ignoring dispersal variation consistently overestimated a population's robustness to local extinctions and underestimated its robustness to local habitat loss. Furthermore, a simplified view of dispersal underestimated the amount of habitat substructure for small populations but overestimated habitat substructure for large populations. Our results demonstrate that considering biologically realistic dispersal alters understanding of landscape connectivity in ecological theory and conservation practice.     

Open and closed seascapes: where does habitat patchiness create populations with high fractions of self-recruitment?

Which populations are replenished primarily by immigrants (open) and which by local production (closed) remains an important question for management with implications for response to exploitation, protection, and disturbance. However, we lack methods for predicting population openness. Here, we develop a model for openness and show that considering habitat isolation explains the existence of surprisingly closed populations in high-dispersal species, including many marine organisms. Relatively closed populations are expected when patch spacing is more than twice the standard deviation of a species'. dispersal kernel. In addition, natural scales of habitat patchiness on coral reefs are sufficient to create both largely open and largely closed populations. Contrary to some previous interpretations, largely closed marine populations do not require mean dispersal distances that are unusually short, even for species with relatively long pelagic larval durations. We predict that habitat patchiness has strong control over population openness for many marine and terrestrial species with a highly dispersive life stage and relatively sedentary adults. This information can be used to make initial predictions about where populations will be more or less resilient to local exploitation and disturbance.     

Implications of Macroalgal Isolation by Distance for Networks of Marine Protected Areas

The global extent of macroalgal forests is declining, greatly affecting marine biodiversity at broad scales through the effects macroalgae have on ecosystem processes, habitat provision, and food web support. Networks of marine protected areas comprise one potential tool that may safeguard gene flow among macroalgal populations in the face of increasing population fragmentation caused by pollution, habitat modification, climate change, algal harvesting, trophic cascades, and other anthropogenic stressors. Optimal design of protected area networks requires knowledge of effective dispersal distances for a range of macroalgae. We conducted a global meta‐analysis based on data in the published literature to determine the generality of relation between genetic differentiation and geographic distance among macroalgal populations. We also examined whether spatial genetic variation differed significantly with respect to higher taxon, life history, and habitat characteristics. We found clear evidence of population isolation by distance across a multitude of macroalgal species. Genetic and geographic distance were positively correlated across 49 studies; a modal distance of 50–100 km maintained F_ST < 0.2. This relation was consistent for all algal divisions, life cycles, habitats, and molecular marker classes investigated. Incorporating knowledge of the spatial scales of gene flow into the design of marine protected area networks will help moderate anthropogenic increases in population isolation and inbreeding and contribute to the resilience of macroalgal forests. Implicaciones del Aislamiento por Distancia de Macroalgas para Redes de Áreas Marinas Protegidas     

Managing Boreal Forest Landscapes for Flying Squirrels

Abstract: Flying squirrel (Pteromys volans) populations have declined severely during the past few decades, and the species has become a focal species in forest management and the conservation debate in Finland. We compared landscape structure around known flying squirrel home ranges with randomly chosen forest sites to determine which landscape patterns characterize the areas occupied by the species in northern Finland. We sought to identify the key characteristics of the landscape that support the remaining flying squirrel populations. We analyzed landscape structure within circular areas with 1- and 3-km radii around 63 forest sites occupied by flying squirrels, and around 96 random sites. We applied stepwise analysis of the landscape structure where landscapes were built up step-by-step by adding patch types in order of their suitability for the flying squirrel. The land-use and forest-resource data for the analysis were derived from multisource national forest inventory and imported to a geographical information system. Landscape patch types were divided into three suitability categories: breeding habitat (mixed spruce-deciduous forests); dispersal habitat ( pine and young forests); and unsuitable habitat ( young sapling stands, open habitats, water). Flying squirrel landscapes contained more suitable breeding habitat patches and were better connected by dispersal habitats than random landscapes. Our results suggest that for the persistence of the flying squirrel, forest managers should 1) maintain a deciduous mixture, particularly in spruce-dominated forests; 2) maintain physical connectivity between optimal breeding habitats; and 3) impose coarse-grained structures on northeastern Finnish landscapes at current levels of habitat availability.     

Conservation of Fragmented Populations

In this paper we argue that landscape spatial structure is of central importance in understanding the effects of fragmentation on population survival. Landscape spatial structure is the spatial relationships among habitat patches and the matrix in which they are embedded. Many general models of subdivided populations make the assumptions that (1) all habitat patches are equivalent in size and quality and (2) all local populations (in the patches) are equally accessible by dispersers. Models that gloss over spatial details of landscape structure can be useful for theoretical developments but will almost always be misleading when applied to real-world conservation problems. We show that local extinctions of fragmented populations are common. From this it follows that recolonization of local extinctions is critical for regional survival of fragmented populations. The probability of recolonization depends on (1) spatial relationships among landscape elements used by the population, including habitat patches for breeding and elements of the inter-patch matrix through which dispersers move, (2) dispersal characteristics of the organism of interest, and (3) temporal changes in the landscape structure. For endangered species, which are typically restricted in their dispersal range and in the kinds of habitat through which they can disperse, these factors are of primary importance and must be explicitly considered in management decisions.     

Modelling the effect of in-stream and overland dispersal on gene flow in river networks

Modelling gene flow across natural landscapes is a current challenge of population genetics. Models are essential to make clear predictions about conditions that cause genetic differentiation or maintain connectivity between populations. River networks are a special case of landscape matrix. They represent stretches of habitat connected according to a branching pattern where dispersal is usually limited to upstream or downstream movements. Because of their peculiar topology, and the increasing concern about conservation issues in hydrosystems, there has been a recent revival of interest in modelling dispersal in river networks. Network complexity has been shown to influence global population differentiation. However, geometric characteristics are likely to interact with the way individuals move across space. Studies have focused on in-stream movements. None of the work published so far took into consideration the ability of many species to disperse overland between branches of the same network though. We predicted that the relative contribution of these two dispersal modalities (in-stream and overland) would affect the overall genetic structure. We simulated dispersal in synthetic river networks using an individual-based model. We tested the effect of dispersal modalities, i.e. the ratio of overland/in-stream dispersal, and two geometric parameters, bifurcation angle between branches and network complexity. Data revealed that if geometrical parameters affected population differentiation, dispersal parameters had the strongest effect. Interestingly, we observed a quadratic relationship between p the proportion of overland dispersers and population differentiation. We interpret this U-shape pattern as a balance between isolation by distance caused by in-stream movements at low values of p and intense migrant exchanges within the same branching unit at high values of p. Our study is the first attempt to model out-of-network movements. It clearly shows that both geometric and dispersal parameters interact. Both should be taken into consideration in order to refine predictions about dispersal and gene flow in river network.     

Site specificity in advertisement calls and responses to playbacks of local and foreign call variants in satin bowerbirds, <Emphasis Type="Italic">Ptilonorhynchus violaceus</Emphasis>

Avian vocalisations often show patterns of geographic variation. Previous work on the satin bowerbird has shown that although spatial variation in this species’ advertisement calls is strongly associated with habitat structure, some variation is apparent within habitat types. Seventeen populations located throughout the species’ distribution were used to examine whether spatial call variation could be influenced by other processes such as random drift or the presence of fine-scale vocal traditions; if this were the case, differing call variants would be expected at geographically discrete sampling sites both within and among habitat types. There were population-specific call variants at each of the sites sampled, with different variants apparent even within habitat types. At most sites, individuals gave only a single variant of advertisement call, and the call variant at one site, sampled after a 5-year interval, appears to have been relatively stable. Playback experiments were conducted at three populations to examine whether local call variants invoked a greater response than several non-local variants differing in their degree of similarity to the local variant. Birds responded strongly to local call variants but not to either of two foreign variants, one of which was similar to their local variant and one of which was very different. A pattern of geographic variation across populations, the fact that local and non-local variants evoke different responses and circumstantial evidence indicating that individuals can learn new calls all suggest that factors affecting song learning and the ability of males to establish and defend a bower site may have contributed to the establishment of geographically variable vocal cultures in this species.     

Population genetics of the blue crabCallinectes sapidus: modest population structuring in a background of high gene flow

To determine the genetic population structure of blue crabs (Callinectes sapidus Rathbun), electrophoretic allozyme analysis was performed on 750 individuals collected from 16 nearshore locations ranging from New York to Texas, USA. Twenty enzymes and non-enzymatic proteins coded by 31 presumptive loci were examined. Twenty-two loci were either monomorphic or polymorphic at less than theP ₉₅ level; alleles for these polymorphic loci were geographically dispersed. Allele frequencies for three of the remaining polymorphic loci were homogeneous over all populations, as were levels of polymorphism and heterozygosity. Phenograms generated by the UPGMA (unweighted pair-group method using arithmetic averages) and distance Wagner methods exhibited no geographic pattern in the clustering of populations. Estimates ofN _em (effective number of migrants per generation between populations) indicated substantial gene flow, with aalues sufficiently high to infer panmixia between all blue crab populations from New York to Texas. However, despite this high level of gene flow, two striking patterns of geographic differentiation occurred: genetic patchiness and clinal variation. Allele frequencies atEST-2, GP-1, IDHP-2, DPEP-1, DPEP-2, andTPEP exhibited genetic patchiness on local and range-wide geographic scales, and allele frequencies atEST-2 varied temporally. Genetic patchiness in blue crabs is likely to be the result of the pre-settlement formation and subsequent settlement of genetically heterogeneous patches of larvae; allele frequencies of those larval patches may then be further modified through ontogeny by localized selection. In the Atlantic Ocean, a regional latitudinal cline ofEST-2 allele frequencies was superimposed on the range-wide genetic patchiness exhibited by that locus. This pattern against a background of high gene flow is highly likely to be maintained by selection. In estuaries along the Atlantic Ocean coast, a combination of low adult long-distance migration and a high retention rate of locally spawned larvae could serve to segregate populations and allow for the development of the geographic cline inEST-2. The Gulf of Mexico showed no apparent cline, perhaps due to long-distance migration of females in some regions of the Gulf, or to masking by genetic patchiness. These results emphasize the importance of both ecological and evolutionary time scales and structuring mechanisms in determining genetic population structure.     

Spatial scaling of avian population dynamics: population abundance, growth rate, and variability

Synchrony in population fluctuations has been identified as an important component of population dynamics. In a previous study, we determined that local-scale (<15-km) spatial synchrony of bird populations in New England was correlated with synchronous fluctuations in lepidopteran larvae abundance and with the North Atlantic Oscillation. Here we address five questions that extend the scope of our earlier study using North American Breeding Bird Survey data. First, do bird populations in eastern North America exhibit spatial synchrony in abundances at scales beyond those we have documented previously? Second, does spatial synchrony depend on what population metric is analyzed (e.g., abundance, growth rate, or variability)? Third, is there geographic concordance in where species exhibit synchrony? Fourth, for those species that exhibit significant geographic concordance, are there landscape and habitat variables that contribute to the observed patterns? Fifth, is spatial synchrony affected by a species' life history traits? Significant spatial synchrony was common and its magnitude was dependent on the population metric analyzed. Twenty-four of 29 species examined exhibited significant synchrony in population abundance: mean local autocorrelation (rho)= 0.15; mean spatial extent (mean distance where rho=0) = 420.7 km. Five of the 29 species exhibited significant synchrony in annual population growth rate (mean local autocorrelation = 0.06, mean distance = 457.8 km). Ten of the 29 species exhibited significant synchrony in population abundance variability (mean local autocorrelation = 0.49, mean distance = 413.8 km). Analyses of landscape structure indicated that habitat variables were infrequent contributors to spatial synchrony. Likewise, we detected no effects of life history traits on synchrony in population abundance or growth rate. However, short-distance migrants exhibited more spatially extensive synchrony in population variability than either year-round residents or long-distance migrants. The dissimilarity of the spatial extent of synchrony across species suggests that most populations are not regulated at similar spatial scales. The spatial scale of the population synchrony patterns we describe is likely larger than the actual scale of population regulation, and in turn, the scale of population regulation is undoubtedly larger than the scale of individual ecological requirements.     

Patch Isolation, Corridor Effects, and Colonization by a Resident Sparrow in a Managed Pine Woodland

The isolation of habitat patches is often cited as having a major impact on the dynamics of small populations occupying patches in a complex landscape. Few studies, however, have provided field data demonstrating that isolation has an identifiable effect on specific populations independent of other factors such as local habitat quality or that landscape factors such as corridors can alleviate such effects. We conducted field surveys of Bachman's Sparrow ( Aimophila aestivalis ) populations in regions, which we call linear landscapes, where suitable habitat patches were isolated to varying degrees from potential sources of dispersing birds. In these linear landscapes, isolated patches of habitat were less likely to be colonized than were nonisolated patches. We also found that corridor configurations of habitat patches improved the ability of sparrows to find and settle in newly created patches. These results suggest that, for species that do not disperse easily through inhospitable landscapes, habitat occupancy at a regional scale can be enhanced by careful landscape design and planning.     

Correlation between pelagic larval duration and realised dispersal: long-distance genetic connectivity between northern New Zealand and the Kermadec Islands archipelago

The extent to which marine populations are “open” (panmixia) or “closed” (self-recruitment) remains a matter of much debate, with recent reports of high levels of genetic differentiation and self-recruitment among populations of numerous species separated by short geographic. However, the interpretation of patterns of gene flow (connectivity) is often based on a stepping stone model of dispersal that can genetically homogenise even distant populations and blur genetic patterns that may better reflect realised dispersal. One way in which realised long-distance dispersal can be accurately determined is by examination of gene flow of taxa between isolated archipelagos and a mainland where there is no possible stepping stone dispersal across the open ocean. We investigated the genetic structuring of populations of the intertidal gastropod Nerita melanotragus from the subtropical Kermadec Islands and temperate New Zealand’s North Island (the mainland), separated by 750 km of open ocean and characterised by contrasting environmental conditions. Analyses of seven microsatellite markers revealed an absence of genetic structuring with low F _ST and Jost’s D values (from 0.000 to 0.007 and from 0.000 to 0.015, respectively) over large geographic distances and no evidence of isolation by distance among all populations. These results indicate that the realised dispersal of N. melanotragus is of at least 750 km, this species exhibits a very “open” form of connectivity and its larvae exhibit sufficient phenotypic plasticity to settle successfully in different environmental conditions, ranging from subtropical to cool temperate.     

Multiscale patterns of genetic structure in a marine snail (Solenosteira macrospira) without pelagic dispersal

The northern Gulf of California (NGC) is one of the most dynamic and productive marine ecosystems in the world, yet knowledge about population connectivity and dispersal patterns is lacking for many of its resident species. Using nuclear and mitochondrial markers, we investigated the effects of open water, geographical distance and suitable habitat on patterns of genetic structure of Solenosteira macrospira, a benthic buccinid whelk with direct development. We collected samples in April 2004, 2005 and May 2007 from the upper NGC (31°34.39″N, 114°44.45″W). Phylogenetic analyses, hierarchical analyses of variance and Bayesian assignment tests substantiated a break between the east and west coasts. Genetic distance between population pairs increased with geographical distance, but only when assuming a U-shaped dispersal pathway over the open water of the NGC. Given S. macrospira’s association with rocky intertidal habitats, and its limited dispersal potential, we assumed that the geographical distribution of rocky habitat would play a significant role in genetic differentiation of S. macrospira. Nevertheless, populations separated by sand were more similar than populations separated by rocks. The influence of open water, geographical distance and suitable habitat (rocks vs. sand) also varied significantly across different genetic markers that presumably evolve at different rates. Specifically, the more rapidly evolving nuclear microsatellites suggested that physical transport processes strongly influence genetic differentiation on contemporary time scales, even in a species with direct benthic development. This underscores the strong, and potentially homogenizing, effect of present-day ocean circulation patterns in the NGC.     

Dispersal depression with habitat fragmentation in the bog fritillary butterfly

Habitat fragmentation is expected to impose strong selective pressures on dispersal rates. However, evolutionary responses of dispersal are not self-evident, since various selection pressures act in opposite directions. Here we disentangled the components of dispersal behavior in a metapopulation context using the Virtual Migration model, and we linked their variation to habitat fragmentation in the specialist butterfly Proclossiana eunomia. Our study provided a nearly unique opportunity to study how habitat fragmentation modifies dispersal at the landscape scale, as opposed to microlandscapes or simulation studies. Indeed, we studied the same species in four landscapes with various habitat fragmentation levels, in which large amounts of field data were collected and analyzed using similar methodologies. We showed the existence of quantitative variations in dispersal behavior correlated with increased fragmentation. Dispersal propensity from habitat patches (for a given patch size), and mortality during dispersal (for a given patch connectivity) were lower in more fragmented landscapes. We suggest that these were the consequences of two different evolutionary responses of dispersal behavior at the individual level: (1) when fragmentation increased, the reluctance of individuals to cross habitat patch boundaries also increased; (2) when individuals dispersed, they flew straighter in the matrix, which is the best strategy to improve dispersal success. Such evolutionary responses could generate complex nonlinear patterns of dispersal changes at the metapopulation level according to habitat fragmentation. Due to the small size and increased isolation of habitat patches in fragmented landscapes, overall emigration rate and mortality during dispersal remained high. As a consequence, successful dispersal at the metapopulation scale remained limited. Therefore, to what extent the selection of individuals with a lower dispersal propensity and a higher survival during dispersal is able to limit detrimental effects of habitat fragmentation on dispersal success is unknown, and any conclusion that metapopulations would compensate for them is flawed.     

Mitochondrial DNA Variability of the Gray Wolf: Genetic Consequences of Population Decline and Habitat Fragmentation

The gray wolf is a large, highly mobile predator whose original geographic range included most of the Northern Hemisphere. High rates of genetic exchange probably characterized even distantly-separated populations in the past, but recent population declines and habitat fragmentation have isolated previously contiguous populations, especially in the Old World. We examine mitochondrial DNA (mtDNA) variability among twenty-six populations of wolves from throughout their geographic range. We find eighteen mtDNA genotypes in gray wolves, seven of these are derived from hybridization with coyotes, four are confined to the New World, six are confined to the Old World and one is shared by both areas. Genetic differentiation among wolf populations is significant but small in magnitude. In the Old World, most localities have a single unique genotype, whereas in the New World several genotypes occur at most localities and three of the five genotypes are nearly ubiquitous. The pattern of genetic differentiation in the gray wolf contrasts with that of another large, highly vagile canid, the coyote, in which genetic differentiation among populations is not significant even among widely separated localities. We suggest that the difference between these two species reflects the rapid, recent increase in coyote numbers and expansion of their geographic range, and the coincident decline in gray wolf populations. Apparent genetic differences among extant wolf populations may be a recent phenomenon reflecting population declines and habitat fragmentation rather than a long history of genetic isolation.     

Isolation by distance in the scleractinian coral <Emphasis Type="Italic">Seriatopora hystrix</Emphasis> from the Red Sea

Pelagic dispersal of larvae in sessile marine invertebrates could in principle lead to a homogeneous gene pool over vast distances, yet there is increasing evidence of surprisingly high levels of genetic differentiation on small spatial scale. To evaluate whether larval dispersal is spatially limited and correlated with distance, we conducted a study on the widely distributed, viviparous reef coral Seriatopora hystrix from the Red Sea where we investigated ten populations separated between ~0.150 km and ~610 km. We addressed these questions with newly developed, highly variable microsatellite markers. We detected moderate genetic differentiation among populations based on both F _ST and R _ST (0.089 vs. 0.136, respectively) as well as considerable heterozygote deficits. Mantel tests revealed isolation by distance effects on a small geographic scale (≤20 km), indicating limited dispersal of larvae. Our data did not reveal any evidence against strictly sexual reproduction among the studied populations.     

Evaluating Effects of Habitat Loss and Land-Use Continuity on Ant Species Richness in Seminatural Grassland Remnants

Abstract:  Seminatural grasslands in Europe are susceptible to habitat destruction and fragmentation that result in negative effects on biodiversity because of increased isolation and area effects on extinction rate. However, even small habitat patches of seminatural grasslands might be of value for conservation and restoration of species richness in a landscape with a long history of management, which has been argued to lead to high species richness. We tested whether ant communities have been negatively affected by habitat loss and increased isolation of seminatural grasslands during the twentieth century. We examined species richness and community composition in seminatural grasslands of different size in a mosaic landscape in Central Sweden. Grasslands managed continuously over centuries harbored species-rich and ecologically diverse ant communities. Grassland remnant size had no effect on ant species richness. Small grassland remnants did not harbor a nested subset of the ant species of larger habitats. Community composition of ants was mainly affected by habitat conditions. Our results suggest that the abandonment of traditional land use and the encroachment of trees, rather than the effects of fragmentation, are important for species composition in seminatural grasslands. Our results highlight the importance of considering land-use continuity and dispersal ability of the focal organisms when examining the effects of habitat loss and fragmentation on biodiversity. Landscape history should be considered in conservation programs focusing on effects of land-use change.     

Fire and forest landscapes in the Georgia Piedmont: an assessment of spatial modeling assumptions

Landscape simulation models are widely used to study the behavior of ecological systems. As computing power has increased, these models have become more complex and incorporated more realistic spatial representations of landscape patterns and ecological processes. The goal of this research was to examine the sensitivity of simulated landscape patterns to fundamental spatial modeling assumptions. The LANDIS simulator was parameterized for forests of the Georgia Piedmont and used to model landscape-scale community dynamics at fire return intervals from 20 to 100 years. A base scenario incorporating localized seed dispersal along with landform-related variation in species establishment rates and disturbance regimes was contrasted with three alternative scenarios. The uniform habitat scenario applied the same set of species establishment coefficients across all landforms. The uniform dispersal scenario removed the effects of seed source abundance and pattern on species establishment. The uniform disturbance scenario assumed identical disturbance regimes on all landforms.At the shortest fire return intervals, fire severities were low and the stand age distribution was dominated by older forests. At longer fire return intervals, fire severities were high and the stand age distribution was skewed toward younger forests. Species composition generally followed a gradient from fire-resistant species at short fire return intervals to fire-sensitive species at longer fire return intervals. However, some species exhibited bimodal distributions with high abundances at both short and long fire return intervals. Landscape responses to fire were similar in the uniform habitat scenario and the base scenario. Communities were less sensitive to fire return interval and had more fire-sensitive species in the uniform dispersal scenario than in the base scenario. Species composition in the uniform disturbance scenario was similar to the base scenario for the longest fire-intervals, but was more sensitive to changes in the fire regime at shorter fire return intervals. In models of Piedmont forest landscapes, accurate spatial representations of dispersal and fire regime heterogeneity are essential for predicting landscape-scale species composition under changing fire regimes. In contrast, the precise spatial representation of species–habitat relationships may be considerably less important.     

Genetic population structure of the seagrass Thalassodendron ciliatum in sandy and rocky habitats in southern Mozambique

The genetic diversity of the dioecious seagrass Thalassodendron ciliatum in six populations in southern Mozambique was studied by means of random amplified polymorphic DNA (RAPD). Samples were taken from sandy and rocky habitats over a span of 880 km. Analysis of molecular variance (AMOVA) showed that most of the genetic variability (71.6% of total genetic variation) was observed within populations. There were also significant genetic differences among populations within each habitat (sandy and rocky). We did not find any significant overall genetic difference between sandy versus rocky populations, indicating that the morphological differences between plants from these two habitats are not maintained by reproductive isolation. There was no significant correlation between geographical and genetic distance, which is discussed in the light of current patterns along the coast. All sampled populations consisted of several genetically distinct individuals, indicating that sexual reproduction is widespread.     

Importance of species traits for species distribution in fragmented landscapes

Knowledge of the relationship between species traits and species distribution in fragmented landscapes is important for understanding current distribution patterns and as background information for predictive models of the effect of future landscape changes. The existing studies on the topic suffer from several drawbacks. First, they usually consider only traits related to dispersal ability and not growth. Furthermore, they do not apply phylogenetic corrections, and we thus do not know how considerations of phylogenetic relationships can alter the conclusions. Finally, they usually apply only one technique to calculate habitat isolation, and we do not know how other isolation measures would change the results. We studied the issues using 30 species forming congeneric pairs occurring in fragmented dry grasslands. We measured traits related to dispersal, survival, and growth in the species and recorded distribution of the species in 215 grassland fragments. We show many strong relationships between species traits related to both dispersal and growth and species distribution in the landscape, such as the positive relationship between habitat occupancy and anemochory and negative relationships between habitat occupancy and seed dormancy. The directions of these relationships, however, often change after application of phylogenetic correction. For example, more isolated habitats host species with smaller seeds. After phylogenetic correction, however, they turn out to host species with larger seeds. The conclusions also partly change depending on how we calculate habitat isolation. Specifically, habitat isolation calculated from occupied habitats only has the highest predictive power. This indicates slow dynamics of the species. All the results support the expectation that species traits have a high potential to explain patterns of species distribution in the landscape and that they can be used to build predictive models of species distribution. The specific conclusions are, however, dependent on the technique used, and we should carefully consider this when comparing among different studies. Since different techniques answer slightly different questions, we should attempt to use analyses both with and without phylogenetic correction and explore different isolation measures whenever possible and compare the results.