Published in: Plant biology (2009), vol. 10, pp. 684-693
Statut: Postprint (Author's version)
Fitness and genetic variation of Viola calaminaria, an endemic metallophyte:
implications of population structure and history
Bizoux J.-P.1, Daïnou K.1,4, Raspe O.2 , Lutts S.3, Mahy G.1
1
Laboratory of Ecology, Gembloux Agricultural University, Gembloux, Belgium.
2
National Botanic Garden of Belgium, Domein van Bouchout, Meise, Belgium.
Biology Unit, Catholic University of Louvain (UCL), Louvain-la-Neuve, Belgium.
4 Present address: Laboratory of Tropical and Subtropical Forestry, Gembloux Agricultural University, Gembloux,
Belgium.
3 Vegetal
Abstract: We investigated variations in genetic diversity and plant fitness in a rare endemic metallophyte of calamine
soils, Viola calaminaria, in relation to population size, population connectivity and population history in order to
evaluate and discuss potential conservation strategies for the species. Mean population genetic diversity (H s = 0.25) of
V. calaminaria was similar to endemic non-metallophyte taxa. Twenty-one per cent of the genetic variation was
partitioned among populations and a low (9%) but significant differentiation was found among geographical regions.
Our results did not support the hypothesis that the acquisition of metal tolerance may result in reduced genetic diversity,
and suggested that strict metallophytes do not exhibit higher inter-population differentiation resulting from scattered
habitats. There were no relationships between population genetic diversity and population size. Significant correlations
were found between plant fitness and (i) population size and (ii) connectivity index. Recently-founded populations
exhibited the same level of genetic diversity as ancient populations and also possessed higher plant fitness. There was
no indication of strong founder effects in recently-established populations. The results suggest that the creation of
habitats through human activities could provide new opportunities for conservation of this species.
Keywords: Genetic variation – heavy metal – natural and anthropogenic population – plant fitnes – RAPD – Viola
calaminaria.
INTRODUCTION
Because soils with elevated concentrations of heavy metals (metalliferous soils) are often phytotoxic, they represent
very harsh and restrictive habitats for plants (Antonovics et al. 1971; Ernst 1990; Brown 2001). Due to these conditions
and the fact that metalliferous sites generally encompass small areas that are geographically isolated from one another,
they often host rare, ecological endemic taxa adapted to high concentrations of heavy metals (Ernst 1990; Brown 2001;
Wolf 2001). Many of these species are rare, making conservation a high priority (Whiting et al. 2004). Because these
metal-contaminated areas are spatially isolated from one another, they may be considered as ecological islands of
various sizes (Wolf et al. 2000).
Genetic theory predicts that population size and isolation are major factors that regulate genetic diversity within, and
genetic differentiation among, populations; these factors can have substantial consequences for population fitness.
Genetic drift and a concomitant increase in inbreeding within any small, isolated population will reduce genetic
diversity within the individual population and increase the likelihood of genetic differentiation among populations.
Higher levels of inbreeding can reduce fitness and may potentially lower viability in small populations (Lande 1988;
Ellstrand & Elam 1993; Ham-rick & Godt 1996; Oostermeijer et al. 2003). Most studies that have examined genetic
diversity and fitness in relation to population size and isolation focused on species that, through habitat fragmentation,
became isolated and experienced a substantial decrease in population size (Reed & Frankham 2003; Reed 2005). Yet
many plant species, such as metallophytes, exist naturally in small and patchy populations. The impact of inbreeding on
fitness may be lower in populations that have been small over many generations or in populations that have declined
gradually (Wolf & Harrison 2001; Leimu & Mutikainen 2005). This phenomenon could be related to the purging of
deleterious alleles by selection (Schemske & Lande 1985; Charlesworth & Charlesworth 1987). An important question
relative to species conservation is whether population size affects fitness and genetic diversity in species occupying
recently fragmented habitats in the same manner as it does for populations that have experienced slow and/or long-term
size reduction (Leimu & Mutikainen 2005).
Published in: Plant biology (2009), vol. 10, pp. 684-693
Statut: Postprint (Author's version)
In contrast to habitat fragmentation and subsequent reduction in available habitat, some human activities have led to the
opposite scenario in which additional habitat is created, allowing for the potential extension of naturally isolated
populations (Krüger et al. 2002). One of the most important human activities that may have such consequences is
mining (Krüger et al. 2002). While mining is often associated with heavy-metal pollution from the surface deposition of
residual ore deposits and overburdening, as well as aerial fallout from smelters and industrial processes, mining may
also increase the number of suitable habitats for those plant species that are naturally metal-tolerant (Mengoni et al.
2000; Krüger et al. 2002). To assess the conservation value of recently added anthropogenic habitats it is important to
consider the influence of colonisation events (e.g. founder effects, bottlenecks) on the genetic diversity of the actual
populations, the genetic structure through modification of metapopulation dynamics, and the effects of potential genetic
drift and lower habitat quality on population fitness components (Krüger et al. 2002; Brock et al. 2007; QuintanaAscencio et al. 2007).
Because endemic and rare species usually have a narrow spatial distribution and are present in small populations, they
are expected to display low-inter-population gene flow and lower total genetic diversity (heterozygosity) than their
more widespread counterparts (Ellstrand & Elam 1993). In endemic metallophytes, additional factors may be important
in shaping population genetic variation and fitness. In particular, the occurrence of major founder effects as a result of
the retention of a limited number of genotypes under severe selection pressure during the evolution of metal tolerance
may reduce the genetic diversity of metallophyte taxa (Vekemans & Lefebvre 1997). In contrast, because metalliferous
sites are generally heterogeneous for several ecological factors (e.g. metal toxicity and drought) on a small spatial scale,
genetic variation may be more important than expected within metallophyte populations under the founder effect
hypothesis (Krüger et al. 2002; Mattner et al. 2002; Pluess & Stocklin 2004).
In eastern Belgium and western Germany, metalliferous sites are characterised by calamine soils with high
concentrations of zinc, cadmium and lead. These sites support metallophytes such as Viola calaminaria (Gingins) Lej., a
species endemic to calamine sites (Lambinon & Auquier 1964; Duvigneaud 1982). Historically, V. calaminaria occurred
on natural metalliferous sites and on former mines (ancient populations). From the end of the 19th century until the
1970s, new populations (recent populations) of V. calaminaria appeared as a result of the increase in habitat availability
resulting from industrial pollution (Bizoux et al. 2004). In recent years, the destruction of species habitat and site
remediation has led to population extinctions or disturbance (Duvigneaud & Saintenoy-Simon 1996; Bizoux et al.
2004).
Viola calaminaria provides a good model to study the influence of scattered habitat on genetic and plant fitness
variation in a rare taxon and to assess the conservation value of recent anthropogenic habitats. The specific aims of this
study were to: (i) assess the genetic and fitness variation of a rare endemic metallophyte at the scale of its distribution
range; (ii) examine the influence of population size and isolation on genetic and fitness variation in a rare species
growing in a harsh environment; and (iii) compare genetic diversity and fitness components between populations
growing on recent, artificially-created habitats and those established on ancient, natural sites. To our knowledge, this is
the first time that genetic diversity and fitness components of an endemic species have been studied jointly in the
context of new anthropogenic habitats.
MATERIALS AND METHODS
Study species
Viola calaminaria is a perennial pansy that reproduces primarily through seeds. Vegetative propagation by means of
rhizomes does not extend beyond 40 cm from the parent plant (Bizoux & Mahy 2007). Predominantly alloga-mous, the
species flowers from April to November (flowering peak in June and July) and is mainly visited by solitary bees,
bumblebees and flies (Syrphideae) (personal observation). Viola calaminaria is endemic to calamine outcrops and has a
narrow range of distribution, which includes eastern areas of Belgium and the vicinity of Aachen (western Germany). In
Belgium, 23 populations are currently known (Bizoux et al. 2004).
The taxonomic status of V. calaminaria remains unresolved. On the basis of cytological criteria, V. calaminaria has
generally been considered an autotetraploid (2n = 52), derived from V. tricolor L. subsp. subalpina Gaud. (2n = 26)
during the last ice age (Heimans 1961; Kakes 1979). Nevertheless, Hildebrandt et al. (2006), on the basis of a
phylogenetic study and new chromosome counts, recommend classifying the plant as V. lutea subsp. calaminaria
(Gingins) Nauenb. Regardless of its taxonomic position, V. calaminaria is undoubtedly a rare calamine-endemic taxon
that is worthy of conservation efforts because of its habitat specificity.
Published in: Plant biology (2009), vol. 10, pp. 684-693
Statut: Postprint (Author's version)
Fig. 1. Location of sampled populations of Viola calaminaria. The circles represent populations sampled for the
genetic and fitness study, squares indicate a population was sampled only for the genetic study, and triangles
represent populations sampled only for the fitness study. See Table 1 for population abbreviations.
RAPD analysis
For the genetic analysis, 12 populations of V. calaminaria were sampled across most of its distribution range in eastern
Belgium and the vicinity of Aachen. Four geographic regions are delimited in the sampled area (Fig. 1): (i) vicinity of
Liege (P1, P2, P3, P4); (ii) Theux (T1, T2); (iii) vicinity of La Calamine (K1, K2, K3, K4), and (iv) vicinity of Stolberg
(S1, S2). Among those populations, four populations were recent anthropogenic sites (P1, P2, P3, P4) and four
populations were ancient natural sites (K1, T1, K3, K4). Historical status could not be confirmed with accuracy for the
other populations. Leaves were collected from 15 individuals per population along transects every 2-3 m, a distance that
is sufficient to avoid sampling closely-related individuals (Bizoux & Mahy 2007). Leaves were collected by hand with
small pliers, placed in a collection tube and frozen directly in liquid nitrogen in the field.
DNA was extracted from frozen leaves using the DNeasyTM extraction kit (Qiagen, Germany). DNA amplification
reactions were performed in 25 µl total volume containing: 1x buffer, 1.5 mM MgCl2, 200 µM of each DNTP (Invitrogen,
UK), 200 µg ml-1 BSA (Aldrich, USA), 1 U Taq DNA polymerase (Amersham Bioscience, UK), 10 pmol of primers
and 20-30 ng of template DNA. Amplifications were performed with a PTC 200 (MJ Research, USA) for an initial
denaturing step of 2 min at 95°C, followed by 44 cycles of 20 s at 94°C, 60 s at 36°C, 60 s at 72°C, and a final step of
10 min at 72°C before cooling at 4°C.
A no template control was included to detect any contamination. Amplification products (12 µl) were subjected to
electrophoresis in 1.8% (w/v) agarose gels (containing ethidium bromide) in 1 x TAE buffer and then photographed
under UV light. Lengths of amplification products were estimated by reference to a 100 base-pair DNA ladder
(GeneRuler, Fermentas GmbH, Germany). Eight primers (Operon technologies, Germany) were selected from an initial
survey of 42 primers: OPA-2, OPA-10, OPB-12, OPH-3, OPL-1, OPL-2, OPM-15 and OPT-15. Band selection was
based on a repeatability test. DNA was amplified twice for 20 individuals and only reproducible bands were considered
for analysis. The amplification products for the different samples were screened for presence/absence (1/0) of the 45
selected bands. Bands of identical size that were amplified with the same primer were considered to be homologous.
Fitness analysis
Two population fitness components, seed set and percentage germination, were examined in 11 Belgian populations
(Fig. 1). Populations sampled for the fitness analysis were not exactly the same as those for the genetic study because
Published in: Plant biology (2009), vol. 10, pp. 684-693
Statut: Postprint (Author's version)
two populations (T2 and K2) were partially destroyed between the two collection periods. Three additional populations
(P5, K5 and K6) were thus sampled to have sufficient populations of various sizes for fitness analysis. The selected
populations included five recent populations (P1, P2, P3, P4, P5) and six ancient ones (T1, K1, K3, K4, K5, K6). For
the seed set experiment, 20 mature fruits were collected along transects (every 2-3 m) in each of the 11 selected
populations in July and September 2004. Well-developed mature seeds in the fruit were then counted to determine the
seed set.
To determine percentage germination, seeds were subsequently collected from open, mature capsules in the 11 selected
populations. A previous test indicated that maximal germination percentages were obtained after a 4-month period of
stratification (Bizoux 2006). Seventy-five seeds were placed in Petri dishes (one dish/population) on filter paper
moistened with distilled water, and the Petri dishes were then held in the dark at 4°C for 4 months. After this
stratification treatment, the Petri dishes were placed in a temperature-controlled room at 22°C with a 16-h light : 8-h
dark photoperiod. The positions of the Petri dishes on benches were randomised and regularly shifted to new positions
to minimise any effect of location. Because available mature seeds were limited in very small populations, the
germination treatment was not replicated.
Population characteristics
The population size of all the selected Belgian populations (except T2) was known from a previous study (Bizoux et al.
2004; Table 1). Population size was measured by the surface covered by V. calaminaria, i.e. the total area of V.
calaminaria patches multiplied by the mean percentage of V. calaminaria cover within patches. As different populations
of V. calaminaria presented the same clonality pattern (Bizoux & Mahy 2007), the surface covered by V. calaminaria
was a good surrogate for estimation of the number of individuals. The spatial connectivity of selected Belgian
populations was measured with the IFM index (for population i: Ii = Σe-dij. Aj where Aj is the areal coverage of the
other sites and d the distance to the other sites; Hanski 1994).
Data analysis
An Analysis of Molecular Variance (Amova, Excoffier et al. 1992) was performed to partition the genetic variance
among three hierarchical levels: (i) within populations, (ii) among populations and (iii) among geographic regions,
using the software Arlequin ver. 2.0. Significance was tested on the basis of 10000 random permutations (Schneider et
al. 2000). In order to test for isolation by distance, the relationship between pairwise F st values derived from the
ANOVA and the pairwise geographical distances for all population pairs was evaluated with a Mantel test, using the
software Arlequin ver. 2.0.
The level of differentiation (Fst) among recent populations (P1, P2, P3, P4) was statistically compared to the level of
differentiation (Fst) among ancient populations (K1, T1, K3, K4) on the basis of inclusion of the zero value (null
hypothesis = no difference) using the confidence interval (CI: 2.5-97.5%) of the Fst distribution of the difference
between Fst among ancient populations and Fst among recent populations computed by Hickory ver 1.00 (Holsinger et
al. 2002). A Bayesian approach based on a Markov chain Monte Carlo (MCMC) procedure allows Fst estimation from
dominant markers without assuming Hardy-Weinberg proportions in populations and allows for the determination of the
Fst distribution with confidence intervals (Holsinger & Wallace 2004).
Genetic population diversity (Hs) and global species genetic diversity (H T) were computed with Hickory ver. 1.0
(Holsinger et al. 2002). The relationships between genetic diversity and (i) population size (n = 9), and (ii) IFM index of
population connectivity (n = 9) were tested using Spearman correlations (Table 2). Average population genetic diversity
(Hs) was compared between recent and ancient populations using a Student's t-test.
For the seed set and the germination rate studies, Spearman correlation was used to test for a relationship between
population size (n = 11), and IFM index (n = 11). Differences in mean seed set between the two collection periods (June
and September), between populations, and between recent and ancient populations were tested by a three-way
hierarchical ANOVA (populations, nested within population type (recent or ancient), nested within collection period).
Differences between mean percentage germination of ancient and recent populations was tested using the Student's ttest. The relationship between genetic diversity (Hs) and plant fitness (n = 8) was examined by Spearman correlation.
All ANOVAS and correlation analyses were performed with MINITAB ver. 13.20 (Minitab, Inc., USA).
Published in: Plant biology (2009), vol. 10, pp. 684-693
Statut: Postprint (Author's version)
Table 1. Viola calaminaria sampled populations with their origin, population size, population IFM Index,
genetic diversity, mean seed set and germination percentage.
population
size (m)
IFM
Index
genetic
diversity (Hs)
(CI 2.5 97.5%)
mean seed set ± SD
(June/September)
germination
percentage
longitude
latitude
origin
P1 (Prayon 1)
5°40.28
50°35.12
Recent
32,242
45,854
0.249 (0.207-0.287)
17.6 ± 9.1/19.5 ± 11
23
P2 (Prayon 2)
5°39
50°34.87
Recent
4590
91,698
0.233 (0.200-0.265)
26.9 ± 8.9/23.8 ± 10.3
55
P3 (Angleur 1)
5°36.59
50°37.07
Recent
3744
105,724
0.283 (0.253-0.314)
28.45 ± 10.5/26.7 ± 10.8
27
P4 (Angleur 2)
5°36.52
50°36.79
Recent
5427
100,344
0.254 (0.226-0.284)
21.9 ± 9.7/23 ± 10.6
59
P5 (Streupas)
5°35.95
50°36.37
Recent
219
102,903
26.3 ± 13.5/25.6 ± 6.1
36
Kl (Kelmis 1)
6°0.53
50°42.89
Ancient
0.4
4444
0.278 (0.249-0.306)
17.2 ± 9.4/16.2 ± 7.3
12
K2 (Kelmis 2)
6°0.47
50°42.67
36
2542
0.242 (0.206-0.277)
K3 (Schmalgraf)
5°59.6
50°41.64
Ancient
150
1796
0.214 (0.174-0.249)
11.9 ± 6.1/12.9 ± 8
33
K4 (Plombières)
0.234 (0.198-0.270)
5°58.07
50°44.04
Ancient
5815
103
11.7 ± 6.7/13.5 ± 8.5
27
K5 (Schmalgraf 2) 5°59.18
50°41.87
Ancient
60
3233
21.4 ± 9.2/1.7 ±11.2
15
K6 (Rabotrath)
5°60.84
50°40.21
Ancient
60
226
20 ± 8/24.9 ± 8.7
24
Tl (Theux 1)
5°49.79
50°31.41
Ancient
25
346
16.3 ± 7.4/ 15.3 ± 7.5
11
T2 (Theux 2)
5°49.6
50°32.43
0.213 (0.175-0.248)
Sl (StoIberg 1 )
6°15.02
50°44.5
0.259 (0.227-0.291)
S2 (StoIberg 2)
6°16.95
50°46.6
0.238 (0.209-0.268)
20/20.3
29.1
Means
0.291 (0.254-0.325)
0.254 (0.225-0.270)
Table 2. Results of correlation tests between population genetic diversity (H s) and (1) population size (n = 9),
and (2) IFM index (n = 9). Correlation between plant fitness (mean seed set and percentage germination) and
(1) population size (n = 11) and (2) IFM index (n = 11). Bold = significant or almost significant.
independent
population
parameter
correlation analysis results (rs/P)
range
Hs
seed set population seed set population
mean (June)
mean (September)
percent
germination
population size
IFM Index
0.4-32,242 m
103-105,724
-0.35/0.36
0.23/0.546
0.16/0.639
0.83/0.002
0.58/0.06
0.46/0.15
0.09/0.779
0.67/0.023
RESULTS
Population genetic diversity
Of the 45 selected bands, which ranged in size from 475 to 1700 bp, only one appeared to be monomorphic. No bands
were rejected using the Lynch & Milligan (1994) criteria. Further analyses were performed on the 44 polymorphic
bands. No population-specific bands were found.
Highly significant (P ≤ 0.001) genetic differentiation was detected in the ANOVA among geographic groups and among
populations within groups. Of the total molecular variance, 9% was due to differentiation among geographic groups,
20.9% to differentiation among populations within geographic groups, and 70.1% to within-population variation.
Pairwise genetic differentiation (Fst) between populations ranged from 0.061 to 0.522 (mean ± SD = 0.287 ± 0.094, c.v.
= 32.8%). All of the 91 pairwise Fst values were significant (P < 0.05). The pair-wise Fst values among the 12
Published in: Plant biology (2009), vol. 10, pp. 684-693
Statut: Postprint (Author's version)
populations were not correlated with pairwise geographic distances (Mantel test: r = 0.037, P = 0.38, Fig. 2). Genetic
differentiation among ancient populations (Fst = 0.26) was higher than genetic differentiation among recent populations
(Fst = 0.18). This difference (0.08) was almost significant (2.5-97.5% CI of the distribution of differences: -0.009-0.180)
The global genetic diversity (HT) at the species level was 0.32. Genetic diversity (H s) of individual populations ranged
from 0.21 to 0.29 (mean ± SD = 0.25 ± 0.026, c.v. = 10.6%) (Table 1). Genetic diversity (H s) was not correlated with
population size and IFM index (Table 2). Average genetic diversity (H s) was not significantly different between ancient
(mean ± SD = 0.254 ± 0.036) and recent (mean ± SD = 0.255 ± 0.021) populations (Student's t-test; t = 0.032, P =
0.976).
Fig. 2. Relationship between pairwise genetic distances and pairwise geographic distances for the 12
populations from the main range of Viola calaminaria.
Population fitness
Three-way hierarchical ANOVAS indicated significant differences in seed set among populations (range: 12.4327.58, P <
0.001) and between ancient (mean ± SD = 16.94 ± 9.06) and recent (mean ± SD = 23.99 ± 10.51) populations (P =
0.003) (Tables 1 and 3). However, there was no significant difference between collection periods (June: means ± SD =
20 ± 10.5, September: means ± SD = 20.3 ± 10.2, P = 0.895). No correlation was found between mean seed set and
population size (Table 2). Mean seed set was positively correlated with IFM index (Fig. 3a) but not with population size
(Table 2).
The percentage germination population mean was 29.1% (range: 10.6-57.9, SD = 15.8) (Table 1). Percentage
germination was marginally significantly higher (Student t-test; t = -2.39, P = 0.054) in recent populations (mean ± SD
= 39.7 ± 16.3) than in ancient populations (mean ± SD = 20.2 ± 9.1). Percentage germination was marginally correlated
with population size (Fig. 3b) but not with IFM index (Table 2). There were no significant correlations between plant
fitness and genetic diversity (seed set June: rs = 0.261, P = 0.531; seed set September: rs = 0.309, P = 0.455; percentage
germination: rs = -0.599, P = 0.117).
Table 3. Result of the three-way hierarchical anova for differences of V. calaminaria seed set between the two
collection periods (June and September), between populations, and between recent and ancient populations.
source of
variation
variance analysis of seed set
DL =
F=
P=
periods
1
0.02
0.895
types of sites 2
8.08
0.003
4.01
<0.001
sites
18
error
418
total
493
Published in: Plant biology (2009), vol. 10, pp. 684-693
Statut: Postprint (Author's version)
Fig. 3. Relationship between (a) population mean seed set and population connectivity index, (b) population
percentage germination and population size.
DISCUSSION
In this study, we used RAPD to assess the genetic structure of Viola calaminaria populations. This technique has the
advantage of being quick and easy, requiring little plant material, not needing previous knowledge of DNA sequences,
and having high resolution (Nybom & Bartish 2000). Its main limitations include sensitivity to reaction conditions. This
can however be largely overcome by uniform application and replicate runs including controls (Hadrys et al. 1992;
Lynch & Milligan 1994). It was widely successfully used in studies of genetic diversity and population genetic
structure, especially of threatened species (Nybom & Bartish 2000).
Genetic diversity and differentiation in an endemic metallophyte taxa
Comparisons to non-metallophyte species with similar life histories showed that V. calaminaria (Hs = 0.254) was not
genetically depauperate. Viola calaminaria genetic diversity was higher than the mean of RAPD studies for endemic
species (Hs = 0.190) and for widespread species (Hs = 0.200) reviewed by Nybom & Bartish (2000).This result was
ongruent with most of the population genetic diversity studies on metallophyte taxa, which showed little evidence of
reduced genetic diversity within populations (Ducousso et al. 1990; Westerbergh & Saura 1992; Vekemans & Lefebvre
1997; Mengoni et al. 2000; Mattner et al. 2002; Baumbach & Hellwig 2003; Dubois et al. 2003). Those results did not
indicate that metal tolerance by plants should be associated with low-genetic diversity resulting from strong selection
and the founder effect at the time of metal tolerance acquisition. Gene flow between non-tolerant and tolerant
populations has been cited as a factor maintaining genetic diversity in metallicous populations (Vekemans & Lefebvre
1997; Mengoni et al. 2000). Recent gene flows from non-tolerant populations are not possible for V. calaminaria as the
species is a strict metallophyte. However, gene introgression following hybridisation with related Viola species may be
a factor that helps to maintain genetic diversity in V. calaminaria. Hybridisation among related Viola species has been
regularly reported (Krahulcova et al. 1996; Neuffer et al. 1999). Kakes (1977) found that Viola guestphali-ca could
Published in: Plant biology (2009), vol. 10, pp. 684-693
Statut: Postprint (Author's version)
hybridise with Viola arvensis in nature, and Hildebrandt et al. (2006) found hybrid V. calaminaria x V. arvensis at
Breinigerberg (Stolberg 1). Hybridisation could also occur with garden pansy, especially in recent populations,
frequently located near urban area.
The level of genetic differentiation among populations (Fst = 0.21) was very similar to the values reported by Nybom &
Bartish (2000) for RAPD studies on endemic species (Fst = 0.20), mixed-mating/outcrossed species (Fst = 0.27 and 0.28,
respectively) or perennials (Fst = 0.25). General conclusions on genetic population differentiation in endemic
metallophytes as compared to non-metallophytes should be treated cautiously, however, as Fst values are also dependent
on the extent of the area studied (Nybom & Bartish 2000); metallophyte endemics are generally restricted in their range.
Only a moderate level of genetic differentiation was observed in the current study, despite the scattered distribution of
the species, which may suggest that strict metallophytes do not exhibit higher inter-population differentiation than nonmetallophyte taxa sharing similar life histories. A similar observation was reported by Mattner et al. (2002) in the
Australian serpentine endemic, Hemigenia exilis, whereas Wolf et al. (2000) reported a much higher level of
differentiation in the serpentine endemic Calystegia collina. In the latter, however, high differentiation (Fst= 0.417) was
probably due more to the effect of clonality than to the effect of a harsh environment (Wolf et al. 2000). We did not find
evidence for a general pattern of isolation by distance at the scale of the main range of V. calaminaria. Mining activities,
resulting in the transport of slag and minerals, may have promoted seed dispersal among distal sites. This artificiallyinduced gene flow may have obscured the effect of spatial isolation. Despite a general lack of isolation by distance, our
results showed weak (Fst = 9%) but significant genetic differentiation of V. calaminaria into regional population groups.
Variation of genetic diversity in relation to population size, isolation and history
Population size was not significantly correlated with genetic diversity in V. calaminaria. A similar lack of correlation
has been reported in other species growing in small and scattered habitats, such as natural metalliferous sites, naturally
salt-contaminated areas or alpine habitats (Krüger et al. 2002; Mattner et al. 2002; Pluess & Stocklin 2004). These
results are in contrast to those usually reported for species in a recently fragmented habitats (e.g. Fischer & Matthies
1998; Hensen & Oberprieler 2005; Hensen et al. 2005) and suggests that naturally rare, scattered species may not suffer
from the genetic impoverishment often associated with small populations. Alternatively, this lack of correlation could be
explained by a misconception in what is considered a 'small' population, and that many are, in fact, large enough to
avoid a reduction of genetic diversity, considering the clonality pattern in V. calaminaria. In fact, a population of 60 m2
may contain at least 100 flowering individuals and the very small population of Kelmis 1 (K1, 0.4 m2) may contain
about 20-30 flowering individuals.
We found no significant difference in genetic diversity between ancient and recent populations and also determined that
recent populations exhibited lower genetic differentiation than ancient ones. Mengoni et al. (2000) and Krüger et al.
(2002) also reported that genetic diversity in recently-colonised habitats on artificial copper or salt deposits was higher
than, or similar to, that of older populations on naturally contaminated soils. The maintenance of genetic diversity may
result from successive colonisation events. Even in the case of non-successive colonisation events, the decrease in
genetic diversity may be rather low if the growth rate of the newly-founded population is very high (Nei et al. 1975).
This is a situation that has been documented in some of the recently-founded populations of V. calaminaria (Bizoux et
al. 2004).
Several models have demonstrated that the genetic consequences of founder events may be complex and dependent on
both the number of individuals involved and the number of source populations from which they originated (Slatkin
1977; Wade & McCauley 1988; McCauley et al. 1995). The pattern observed in our study is also more consistent with
the 'migrant pool' model (Slatkin 1977), in which colonising individuals are drawn at random from all possible source
populations, than with the 'propagule source' model (Slatkin 1977), in which the founding group originated from just
one of the possible source populations (McCauley et al. 1995). Such a situation may have resulted from humanmediated transport of seeds from different areas at several different times to the newly colonised sites during mining
activities, a pattern also suggested by the lack of any detected spatial isolation.
Variation in fitness in relation to population size, isolation and history
In contrast to genetic diversity, we found that population fitness was strongly influenced by some of the population
parameters that were examined though germination percentage results and should be interpreted with cautious due to
the lack of repetition for individual populations.
Significant relationships were found between fitness components and population spatial isolation, but population size
and isolation did not affect the same fitness components. We did not identify any correlation between population size
Published in: Plant biology (2009), vol. 10, pp. 684-693
Statut: Postprint (Author's version)
and seed set; we did find a positive trend between population size and percentage germination. The relation between
genetic diversity and percentage germination was not significant, which suggests that inbreeding was probably not the
main reason for the lower germination rate in small populations.
Population connectivity might also influence plant fitness. Several studies have examined the impact of patchiness and
population isolation on plant fitness (Price & Waser 1979; Sih & Baltus 1987; Rathcke & Jules 1993; Broyles et al.
1994). These studies underlined the importance of pollinator visitation rate, which depends on pollinator movements
and distance between population patches. The positive correlation we observed between population connectivity (IFM
index) and seed set suggests that the ease or likelihood that pollen will flow from one population patch to another
influences fitness in V. calaminaria. Pollen exchange, and its relative impact on fitness, could be more important
between large close populations than between isolated smaller ones.
We also showed that population history (colonisation events) could influence species fitness. Ancient populations
displayed significantly lower seed set and percentage germination than newly-established populations. Although the
reason for this difference is not readily evident from our data, it is possible that the important geographic and genetic
differentiation of ancient populations, accentuated by extinction of some of them, could negatively influence the fitness
of these populations (Wolf & Harrison 2001). Also, rapid demographic expansion, more pronounced recent gene flow,
and creation of recent populations from several sources may have promoted fitness, as found by Leimu & Mutikainen
(2005) in Vincetoxicum hirundinaria. Alternatively, we can conclude that the apparent historical impact may, in fact, be
a spatial phenomenon since a majority of the recent populations are situated in the same geographical region.
CONCLUSIONS
Implications for V. calaminaria conservation
Our results add to the growing body of evidence that suggests that the prediction of population genetic diversity and
differentiation drawn from the extensive literature on habitat fragmentation should not be applied directly to the
conservation of rare, naturally scattered taxa. First, our study illustrated that rare taxa are not, by default, genetically
impoverished. Second, as reported in other studies, we did not find a relationship between population size/isolation and
genetic diversity.
Our data also show that habitats created by recent anthropogenic activity may be of value for the conservation of rare
species when they are adapted to harsh environments (Krüger et al. 2002; Brock et al. 2007) or to human-driven
communities (Quintana-Ascencio et al. 2007). In increasingly disturbed environments, this information indicates that, at
least for some species, conservation strategies should not focus solely on traditional and natural habitats but also
consider the potential benefits offered by modified landscapes.
Acknowlegements
We are grateful to Richard Ennos and Claude Lefebvre for their comments, suggestions and language improvement on a
previous version. J.-P. Bizoux holds a Ph.D. fellowship from the Fund for Research Training in Industry and Agriculture
(FRIA). The research was supported by the FNRS (FRFC programme).
References
Antonovics J., Bradshaw A.D., Turner R.G. (1971) Heavy metal tolerance in plants. Advances in Ecological Research, 7, 1-85.
Baumbach H., Hellwig F.H. (2003) Genetic variation within and among metal-tolerant and non-tolerant populations of Armeria maritima (Mill.)
Willd. s.l. (Plumbaginaceae) in Central and Northeast Germany. Plant Biology, 52, 186-193.
Bizoux J.-P. (2006) Biologie de la conservation d'une metallophyte endémique:Viola calaminaria. Ph.D. thesis, Faculté universitaire de sciences
agronomiques de Gembloux, Belgium.
Bizoux J.-P., Mahy G. (2007) Within-population genetic structure and clonal diversity of a threatened endemic metallophyte, Viola calaminaria
(Violaceae). American Journal of Botany, 94, 887-895.
Bizoux J.-P., Brevers F., Meerts P., Graitson E., Mahy G. (2004) Ecology and conservation of Belgian populations of Viola calaminaria, a
metallophyte with a restricted geographic distribution. Belgian Journal of Botany, 1371, 91104.
Brock J., Aboling S., Stelzer R., Esch E., Papenbrock J. (2007) Genetic variation among different populations of Aster tripolium grown on naturally
and anthropogenic salt-contaminated habitats: implications for conservation strategies. Journal of Plant Research, 120, 99-112.
Brown G. (2001) The heavy metal vegetation of north-western mainland Europe. Botanische Jahrbuucher fuur Systematik, 1231, 63-110.
Broyles S.B., Schnabel A., Wyatt R. (1994) Evidence for long distance pollen dispersal in milkweeds (Asclepias exeltata). Evolution, 48, 1032-1040.
Charlesworth D., Charlesworth B. (1987) Inbreeding and its evolution consequences. Annual Review of Ecology and Systematics, 18, 237-268.
Published in: Plant biology (2009), vol. 10, pp. 684-693
Statut: Postprint (Author's version)
Dubois P.O., Cheptou C., Petit P., Meerts M., Poncelet X., Vekemans C., Lefebvre C., Escarre J. (2003) Genetic structure and mating systems of
metallicolous and nonme-tallicolous populations of Thlaspi caerulescens. New Phytologist, 157, 633-641.
Ducousso A., Petit D., Valero M., Vernet P. (1990) Genetic variation between and within populations of a perennial grass: Arrenatherum elatius.
Heredity, 65, 179-188.
Duvigneaud J. (1982) La halde calaminaire du Rocheux à Theux. une nouxelle réserve naturelle d'Ardenne et Gaume. Parc nationaux,37, 117-138.
Duvigneaud J., Saintenoy-Simon J. (1996) Le site calaminaire de Plombieres. Ardenne et Gaume, 5101, 5-9.
Ellstrand N.C., Elam D.R. (1993) Population genetic consequences of small population size - Implications for plant conservation. Annual Review of
ecology and systematics, 24, 217-242.
Ernst W.H.O. (1990) Mine vegetation in Europe. In: Shaw A.J. (Ed.), Heavy Metal Tolerance in Plants: Evolutionary Aspects. CRC Press, Boca
Raton, FL: 285-300.
Excoffier L., Smouse P.E., Quattro M. (1992) Analysis of molecular variance inferred from metric distances among DNA haplotypes: application to
human mitochondrial DNA restriction data. Genetics, 131, 479-491.
Fischer M., Matthies D. (1998) RAPD variation in relation to population size and plant fitness in the rare Gentianella ger-manica (Gentianaceae).
American Journal of Botany, 856, 811-819.
Hadrys H., Blick M., Schierwater B. (1992) Applications of random amplified polymorphic DNA (RAPD) in molecular ecology. Molecular Ecology,
1, 55-63.
Hamrick J.L., Godt M.J.W. (1996) Effects of life history traits on genetic diversity in plant species. Philosophical Transactions of the Royal Society of
London Series B-Biological Sciences, 351, 1291-1298.
Hanski I. (1994) A practical model of metapopulation dynamics. Journal of Animal Ecology, 63, 151-162.
Heimans J. (1961) Taxonomical, phytogeographical and ecological problems round Viola calaminaria, the zinc violet. Publicaties Natuurhist
Genootschap Limburg, 12, 55-71.
Hensen I., Oberprieler C. (2005) Effects of population size on genetic diversity and seed production in the rare Dictamnus albus (Rutaceae) in central
Germany. Conservation Genetics, 61, 63-73.
Hensen I., Oberprieler C., Wesche K. (2005) Genetic structure, population size, and seed production of Pulsatilla vulgaris Mill. (Ranunculaceae) in
Central Germany. Flora, 200, 3-14.
Hildebrandt U., Hoef-Emden K., Backhausen S., Bothe H., Bozek M., Siuta A., Kuta E. (2006) The rare, endemic zinc violets of Central Europe
originate from Viola lutea Huds. Plant Systematics and Evolution, 257, 205-222.
Holsinger K.E., Wallace L.E. (2004) Bayesian approaches for the analysis of population genetic structure: an example from Platanthera leucophaea
(Orchidaceae). Molecular Ecology, 134, 887-894.
Holsinger K.E., Lewis P.O., Dey D.K. (2002) A Bayesian approach to inferring population structure from dominant markers. Molecular Ecology, 11,
1157-1164.
Kakes P. (1977) Genecological investigations on zinc plants II. Introgression in a small population of the zinc violet Viola calaminaria subsp.
westfalica Lej. Ernst. Acta Botanica Neerlandica, 26, 385-400.
Kakes P. (1979) Genecological investigations on zinc plants III. Cytology of hybrids between Viola calaminaria (Lej.) Ernst and Viola arvensis Murr.
Genetica, 512, 135-142.
Krahulcova A., Krahulec F., Kirschner J. (1996) Introgressive hybridization between a native and an introduced species: Viola lutea subsp. sudetica
versus V. tricolor. Folia Geobotanica et Phytotaxonomica, 312, 219-244.
Krüger A.M., Hellwig F.H., Oberprieler C. (2002) Genetic diversity in natural and anthropogenic inland populations of salt-tolerant plants: random
amplified polymorphic DNA analyses of Aster tripolium L. (Compositae) and Salicornia ramosissima Woods (Chenopodiaceae). Molecular
Ecology, 119, 1647-1655.
Lambinon J., Auquier P. (1964) La flore et la vegetation des terrains calaminaires de la Wallonie septentrionale et de la Rhenanie aixoise. Types
chorologiques et groupes ecologiques. Natura Mosana, 16, 113-130.
Lande R. (1988) Genetics and demography in biological conservation. Science, 241, 1455-1460.
Leimu R., Mutikainen P. (2005) Population history, mating system, and fitness variation in a perennial herb with a fragmented distribution.
Conservation Biology, 192, 349-356.
Lynch M., Milligan B.G. (1994) Analysis of population genetic structure with RAPD markers. Molecular Ecology, 32, 91-99.
Mattner J., Zawko G., Rossetto M., Krauss S.L., Dixon K.W., Sivasithamparam K. (2002) Conservation genetics and implications for restoration of
Hemigenia exilis (Lamiaceae), a serpentine endemic from Western Australia. Biological Conservation, 1071, 37-45.
McCauley D.E., Raveill J., Antonovics J. (1995) Local founding events as determinants of genetic structure in a plant meta-population. Heredity, 75,
630-636.
Mengoni A., Gonnelli C., Galardi F., Gabbrielli R., Bazzicalupo M. (2000) Genetic diversity and heavy metal tolerance in populations of Silene
paradoxa L. (Caryophyllaceae): a random amplified polymorphic DNA analysis. Molecular Ecology, 99, 1319-1324.
Nei M., Maruyama T., Chakraborty R. (1975) The bottleneck effect and genetic variability in populations. Evolution, 29, 1.
Neuffer B., Auge H., Mesch H., Amarell U., Brandl R. (1999) Spread of violets in polluted pine forests: morphological and molecular evidence for the
ecological importance of interspecific hybridization. Molecular Ecology, 83, 365-377.
Nybom H., Bartish I. (2000) Effects of life history traits and sampling strategies on genetic diversity estimates obtained with RAPD markers in plants.
Perspectives in Plant Ecology, Evolution and Systematics, 32, 93-114.
Oostermeijer J.G.B., Luijten S.H., den Nijs J.C.M. (2003) Integrating demographic and genetic approaches in plant conservation. Biological
Conservation, 1133, 389-398.
Pluess A.R., Stocklin J. (2004) Population genetic diversity of the clonal plant Geum reptans (Rosaceae) in the Swiss Alps. American Journal of
Botany, 9112, 2013-2021.
Price M.V., Waser N.M. (1979) Pollen dispersal and optimal outcrossing in Delphinium nelsonii. Nature, 277, 294-296.
Published in: Plant biology (2009), vol. 10, pp. 684-693
Statut: Postprint (Author's version)
Quintana-Ascencio P.F., Weekley C.W., Menges E.S. (2007) Comparative demography of a rare species in Florida scrub and road habitats. Biological
Conservation, 137, 263-270.
Rathcke B.J., Jules E.S. (1993) Habitat fragmentation and plant pollinator interactions. Current Science, 65, 273-277.
Reed D.H. (2005) Relationship between population size and fitness. Conservation Biology, 19, 563-568.
Reed D.H., Frankham R. (2003) Correlation between fitness and genetic diversity. Conservation Biology, 171, 230-237.
Schemske D.W., Lande R. (1985) The evolution of self-fertilization and inbreeding depression in plants. II. Empirical observations. Evolution, 39, 4152.
Schneider S., Roessli D., Excoffier L. (2000) Arlequin ver 2.000: A Software for Population Genetics Data Analysis. Genetics and Biometry
Laboratory, University of Geneva, Geneva.
Sih A., Baltus M.S. (1987) Patch size, pollinator behavior, and pollinator limitation in catnip. Ecology, 686, 1679-1690.
Slatkin M. (1977) Gene flow and genetic drift in a species subject to frequent extinctions. Theoretical Population Biology, 12, 253-262.
Vekemans X., Lefebvre C. (1997) On the evolution of heavy-metal tolerant populations in Armeria maritima: Evidence from allozyme variation and
reproductive barriers. Journal of Evolutionary Biology, 102, 175-191.
Wade M.J., McCauley D.E. (1988) Extinction and recolonization - their effects on the genetic differentiation of local populations. Evolution, 425,
995-1005.
Westerbergh A., Saura A. (1992) The effect of serpentine on the population structure of Silene dioica (Caryophyllaceae). Evolution, 46, 1537-1548.
Whiting S.N., Reeves R.D., Richards D., Johnson M.S., Cooke J.A., Malaisse F., Paton A., Smith J.A.C., Angle J.S., Chaney R.L., Ginocchio R.,
Jaffre T., Johns R., McIntyre T., Purvis O.W., Salt D.E., Schat H., Zhao F.J., Baker A.J.M. (2004) Research priorities for conservation of
metallophyte biodiversity and their potential for restoration and site remediation. Restoration Ecology, 121, 106-116.
Wolf A. (2001) Conservation of endemic plants in serpentine landscapes. Biological Conservation, 1001, 35-44.
Wolf A.T., Harrison S.P. (2001) Effects of habitat size and patch isolation on reproductive success of serpentine morning glory. Conservation Biology,
15, 111-121.
Wolf A.T., Howe R.W., Hamrick J.L. (2000) Genetic diversity and population structure of the serpentine endemic Calyste-gia collina
(Convolvulaceae) in northern California. American Journal of Botany, 87, 1138-1146.