| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Population Biology |
Plant Biology and Nature Management, Vrije Universiteit Brussel, Pleinlaan 2, B-1050 Brussels, Belgium
Received for publication May 5, 2005. Accepted for publication July 15, 2006.
ABSTRACT
Many rare species are threatened by habitat fragmentation; however, less is known about effects of fragmentation on common species, despite their potential role in ecosystem productivity and functioning. We identified key factors and processes influencing gene flow in a large population of Primula elatior, a common distylous perennial herb, at an early stage of the fragmentation process, i.e., when fragmentation is taking place. Using 19 allozyme loci, we investigated genetic variation and fine-scale spatial genetic structure (SGS) at seedling and adult life stages in relation to fragmentation history (recent bottlenecks), selection, clonal propagation, sexual reproduction (seed and pollen dispersal, distyly), and patchy structure (patch size, plant density, and morph ratio). The main factors contributing to the strong SGS are seed and (to a lesser extent) pollen dispersal, through a spatial Wahlund effect and biparental inbreeding. Significant differences in allele frequencies between seedlings and adults indicate a temporal Wahlund effect. Patch plant density and biased morph ratio also affect the genetic patterns. Our results show that if P. elatior populations evolve into patchworks of small, isolated remnants, genetic erosion, reduced gene flow, and increased inbreeding can be expected, suggesting that such common plant species might require large population sizes to remain viable.
Key Words: allozymes • Flanders • habitat fragmentation • pollen and seed dispersal • Primulaceae • spatial autocorrelation • spatial genetic structure • temporal Wahlund effect
Due to habitat destruction, increasing urbanization, and intensive agricultural practices, many widely distributed plant species occur in highly fragmented (semi-) natural habitats, surrounded by human-altered environments. Commonness is not necessarily a guarantee for long-term survival in fragmented habitats. As reported for rare, endangered, species (e.g., Young et al., 2002
; Oostermeijer et al., 2003
), populations of widespread species may experience bottlenecks, decreasing population sizes, and increasing isolation, and as a result, show genetic erosion and reduced reproductive success (Van Rossum et al., 2002
; Hooftman et al., 2004
). These can subsequently lead to population collapses, and those species can in turn become endangered. Common species may be important components of ecosystem functioning and productivity, and their decline might negatively affect other, already vulnerable species (e.g., Lienert and Fischer, 2003
). Common species can therefore play an important role in the sustainable conservation of biodiversity.
In this context, it is crucial to identify the key processes and factors for the maintenance of viable populations in fragmented habitats and therefore to investigate gene flow patterns, colonization abilities, and recruitment requirements (e.g., Hamrick and Nason, 1996
; Oostermeijer et al., 2003
). Studies of such processes and factors from a sustainable conservation perspective have generally concerned (locally) rare, endangered or declining species (e.g., Jacquemyn et al., 2003
; Oostermeijer et al., 2003
; Ishihama et al., 2005
), and common habitat specialists (e.g., Lienert and Fischer, 2003
), rarely widespread species (Hooftman et al., 2004
). Moreover, most of these studies concerned populations that have already experienced strong fragmentation. For many widely distributed species, populations are still large and in an early stage of fragmentation, i.e., that from continuous they become subdivided into patches of individuals, but have not evolved yet in a patchwork of small, isolated remnants. It is unsure whether we might use models built for declining species with extremely fragmented populations to predict responses to habitat fragmentation and long-term viability for still widespread congeners. So, there is a need to examine populations when the fragmentation process is taking place.
The study of fine-scale spatial genetic structure (SGS) can give a relevant insight into some of the key processes and factors for the maintenance of viable populations, especially those related to gene flow (Escudero et al., 2003
; Ishihama et al., 2005
). Indeed, SGS within plant populations is mainly the result of the balance and interplay between local genetic drift and gene flow through sexual reproduction, i.e., through seed and pollen dispersal (e.g., Heywood, 1991
). The extent of structuring depends on the relative contribution of seed and pollen dispersal to gene flow, the level of self-fertilization, and the availability of compatible mates, plant density, and pollinator abundance and foraging behavior (Richards, 1997
; Chung et al., 2004
; Van Rossum et al., 2004a
; Vekemans and Hardy, 2004
). In self-incompatible species, disassortative mating and negative frequency-dependent selection can be expected (Barrett, 1992
; Richards, 1997
), and as a result, one might also expect an absence of spatial structure for the self-incompatibility locus, and for heterostylous species, a random spatial distribution of the morph types (Husband and Barrett, 1992
). Clonal propagation can also contribute to the patterns of genetic structure. A nonrandom spatial arrangement of clones in a population can substantially increase geitonogamy, local consanguineous matings and local genetic drift (Heywood, 1991
; Charpentier, 2002
). Populations at an early stage of fragmentation, i.e., with a patchy spatial distribution of the individuals, are expected to have a more pronounced SGS than continuous populations as a result of changing pollination and seed dispersal patterns that can lead to increased levels of local genetic drift and inbreeding (Doligez et al., 1998
; Chung et al., 2000
).
Population genetic structure can also vary among life stages, as a result of demographic and evolutionary processes (Hamrick and Nason, 1996
; Chung et al., 2003a
). A temporal variation in population reproductive rates may lead to a temporal Wahlund effect, i.e., differences in allelic frequencies among generations of produced seeds (Tonsor et al., 1993
; Chung et al., 2003b
). Selection in favor of heterozygotes (heterosis) or against homozygotes (inbreeding depression) may also contribute to a change in population structure over time, with higher levels of homozygosity expected in young recruits than old adults (Tonsor et al., 1993
; Doligez and Joly, 1997
). Investigating the temporal aspects of population genetic variation and structure may therefore be necessary to reveal a more dynamic picture of population genetic structure, especially for perennial species with overlapping generations (Loiselle et al., 1995
; Chung et al., 2003a
).
Despite the fragmentation of its habitat—ancient deciduous forests—over the last few centuries, the perennial herb Primula elatior (L.) Hill (Primulaceae) still remains a widely distributed species in Flanders (northern Belgium) (Tack et al., 1993
), often in large populations. However, its decline in genetic diversity, reproductive success, and colonization abilities as a result of habitat fragmentation (Van Rossum et al., 2002
; Jacquemyn et al., 2002
) suggest that this species might be in a transition phase to the status of common to declining.
To identify key factors and processes related to gene flow in populations of P. elatior and their response to any fragmentation, we used allozymes to investigate within-population genetic variation and fine-scale SGS at seedling and adult life stages in a population of P. elatior. This large population is located within an isolated forest fragment and is in an early stage of population fragmentation process, as expressed by its patchy spatial structure, due to habitat degradation. We address the following questions: (1) Is there evidence of recent bottlenecks in the studied population, that could have resulted from ancient forest fragmentation due to anthropogenic disturbance? (2) Do genetic variation and structure change across life stages? (3) What are the relative contributions of sexual reproduction (i.e., relative contributions of seed and pollen dispersal, distyly) and clonal propagation (if P. elatior clonally propagates) to fine-scale SGS? (4) Does the patchy structure, through variation of patch size, plant density, and morph ratio influence population genetic variation and structure? Finally, we discuss the implications of our results for the sustainable conservation of common species in fragmented habitats.
MATERIALS AND METHODS
The species
Primula elatior is a diploid, long-lived, rosette-forming perennial that mainly occurs in moist deciduous forests. It is distributed over western and central Europe, extending northward to Denmark and eastward to central Asia (Valentine, 1948
). In early spring, it produces umbels with pale yellow flowers, which are primarily pollinated by Hymenoptera (mostly bumblebees) and Diptera (Schou, 1983
). It is an obligate outcrosser because it has a self-incompatibility system (distyly) characterized by two genetically determined flower morphs (pin or thrum) that occur on separate individuals. This prevents intramorph pollination (Richards, 1997
).
Population studied and sampling procedure
The study site, the Ename forest (3°38' E, 50°51' N), is situated near Oudenaarde, in Flanders (Belgium). It is an ancient forest, at least existing since the 14th century. However, it underwent phases of partial clearing and reforestation through centuries, resulting in a fragmented patchwork (Tack et al., 1993
). The studied population of P. elatior ("Ename") is located in an isolated forest fragment (ca. 5 ha) surrounded by intensively used pastures and habitations. This fragment was almost completely cleared during the 1850s for agricultural use; only a hedge subsisted in the moistest part (Tack et al., 1993
). Reforestation took place at the beginning of the 20th century. The intensive fertilization of the pastures located above the forest fragment has resulted in a progressive ruderalization and closing of the vegetation (invasion by Rubus fruticosus sensu lato and by nitrophytes such as Urtica dioica). This degradation of habitat quality has negatively affected P. elatior at least over the last five years, leading to population decline and to an increased subdivision of the population (F. Van Rossum, personal observations). The population of P. elatior is still large (in 2001: 835 flowering plants for a total of 1634 individuals), but shows a spatially patchy pattern (Fig. 1). On the field, 14 patches or subpopulations (A to N) could be clearly distinguished according not only to spatial isolation, but also to separation by unsuitable conditions for the species (degraded vegetation, bushes) and to heterogeneity in plant density (e.g., I and J).
|
Allozyme electrophoresis
Extraction and electrophoretic procedures followed those described in Van Rossum and Triest (2003)
. Twenty-seven putative loci encoded by 15 enzyme systems revealed a clear pattern for genetic analyses. Nineteen of the 27 surveyed loci were polymorphic and lacked any linkage disequilibrium among each other. A total of 60 alleles were detected: acid phosphatase (Acp-1 and -2; two and three alleles, respectively), aconitase (Aco-2, 2 alleles), alcohol dehydrogenase (Adh, 3 alleles), esterase (Est, 6 alleles), glutamate dehydrogenase (Gdh, 4 alleles), glucose-6-phosphate-isomerase (Pgi-1, 2 alleles), glutamate-oxaloacetate transaminase (Got-1 to -4; 2, 2, 3 and 3 alleles, respectively), isocitrate dehydrogenase (Idh, 2 alleles), leucine aminopeptidase (Lap-2, 3 alleles), malate dehydrogenase (Mdh-1 and -4; 2 and 3 alleles, respectively), phosphoglucomutase (Pgm-1, 3 alleles), 6-phospho-d-gluconate (Pgd-2, 2 alleles), shikimate dehydrogenase (Skd, 3 alleles), and triose phosphate isomerase (Tpi-1, 2 alleles). The eight monomorphic loci were Acp-3, Lap-1, Mdh-2, Mdh-3, Pgd-1, Pgm-2, superoxide dismutase (Sod), and Tpi-2.
Detection of recent bottlenecks
In order to infer recent (within the past few dozen generations) population bottlenecks from allozyme allele frequencies, we performed Wilcoxon signed-ranks tests (with 1000 simulations iterations) under the infinite allele model (IAM) using BOTTLENECK (Cornuet and Luikart, 1996
). This software tests for population heterozygosity excess, i.e., for a higher expected heterozygosity (He, calculated from the sum of squared allele frequencies) than the equilibrium heterozygosity (Heq) estimated from the observed number of alleles under the assumption of mutation-drift equilibrium (Piry et al., 1999
).
Genetic variation and structure within population
The following measures of genetic variation were calculated for each locus and as means over loci, at the population level and for each life stage (adults and seedlings), morph type (pin and thrum), and subpopulation: allelic richness (A) for a fixed sample size (g genes), based on the rarefaction method (El Mousadik and Petit, 1996
), observed heterozygosity (Ho), expected heterozygosity (He), and Wright's inbreeding coefficient (FIS, = FIT at the population level) both corrected for small sample size. At the population level, the FIT value was decomposed into its within- and among-subpopulation component using Weir and Cockerham's (1984)
estimators of FIS and FST (
), respectively, to separate the contributions of local biparental inbreeding and isolation by distance to SGS (Kalisz et al., 2001
). Allelic richness and Weir and Cockerham's (1984)
estimators were calculated using FSTAT (Goudet, 2001
) and the other variables using GEN-SURVEY (Vekemans and Lefèbvre, 1997
). The significance of the FIT, FIS, and FST values calculated over all loci was tested by randomization tests using FSTAT and sequential Bonferroni-type correction (Rice, 1989
).
We assessed fine-scale SGS using spatial autocorrelation analyses. These were performed on the 19 polymorphic loci with kinship coefficients (Loiselle et al., 1995
) using SPAGeDi (Hardy and Vekemans, 2002
). To test for isolation by distance, the multilocus kinship coefficient for each pair of individuals was plotted against the logarithm of the geographical distance separating them. The slope (b) of this linear regression provides a good estimator of the degree of SGS at this scale. For graphical representation of kinship, average multilocus kinship coefficients per distance intervals (Fj) were computed for 14 distances classes (upper bound distance in meters): 0.25, 0.5, 1, 2, 3, 4, 5, 7, 10, 20, 30, 50, 100, and 250. Standard errors were estimated using a jackknife procedure over the loci. We tested the significance of the kinship coefficients and of the b estimates by comparing the observed values with those obtained after 2000 random permutations of the individuals among positions. We also quantified SGS using the Sp statistics, calculated as –b/(1 – F1), where F1 is the mean Fj for the first distance interval. F1 can be considered as an approximation of the kinship between pairs of neighbors, provided the first distance class contains enough pairs of individuals to get a reasonably precise F1 value, and 1/Sp can be an estimate of neighborhood size, Nb = – (1 – FIS)/b (Vekemans and Hardy, 2004
).
Genetic variation and SGS within and between life stages
Separate spatial autocorrelation analyses were performed within seedlings, within adults, and between the two life stages. Differences in genetic variation and SGS between seedlings and adults were tested with pairwise nonparametric Wilcoxon matched pairs tests on A, Ho, He, FIS, and on the b values, with the 19 polymorphic loci as replicates. Genetic differentiation between seedlings and adults was estimated by Weir and Cockerham's (1984)
estimator of FST (), and by comparing the slope (b) values between adults and seedlings with those within the same life stage (adults or seedlings) using Wilcoxon matched pairs tests. The significance of the FST value was tested by a randomization test using FSTAT. As seedlings could not be sampled in subpopulations A, B, H, and K (Fig. 1), an analysis of molecular variance (AMOVA) was also performed on eight subpopulations (with n 
10) using ARLEQUIN (Schneider et al., 2000
), to distinguish life-stage (temporal) from subpopulation (spatial) effects. F statistics were used to partition the genetic diversity into its between-subpopulation component (= FCT) and its between-life stage (within subpopulation) component (= FSC). The significance of the FST, FSC, and FCT values was tested using a nonparametric permutation approach (Excoffier et al., 1992
).
Extent of clonality and genotypic diversity
To investigate the capacity of P. elatior for clonal propagation, we tested whether adult plants (= ramets) of the same multilocus genotype (including morph type) could be clones generated by asexual propagation or whether they could be identical genotypes (for the markers used) produced via sexual reproduction. The probability of obtaining a second encounter of the same multilocus genotype (= genet) in the sampled adults (pse) was calculated according to Parks and Werth (1993)
. It may be assumed over 95% confidence that the multiple occurrence of a multilocus genotype with pse < 0.05 is likely to be generated by clonal propagation of a single genet. If pse > 0.05, ramets are likely to be accounted for by sexual reproduction (Parks and Werth, 1993
). The proportion of distinguishable multilocus genotypes was calculated as G/N (Ellstrand and Roose, 1987
), where G is the number of distinct multilocus genotypes (or genets) and N the number of ramets. Multilocus genotype diversity (DG) was calculated as a modification of the Simpson index (Pielou, 1969
). Adult plants with missing data were excluded.
In case of a significant structuring effect of clonal growth, a higher SGS can be expected for the ramets than the genets (sexually reproduced genotypes) (Heywood, 1991
). Therefore, to perform a spatial autocorrelation analysis including genets only, one ramet per genotype for which multiples could be considered as putative clones (see results) was randomly chosen. Wilcoxon matched pairs tests between ramets and genets were performed on the slopes (b) of the regression (replicates = 19 polymorphic loci).
Sexual reproduction: relative contributions of pollen and seed dispersal and distyly
The relative contributions of seed and pollen dispersal to total gene flow were inferred by analyzing the shape of the regression between the kinship coefficients and the logarithm of the distances obtained from the spatial autocorrelation analysis, using the methods described in Heuertz et al. (2003)
. The shape of the regression can be described by the k value, calculated using the coefficients of the term of second and third power of a cubic regression between the residuals of the regression and the logarithm of the distance (for details, see Vekemans and Hardy, 2004
). A concave shape (k > 0) indicates leptokurtic gene flow, suggesting more restricted seed dispersal than pollen dispersal (
s << p). For convex curvature (k < 0), s p.
In order to test for negative frequency-dependent selection and therefore for an absence of spatial structure for the self-incompatibility supergene locus (S-locus), we performed a separate spatial autocorrelation analysis on the S-locus, with the heterozygous thrum morph (Ss) and the homozygous pin morph (ss) coded as 12 and 11, respectively. The two morphs differ in several traits, such as pollen size and amount (Schou, 1983
). It is not known whether these differences may affect genetic variation and structure. Separate spatial autocorrelation analyses were performed on allozyme data for pairs of adults within pin and thrum morphs and between the two morphs (the two first distance classes were merged). Differences between the morph types in A, Ho, He, FIS, and b values were tested by Wilcoxon matched pairs tests (replicates = 19 polymorphic loci). Genetic differentiation between pin and thrum morphs was estimated by Weir and Cockerham's (1984)
estimator of FST (), and by comparing the b values between morph types with those within the same morph (pin or thrum) using Wilcoxon matched pairs tests.
Effects of subpopulation size, plant density, and morph ratio
We examined the relationships between within-subpopulation genetic variation (A, Ho, He, and FIS = dependent variables), subpopulation (patch) size (number of flowering individuals), plant density, and morph bias, by performing multiple stepwise regression analyses (with forward stepwise procedure) on all individuals, and on adults and seedlings (10 subpopulations) separately. Patch size, plant density, and allelic richness were log-transformed to achieve homoscedasticity and normality. The relationships between the predictor variables (patch size, plant density, and morph bias) were also examined using Pearson's correlation coefficients. Because multiple tests were involved, the Bonferroni correction was applied to test for significance (Rice, 1989
). Besides, spatial autocorrelation analysis was performed for six subpopulations (E, F, G, I, J, N) having similar range of distance classes and large sample sizes (n 40). Spearman rank correlation analyses were performed between the slope (b) of the regression and patch size, plant density and morph bias.
Voucher specimens
Voucher specimens of pin and thrum plants from the population studied herein were deposited at the herbarium of the National Botanic Garden of Belgium (BR) (see Appendix).
RESULTS
Detection of recent bottlenecks
No evidence of recent bottlenecks was found in the Ename population; no significant (P < 0.05) excess of heterozygosity (He > Heq) was found under IAM when tested by Wilcoxon signed-ranks tests.
Genetic variation and SGS within population
At the population level, the estimates of within-population genetic variation gave A = 2.13, Ho = 0.105, and He = 0.136 (Table 1). The FIT value (=0.229) was significantly positive (P = 0.002), and was decomposed into Weir and Cockerham's FIS = 0.181 (P = 0.001) and FST = 0.062 (P = 0.001). We found a significant linear relationship between decreasing pairwise kinship coefficients and the logarithm of increasing geographical distance in the entire population (Fig. 2). The slopes (b) of the regression between pairwise kinship coefficients and spatial distances were significant (P < 0.05) for 18 of the 19 loci (not significant for Lap-2), with a mean b value of –0.0139. The mean FIS value was similar to the average kinship coefficient between plants separated by the first distance interval (F1 = 0.227). The estimate of neighborhood size (Nb) was 55; the Sp statistic was 0.0180.
|
|
Extent of clonality and genotypic diversity
In total, 250 (of 266 genotyped adults) distinct multilocus genotypes were identified. Only 14 genets occurred twice or three times. For four of them, ramets were located at short distances (from 0.06 to 1.27 m) and were likely to be accounted for by clonal propagation (pse < 0.05). They could, therefore, be considered as putative clones. The estimates of clonal diversity were high: values of G/N and DG were 0.940 and 0.999, respectively. Removing the multiples of the genets considered as putative clones in the spatial autocorrelation analysis did not change the b values of the regression (mean b = –0.0167), when compared with the ramets (mean b = –0.0170) (Wilcoxon matched pairs test Z = 0.16, P > 0.05).
Sexual reproduction: relative contributions of pollen and seed dispersal and distyly
The shape of the linear regression between the kinship coefficients and the logarithm of the distances (Fig. 2) was found to be concave (k = 1.88; cubic regression R2 = 0.842, P < 0.001), suggesting that seed dispersal was more restricted than pollen dispersal (s << p), according to Heuertz et al. (2003)
.
No significant SGS was found for the self-incompatibility locus (b = –0.0006, P > 0.10) (Fig. 2). Based on allozymes, pin and thrum individuals had significant (P < 0.001) FIS values (0.200 and 0.250, respectively) and SGS (b = –0.0149, and –0.0186, respectively) (Table 1, Fig. 2). The two morph types did not differ in A, Ho, He, FIS, and b (Wilcoxon matched pairs tests Z < 1.65, P > 0.05) and were not genetically differentiated (FST = 0.006, P = 0.09). The slope of the regression for pairs of individuals between morph types (b = –0.0160, P < 0.001) was not significantly different from those for pairs of individuals within morph (pin or thrum) (Z = 0.44 and 0.24, respectively, P > 0.05).
Effects of patch size, plant density, and morph bias on genetic variation and structure
At the subpopulation level, A (for g = 12) ranged from 1.27 to 1.52, Ho from 0.043 to 0.155, He from 0.084 to 177, and FIS from –0.058 to 0.480. Patch size and morph bias were negatively correlated (r = –0.750, P = 0.002). There was no significant correlation between patch size and plant density (r = –0.292, P > 0.10) or between morph bias and plant density (r = –0.437, P > 0.10). A significant multiple stepwise regression model after forward selection and Bonferroni correction was found for A (multiple regression coefficient R = 0.772, P = 0.007). The standard partial regression (ß) coefficient indicated that A was negatively related to plant density (ß = –0.560, P = 0.017) (Fig. 3) and showed a trend toward a positive relationship with patch size (ß = 0.391, P = 0.077). The regression model for FIS (R = 0.674, P = 0.036) showed a trend for positive and negative relationships with morph bias and plant density, respectively (ß = 0.687, P = 0.018 and ß = –0.570, P = 0.042, respectively). No significant model was found for Ho (R = 0.464, P = 0.095) and He (R = –0.324, P > 0.10). When the life stages were analyzed separately, the multiple regression model after forward selection and Bonferroni correction showed a significantly negative relationship between allelic richness and morph bias (R = –0.780, P = 0.013) and a trend for positive relationship between FIS and morph bias (R = 0.642, P = 0.045) for the seedlings. The other regression analyses for seedlings and adults were not significant (P > 0.10, results not shown).
|
|
Detection of recent bottlenecks
No evidence of recent bottlenecks was found. During the 1850s, most of the Ename forest was cleared for agriculture use, including the studied forest fragment and its surrounding (Tack et al., 1993
). The maintenance of a wooded linear element in this fragment may have protected the population from severe bottlenecks.
Gene flow within population appears to be restricted
We detected a significant pattern of fine-scale SGS consistent with the model of isolation by distance: individuals separated by short distances are more likely to be genetically related than those that are farther apart. Clonal growth is observed, but only for a few genets, and on very short distances. It contributes neither to the FIS value nor to fine-scale SGS, and genotypic diversity (DG > 0.90) is higher than the average (0.61) reported by Ellstrand and Roose (1987)
in their review on clonally reproducing species. By contrast, sexual reproduction contributes significantly to SGS. Analysis of the relative contributions of seed and pollen dispersal to total gene flow indicates that seed dispersal is more restricted than pollen dispersal (s << p). Seeds of P. elatior have no particular mechanism for long-distance dispersal (Valentine, 1948
). Hence, seed dispersal is expected to be restricted. Besides, in the close relative Primula veris, pollen carryover has been reported to be leptokurtic, most of the pollen being deposited over short distances, more precisely <6 m (Richards, 1997
). Because P. elatior has life-history traits very similar to P. veris (Barrett, 1992
; Richards, 1997
), a similar pattern of pollen dispersal can be expected in P. elatior. This is consistent with the low and high estimate of neighborhood size (Nb = 55) and the Sp statistic (= 0.018), respectively. These values were similar to those reported for insect-pollinated, self-incompatible herbaceous species (e.g., Richards, 1997
; Vekemans and Hardy, 2004
; Ishihama et al., 2005
). The high, significant, heterozygote deficiency (FIT = 0.229) is quite unexpected for obligate outcrossers. This can be explained, on the one hand, by the significant within-population substructure (Wahlund effect), as shown by the significant Weir and Cockerham's estimator of FST, and on the other hand, by biparental inbreeding as a result of restricted seed and (to a lesser extent) pollen dispersal, as indicated by the high local, subpopulation-level inbreeding value (FIS = 0.181) and by the high average kinship coefficient of the first distance interval (F1 = 0.227) (Kalisz et al., 2001
; Vekemans and Hardy, 2004
).
Despite contrasting differences between pin and thrum morphs for pollen traits (Schou, 1983
), the two morphs do not differ in their patterns of genetic variation and structure and are not differentiated. The lack of SGS found for the self-incompatibility locus, is consistent with the existence of negative frequency-dependent selection for this locus (Barrett, 1992
).
Genetic variation and SGS within and between life stages
In the absence of a persistent seed bank, as seems to be the case for P. elatior, seedlings belong to the same cohort and represent seed dispersal at one generation. The two life stages show a clumped spatial distribution (Fig. 1) and similar patterns of genetic structure, suggesting that the initial fine-scale SGS developed in the early life stages has been carried over into adults (Chung et al., 2003a
) and that the overall patterns of seed dispersal are similar over time (Chung et al., 2003b
). There is a lack of difference in genetic variation (in particular in Ho) and in FIS between seedlings and adults, and therefore, no evidence of a selection process, e.g., against homozygotes or in favor of heterozygotes.
A significant autocorrelation between seedlings and adults indicates that the seedlings may have been produced by the local adults rather than being immigrants. However, significant genetic differentiation between seedlings and adults, and higher SGS between individuals within each life stage than between both age groups were detected, suggesting a temporal variation of demographic processes and reproductive patterns (Chung et al., 2003b
). These may have resulted in allele frequency differences among generations of produced seeds, i.e., in a temporal Wahlund effect (Tonsor et al., 1993
; Hamrick and Nason, 1996
). This temporal Wahlund effect indicated by the genetic differentiation (FST) between seedlings and adults might therefore partly explain the observed heterozygote deficiency in the Ename population (Chung et al., 2003b
). However, the FST value is small (= 0.016), and the between-life stage component of variation only explains 2.9% of the total variance in the AMOVA performed on eight subpopulations. This indicates that the overall heterozygote deficiency and SGS are mainly attributable to biparental inbreeding and to a spatial Wahlund effect. The stronger kinship between neighboring seedlings than between seedlings and their neighboring adults (Fig. 2) also suggests that neighboring seedlings might be half siblings with pollen donors that are not necessarily the very nearest adult neighbors and that the seed shadows might not overlap much because of a highly localized seed dispersal (Hamrick and Nason, 1996
; Chung et al., 2003a
, 2004
).
Effects of patch size, plant density, and morph bias on genetic variation and structure
Despite a large population size and short distances between neighbor patches (<15 m), there is locally a slight variation of the strength of some genetic processes in our P. elatior population. Local plant density negatively affects subpopulation allelic richness. However, we found no clear relationship between subpopulation plant density and fine-scale SGS in the Ename population. These results do not meet common expectations, i.e., higher genetic diversity at higher densities (Gram and Sork, 1999
) and more pronounced genetic structure at low plant density as a result of a lower number of potential mates (Vekemans and Hardy, 2004
). The faster loss of rare alleles at the highest densities may be explained by increased biparental inbreeding, which may be ascribed to the combination of restricted seed dispersal and pollinators behavior changing with plant density (Heywood, 1991
; Van Rossum et al., 2004b
). Increased biparental inbreeding was reported at higher plant densities for P. elatior (Van Rossum et al., 2002
). However, at low plant densities, patterns of genetic structure vary according to the positioning. Subpopulations at the edge show higher SGS than the subpopulations located inside the population. Patches at edges may receive less pollen from other patches, leading to increased levels of intrapatch pollination and biparental inbreeding.
Subpopulations of small size show skewed pin-thrum ratios. Such a relationship was also found at the population level in P. elatior (Jacquemyn et al., 2002
). The loss of individuals through demographic stochasticity in small patches can lead to morph bias and to reproductive failure from a lack of compatible mates (Allee effect) and thus counter negative frequency-dependent selection (Kéry et al., 2003
). There is a loss of allelic richness and a trend for higher subpopulation FIS values with increasing morph bias within our P. elatior population, in particular for seedlings. By reducing the number of potential mates (Allee effect), and therefore effective population size, a skewed morph ratio may result in higher levels of genetic drift and inbreeding.
Implications for conservation
The study of fine-scale genetic variation and structure of a large population of the common self-incompatible P. elatior at an early stage of fragmentation indicates strong local SGS, restricted seed and pollen dispersal, biparental inbreeding, and effects of local variation of demographic factors, such as plant density and morph bias related to patch size and position. Such traits indicate that gene flow processes may be disrupted in fragmented habitats. Indeed, if patchy populations of P. elatior evolve under further habitat fragmentation and deterioration into a patchwork of small, isolated remnants, we may expect a strong decline in genetic diversity, a high disruption of gene flow processes, and increased inbreeding levels. These processes might lead to inbreeding depression and reproductive failure, and therefore to a high extinction risk of small remnant populations (e.g., Kwak et al., 1998
; Oostermeijer et al., 2003
). That contrasts with the patterns observed for the closely related species Primula vulgaris. In Flanders, P. vulgaris is rare and declining, and its highly fragmented distribution consists of patchworks of small (mostly <100 flowering individuals) remnants of formerly large populations (Endels et al., 2002
). However, these small populations show less genetic erosion than highly fragmented populations of the commoner congener P. veris (Van Rossum et al., 2004b
), and populations appear still to be connected by moderate levels of gene flow (Van Rossum and Triest, 2003
). We may therefore propose the hypothesis that populations of common species need to be of large size (and with large fragments) to remain viable. Whether patches of previously large populations could be functional as stepping stones and whether corridors have a role in connectivity still merit further investigation.
APPENDIX
Voucher information for the taxon used in this study. All voucher specimens are deposited in the herbarium of the National Botanic Garden of Belgium (BR).
Taxon—locality; collector and collection number.
Primula elatior (L.) Hill—Bos t' Ename, Ename (Oudenaarde), Belgium (IFBL E3.21.33); Ludwig TRIEST, BR - S.P. 1 229 960.
Primula elatior (L.) Hill—Bos t' Ename, Ename (Oudenaarde), Belgium (IFBL E3.21.33); Ludwig TRIEST, BR - S.P. 1 229 961.
Primula elatior (L.) Hill—Bos t' Ename, Ename (Oudenaarde), Belgium (IFBL E3.21.33); Ludwig TRIEST, BR - S.P. 1 229 962.
Primula elatior (L.) Hill—Bos t' Ename, Ename (Oudenaarde), Belgium (IFBL E3.21.33); Ludwig TRIEST, BR - S.P. 1 229 963.
Primula elatior (L.) Hill—Bos t' Ename, Ename (Oudenaarde), Belgium (IFBL E3.21.33); Ludwig TRIEST, BR - S.P. 1 229 964.
Primula elatior (L.) Hill—Bos t' Ename, Ename (Oudenaarde), Belgium (IFBL E3.21.33); Ludwig TRIEST, BR - S.P. 1 229 965.
FOOTNOTES
1 The authors thank G. Tack for sampling authorization in Ename forest, D. Parmentier for field assistance, and O. J. Hardy and two anonymous referees for comments on the manuscript. This work was funded by the Ministry of the Flemish Community (contract VLINA 98/03) and the Vrije Universiteit Brussel (OZR Funding). 
2 Author for correspondence (fabienne.vanrossum{at}br.fgov.be
; fax: +32 2 2600945; present address: Department of Vascular Plants, National Botanic Garden of Belgium, Domein van Bouchout, B-1860 Meise, Belgium
LITERATURE CITED
Barrett S. C. H.. 1992. Evolution and function of heterostyly. Monograph on theoretical and applied genetics, vol. 15 Springer-Verlag, Berlin, Germany.
Charpentier A.. 2002. Consequences of clonal growth for plant matings. Evolutionary Ecology 15: 521-530.[CrossRef][ISI]
Chung M. G. Chung J. M. Chung M. Y. Epperson B. K.. 2000. Spatial distribution of allozyme polymorphisms following clonal and sexual reproduction in populations of Rhus javanica (Anacardiaceae). Heredity 84: 178-185.
Chung M. Y. Epperson B. K. Chung M. G.. 2003a. Genetic structure of age classes in Camellia japonica (Theaceae). Evolution 57: 62-73.[ISI][Medline]
Chung M. Y. Epperson B. K. Chung M. G.. 2003b. Temporal aspects of the fine-scale genetic structure in a population of Cinnamomum insularimontanum (Lauraceae). Heredity 90: 98-106.[CrossRef][ISI][Medline]
Chung M. Y. Nason J. D. Chung M. G.. 2004. Spatial genetic structure in populations of the terrestrial orchid Cephalanthera longibracteata (Orchidaceae). American Journal of Botany 91: 52-57.
Cornuet J.-M. Luikart G.. 1996. Description and power analysis of two tests for detecting recent population bottlenecks from allele frequency data. Genetics 144: 2001-2014.[Abstract]
Doligez A. Baril C. Joly H. I.. 1998. Fine-scale spatial genetic structure with nonuniform distribution of individuals. Genetics 148: 905-919.
Doligez A. Joly H. I.. 1997. Genetic diversity and spatial genetic structure within a natural stand of a tropical forest tree species, Carapa procera (Meliaceae), in French Guiana. Heredity 79: 72-82.[CrossRef][ISI]
Ellstrand N. C. Roose M. L.. 1987. Patterns of genotypic diversity in clonal plant species. American Journal of Botany 74: 123-131.[CrossRef][ISI]
El Mousadik A. Petit R. J.. 1996. High level of genetic differentiation for allelic richness among populations of the argan tree (Argania spinosa (L.) Skeels) endemic to Morocco. Theoretical and Applied Genetics 9: 832-839.
Endels P. Jacquemyn H. Brys R. Hermy M. De Blust G.. 2002. Temporal changes (1986–1999) in populations of primrose (Primula vulgaris Huds.) in an agricultural landscape and implications for conservation. Biological Conservation 105: 11-25.[CrossRef][ISI]
Escudero A. Iriondo J. M. Torres M. E.. 2003. Spatial analysis of genetic diversity as a tool for plant conservation. Biological Conservation 113: 351-365.[CrossRef][ISI]
Excoffier L. Smouse P. Quattro J.. 1992. Analysis of molecular variance inferred from metric distances among DNA haplotypes: application to human mitochondrial DNA restriction data. Genetics 131: 479-491.[Abstract]
Goudet J.. 2001. FSTAT, a program to estimate and test gene diversities and fixation indices, version 2.9.3.2 Website, http://www.unil.ch/izea/softwares/fstat.html.
Gram W. K. Sork V. L.. 1999. Population density as a predictor of genetic variation for woody plant species. Conservation Biology 13: 1079-1087.[CrossRef][ISI]
Hamrick J. L. Nason J. D.. 1996. Consequences of dispersal in plants. In O. E. Rhodes Jr., R. K. Chesser, and M. H. Smith [eds.] Population dynamics in ecological space and time 203-236 University Chicago Press, Chicago, Illinois, USA.
Hardy O. J. Vekemans X.. 2002. SPAGeDi: a versatile computer program to analyse spatial genetic structure at the individual or population levels. Molecular Ecology Notes 2: 618-620.[CrossRef][ISI]
Heuertz M. Vekemans X. Hausman J. F. Palada M. Hardy O. J.. 2003. Estimating seed versus pollen dispersal from spatial genetic structure in the common ash. Molecular Ecology 12: 2483-2495.[CrossRef][Medline]
Heywood J. S.. 1991. Spatial analysis of genetic variation in plant populations. Annual Review of Ecology and Systematics 22: 335-355.[CrossRef][ISI]
Hooftman D. A. Billeter R. C. Schmid B. Diemer M.. 2004. Genetic effects of habitat fragmentation on common species of Swiss fen meadows. Conservation Biology 18: 1043-1051.[CrossRef][ISI]
Husband B. C. Barrett S. C. H.. 1992. Pollinator visitation in populations in tristrylous Eichhornia paniculata in northeasthern Brazil. Oecologia 89: 365-371.[ISI]
Ishihama F. Ueno S. Tsumura Y. Washitani I.. 2005. Gene flow and inbreeding depression inferred from fine-scale genetic structure in an endangered heterostylous perennial, Primula sieboldii. Molecular Ecology 14: 983-990.[CrossRef][Medline]
Jacquemyn H. Brys R. Hermy M.. 2002. Patch occupancy, population size and reproductive success of a forest herb (Primula elatior) in a fragmented landscape. Oecologia 130: 617-625.[CrossRef][ISI]
Jacquemyn H. Van Rossum F. Brys R. Endels P. Hermy M. Triest L. De Blust G.. 2003. Effects of agricultural land use and fragmentation on genetics, demography and population persistence of the rare Primula vulgaris, and implications for conservation. Belgian Journal of Botany 136: 5-22.[ISI]
Kalisz S. Nason J. D. Hanzawa F. M. Tonsor S. J.. 2001. Spatial population genetic structure in Trillium grandiflorum: the roles of dispersal, mating, history, and selection. Evolution 55: 1560-1568.[CrossRef][ISI][Medline]
Kéry M. Matthies D. Schmid B.. 2003. Demographic stochasticity in population fragments of the declining distylous perennial Primula veris (Primulaceae). Basic and Applied Ecology 4: 197-206.[CrossRef][ISI]
Kwak M. M. Velterop O. van Andel J.. 1998. Pollen and gene flow in fragmented habitats. Applied Vegetation Science 1: 37-54.
Lienert J. Fischer M.. 2003. Habitat fragmentation affects the common wetland specialist Primula farinosa in north-east Switzerland. Journal of Ecology 91: 587-599.[CrossRef]
Loiselle B. A. Sork V. L. Nason J. Graham C.. 1995. Spatial genetic structure of a tropical understory shrub, Psychotria officinalis (Rubiaceae). American Journal of Botany 82: 1420-1425.[CrossRef][ISI]
Oostermeijer J. G. B. Luijten S. H. den Nijs J. C. M.. 2003. Integrating demographic and genetic approaches in plant conservation. Biological Conservation 113: 389-398.[CrossRef][ISI]
Parks J. C. Werth C. R.. 1993. A study of spatial features of clones in a population of bracken fern, Pteridium aquilinum (Dennstaedtiaceae). American Journal of Botany 80: 537-544.[CrossRef][ISI]
Pielou E. C.. 1969. An introduction to mathematical ecology Wiley Interscience, New York, New York, USA.
Piry S. Luikart G. Cornuet J.-M.. 1999. BOTTLENECK: a computer program for detecting recent reductions in the effective population size using allele frequency data. Journal of Heredity 90: 502-503.
Rice W. R.. 1989. Analysing tables of statistical tests. Evolution 43: 223-225.[CrossRef][ISI]
Richards A. J.. 1997. Plant breeding systems, 2nd ed Chapman and Hall, Cambridge, UK.
Schneider S. Roessli D. Excoffier L.. 2000. Arlequin, version 2000: a software for population genetics data analysis Genetics and Biometry Laboratory, University of Geneva, Geneva, Switzerland.
Schou O.. 1983. The distyly in Primula elatior (L.) Hill (Primulaceae), with a study of flowering phenology and pollen flow. Botanical Journal of the Linnean Society 86: 261-274.
Tack G. Van den Bremt P. Hermy M.. 1993. Bossen van Vlaanderen. Een historische ecologie Davidsfonds, Leuven, Belgium.
Tonsor S. J. Kalisz J. Fisher J. Holtsford T. P.. 1993. A life-history based study of population genetic structure: seed bank to adults in Plantago lanceolata. Evolution 47: 833-843.[CrossRef][ISI]
Valentine D. H.. 1948. Studies in British Primulas. II. Ecology and taxonomy of primrose and oxlip (Primula vulgaris Huds. and P. elatior Schreb). New Phytologist 47: 111-130.[CrossRef][ISI]
Van Rossum F. Bonnin I. Fénart S. Pauwels M. Petit D. Saumitou-Laprade P.. 2004a. Spatial genetic structure of a metallicolous population of Arabidopsis halleri, a clonal, self-incompatible, and heavy metal tolerant species. Molecular Ecology 13: 2959-2967.[CrossRef][Medline]
Van Rossum F. Campos De Sousa S. Triest L.. 2004b. Genetic consequences of habitat fragmentation in an agricultural landscape on the common Primula veris, and comparison with its rare congener, P. vulgaris. Conservation Genetics 5: 231-245.[CrossRef][ISI]
Van Rossum F. Echchgadda G. Szabadi I. Triest L.. 2002. Commonness and long-term survival in fragmented habitats: Primula elatior as a study case. Conservation Biology 16: 1286-1295.[CrossRef][ISI]
Van Rossum F. Triest L.. 2003. Spatial genetic structure and reproductive success in fragmented and continuous populations of Primula vulgaris. Folia Geobotanica 38: 239-254.[CrossRef][ISI]
Vekemans X. Hardy O. J.. 2004. New insights from fine-scale spatial genetic structure analyses in plant populations. Molecular Ecology 13: 912-935.
Vekemans X. Lefèbvre C.. 1997. On the evolution of heavy-metal tolerant populations in Armeria maritima: evidence from allozyme variation and reproductive barriers. Journal of Evolutionary Biology 10: 175-191.[CrossRef][ISI]
Weir B. S. Cockerham C. C.. 1984. Estimating F-statistics for the analysis of population structure. Evolution 38: 1358-1370.[CrossRef][ISI]
Young A. G. Hill J. H. Murray B. G. Peakall R.. 2002. Breeding system, genetic diversity and clonal structure in the sub-alpine forb Rutidosis leiolepis F. Muell. (Asteraceae). Biological Conservation 106: 71-76.[CrossRef][ISI]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
, seedlings. The scales indicate the geographical distance in meters
