Skip to main content
NCBI home page
Annals of Botany logoLink to Annals of Botany
. 2012 May 29;110(1):91–99. doi: 10.1093/aob/mcs084

Estimation of heritability, evolvability and genetic correlations of two pollen and pistil traits involved in a sexual conflict over timing of stigma receptivity in Collinsia heterophylla (Plantaginaceae)

Josefin A Madjidian 1,2,*, Stefan Andersson 1, Åsa Lankinen 1,2
PMCID: PMC3380587  PMID: 22645118

Abstract

Background and Aims

Heritable genetic variation is crucial for selection to operate, yet there is a paucity of studies quantifying such variation in interactive male/female sexual traits, especially those of plants. Previous work on the annual plant Collinsia heterophylla, a mixed-mating species, suggests that delayed stigma receptivity is involved in a sexual conflict: pollen from certain donors fertilize ovules earlier than others at the expense of reduced maternal seed set and lower levels of pollen competition.

Methods

Parent–offspring regressions and sib analyses were performed to test for heritable genetic variation and co-variation in male and female interactive traits related to the sexual conflict.

Key Results

Some heritable variation and evolvability were found for the female trait (delayed stigma receptivity in presence of pollen), but no evidence was found for genetic variation in the male trait (ability to fertilize ovules early). The results further indicated a marginally significant correlation between a male's ability to fertilize early and early stigma receptivity in offspring. However, despite potential indirect selection of these traits, antagonistic co-evolution may not occur given the lack of heritability of the male trait.

Conclusions

To our knowledge, this is the first study of a plant or any hermaphrodite that examines patterns of genetic correlation between two interactive sexual traits, and also the first to assess heritabilities of plant traits putatively involved in a sexual conflict. It is concluded that the ability to delay fertilization in presence of pollen can respond to selection, while the pollen trait has lower evolutionary potential.

Keywords: Antagonistic trait, co-evolution, Collinsia heterophylla, evolvability, genetic correlation, genetic variation, heritability, mate choice, pollen competition, sexual conflict, stigma receptivity

INTRODUCTION

A fundamental requirement for evolution by natural selection is the presence of heritable variation in traits with direct or indirect effects on fitness (Darwin, 1859; Fisher, 1958). In flowering plants, much effort has been devoted to the estimation of genetic variation in characters describing morphology, phenology or life history [reviewed in Geber and Griffen (2003) and in Ashman and Majetic (2006)]. Few studies have considered traits directly involved in sexual interactions, such as those that take place during pollen germination on the stigma and during the growth of pollen tubes in the stylar tissue of the maternal parent (Skogsmyr and Lankinen, 2000; Lamborn et al., 2005; Jolivet and Bernasconi, 2007; Lankinen et al., 2007, 2009; see also review by Mazer et al., 2010).

Selection operating on characters or abilities that result in reproductive advantages over other individuals of the same sex is referred to as sexual selection, as opposed to natural selection on enhanced survival or fecundity (Darwin, 1871; Skogsmyr and Lankinen, 2002). In plants, pollen competition for fertilization is a main driver of evolution of pollen sexual traits as it provides an opportunity for selective processes that sort among pollen donors and individual pollen genotypes (Mulcahy, 1979; Wilson and Burley, 1983; Bernasconi, 2003; Bernasconi et al., 2003; Armbruster and Rogers, 2004; Mazer et al., 2010). Pollen competition is further believed to be beneficial for the female, as enhanced pollen competition often leads to increased offspring fitness (e.g. Mulcahy, 1971; Marshall and Whittaker, 1989; Jóhannsson and Stephenson, 1997; Skogsmyr and Lankinen, 2000; Quesada et al., 2001; Paschke et al., 2002; Lankinen and Armbruster, 2007). Female choice for competitive pollen can be achieved by, for example, growing a long style (Mulcahy, 1983; Ruane, 2009), having a large stigmatic surface (Armbruster, 1996; Rodrigo et al., 2009) or delaying stigma receptivity (Wilson and Burley, 1983; Galen et al., 1986; Herrero, 2003). However, it is important to consider that pollen–pollen competition and female choice are interactive processes (Mulcahy, 1983; Lankinen and Skogsmyr, 2001; cf. cryptic female choice in animals, e.g. Eberhard, 1996); male traits enhancing pollen competitive ability are not independent of selection on pistil traits and conversely, female traits intensifying pollen competition are not independent of selection on pollen competitive ability, as suggested in some species (Armbruster, 1996; Dahl and Fredrikson, 1996; Herrero and Hormaza, 1996). To fully understand how sexual traits (co-)evolve in plant populations, it is crucial to assess their heritability as well as the degree of genetic correlation between them.

Co-evolution between the sexes is a cornerstone of sexual conflict theory, a theory that has developed from sexual selection research (Parker, 1979). This theory states that a trait will be under positive selection whenever the benefits of the trait to the bearer exceed the cost imposed on the partner (Parker, 1979). Such a trait is referred to as a sexually antagonistic trait, and has the potential to cause a counter-response in the form of a sexually antagonistic trait in the partner, resulting in co-evolution between sexes (Holland and Rice, 1998; reviewed in Arnqvist and Rowe, 2005). The counter-response to an antagonistic trait could be seen as a screening mechanism to avoid costly partners, similar to the screening performed by female choice mechanisms (Kokko et al., 2003). In this context, it is highly relevant to consider the role played by genetic factors, e.g. genetic correlations between sexually antagonistic traits of males and females (Moore and Pizzari, 2005; Agrawal and Stinchcombe, 2009). If the antagonistic traits are positively genetically correlated, then selection for harmful, sexually antagonistic traits in one sex can be expected to increase the resistance to these traits in the opposite sex, which in turn could select for even more harmful traits in the first sex. On the other hand, if the traits are negatively correlated, selection for harmful traits in one sex will increase the frequency of less-resistant partners, leading to an additional indirect cost of harm (Moore and Pizzari, 2005). Empirical results in support of these expectations are scarce, and virtually nothing is known about heritabilities and genetically based correlations of traits putatively involved in sexual conflict (Moore and Pizzari, 2005). To our knowledge, only one recent study, on the seed beetle Callosobruchus maculatus by Gay et al. (2010), has assessed heritable variation in two interactive sexually antagonistic traits.

Despite increasing interest among plant evolutionary biologists (Bernasconi et al., 2004; Prasad and Bedhomme, 2006; Lankinen and Larsson, 2007; Bedhomme et al., 2009), only a few efforts have been made to apply sexual conflict theory to plants (but see Härdling and Nilsson, 2001; Lankinen et al., 2006; Lankinen and Kiboi, 2007; Madjidian and Lankinen, 2009), as opposed to the more common application of traditional sexual selection theory (e.g. Queller, 1983; Snow and Spira, 1991). One exception concerns the annual hermaphroditic plant Collinsia heterophylla, whose flowers possess delayed stigma receptivity (Armbruster et al., 2002), a trait that has been suggested to enhance both progeny quantity and quality by intensifying pollen competition (Lankinen and Armbruster, 2007; Lankinen and Madjidian, 2011). Studies of C. heterophylla indicate that delayed stigma receptivity may be involved in a sexual conflict: pollen from certain donors are able to fertilize ovules at early stages of floral development, preventing a ‘fair start’ of pollen in the race towards the ovules and, in turn, causing reduced seed set for the maternal parent (Lankinen and Kiboi, 2007). Further maternal plants have been shown to be released from the cost of early fertilization when crossed with donors from a distant geographic location, a pattern consistent with a history of sexually antagonistic co-evolution within local populations (Madjidian and Lankinen, 2009).

In this study, we test for genetic variation in three traits potentially related to sexual conflict over timing of stigma receptivity in C. heterophylla: (1) the female-controlled timing of stigma receptivity in the absence of pollen (henceforth innate onset of stigma receptivity); (2) the ability of the stigma to delay receptivity when pollen is present (henceforth pistil onset of stigma receptivity); and (3) the ability of deposited pollen to germinate early on the stigmatic surface (henceforth pollen onset of stigma receptivity). Pistil and pollen onset of stigma receptivity (2 and 3) were determined from manually pollinated flowers and therefore represent potential sexually antagonistic traits expressed under direct male–female interaction. Apart from assessing heritability of the three traits, we examined the pattern of genetic correlation between them. Because C. heterophylla is a self-compatible hermaphrodite, i.e. both sexual functions are present in the same individual, we were particularly interested in how traits related to the male or female function in the parent generation covaried with the traits related to the opposite function in the progeny generation. Given the antagonistic co-evolution inferred from previous experiments (Madjidian and Lankinen, 2009), we predicted positive correlations between genetically variable pollen and pistil traits (cf. Moore and Pizzari, 2005).

MATERIALS AND METHODS

Study species

Collinsia heterophylla Graham (Plantaginaceae), known as ‘Chinese houses’, is a diploid (2n = 14), annual hermaphrodite, native to the California Floristic Province, North America (Newsom, 1929; Neese, 1993). The species flowers between March and June depending on elevation and latitude, and the main pollinators are native species of bees in the genera Osmia, Bombus and Anthophora (Armbruster et al., 2002). Flowers are arranged in whorls on spikes and have a zygomorphic corolla with two lips. Corolla colour can be white to pale purple on the upper lip and pale to dark purple on the lower lip. Each flower has four epipetalous stamens and one pistil, containing up to 20 ovules (Armbruster et al., 2002; Madjidian and Lankinen, 2009). Up to approx. 60 deposited pollen grains can be in direct contact with the stigmatic surface at the same time (Å. Lankinen and W. S. Armbruster, unpubl. res.) and the stigma can hold many more pollen grains in additional layers. Thus, it is quite likely that pollen competition can occur in this species.

Collinsia heterophylla possesses a mixed mating system, involving both outcrossing and selfing in the same individual (Armbruster et al., 2002). When a flower opens, the anthers are undehisced, the pistil is short and the stigma is unreceptive. Anthers dehisce one at a time over 3–4 d, while the style elongates and the stigma becomes receptive (Lankinen et al., 2007). Self-pollination can occur at a late developmental stage, when the elongated style comes into contact with the dehisced anthers (Kalisz et al., 1999; Armbruster et al., 2002). On average, anther–stigma contact takes place at about the same time as stigmas become fully receptive (Lankinen et al., 2007), reducing the probability that self-pollen will be deposited on unreceptive stigmas. Ovaries develop into dry dehiscent capsules. Estimates of the mean population outcrossing rate ranged from 0·32 to 0·64 in an allozyme-based study (Charlesworth and Mayer, 1995) and from 0·62 to 1·00 in an early investigation based on morphological markers (Weil and Allard, 1964).

This study involves plants collected as seeds from a natural population in California, Mariposa County (situated at 37·5019N, 120·1236W), sampled by collecting seeds from about 50 open-pollinated maternal individuals. Two generations of random outcrossing performed on fully receptive pistils were used to establish a completely outbred base population for the present study. All pollinations were based on plants raised from cold-stratified seeds and grown under pollinator-proof conditions during the winter and spring of 2009 and 2010.

Traits measured

Pollen and pistil onset of stigma receptivity was estimated by performing hand-pollinations at particular stages of floral development. Pistil onset of stigma receptivity refers to the earliest floral stage at which flowers on a given recipient set seed with pollen from other individuals, while pollen onset of stigma receptivity refers to the earliest floral stage during which pollen from a given pollen donor resulted in seed maturation on other individuals. One-donor crosses were performed on emasculated flowers at each of four developmental stages, designated F1–F4 to indicate the number of days after anthesis (Armbruster et al., 2002). One-donor crosses rather than mixed-donor crosses were used to evaluate innate pollen onset of stigma receptivity in the absence of competitors. A hand-pollination was carried out by placing pollen grains from a flower on a microscope slide and then applying the pollen onto a stigma until completely covered. Four hours after the pollination, the pistil was cut off to ensure that any resulting seeds had been fertilized by pollen grains that germinated immediately after being applied, thus indicating that the stigma was receptive at the stage of pollination (following Lankinen and Kiboi, 2007).

Innate onset of stigma receptivity, i.e. the timing of stigma receptivity in the absence of pollen deposition, was estimated by excising pistils at each of the four floral stages (F1–F4) and subjecting these pistils to a peroxidase test in vitro (Kearns and Inouye, 1993). Innate onset of stigma receptivity of an individual was defined as the earliest stage at which excised pistils showed signs of peroxidase activity at the stigmatic surface (following Lankinen et al., 2007). Timing of excision is strongly correlated with the occurrence of pollen tubes in the pistil – although some stigmas show signs of peroxidase activity in the absence of pollen tubes, the opposite does not occur (Lankinen et al., 2007).

Experimental procedures

Heritabilities and genetic correlations were inferred from the degree of resemblance between parents and offspring (parent–offspring regression) or between plants in different progeny families (sib analyses). As a first step, we assigned 20 plants as pollen donors and 20 plants as pollen recipients. Each recipient was mated with three donors and each donor was mated with three recipients in an incomplete factorial design, with matings restricted to plants in different maternal families (Fig. 1). These crosses provided parental data on pollen and pistil onset of stigma receptivity, but also yielded seeds with known parentage for the progeny population. Pollinations in a given recipient–donor combination were replicated three times per floral stage, resulting in a total of 720 pollinations (20 recipients × 3 donors × 4 stages × 3 replicates).

Fig. 1.

Fig. 1.

Open in a new tab

Experimental design for assessing heritabilities and genetic correlations of pistil, pollen and innate onset of stigma receptivity in Collinsia heterophylla. In the parental generation, pistil onset of stigma receptivity was estimated by averaging the stage of first seed production across three different pollen donors, while pollen onset of stigma receptivity was estimated by averaging the stage of first seed production across three different recipients. Crosses were done in an incomplete factorial design, so that each progeny individual could be assigned to a paternal and maternal half-sib family. In the progeny generation, 40 and 16 plants were assigned as pollen donors and recipients, respectively. Pistil and pollen onset of stigma receptivity for individual progeny was obtained by averaging the stage of first seed production across five donors and two recipients, respectively. All donors and recipients in the progeny generation were also scored for innate onset of stigma receptivity.

A progeny generation was established by sowing seeds from flowers pollinated at the latest stage (F4) in the parent generation. We chose to use seeds only from this stage to minimize environmental and maternal sources of variation and because this was how we generated seeds for the parental generation. Seeds from each recipient–donor combination were sown and the resulting plants were subjected to experimental pollinations to provide data on pollen and pistil onset of stigma receptivity. The latter data were obtained from a sample of 16 pollen recipients and 40 pollen donors, the excess of donors reflecting our interest in assessing the male trait more thoroughly. Each recipient represented a single, unique mother plant (derived from a separate full-sib family), whereas the sample of donors contained pairs of plants, each sired by the same father but having different mothers (Fig. 1). Each donor was mated with two recipients and each recipient was mated with five donors, with three replicate pollinations per floral stage and parent combination. To be able to perform all necessary crosses for a particular recipient–donor combination, we distributed the pollinations over the focal recipient and one of its full-siblings and treated these pollinations as if they had been carried out on the same maternal individual. A total of 960 hand-pollinations were performed in the progeny generation (16 recipients × 5 donors × 4 stages × 3 replicates). The 60 plants used in crosses were also scored for innate pistil onset of stigma receptivity (based on two excised pistils per floral stage).

Statistical analyses

To assess the genetic component of variation in pollen and pistil onset of stigma receptivity, we regressed mean pollen or pistil onset of stigma receptivity for each progeny individual (averaged over replicates) on the corresponding mean for the male or female plant in the parent generation (based on n = 16 and 20 for pistil and pollen onset, respectively). The narrow-sense heritability (h2) for the two measures was obtained as twice the slope of the parent–offspring regression (Falconer and McKay, 1996). To deal with problems arising from small sample sizes, we analysed the statistical significance of the regressions with bootstrap analyses (Manly, 1991). A bootstrap procedure with 10 000 re-samplings was used to construct bias-corrected confidence intervals (CI) around each estimate, performed in R based on the method described in Crawley (2007). For innate onset of stigma receptivity (for which no parental data were available) and pollen onset of stigma receptivity, it was also possible to estimate h2 as four times the proportion of variance attributed to sire and dam in the progeny generation, based on variance components from two-way ANOVAs with both paternal and maternal parent as random factors (Falconer and McKay, 1996).

As a complementary approach, we also estimated trait evolvabilities (IA; Houle, 1992; Hansen et al., 2003). IA scales the additive (selectable) genetic variance by the squared trait mean (rather than by the total phenotypic variance as in the case of h2) and was calculated directly from h2 as: IA = 100 h2 VP/Z2, where VP is the total variance and Z is the trait mean (Hansen et al., 2003).

Genetic correlations between traits were quantified as Pearson correlation coefficients, based on family means in the progeny generation. Because 5–10 % of genes in plants are specifically expressed in pollen (Borg et al., 2009), it is possible that the paternal control of the progeny involves genetic factors other than those involved in mediating the maternal control (Skogsmyr and Lankinen, 2002). Therefore, we calculated correlation coefficients based on both paternal and maternal progeny means to test for genetically based associations between the three measures of stigma receptivity. Correlations involving female onset of stigma receptivity were not strictly genetic, as each data point for this variable was represented by a single value, rather than a mean across half-sibs (as was the case for innate and pollen onset of stigma receptivity). When possible, we also tested for genetic correlations by regressing the progeny mean for each trait on the mean for another trait in the parent generation, e.g. pollen onset of stigma receptivity of progeny vs. pistil onset of receptivity of the maternal parent (Falconer and McKay, 1996).

Analyses were performed in Statistica 7 for Windows (Statistica, 2005), except for the bootstrapping procedures.

RESULTS

Heritabilities and evolvabilities

Both pollen and pistil onset of stigma receptivity showed non-significant parent–offspring regression according to parametric analyses (pollen onset: r = 0·045, b = 0·026, P = n.s., n = 20; pistil onset: r = 0·412, b = 0·140, P = n.s., n = 16). Bootstrapping the parent–offspring regression coefficient for pistil onset of stigma receptivity resulted in a 95 % CI that excluded zero (0·028, 0·362); thus, the narrow-sense heritability for this variable (h2 = 0·28; Table 1) may be considered significant at the 5 % level. For pollen onset of stigma receptivity, the 95 % CI of the regression coefficient included zero (–0·209, 0·278), verifying the non-detectable heritability observed for this variable in the parametric analysis (h2 = 0·05). Both traits were more variable in the parental than in the progeny generation (F-test comparing variances: pollen onset of stigma receptivity, F19,19 = 3·56, P < 0·01; pistil onset of stigma receptivity, F19,15 = 7·54, P < 0·001; Fig. 2).

Table 1.

Narrow-sense heritability (h2) of pistil onset of stigma receptivity and pollen onset of stigma receptivity, calculated from parent–offspring regressions, and genetic correlations between variables, quantified as the Pearson correlation coefficient based on paternal or maternal family means from the progeny generation

Paternal mean correlations
 Maternal mean correlations
Variable h2 (1) Pollen onset (2) Pistil onset (3) Innate onset (1) Pollen onset (2) Pistil onset (3) Innate onset
(1) Pollen onset of stigma receptivity 0·051, n = 20, P = 0·84 – –0·300, n = 20, P = 0·26 0·185, n = 20, P = 0·44 – 0·104, n = 16, P = 0·70 –0·207, n = 16, P = 0·44
(2) Pistil onset of stigma receptivity 0·279, n = 16, P = 0·11 – – –0·428, n = 16, P = 0·10 – – 0·183, n = 16, P = 0·18
(3) Innate onset of stigma receptivity – – – – – – –
Open in a new tab

Fig. 2.

Fig. 2.

Open in a new tab

Regression of offspring value on parent value for (A) pollen onset of stigma receptivity, and (B) pistil onset of stigma receptivity. Continuous line = regression line, dashed line = line through the origin.

Two-way ANOVA on progeny data confirmed the apparent lack of heritable variation in pollen onset of stigma receptivity: neither the paternal nor the maternal parent exerted a significant influence on the progeny phenotype (paternal parent: F16,44 = 0·65, P = n.s.; maternal parent F9,44 = 0·78, P = n.s.). Because of weakly negative variance estimates for both sire (–0·044) and dam (–0·032), the h2 value for pollen onset of stigma receptivity can be set to 0. Similarly, innate onset of stigma receptivity showed no heritable variation in the progeny generation: the variance attributable to sire or dam was non-significant (paternal parent: F9,21= 0·71, P = n.s.; maternal parent: F18,21 = 1·08, P = n.s.) and much lower (paternal variance = –0·066; maternal variance = 0·018) than the error variance (0·50); based on the low sire and dam variances (which yielded a negative value when averaged), we set the progeny-based estimate of h2 to 0.

The evolvability, IA, based on the non-zero h2 values from the parent–offspring analyses, was calculated to be 2·1 % for pistil onset of stigma receptivity (IA = 100 h2 VP/Z2 = 100 × 0·279 × 0·29/1·962) and 0·38 % for pollen onset of stigma receptivity (IA = 100 × 0·051 × 0·29/1·962).

Genetic correlations

Family-mean correlations, calculated from data obtained in the progeny generation, revealed low to moderate (though non-significant) associations between the three measures of onset of stigma receptivity (Table 1). Interestingly, all six correlation coefficients had different signs depending on whether the analyses compared paternal or maternal families. For comparative purposes, we also correlated the pistil or pollen onset of stigma receptivity for the maternal or paternal progeny family with the innate onset of stigma for individual recipients (rather than averaging across siblings as in the original analyses). Consistent with the results from the family-mean analyses, none of the correlations reached significance (pistil vs. innate onset: r = 0·037, P = n.s.; pollen vs. innate onset: r = 0·098, P = n.s.).

As a final step, we tested for genetic associations by regressing the progeny mean for one trait on the parent mean for another trait. There was a marginally significant tendency for fathers whose pollen induced early onset of stigma receptivity to produce offspring with an earlier innate onset of stigma receptivity than fathers whose pollen induced later onset of stigma receptivity (Table 2 and Fig. 3). Remaining parent–offspring comparisons revealed low, non-significant associations between different measures of onset of stigma receptivity (Table 2), confirming the results of the progeny-based analyses (Table 1).

Table 2.

Pearson correlation coefficients (r) and regression coefficients (b) between innate, pistil and pollen onset of stigma receptivity, based on parent–offspring comparisons.

Trait combination n r b P
♂ Pollen onset–innate onset 20 0·430 0·477 0·056
♂ Pollen onset–pistil onset 16 –0·101 –0·051 0·711
♀ Pistil onset–innate onset 16 0·102 0·088 0·706
♀ Pistil onset–pollen onset 16 0·162 0·062 0·549
Open in a new tab

♂ indicates correlations of paternal progeny and ♀ indicates correlations of maternal progeny.

The value close to being significant is italized.

Fig. 3.

Fig. 3.

Open in a new tab

Parent–offspring regression plot demonstrating a marginally significant relationship (P = 0·056; Table 2) between pollen onset of stigma receptivity and innate onset of stigma receptivity.

DISCUSSION

Although genetic variation is crucial for selection to operate, there is still a paucity of studies quantifying genetic variation of traits directly involved in sexual interactions, especially those occurring in plants, but also those operating in other organisms (Moore and Pizzari, 2005). To our knowledge, this is the first study of a plant or any hermaphrodite that examines patterns of genetic correlation between two interactive sexual traits, and also the first to assess heritabilities of plant traits putatively involved in a sexual conflict. The female trait (the ability to withstand early fertilization of early-arriving pollen) showed modest though significant heritability, while the male trait (the ability of pollen to fertilize ovules early) did not. There was further a tendency for a positive correlation between the male trait and the female-controlled timing of stigma receptivity in the absence of pollen (innate onset of stigma receptivity).

Even though sexual selection is well documented in both animals and plants (Andersson, 1994; Skogsmyr and Lankinen, 2002; Moore and Pannell, 2011), most studies on mate choice have focused on assessing the potential for mate choice, not on elucidating the selective and genetic factors that lead to its origin and maintenance in natural populations (Jennions and Petrie, 1997; Maklakov and Arnqvist, 2009). In plants, where sexual selection is believed to be of particular importance at the prezygotic stage during pollen competition (Bernasconi et al., 2004), few empirical studies have considered the interaction between pistil and pollen traits, and how these (co)vary at the genetic level.

Previous plant studies indicate that genetic differences can explain significant portions of the variation in measures of pollen performance such as pollen-tube growth rate [49 % in Viola tricolor (Skogsmyr and Lankinen, 2000), 36 % in Collinsia heterophylla (Lankinen et al., 2009), but only 9 % in Oenothera organensis (Havens, 1994)], pollen grain size (19–40 % in Mimulus guttatus; Lamborn et al., 2005) and pollen germination rate (36 % in Silene latifolia; Jolivet and Bernasconi, 2007). In a previous study of C. heterophylla, Lankinen et al. (2007) documented heritable variation in a potential mate choice trait (innate timing of stigma receptivity); however, this analysis was based on a small sample of selfed progeny families, making it difficult to generalize the results. All these studies measured traits in the absence of any influence of the mating partner, though a few assessed pollen traits both in vitro and in vivo (e.g. Jolivet and Bernasconi, 2007). The current investigation was primarily designed to assess genetic variation in sexually antagonistic traits expressed under direct male–female interactions, i.e. during pollen germination on the stigma and growth of pollen tubes in the style. Overall, our results indicate low, non-significant levels of genetic variation in the three traits considered, the only exception being pistil onset of stigma receptivity, whose narrow-sense heritability (h2 = 0·28) reached significance according to bootstrap analysis. Similar results, though in the opposite direction concerning male and female traits, were obtained in a recent investigation of a seed beetle (Gay et al., 2010). This study used a nested crossing design (involving 100 fathers and several hundred sons) to document modest, though significant, heritability for a harmful male sexually antagonistic trait (male spine length, h2 = 0·32) and low, non-significant heritability for the ability of females to withstand the harm inflicted by this male trait (h2 = 0·15).

Quantitative fitness traits are polygenic and capture genetic variation across a large number of loci, leading to potentially high levels of genetic variance (Houle, 1992). However, because the traditional heritability statistic (h2) scales the additive genetic variance by the total variance (Falconer and McKay, 1996), the additive portion of variation can appear low if environmental or non-additive sources of variance make a major contribution to the total phenotypic variance. For this reason, it has been proposed that heritability is a poor estimate of adaptive potential and that the additive genetic variance should be scaled differently to provide a more informative measure of evolvability (Houle, 1992). Hansen et al. (2003) suggested that evolvability could be measured as IA (the additive variance scaled by the square of the trait mean) to express the proportional response to directional selection when a particular unit of selection strength is applied to the trait considered. On the basis of this parameter, we find that pistil onset of stigma receptivity would be expected to change by 2·1 % per generation, a high rate compared with those reported in other studies (e.g. 0·01–1·71 % for floral traits of Dalechampia; Hansen et al., 2003). Pollen onset of stigma receptivity would be expected to change by 0·38 %, as compared with the estimate of 3·8 % for pollen tube growth rate, another potentially important male trait in C. heterophylla (Lankinen et al., 2009).

As mentioned above, the ability to detect additive genetic variance may be reduced for environmentally sensitive traits, especially if genotype–environment interactions exist in the population and environmental conditions differ between generations, as may happen in genetic studies based on parent–offspring regression (Falconer and Mackay, 1996). In the present study, both pistil and pollen onset of stigma receptivity were more variable in the parent generation than in the progeny generation. Although both generations were raised in the same greenhouse during the same period of the year from seeds produced at the same floral stage, the winter during the second generation was unusually harsh (J. Madjidian and Å. Lankinen, unpubl. res.), potentially also affecting the greenhouse environment. Thus, we cannot exclude the possibility that some uncontrolled environmental factor such as light or temperature had a high impact in the second year, reducing the ability to detect similarity between related plants in the progeny generation and thus lowering the genetic variance estimated from the sib analyses and the parent–offspring regressions.

Genetic correlations can promote evolutionary responses of characters not under direct selection (Armbruster and Schwaegerle, 1996) and may therefore have profound effects on the evolutionary outcome of male–female interactions (Moore and Pizzari, 2005). Although genetic correlations in the progeny generation failed to reach significance in the current study, we detected a marginally significant positive link between pollen onset of stigma receptivity in the parent generation and innate onset of stigma receptivity in the progeny generation, the latter being one component of the ability of females to delay stigma receptivity. Because early pollen onset vs. late pistil onset represent potential male and female antagonistic traits, respectively, the link between male and female influence on timing of stigma receptivity would indicate a negative association between sexually antagonistic traits. Thus, selection of pollen with a capacity of early fertilization of ovules would result in offspring less protected from costs of early fertilization in this study system. Our result is again the opposite to that found in seed beetles, which showed a positive genetic correlation between the two antagonistic traits (male damage by spines and female susceptibility to harm, with more scars on resistant females), suggesting that if selection favours more harmful males, the prevalence of susceptible females will increase (Gay et al., 2010). However, we urge caution in the interpretation of the observed parent–offspring correlation, given the lack of detectable genetic variation in both pollen onset of stigma receptivity and innate onset of stigma receptivity [although note that Lankinen et al. (2007) detected heritable variation in the latter trait]. It is further possible that other pistil traits might affect the ability to withstand early germination of pollen grains, e.g. the extent to which the stigma becomes receptive in a gradual or abrupt fashion.

Because our studies on C. heterophylla point to high evolutionary potential and selection for delayed stigma receptivity (Lankinen and Armbruster, 2007; Lankinen and Kiboi, 2007; Lankinen and Madjidian, 2011), it is puzzling that timing of stigma receptivity is so variable within populations (Lankinen et al., 2007). This variation may be caused by environmental factors (e.g. temporal or spatial variation, or maternal effects related to timing of fertilization) or by factors related to the mating system, e.g. a trade-off between the benefit of late stigma receptivity in terms of maternal choice vs. the benefit of early stigma receptivity as a reproductive strategy under poor pollination conditions (Goodwillie et al., 2005). On the other hand, pollen competition may be beneficial also in the absence of pollinators: experimental evidence suggests that increased competition between self-pollen can reduce inbreeding depression (Lankinen and Armbruster, 2007), which by itself may promote a mixed mating system in the study species (Armbruster and Rogers, 2004). It has further been suggested that sexual conflict can maintain variation within populations rather than causing population divergence (Gavrilets and Hayashi, 2005). For example, females can diversify genetically into separate groups, ‘trapping’ the males at a state characterized by reduced mating success (Gavrilets and Waxman, 2002). Another recent model indicates that sexual conflict can maintain variation in sexually antagonistic traits through genetic correlations established by assortative mating between mating morphs (Härdling and Karlsson, 2009). Even though we lack knowledge of how important the sexual conflict is in natural populations of C. heterophylla, we can hypothesize that male–female co-evolution could contribute to the large variation seen in timing of stigma receptivity. For example, the cost of reduced seed production in partly receptive pistils (Lankinen and Kiboi, 2007) may favour pistils that accept rather than withstand early fertilization (Arnqvist and Rowe, 2005), resulting in continued production of plants with early ovule fertilization.

This investigation assessed heritabilities and genetic correlations of male and female interactive traits putatively involved in sexual conflict in Collinsia heterophylla, and identified one female trait – pistil onset of stigma receptivity – that could respond to selection driven by male–female interactions. The male trait – pollen onset of stigma receptivity – showed no detectable heritability compared with the female trait, presumably indicating that the evolutionary potential is higher for the female trait than for the male trait, which would slow down male–female co-evolution. In future studies it would be of interest to pursue the role of mating system and hermaphroditism. Because delayed stigma receptivity is beneficial both in terms of enhanced competition between outcross pollen (Lankinen and Madjidian, 2011) and enhanced competition between self-pollen (Lankinen and Armbruster, 2007), it is possible that antagonistic selection during pollen competition may be associated with the persistence of a mixed mating system (Armbruster and Rogers, 2004). Furthermore, the significance of sexual conflict regimes may differ with mating system (Bedhomme et al., 2009; Mazer et al., 2010), and it has been suggested that hermaphrodites are inherently more likely to accept higher mating costs as long as paternity outweighs the fecundity cost paid by the female function (Michiels and Koene, 2006). To increase our general knowledge about co-evolutionary dynamics of male and female sexual traits, it is therefore necessary to assess genetic variation and co-variation in additional taxa, preferably with different mating systems. Further studies of sexual selection in plants, with a focus on fitness costs in relation to sexual conflict (Jolivet and Bernasconi, 2007), may increase our understanding of plant reproductive processes, particularly those occurring during pollen–pistil interactions (Hiscock, 2011).

ACKNOWLEDGEMENTS

This work was supported by the Swedish Research Council. We thank K. Eriksson and S. Hydbom for assistance in the greenhouse, and K. Karlsson Green, H. G. Smith and two anonymous referees for constructive comments on previous versions of the text.

LITERATURE CITED

  1. Agrawal AF, Stinchcombe JR. How much do genetic covariances alter rate of adaptation? Proceedings of the Royal Society London Series B. 2009;276:1183–1191. doi: 10.1098/rspb.2008.1671. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Andersson M. Sexual selection. Princeton, NJ: Princeton University Press; 1994. [Google Scholar]
  3. Armbruster WS. Evolution of floral morphology and function: an integrative approach to adaptation, constraint, and compromise in Dalechampia (Euphorbiaceae) In: Loyd DG, Barrett SCH, editors. Floral biology. New York, NY: Chapman & Hall; 1996. pp. 241–272. [Google Scholar]
  4. Armbruster WS, Rogers DG. Does pollen competition reduce the cost of inbreeding? American Journal of Botany. 2004;91:1939–1943. doi: 10.3732/ajb.91.11.1939. [DOI] [PubMed] [Google Scholar]
  5. Armbruster WS, Schwaegerle KE. Causes of covariation of phenotypic traits among populations. Journal of Evolutionary Biology. 1996;9:261–276. [Google Scholar]
  6. Armbruster WS, Mulder CPH, Baldwin BG, Kalisz S, Wessa B, Nute H. Comparative analysis of late floral development and mating-system evolution in tribe Collinsiae (Scrophulariaceae s.l.) American Journal of Botany. 2002;89:37–49. doi: 10.3732/ajb.89.1.37. [DOI] [PubMed] [Google Scholar]
  7. Arnqvist G, Rowe L. Sexual conflict. Princeton, NJ: Princeton University Press; 2005. [Google Scholar]
  8. Ashman T-L, Majetic CJ. Genetic constraints on floral evolution: a review and evaluation patterns. Heredity. 2006;96:343–352. doi: 10.1038/sj.hdy.6800815. [DOI] [PubMed] [Google Scholar]
  9. Bedhomme S, Bernasconi G, Koene JM, et al. How does breeding system variation modulate sexual antagonism? Biology Letters. 2009;5:717–720. doi: 10.1098/rsbl.2009.0401. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Bernasconi G. Seed paternity in flowering plants: an evolutionary perspective. Perspectives in Plant Ecology, Evolution and Systematics. 2003;6:149–158. [Google Scholar]
  11. Bernasconi G, Paschke M, Schmid B. Diversity effects in reproductive biology. Oikos. 2003;102:217–220. [Google Scholar]
  12. Bernasconi G, Ashman T-L, Birkhead TR, et al. Evolutionary ecology of the prezygotic stage. Science. 2004;303:971–975. doi: 10.1126/science.1092180. [DOI] [PubMed] [Google Scholar]
  13. Borg M, Brownfield L, Twell D. Male gametophyte development: a molecular perspective. Journal of Experimental Biology. 2009;60:1465–1478. doi: 10.1093/jxb/ern355. [DOI] [PubMed] [Google Scholar]
  14. Charlesworth D, Mayer SS. Genetic variability of plant characters in the partial inbreeder Collinsia heterophylla (Scrophulariaceae) American Journal of Botany. 1995;82:112–120. [Google Scholar]
  15. Crawley MJ. The R Book. Chichester: John Wiley & Sons; 2007. [Google Scholar]
  16. Dahl ÅE, Fredriksson M. The timetable for development of maternal tissues sets the stage for male genomic selection in Betula pendula (Betulaceae) American Journal of Botany. 1996;83:895–902. [Google Scholar]
  17. Darwin C. In: On the origin of species: the illustrated edition. Quammen D, editor. New York, NY: Sterling; 1859. [Google Scholar]
  18. Darwin C. The descent of man and selection in relation to sex. London: John Murray; 1871. [Google Scholar]
  19. Eberhard WG. Female control: sexual selection by cryptic female choice. Princeton, NJ: Princeton University Press; 1996. [Google Scholar]
  20. Falconer DS, McKay TFC. Introduction to quantitative genetics. Harlow: Longman; 1996. [Google Scholar]
  21. Fisher RA. The genetical theory of natural selection. 2nd edn. New York, NY: Dover; 1958. [Google Scholar]
  22. Galen C, Shykoff JA, Plowright RC. Consequences of stigma receptivity schedules for sexual selection in flowering plants. American Naturalist. 1986;127:462–476. [Google Scholar]
  23. Gavrilets S, Hayashi TI. Speciation and sexual conflict. Evolutionary Ecology. 2005;19:167–198. [Google Scholar]
  24. Gavrilets S, Waxman D. Sympatric speciation by sexual conflict. Proceedings of the National Academy of Sciences of the USA. 2002;99:10533–10538. doi: 10.1073/pnas.152011499. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Gay L, Brown E, Tregenza T, et al. The genetic architecture of sexual conflict: male harm and female resistance in Callosobruchus maculatus. Journal of Evolutionary Biology. 2010;24:449–456. doi: 10.1111/j.1420-9101.2010.02182.x. [DOI] [PubMed] [Google Scholar]
  26. Geber MA, Griffen LR. Inheritance and natural selection on functional traits. International Journal of Plant Sciences. 2003;164:S21–S42. [Google Scholar]
  27. Goodwillie C, Kalisz S, Eckert CG. The evolutionary enigma of mixed mating systems in plants. Annual Review of Ecology, Evolution, and Systematics. 2005;36:47–79. [Google Scholar]
  28. Hansen TF, Pélabon C, Armbruster WS, Carlson ML. Evolvability and genetic constraint in Dalechampia blossoms: components of variance and measures of evolvability. Journal of Evolutionary Biology. 2003;16:756–766. doi: 10.1046/j.1420-9101.2003.00556.x. [DOI] [PubMed] [Google Scholar]
  29. Härdling R, Karlsson K. The dynamics of sexually antagonistic coevolution and the complex influences of mating system and genetic correlation. Journal of Theoretical Biology. 2009;260:276–282. doi: 10.1016/j.jtbi.2009.05.024. [DOI] [PubMed] [Google Scholar]
  30. Härdling R, Nilsson P. A model of triploid endosperm evolution driven by parent–offspring conflict. Oikos. 2001;92:417–423. [Google Scholar]
  31. Havens K. Clonal repeatability of in vitro pollen tube growth rates in Oenothera organensis (Onagraceae) American Journal of Botany. 1994;81:161–165. [Google Scholar]
  32. Herrero M. Male and female synchrony and the regulation of mating in flowering plants. Proceedings of the Royal Society London Series B. 2003;358:1019–1024. doi: 10.1098/rstb.2003.1285. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Herrero M, Hormaza JI. Pistil strategies controlling pollen tube growth. Sexual Plant Reproduction. 1996;9:343–347. [Google Scholar]
  34. Holland B, Rice WR. Perspective: chase-away sexual selection: antagonistic seduction versus resistance. Evolution. 1998;52:1–7. doi: 10.1111/j.1558-5646.1998.tb05132.x. [DOI] [PubMed] [Google Scholar]
  35. Hiscock SJ. Sexual plant reproduction. Annals of Botany. 2011;4:585–587. doi: 10.1093/aob/mcr217. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Houle D. Comparing evolvability and variability of quantitative traits. Genetics. 1992;130:195–204. doi: 10.1093/genetics/130.1.195. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Jennions MD, Petrie M. Variation in mate choice and mating preferences: a review of causes and consequences. Biological Reviews. 1997;72:283–327. doi: 10.1017/s0006323196005014. [DOI] [PubMed] [Google Scholar]
  38. Jóhannsson MH, Stephenson AG. Effects of pollination intensity on the vigor of the sporophytic and gametophytic generation of Cucurbita texana. Sexual Plant Reproduction. 1997;10:236–240. [Google Scholar]
  39. Jolivet C, Bernasconi G. Within/between population crosses reveal genetic basis for siring success in Silene latifolia (Caryophyllaceae) Journal of Evolutionary Biology. 2007;20:1361–1374. doi: 10.1111/j.1420-9101.2007.01344.x. [DOI] [PubMed] [Google Scholar]
  40. Kalisz S, Vogler B, Fails M, et al. The mechanism of delayed selfing in Collinsia verna (Scrophulariaceae) American Journal of Botany. 1999;86:1239–1247. [PubMed] [Google Scholar]
  41. Kearns CA, Inouye DW. Techniques for pollination biologists. Boulder, CO: University Press; 1993. [Google Scholar]
  42. Kokko H, Brooks R, McNamara JM, Houston AI. The sexual selection continuum. Proceedings of the Royal Society London Series B. 2002;269:1331–340. doi: 10.1098/rspb.2002.2020. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Kokko H, Brooks R, Jennions MD, Morley J. The evolution of mate choice and mating biases. Proceedings of the Royal Society London B. 2003;270:653–664. doi: 10.1098/rspb.2002.2235. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Lamborn E, Cresswell JE, Macnair MR. The potential for adaptive evolution of pollen grain size in Mimulus guttatus. New Phytologist. 2005;167:289–296. doi: 10.1111/j.1469-8137.2005.01403.x. [DOI] [PubMed] [Google Scholar]
  45. Lankinen Å, Armbruster WS. Pollen competition reduces inbreeding depression in Collinsia heterophylla (Plantaginaceae) Journal of Evolutionary Biology. 2007;20:737–749. doi: 10.1111/j.1420-9101.2006.01233.x. [DOI] [PubMed] [Google Scholar]
  46. Lankinen Å, Larsson M. Conflicting selection pressures on reproductive functions and speciation in plants. Evolutionary Ecology. 2009;23:147–157. [Google Scholar]
  47. Lankinen Å, Kiboi S. Pollen donor identity affects timing of stigma receptivity in Collinsia heterophylla (Plantaginaceae): a sexual conflict during pollen competition? American Naturalist. 2007;170:854–863. doi: 10.1086/522839. [DOI] [PubMed] [Google Scholar]
  48. Lankinen Å, Madjidian JA. Enhancing pollen competition by delaying stigma receptivity: pollen deposition schedules affect siring ability, paternal diversity and seed production in Collinsia heterophylla (Plantaginaceae) American Journal of Botany. 2011;98:1–10. doi: 10.3732/ajb.1000510. [DOI] [PubMed] [Google Scholar]
  49. Lankinen Å, Skogsmyr I. Evolution of pistil length as a choice mechanism for pollen quality. Oikos. 2001;92:81–90. [Google Scholar]
  50. Lankinen Å, Hellriegel B, Bernasconi G. Sexual conflict over floral receptivity. Evolution. 2006;60:2454–2465. [PubMed] [Google Scholar]
  51. Lankinen Å, Armbruster WS, Antonsen L. Delayed stigma receptivity in Collinsia heterophylla (Plantaginaceae): genetic variation and adaptive significance in relation to pollen competition, delayed self-pollination, and mating-system evolution. American Journal of Botany. 2007;94:1183–1192. doi: 10.3732/ajb.94.7.1183. [DOI] [PubMed] [Google Scholar]
  52. Lankinen Å, Maad J, Armbruster WS. Pollen-tube growth rates in Collinsia heterophylla (Plantaginaceae): one-donor crosses reveal heritability but no effect on sporophytic–offspring fitness. Annals of Botany. 2009;103:941–950. doi: 10.1093/aob/mcp014. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Madjidian JA, Lankinen Å. Sexual conflict and sexually antagonistic coevolultion in an annual plant. PLoS One. 2009;4:e5477. doi: 10.1371/journal.pone.0005477. http://dx.doi.org/10.1371/journal.pone.0005477 . [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Maklakov AA, Arnqvist G. Testing for direct and indirect effects of mate choice by manipulating female choosiness. Current Biology. 2009;19:1903–1906. doi: 10.1016/j.cub.2009.08.058. [DOI] [PubMed] [Google Scholar]
  55. Manly BFJ. Randomization and Monte Carlo methods in biology. London: Chapman and Hall; 1991. [Google Scholar]
  56. Marshall DL, Whittaker KL. Effects of pollen donor quality in wild radish, Raphanus sativus. American Journal of Botany. 1989;76:1081–1088. [Google Scholar]
  57. Mazer SJ, Hove AA, Miller BS, Barbet-Massin M. The joint evolution of mating system and pollen performance: predictions regarding male gametophytic evolution in selfers vs. outcrossers. Perspectives in Plant Ecology, Evolution and Systematics. 2010;12:31–41. [Google Scholar]
  58. Michiels NK, Koene JM. Sexual selection favours harmful mating in hermaphrodites more than in gonochorists. Integrative and Comparative Biology. 2006;46:473–480. doi: 10.1093/icb/icj043. [DOI] [PubMed] [Google Scholar]
  59. Moore AJ, Pizzari T. Quantitative genetic models of sexual conflict based on interacting phenotypes. American Naturalist. 2005;165:S88–S97. doi: 10.1086/429354. [DOI] [PubMed] [Google Scholar]
  60. Moore JC, Pannell JR. Sexual selection in plants. Current Biology. 2011;21:R176–R182. doi: 10.1016/j.cub.2010.12.035. [DOI] [PubMed] [Google Scholar]
  61. Mulcahy DL. A correlation between gametophytic characters in Zea mays L. Science. 1971;171:1155–1156. doi: 10.1126/science.171.3976.1155. [DOI] [PubMed] [Google Scholar]
  62. Mulcahy DL. The rise of the angiosperms: a genecological factor. Science. 1979;206:20–23. doi: 10.1126/science.206.4414.20. [DOI] [PubMed] [Google Scholar]
  63. Mulcahy DL. Models of pollen-tube competition in Geranium maculatum. In: Real L, editor. Pollination biology. New York, NY: Academic Press; 1983. pp. 151–161. [Google Scholar]
  64. Neese EC. Collinsia. In: Hickman JC, editor. The Jepson manual: higher plants of California. Berkeley, CA: University of California Press; 1993. pp. 1024–1027. [Google Scholar]
  65. Newsom VM. A revision of the genus Collinsia. Botanical Gazette. 1929;87:260. [Google Scholar]
  66. Parker GA. Sexual selection and sexual conflict. In: Blum MS, Blum NA, editors. Sexual selection and reproductive competition in insects. London: Academic Press; 1979. pp. 123–166. [Google Scholar]
  67. Paschke M, Abs C, Schmid B. Effects of population size and pollen diversity on reproductive success and offspring size in the narrow endemic Cochlearia bavarica (Brassicaceae) American Journal of Botany. 2002;89:1250–1259. doi: 10.3732/ajb.89.8.1250. [DOI] [PubMed] [Google Scholar]
  68. Prasad NG, Bedhomme S. Sexual conflict in plants. Journal of Genetics. 2006;85:161–164. doi: 10.1007/BF02935325. [DOI] [PubMed] [Google Scholar]
  69. Queller DC. Sexual selection in a hermaphroditic plant. Nature. 1983;305:706–707. [Google Scholar]
  70. Quesada M, Fuchs EJ, Lobo JA. Pollen load size, reproductive success, and progeny kinship of naturally pollinated flowers of the tropical dry forest tree Pachira quinata (Bombacaceae) American Journal of Botany. 2001;88:2113–2118. [PubMed] [Google Scholar]
  71. R Development Core team. 2011 R: a language and environment for statistical computing, version 2·12·2. Vienna: R Foundation for Statistical Computing. http://www.r-project.org/ [Google Scholar]
  72. Ruane LG. Post-pollination processes and non-random mating among compatible mates. Evolutionary Ecology Research. 2009;11:1031–1051. [Google Scholar]
  73. Rodrigo J, Herrero M, Hormaza JI. Pistil traits and flower fate in apricot (Prunus armeniaca) Annals of Applied Biology. 2009;154:365–375. [Google Scholar]
  74. Skogsmyr I, Lankinen Å. Potential selection for female choice in Viola tricolor. Evolutionary Ecology Research. 2000;2:965–979. [Google Scholar]
  75. Skogsmyr I, Lankinen Å. Sexual selection: an evolutionary force in plants? Biological Reviews. 2002;77:537–562. doi: 10.1017/s1464793102005973. [DOI] [PubMed] [Google Scholar]
  76. Snow AA, Spira TP. Differential pollen-tube growth rates and nonrandom fertilization in Hibiscus moscheutos (Malvaceae) American Journal of Botany. 1991;78:1419–1426. [Google Scholar]
  77. Statistica. STATISTICA 7·1. Chicago, IL: Statsoft Inc; 2005. [Google Scholar]
  78. Weil J, Allard RW. The mating system and genetic variability in natural populations of Collinsia heterophylla. Evolution. 1964;18:515–525. [Google Scholar]
  79. Wilson MF, Burley N. Mate choice in plants: tactics, mechanisms, and consequences. Princeton, NJ: Princeton University Press; 1983. [Google Scholar]

Articles from Annals of Botany are provided here courtesy of Oxford University Press

RESOURCES