• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of annbotAboutAuthor GuidelinesEditorial BoardAnnals of Botany
Ann Bot. Jan 2012; 109(1): 227–235.
Published online Oct 21, 2011. doi:  10.1093/aob/mcr272
PMCID: PMC3241589

Biased morph ratios and skewed mating success contribute to loss of genetic diversity in the distylous Pulmonaria officinalis


Background and Aims

In heterostylous plant species, skewed morph ratios are not uncommon and may arise from a range of factors. Despite the recognized importance of skewed morph ratios on overall reproductive success within populations, little is known about the impact of skewed morph ratios on population genetic diversity and differentiation in heterostylous species. This study specifically aimed to clarify the effect of population size and morph bias on population genetic diversity and differentiation in the temperate forest herb Pulmonaria officinalis. This species is characterized by a distylous breeding system and shows morph-specific differences in reproductive success.


Genetic diversity was determined for 27 P. officinalis populations in northern Belgium by using eight recently developed microsatellite markers. Multiple regressions were used to assess the relationship between genetic diversity, morph bias and population size, and FST-values were calculated for short- and long-styled morphs separately to study genetic differentiation as a function of morph type.

Key Results

For all genetic measures used, morph bias was more important in explaining patterns of genetic diversity than population size, and in all cases patterns of population genetic diversity followed a quadratic function, which showed a symmetrical decrease in genetic diversity with increasing morph bias. However, probably due to the reproductive advantage of L-morphs relative to S-morphs, maximum genetic diversity was found in populations showing an excess of L-morphs (60·7 % L-morph). On the other hand, no significant difference in pairwise genetic distances between populations was observed between L- (0·107) and S-morphs (0·106).


Our results indicate that significant deviations from equal morph ratios not only affect plant reproductive success but also population genetic diversity of heterostylous plant species. Hence, when defining conservation measures for populations of heterostylous plant species, morph ratios should be considered as an important trait affecting their long-term population viability.

Keywords: Boraginaceae, distyly, FST, genetic diversity, microsatellite analysis, morph bias, Pulmonaria officinalis, skewed mating success


Heterostyly is a sexual polymorphism that evolved independently in at least 28 flowering plant families (Barrett, 2002). In contrast to dioecious plants, heterostylous plants are characterized by hermaphroditic flowers that show marked differences in floral design between two (distyly) or three (tristyly) distinct mating types. Short-styled (S), long-styled (L) and mid-styled (M) plants (in the case of tristyly) coexist in one population with their reproductive organs reciprocally placed so that cross-pollination between mating groups is facilitated. Flowers exhibit herkogamy and in accordance to the style morph, their stigmas are placed beneath (short), above (long) or between anther whorls (mid) (Darwin, 1877; Barrett, 2002). In most heterostylous species other ancillary polymorphisms can be observed as well, often related to features of the stigma and pollen (e.g. Ganders, 1979).

Besides these morphological differences, most heterostylous species are characterized by a sporophytically controlled, dialellic, heteromorphic incompatibility system (Ganders, 1979; Barrett, 1988, 2002; Barrett and Shore, 2008). This physiological barrier restricts mating opportunities to crossing between plants of the opposite style morph (legitimate pollination). According to Darwin (1877), heterostyly and the associated self-incompatibility system should function as an effective outcrossing mechanism because of pollen segregation on the insect pollinator's body and precise deposition of this pollen on the stigma of the opposite morph. Due to this outcross mating system, populations of heterostylous species can be expected to have high levels of neutral genetic diversity compared with species with a mixed mating system or selfing species (Awadalla and Ritland, 1997; Liu et al., 1998; Jacquemyn et al., 2004). Between-population genetic diversity (FST), on the other hand, is expected to be low in heterostylous species compared with selfing species because of higher levels of pollen flow and less genetic drift (Hamrick and Godt, 1996; Liu et al., 1998; Ingvarsson, 2002; Jacquemyn et al., 2004; Duminil et al., 2009).

Disassortative mating in heterostylous species, enforced by reciprocal herkogamy and associated ancillary polymorphisms (morphology) as well as by the heteromorphic incompatibility system (physiology), is expected to lead to equal morph frequencies (isoplethy) in optimal populations, due to negative frequency-dependent selection and the simple inheritance of heterostyly (Finney, 1952; Ganders, 1979; Heuch, 1979; Barrett and Shore, 2008). However, in several heterostylous species, the heteromorphic incompatibility system has been shown to be subject to leakage, and in some genera breakdown of heterostyly through the loss of self-incompatibility has been observed (Ganders, 1979; Barrett, 1989, 1993). Breakdown of self-incompatibility often occurs asymmetrically among style morphs, and may cause asymmetric patterns of mating and seed-set when one morph is capable of producing more seeds after selfing and intra-morph (illegitimate) pollination than the other(s) (Kéry et al., 2003; Brys et al., 2008b; Hodgins and Barrett, 2008). In this case, illegitimate mating and morph-specific differences in reproductive success can lead to biased morph ratios, even if the population is at equilibrium (Barrett and Hodgins, 2006). Besides resulting from asymmetric mating, deviations from equal morph ratios in heterostylous populations may also arise from demographic stochasticity or founding events. In both cases, morph bias is often, but not necessarily, related to population size (Endels et al., 2002; Jacquemyn et al., 2002; Brys et al., 2003, 2007, 2008b; Kéry et al., 2003; Wang et al., 2005; Van Rossum and Triest, 2006b).

Skewed morph ratios may affect population genetic diversity for several reasons. First, deviations from equal morph ratios may lead to high levels of genetic drift by reducing the number of available mates and thus the effective population size (Husband and Barrett, 1992; Ingvarsson, 2002; Charlesworth, 2003; Duminil et al., 2009). Moreover, because morph bias and population size are often related, morph-biased populations are expected to suffer from reduced genetic diversity due to their small size (Leimu et al., 2006; Aguilar et al., 2008). Second, in distylous plant species in which successful mating is asymmetric and outcrossing is restricted to one of the morphs, within-population genetic diversity is expected to decrease asymmetrically in more biased populations. Populations biased towards the self-incompatible morph will suffer more from mate scarcity whereas populations with a higher proportion of the self-compatible morph will become more affected by inbreeding. However, there are only a few studies that have elucidated the relative importance of morph bias in affecting population genetic diversity in heterostylous species, particularly in species with a weak self-incompatibility system (e.g. Van Rossum and Triest, 2006a).

In this study, we investigated the impact of skewed mating success and morph bias on population genetic diversity of the distylous Pulmonaria officinalis (Boraginaceae). Previous research showed that 27 of 35 (77 %) surveyed populations in northern Belgium were L-biased as a result of the weakened heteromorphic incompatibility system of the L-morph (Brys et al., 2008a). More specifically, we asked the following questions: (1) Do the observed differences in mating success between L- and S-morphs lead to asymmetric patterns of within-population genetic diversity? (2) To what extent is within-population genetic diversity affected by population size and morph bias? (3) How do differences in self-incompatibility and mating success between L- and S-morphs affect patterns of genetic differentiation among populations?


Study species

Pulmonaria officinalis L., common lungwort, is a semi-evergreen, long-lived, clonal herb that grows in the understorey of species-rich mixed and open forests on relatively humid, wet and loamy soils (Hegi, 1927; Bennett, 2003). This forest species has a central European distribution, occurring from southern Sweden and Denmark in the north to Italy and Bulgaria in the south (Merxmüller and Sauer, 1972). The fragmented populations in Belgium and Britain are located outside the native distribution range of the species and have become naturalized, most probably after the species' introduction during medieval times. Recent development of eight polymorphic microsatellite markers revealed that P. officinalis is tetraploid as all loci contained more than two and a maximum of four alleles (Molecular Ecology Resources Primer Development Consortium, 2011). Moreover, as the microsatellite profiles showed no differentiation of allele sets and allele combinations were completely random (no fixed heterozygosity), this also indicates that P. officinalis is an auto- rather than an allotetraploid (Catalán et al., 2006).

From March until the end of April, the species produces purple, tube-shaped, distylous flowers characterized by a range of ancillary polymorphisms in corolla tube shape, anther size and pollen size. Flowers of the L-morph have smaller pollen grains and produce significantly more pollen than those of the S-morph (Brys et al., 2008a). The stigmatic surface also shows morph-specific differences in papillae size, crown area and inter-papillae distances. Short-styled pollen showed higher germination rates and higher in vitro pollen tube elongation, which is needed for traversing the long styles of the L-morph (Brys et al., 2008a).

In Belgium, flowers of P. officinalis are mainly pollinated by generalist insect species, including Bombus terrestris, B. pascuorum, B. pratorum and Bombylius major, whereas in the centre of its distribution range (Central Europe) flowers are mainly pollinated by specialist Anthophora species. These species have long proboscises that perfectly fit within the deep corolla tube of P. offinicalis flowers (Oberrath et al., 1995; Brys et al., 2008a). Within the study populations, only one Anthophora species (A. plumipes) was sporadically observed on P. officinalis flowers. Comparative analyses of pollen transport showed that this species was a much more efficient pollinator than the generalist Bombus species as it transported proportionally more S-pollen on its head and more L-pollen on its proboscis (Brys et al., 2008a).

Because of higher pollen deposition rates and a weaker heteromorphic incompatibility system in the L-morph, reproductive success tends to be higher in L- than in S-morphs (Brys et al., 2008a). Higher pollen deposition in the L-morph was attributed to a better accessibility of their stigmatic surface for generalist pollinators with limited proboscis lengths as well as to the floral hairs that grow at the long-styled corolla entrance and in which L-pollen are captured as pollinators pull back their proboscises. This, in combination with the fact that illegitimate pollination yielded approximately 2·5 (intra-morph pollination) to 10 (selfing) times more seeds in the L-morph than in the S-morph (seed-set after selfing = 1·6 %), results in skewed mating success among morphs (Brys et al., 2008a).

Study area and population sampling

The study area is situated in the north of Belgium. In this region, 35 P. officinalis populations were identified during previous research (Brys et al., 2008a, b). Most populations (n = 26) are located in the north-western part of Belgium near Brakel, six are located in the centre of northern Belgium and three in eastern Belgium near Voeren. Distances between populations vary between 478 m and 159 km (mean: 31 km). During the flowering season of 2010 (March–April), we visited a subsample of 27 P. officinalis populations. For each population, population size was determined by counting the total number of flowering ramets. However, because this species reproduces partially by vegetative propagation it was impossible to estimate the total number of genets per population. For each population, the style morph (L/S) was determined for at least 400 individuals (if possible) and morph bias was calculated as the difference in the number of individuals of the L- and the S-morph, divided by the total number of individuals (Brys et al., 2008b). Hence, morph bias varied between –1 (only S-morph) and 1 (only L-morph) with 0 for isoplethy. Table 1 summarizes population sizes and morph bias of the 27 sampled populations.

Table 1.
Characteristics of 27 Belgian Pulmonaria officinalis populations sampled for microsatellite analysis during the flowering season of 2010: six different measures of genetic diversity were calculated for each population

In each of the selected populations, 20 plants were sampled for genetic analysis. Individuals were sampled from the entire area occupied by each population to avoid the effects of population substructure. To investigate the effect of morph type on population genetic diversity, leaves of ten S-plants and ten L-plants were collected. In small and/or biased populations with fewer than ten individuals per morph type, we collected leaves from all individuals available. Young leaf material was collected and dried in silica gel. Before DNA extraction, leaves were homogenized with a mill (Retsch MM 200, Haan, Germany) to a fine powder. Total DNA was extracted from 20 mg silica-dried leaf tissue of 526 plants in total following the NucleoSpin Plant II protocol (Macherey-Nagel, Düren, Germany). DNA quality and concentration were estimated using a NanoDrop ND-2000 spectrophotometer (Thermo Scientific, Wilmington, DE, USA).

Microsatellite analysis

Eight recently developed microsatellites (Molecular Ecology Resources Primer Development Consortium, 2011) were amplified in two multiplex PCRs in a 2720 Thermal Cycler (Applied Biosystems, CA, USA) in a total volume of 10 µL containing 5 µL Qiagen Multiplex PCR Master Mix, 3 µL RNAse-free water, 1 µL template DNA and 1 µL of one of the two multiplexed primer combinations. The two multiplexes had equal thermocycler profiles with initial denaturation of 15 min at 95 °C; 25 cycles of 30 s at 95 °C, 1·5 min at 59 °C and 1 min at 72 °C; and a final elongation of 30 min at 60 °C. Then, 1 µL of the PCR reaction was added to a solution of 8·8 µL formamide and 0·2 µL of the Applied Biosystems' GeneScan 500 LIZ size standard. Fragments were sized on an ABI Prism, 3130 Genetic Analyzer (Applied Biosystems) and scored with GeneMapper Software v4·0 (Applied Biosystems).

Data analysis

Measures of genetic diversity

Due to the tetraploid nature of P. officinalis (Molecular Ecology Resources Primer Development Consortium, 2011), standard genetic analyses cannot be used to estimate measures of genetic diversity. Therefore, we used the software autotet to estimate measures of population genetic diversity (Thrall and Young, 2000). For each population six measures of genetic diversity were calculated: allelic richness (A), allelic richness within individuals (AI) (i.e. the average number of alleles per individual), genotypic richness (G), observed (HO) and expected (HE) heterozygosity, and the fixation or inbreeding coefficient (FIS). Observed heterozygosity was calculated by weighting the five possible classes of genotypes (AAAA, AAAB, AABB, AABC, ABCD) inversely to the probability of any of their alleles being identical by descent. autotet computes expected heterozygosity and fixation coefficients under the assumption of both random chromatid segregation where sister chromatids can segregate into the same gamete (i.e. double reduction, with a maximum probability of α = 1/7) and random chromosomal segregation (Thrall and Young, 2000). For all eight loci, chi-square goodness-of-fit tests for observed-to-expected genotype frequencies were used to test whether populations showed significant deviation from Hardy–Weinberg equilibrium. These six measures were calculated for each population but also separately for both morphs within the population. A paired-sample t-test was then used to test whether one morph is genetically more diverse than the other morph. Finally, we also tested for linkage disequilibrium by using the program LD4X run in R (R Development Core Team, 2010), which is especially developed for autotetraploid species (Julier, 2009). Linkage disequilibrium or allelic association implies a non-random association of alleles at different loci. First, the software builds 5 × 5 contingency tables for each pair of alleles with allelic dosages from 0 to 4. Secondly, Fisher's exact tests are performed to detect linkage between pairs of alleles from the eight microsatellite loci applied in the analysis. The program applies a Bonferroni correction by adjusting the probability threshold to the number of comparisons tested for each pair of loci.

Relationship between morph bias and genetic diversity

To assess the relationship between measures of genetic diversity (A, AI, G, HO, HE, FIS), morph bias and population size, multiple regressions were conducted in SAS v9·1 using the ‘reg’ procedure. Population size was log-transformed to satisfy the normality assumption. Morph bias, ranging from –1 to 1, was preferred as an independent variable to the absolute value of morph bias (Van Rossum and Triest, 2006a), because we were particularly interested in the distinction between L-morph-biased and S-morph-biased populations in our dataset. Both the linear and the quadratic terms of morph bias were entered in the models. For all models tested, Akaike's information criterion (AIC) was used to check statistical model fit. The competing model with the smallest AIC value was considered the best model. In case the variation in the data was best described by the quadratic function, the morph ratio for which genetic diversity was at its maximum was determined by calculating the first derivative of the curve equation.

Genetic differentiation

Pairwise genetic distances (FST) among the 27 populations were estimated using polysat, a newly developed package in R that calculates FST values from polyploid microsatellite data scored in GeneMapper (Clark and Jasieniuk, 2011). polysat provides a separate set of functions which are applied depending on the species' inheritance mode (autopolyploidy versus allopolyploidy). Allele frequencies were estimated using the ‘simpleFreq’ method, which assumes that all alleles in partial heterozygotes have an equal chance of being present in more than one copy (Clark and Jasieniuk, 2011). On the basis of these allele frequencies, pairwise FST values were calculated between all 27 populations for all individuals, and for L- and S-morphs separately. To investigate whether pairwise FST values were significantly different between L- and S-morphs, a paired-sample t-test was used. Finally, a Mantel test in GenAlEx 6 (Peakall and Smouse, 2006) was used to test the hypothesis that pairwise genetic distances of S-morphs were significantly related to those of L-morphs. The deviation from equal pairwise FST values for both morphs was tested statistically by calculating the 95 % confidence interval around the slope of this relationship.


Measures of genetic diversity

The total number of alleles observed per locus in the overall sample of 526 individuals ranged from three to 15, with an overall total of 53 alleles scored over eight loci. For loci Puof-04, Puof-12 and Puof-15, we found only 15, three and six alleles, respectively, which is less than reported by Molecular Ecology Resources Primer Development Consortium (2011) because no German populations were included for this analysis. For the loci Puof-08, Puof-13 and Puof-19, by contrast, we found more alleles (seven, eight and four, respectively) due to our larger sample size. Significant departures from Hardy–Weinberg equilibrium were observed in all single-locus exact tests of the pooled population data. Chi-square goodness-of-fit tests for observed-to-expected genotype frequencies showed for all loci except for one (Puof-12), a negative deviation from Hardy–Weinberg equilibrium. The χ2-values of all single-locus exact tests calculated for each population separately are summarized in Supplementary Data Table S1 (available online).

Among the 27 sampled populations, 1163 of the 29 926 allelic combinations were significantly correlated at P < 0·05. After Bonferroni correction, only 339 combinations (1·1 %) remained significant in 23 of 27 populations, illustrating that linkage disequilibrium is very limited among the developed microsatellite markers. Allelic richness (A) and average allelic richness per individual (AI) ranged from 1·88 to 4·38 and 2·66 to 3·25, respectively, whereas genotypic diversity (G) varied between 1·88 and 7·38 (Table 1). Expected heterozygosity (HE) varied between 0·35 and 0·71 (Table 1). Negative overall FIS values were observed in all populations and varied between –1·00 and –0·15 (Table 1). None of these six measures differed significantly between the two morphs (0·145 ≤ P ≤ 0·823). A significant difference between morphs could have biased the overall population genetic diversity measurement as both morphs were equally sampled.

Morph bias and genetic diversity

Consistent with previous analyses (Brys et al., 2008b), the average morph bias of the studied populations was 0·17 and deviated significantly from zero (t = 2·216, P = 0·036), indicating that the majority of the studied populations were dominated by individuals of the L-morph. Comparison of regression models with linear and quadratic terms of morph bias and population size showed that, for all measures of genetic diversity used, morph bias2 was the only significant factor (P > 0·05) associated with within-population genetic diversity and thus was a better descriptor of the variability in genetic diversity than population size (Fig. 1, Table 2). AIC fit values for all five competing models are reported in supplementary Table S2 (Supplementary Data). Variation in five of six genetic measures was best explained by the multiple regression model that included the linear and quadratic term of morph bias. Only in the case of genotypic richness (G) did the best model include population size. Genetic diversity was low in strongly morph-biased populations, and was highest in populations that showed slight L-morph excess. Interestingly, calculation of the morph ratio corresponding to maximum genetic diversity from the obtained regression equations showed that, on average, genetic diversity was highest in populations with a morph bias of 0·213 (± 0·013), corresponding to an L-biased population with 60·7 % (± 0·7) L-morph individuals. However, the shift of the curve translated in the coefficient of the variable ‘morph bias’ did not significantly differ from 0 (Table 2).

Fig. 1.
Effect of morph bias on six measures of genetic diversity (autotet). Morph bias varies from –1·0 (only S-morphs) to 1·0 (only L-morphs) through isoplethy (0·0). Each of the six quadratic models was found to be significant: ...
Table 2.
Regression coefficient estimates and the level of significance (*P < 0·05, **P < 0·01, ***P < 0·001) obtained by multiple polynomial regressions for six measures of genetic diversity (A, AI, G, HO, HE, ...

Genetic differentiation of populations between morphs

Pairwise FST values varied between 0·009 and 0·267 (mean = 0·106). No significant difference was found in average pairwise FST values between the L- (0·107) and S-morph (0·106) (P = 0·722). The histogram in Fig. 2A shows a normal distribution of 300 differences in FST values between both morphs centred around zero and a maximum difference of 0·12. Figure 2B shows the correlation between the pairwise population FST values of L- and S-morphs, separately. The right and left tails of the normal distribution in Fig. 2A correspond to the cloud of points above and beneath the dashed line in Fig. 2B, respectively. Both series of FST values were found to be highly correlated (r = 0·761, P < 0·001). As the 95 % confidence interval around the slope (0·715–0·869) excluded the slope value 1 (= equal pairwise FST values), the deviation from the 1 : 1 dashed line (Fig. 2B) was found to be significant.

Fig. 2.
(A) Frequency distribution of the difference between the 300 pairwise population FST values of the L- and the S-morph. (B) Correlation between pairwise population FST values for the L- (x-axis) and the S-morph (y-axis) (continuous line). The 1 : 1 ...


Skewed morph frequencies are not unusual in populations of heterostylous species and variable deviations from isoplethy have been observed before in several species. Figure 3 shows the average L-morph frequency and variation in morph ratios for 12 distylous species. Although the exact causes of skewed morph ratios are not always easy to untangle (Kéry et al., 2003), it has been shown that skewed morph ratios can have a pronounced negative effect on overall reproductive success within heterostylous populations and thus can contribute to patterns of genetic diversity. Kéry et al. (2000) and Brys et al. (2007), for example, found a marked decrease in sexual reproduction in the distylous species Primula veris and Hottonia palustris when populations became more biased in either morph. Wang et al. (2005) and Brys et al. (2008b) disentangled the effects of morph ratio and population size on reproductive success by manipulating morph ratios in experimental populations and confirmed a significant negative effect of morph bias on population fecundity, especially in individuals belonging to the morph type in the majority.

Fig. 3.
Average L-morph frequency and the variation in morph ratios of populations from 12 different distylous species. The dashed line indicates isoplethy (= equal morph frequencies).

In line with our expectations, we showed that morph bias had a pronounced negative effect on within-population genetic diversity. All six measures of genetic diversity were significantly affected by morph bias (Table 2), and AIC values confirmed the higher predictive value of morph bias over population size in five of six response variables (Table S2). Although previous meta-analyses have shown a clear positive correlation of population size and genetic diversity in many plant species (Honnay and Jacquemyn, 2007), our results show that morph bias is a more important demographic variable to explain population genetic diversity in this distylous species. In contrast to our results, Van Rossum and Triest (2006a) did not find a negative relationship between morph bias and each of four genetic diversity parameters (A, HO, HE, FIS) in 24 populations of the distylous Primula veris. In Primula elatior, however, a similar significantly negative effect of morph bias on allelic richness (A) was found in seedlings during an early stage of population fragmentation, which the authors attributed to limited availability of compatible mates and increased inbreeding (Van Rossum and Triest, 2006b).

The effect of morph bias was best described by a negative quadratic function, which showed a decrease in genetic diversity with an increasing deviation from the isoplethic situation (Fig. 1). Interestingly, maximum genetic diversity was found in populations in which 60·7 % of all plants belonged to the L-morph, which is very close to the equilibrium morph ratio reported by Brys et al. (2008b). Due to the strict heteromorphic self-incompatibility system in most heterostylous species, the dominant morph in biased populations is generally characterized by a fitness disadvantage because of fewer mating opportunities (fewer pollen donors/recipients) compared with the rare morph. Such negative frequency-dependence will lead to equilibrium morph ratios in which all morphs have an equivalent fitness (Fisher, 1930; Thompson et al., 2003). However, in species that are lacking strict self-incompatibility and are characterized by asymmetric illegitimate pollination patterns, equilibrium morph ratios may significantly deviate from a 1 : 1 ratio due to the higher reproductive success of one of the two morphs. In P. officinalis, for example, significantly higher rates of illegitimate pollination in the L-morph individuals resulted in a 1·36-fold higher female fecundity compared with the S-morph, which resulted in an equilibrium L/S-morph ratio of 1·94 : 1 (i.e. 66 % L-plants) (Brys et al., 2008b). This L-biased equilibrium ratio was also reflected in patterns of genetic diversity, which showed that maximum genetic diversity occurred at a morph ratio of 60 % L-plants, and decreased with increasing deviations from the equilibrium morph ratio (Fig. 1).

In species with strict self-incompatibility, where only mating between plants of the opposite morph is possible, an increasing morph bias leads to smaller effective population sizes (Husband and Barrett, 1992), which, in turn, leads to higher genetic drift and lowers genetic diversity. However, in species in which partial selfing is possible, reductions in genetic diversity may also result from selfing, as it further lowers the effective population size (Ingvarsson, 2002; Charlesworth, 2003; Siol et al., 2007; Duminil et al., 2009). This may have contributed to the lower genetic diversity in the studied P. officinalis populations, particularly in L-biased populations. However, the inbreeding coefficient decreased rather than increased with increasing morph bias (Fig. 1F), implying an increase in the proportion of heterozygotes as morph ratios became more skewed. Moreover, because patterns were more or less similar for L- and S-biased populations, this does not correspond to the observation that selfing can only occur in the L-morph. Therefore, partial selfing is unlikely to have contributed substantially to the reduced genetic diversity in morph-biased populations.

Decreasing inbreeding coefficients with increasing morph bias have also been reported by Van Rossum and Triest (2006a) for the distylous Primula veris. Overdominance (i.e. higher fitness of the heterozygous genotype compared with the corresponding homozygous genotype) can result in an excess of heterozygotes in biased populations (David, 1998; Hansson and Westerberg, 2002). In small, inbreeding populations of the self-incompatible perennial Arnica montana (Asteraceae), Luijten et al. (2000) also found more heterozygotes. This ‘heterozygosity paradox’, which was first described by Brown (1979), was also found by Barrett and Husband (1990). They reported an excess of heterozygotes in small Jamaican populations of the tristylous species Eichhornia paniculata that showed low outcrossing rates. Van Rossum and Triest (2006b) found, in contrast to our results, an increasing heterozygote deficiency in skewed populations of the self-incompatible Primula elatior (FIS-values: –0·058 to 0·480), which indicates high levels of genetic drift in morph-biased populations.

Genetic diversity and its distribution within and among populations of a particular species is known to be strongly influenced by mating system (Hamrick and Godt, 1996; Ingvarsson, 2002; Duminil et al., 2009). Whereas in the studied P. officinalis populations, plants of the S-morph are more bound to outcrossing, plants of the L-morph most likely exhibit a mixed-mating system that allows certain amounts of intramorph- and self-fertilization (Brys et al., 2008a, b). Hence, when comparing genetic differentiation among populations for S- and L-morphs separately, we expected to obtain lower FST-values for the outcrossing S-morph. Ness et al. (2010) found this pattern when comparing FST-values among selfing and outcrossing populations of the tristylous species Eichhornia paniculata. In P. officinalis, however, no significant difference was found in the average population genetic distances between the L-morph (0·107) and S-morph (0·106) (Fig. 2A). However, the relationship between pairwise genetic distance of the S- and L-morph did not coincide with the dashed line that indicates no difference in pairwise FST-values between the two morphs (Fig. 2B). When FST-values were high (>0·1), pairwise genetic distances between populations were higher for the L-morph than for the S-morph, whereas at high levels of gene flow (FST < 0·1), pairwise genetic distances were higher for the S-morph than for the L-morph. This difference can be explained by differences in mating system between the S-morph (outcrossing) and L-morph (mixed-mating). Anthers of the L-morph produce nearly twice as many pollen grains and long styles receive significantly more pollen than short styles, both resulting in higher gene flow among the L-morphs of different populations when populations are close to each other. However, in spatially isolated populations, for which levels of gene flow are low, pairwise genetic distances are expected to be higher for the L-morph than for the S-morph because of higher levels of inbreeding, as some of the pollen of the L-morph might be used for self-fertilization, which promotes genetic differentiation between populations (Charlesworth and Pannell, 2001; Ingvarsson, 2002).

To conclude, this study has shown that morph ratio exerted a non-negligible effect on population genetic diversity and differentiation of a distylous plant species with a weak self-incompatibility system. The effect of morph bias on genetic diversity was clearly shaped by the skewed mating success between the morphs, as shown by the shifted top of the curve towards an equilibrium morph composition with a majority of L-individuals. Because of the negative effect of morph bias on fecundity, seedling recruitment and genetic diversity in distylous populations, morph ratio imbalance should definitely be considered in conservation management as it also affects the long-term survival of rare and threatened species (Endels et al., 2002).


This work was supported by the Flemish Fund for Scientific Research (FWO) [project G.0500·10 and a postdoctoral grant (R.B.)] and the European Research Council (ERC starting grant 260601 – MYCASOR) (H.J.).


  • Aguilar R, Quesada M, Ashworth L, Herrerias-Diego Y, Lobo J. Genetic consequences of habitat fragmentation in plant populations: susceptible signals in plant traits and methodological approaches. Molecular Ecology. 2008;17:5177–5188. [PubMed]
  • Awadalla P, Ritland K. Microsatellite variation and evolution in the Mimulus guttatus species complex with contrasting mating systems. Molecular Biology and Evolution. 1997;14:1023–1034. [PubMed]
  • Barrett SCH. The evolution, maintenance and loss of self-incompatibility systems. In: Lovett-Doust J, Lovett-Doust L, editors. Reproductive strategies of plants: patterns & strategies. New York: Oxford University Press; 1988. pp. 98–124.
  • Barrett SCH. The evolutionary breakdown of heterostyly. In: Linhart Y, Bock J, editors. The evolutionary ecology of plants. Boulder, CO: Westview Press; 1989. pp. 151–169.
  • Barrett SCH. The evolutionary biology of tristyly. In: Futuyma D, Antonovics J, editors. Oxford surveys in evolutionary biology. Oxford: Oxford University Press; 1993. pp. 283–326.
  • Barrett SCH. The evolution of plant sexual diversity. Nature Reviews. 2002;3:274–284. [PubMed]
  • Barrett SCH, Hodgins KA. Floral design and the evolution of asymmetrical mating. In: Harder LD, Barrett SCH, editors. The ecology and evolution of flowers. Oxford: Oxford University Press; 2006. pp. 239–255.
  • Barrett SCH, Husband BC. Variation in outcrossing rates in Eichhornia paniculata: the role of demographic and reproductive factors. Plant Species Biology. 1990;5:41–55.
  • Barrett SCH, Shore JS. New insights on heterostyly: comparative biology, ecology and genetics. In: Franklin-Tong V, editor. Self-incompatibility in flowering plants: evolution, diversity and mechanisms. Berlin: Springer-Verlag; 2008. pp. 3–32.
  • Bennett M. Pulmonarias and the Borage family. London: BT Batsford; 2003.
  • Brown AHD. Enzyme polymorphism in plant populations. Theoretical Population Biology. 1979;15:1–42.
  • Brys R, Jacquemyn H, Endels P, Hermy M, De Blust G. The relationship between reproductive success and demographic structure in remnant populations of Primula veris. Acta Oecologica. 2003;24:247–253.
  • Brys R, Jacquemyn H, Hermy M. Impact of mate availability, population size, and spatial aggregation of morphs on sexual reproduction in a distylous, aquatic plant. American Journal of Botany. 2007;94:119–127. [PubMed]
  • Brys R, Jacquemyn H, Hermy M, Beeckman T. Pollen deposition rates and the functioning of distyly in the perennial Pulmonaria officinalis (Boraginaceae) Plant Systematics and Evolution. 2008a;273:1–12.
  • Brys R, Jacquemyn H, Beeckman T. Morph-ratio variation, population size and female reproductive success in distylous Pulmonaria officinalis (Boraginaceae) Journal of Evolutionary Biology. 2008b;21:1281–1289. [PubMed]
  • Catalán P, Segarra-Moragues JG, Palop-Esteban M, Moreno C, González-Candelas F. A Bayesian approach for discriminating among alternative inheritance hypotheses in plant polyploids: the allotetraploid origin of genus Borderea (Dioscoreaceae) Genetics. 2006;172:1939–1953. [PMC free article] [PubMed]
  • Charlesworth D. Effect of inbreeding on the genetic diversity of populations. Philosophical Transactions of the Royal Society B. 2003;358:1051–1070. [PMC free article] [PubMed]
  • Charlesworth D, Pannell JR. Mating system and population structure in the light of the coalescence theory. In: Silvertown J, Antonovics J, editors. Integrating ecology and evolution in a spatial context. London: Blackwell Science; 2001. pp. 73–95.
  • Clark LV, Jasieniuk M. POLYSAT: an R package for polyploidy microsatellite analysis. Molecular Ecology Resources. 2011;11:562–566. [PubMed]
  • Darwin C. The different forms of flowers on plants of the same species. London: John Murray; 1877.
  • David P. Heterozygosity–fitness correlations: new perspectives on old problems. Heredity. 1998;80:531–537. [PubMed]
  • Duminil J, Hardy OJ, Petit RJ. Plant traits correlated with generation time directly affect inbreeding depression and mating system and indirectly genetic structure. BMC Evolutionary Biology. 2009;9(177) http://dx.doi.org/10.1186/1471-2148-9-177 . [PMC free article] [PubMed]
  • Endels P, Jacquemyn H, Brys R, Hermy M, de Blust G. Temporal changes (1986–1999) in populations of primrose (Primula vulgaris Huds.) in an agricultural landscape and implications for conservation. Biological Conservation. 2002;105:11–25.
  • Finney DJ. The equilibrium of a self-incompatible polymorphic species. Genetica. 1952;26:33–64. [PubMed]
  • Fisher RA. The genetical theory of natural selection. New York: Dover; 1930.
  • Ganders FR. The biology of heterostyly. New Zealand Journal of Botany. 1979;17:607–35.
  • Hamrick JL, Godt MJW. Effects of life history traits on genetic diversity in plant species. Philosophical Transactions oft he Royal Society B: Biological Sciences. 1996;351:1291–1298.
  • Hansson B, Westerberg L. On the correlation between heterozygosity and fitness in natural populations. Molecular Ecology. 2002;11:2467–2474. [PubMed]
  • Hegi G. Illustrierte Flora von Mitteleuropa. München: JF Lehmanns Verlag; 1927. Pulmonaria officinalis L; pp. 2213–2216. Band V, Teil 3.
  • Heuch I. Equilibrium populations of heterostylous plants. Theoretical Population Biology. 1979;15:43–57.
  • Hodgins KA, Barrett SCH. Asymmetrical mating patterns and the evolution of biased morph ratios in a tristylous daffodil. Genetic Research. 2008;90:3–15. [PubMed]
  • Honnay O, Jacquemyn H. Susceptibility of common and rare plant species to the genetic consequences of habitat fragmentation. Conservation Biology. 2007;21:823–831. [PubMed]
  • Husband BC, Barrett SCH. Effective population size and genetic drift in tristylous Eichhornia paniculata (Pontederiaceae) Evolution. 1992;46:1875–1890.
  • Ingvarsson PK. A metapopulation perspective on genetic diversity and differentiation in partially self-fertilizing plants. Evolution. 2002;56:2368–2373. [PubMed]
  • Jacquemyn H, Brys R, Hermy M. Patch occupancy, population size and reproductive success of a forest herb (Primula elatior) in a fragmented landscape. Oecologia. 2002;130:617–625.
  • Jacquemyn H, Honnay O, Galbusera P, Roldán-Ruiz I. Genetic structure of the forest herb Primula elatior in a changing landscape. Molecular Ecology. 2004;13:211–219. [PubMed]
  • Julier B. A program to test linkage disequilibrium between loci in autotetraploid species. Molecular Ecology Resources. 2009;9:746–748. [PubMed]
  • Kéry M, Matthies D, Spillmann H-H. Reduced fecundity and offspring performance in small populations of the declining grassland plants Primula veris and Gentiana lutea. Journal of Ecology. 2000;88:17–30.
  • Kéry M, Matthies D, Schmid B. Demographic stochasticity in population fragments of the declining distylous perennial Primula veris (Primulaceae) Basic Applied Ecology. 2003;5:197–206.
  • Leimu R, Mutikainen P, Koricheva J, Fischer M. How general are positive relationships between plant population size, fitness and genetic variation? Journal of Ecology. 2006;94:942–952.
  • Liu F, Zhang L, Charlesworth D. Genetic diversity in Leavenworthia populations with different inbreeding levels. Proceedings of the Royal Society B. 1998;265:293–301. [PMC free article] [PubMed]
  • Luijten SH, Dierick A, Oostermeijer JGB, Raijmann LEL, Den Nijs HCM. Population size, genetic variation, and reproductive success in a rapidly declining, self-incompatible perennial (Arnica montana) in The Netherlands. Conservation Biology. 2000;14:1776–1787.
  • Agata K, Alasaad S, Almeida-Val VMF, et al. Molecular Ecology Resources Primer Development Consortium. Permanent Genetic Resources added to Molecular Ecology Resources Database 1 December 2010–31 January 2011. Molecular Ecology Resources. 2011;11:586–589. [PubMed]
  • Merxmüller H, Sauer W. Pulmonaria L. In: Tutin TG, Heywood VH, Burges NA, et al., editors. Flora Europaea. Vol. 3. London: Cambridge University Press; 1972. 100.
  • Ness RW, Wright SI, Barrett SCH. Mating-system variation, demographic history and patterns of nucleotide diversity in the tristylous plant Eichhornia paniculata. Genetics. 2010;184:381–392. [PMC free article] [PubMed]
  • Oberrath R, Zanke C, Böhning-Gaese K. Triggering and ecological significance of floral color-change in Lungwort (Pulmonaria sp.) Flora. 1995;190:155–159.
  • Peakall R, Smouse PE. GENALEX 6: Genetic analysis in Excel. Population genetic software for teaching and research. Molecular Ecology Notes. 2006;6:288–295. [PMC free article] [PubMed]
  • R Development Core Team. R: A language and environment for statistical computing. Vienna: R Foundation for Statistical Computing; 2010. http://www.r-project.org/
  • Siol M, Bonnin I, Olivieri I, Prosperi JM, Ronfort J. Effective population size associated with self-fertilization: lessons from temporal changes in allele frequencies in the selfing annual Medicago truncatula. Journal of Evolutionary Biology. 2007;20:2349–2360. [PubMed]
  • Thompson JD, Barrett SCH, Baker AM. Frequency-dependent variation in reproductive success in Narcissus: implications for the maintenance of stigma-height dimorphism. Proceedings of the Royal Society B. 2003;270:949–953. [PMC free article] [PubMed]
  • Thrall PH, Young A. Autotet: a program for analysis of autotetraploid genotypic data. Journal of Heredity. 2000;91:348–349. [PubMed]
  • Van Rossum F, Triest L. Within-population genetic variation in the distylous Primula veris: does floral morph anisoplethy matter in fragmented habitats? Perspectives in Plant Ecology, Evolution and Systematics. 2006a;7:263–273.
  • Van Rossum F, Triest L. Fine-scale genetic structure of the common Primula elatior (Primulaceae) at an early stage of population fragmentation. American Journal of Botany. 2006b;93:1281–1288. [PubMed]
  • Wang Y, Wang Q, Guo Y, Barrett SCH. Reproductive consequences of interactions between clonal growth and sexual reproduction in Nymphoides peltata: a distylous aquatic plant. New Phytologist. 2005;165:329–336. [PubMed]

Articles from Annals of Botany are provided here courtesy of Oxford University Press
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...