Phenotypic selection on flowering phenology and pollination efficiency traits between Primula populations with different pollinator assemblages

Floral traits have largely been attributed to phenotypic selection in plant–pollinator interactions. However, the strength of this link has rarely been ascertained with real pollinators. We conducted pollinator observations and estimated selection through female fitness on flowering phenology and floral traits between two Primula secundiflora populations. We quantified pollinator‐mediated selection by subtracting estimates of selection gradients of plants receiving supplemental hand pollination from those of plants receiving open pollination. There was net directional selection for an earlier flowering start date at populations where the dominant pollinators were syrphid flies, and flowering phenology was also subjected to stabilized quadratic selection. However, a later flowering start date was significantly selected at populations where the dominant pollinators were legitimate (normal pollination through the corolla tube entrance) and illegitimate bumblebees (abnormal pollination through nectar robbing hole which located at the corolla tube), and flowering phenology was subjected to disruptive quadratic selection. Wider corolla tube entrance diameter was selected at both populations. Furthermore, the strength of net directional selection on flowering start date and corolla tube entrance diameter was stronger at the population where the dominant pollinators were syrphid flies. Pollinator‐mediated selection explained most of the between‐population variations in the net directional selection on flowering phenology and corolla tube entrance diameter. Our results suggested the important influence of pollinator‐mediated selection on floral evolution. Variations in pollinator assemblages not only resulted in variation in the direction of selection but also the strength of selection on floral traits.     

Evolutionary stasis despite selection on a heritable trait in an invasive zooplankton

Invasive species are one of the greatest threats to ecosystems, and there is evidence that evolution plays an important role in the success or failure of invasions. Yet, few studies have measured natural selection and evolutionary responses to selection in invasive species, particularly invasive animals. We quantified the strength of natural selection on the defensive morphology (distal spine) of an invasive zooplankton, Bythotrephes longimanus, in Lake Michigan across multiple months during three growing seasons. We used multiple lines of evidence, including historic and contemporary wild‐captured individuals and palaeoecology of retrieved spines, to assess phenotypic change in distal spine length since invasion. We found evidence of temporally variable selection, with selection for decreased distal spine length early in the growing season and selection for increased distal spine length later in the season. This trend in natural selection is consistent with seasonal changes in the relative strength of non‐gape‐limited and gape‐limited fish predation. Yet, despite net selection for increased distal spine length and a known genetic basis for distal spine length, we observed little evidence of an evolutionary response to selection. Multiple factors likely limit an evolutionary response to selection, including genetic correlations, trade‐offs between components of fitness, and phenotypic plasticity.     

Quantifying hummingbird preference for floral trait combinations: The role of selection on trait interactions in the evolution of pollination syndromes

Darwin recognized the flower's importance for the study of adaptation and emphasized that the flower's functionality reflects the coordinated action of multiple traits. Here we use a multitrait manipulative approach to quantify the potential role of selection acting on floral trait combinations underlying the divergence and maintenance of three related North American species of Silene (Caryophyllaceae). We artificially generated 48 plant phenotypes corresponding to all combinations of key attractive traits differing among the three Silene species (color, height, inflorescence architecture, flower orientation, and corolla‐tube width). We quantified main and interaction effects of trait manipulation on hummingbird visitation preference using experimental arrays. The main effects of floral display height and floral orientation strongly influenced hummingbird visitation, with hummingbirds preferring flowers held high above the ground and vertically to the sky. Hummingbirds also prefer traits in a nonadditive manner as multiple two‐way and higher order interaction effects were important predictors of hummingbird visitation. Contemporary trait combinations found in hummingbird pollinated S. virginica are mostly preferred. Our study demonstrates the likelihood of pollination syndromes evolving due to selection on trait combinations and highlights the importance of trait interactions in understanding the evolution of complex adaptations.     

Where have all the blue flowers gone: pollinator responses and selection on flower colour in New Zealand Wahlenbergia albomarginata

Although pollinators are thought to select on flower colour, few studies have experimentally decoupled effects of colour from correlated traits on pollinator visitation and pollen transfer. We combined selection analysis and phenotypic manipulations to measure the effect of petal colour on visitation and pollen export at two spatial scales in Wahlenbergia albomarginata. This species is representative of many New Zealand alpine herbs that have secondarily evolved white or pale flowers. The major pollinators, solitary bees, exerted phenotypic selection on flower size but not colour, quantified by bee vision. When presented with manipulated flowers, bees visited flowers painted blue to resemble a congener over white flowers in large, but not small, experimental arrays. Pollen export was higher for blue flowers in large arrays. Pollinator preference does not explain the pale colouration of W. albomarginata, as commonly hypothesized. Absence of bright blue could be driven instead by indirect selection of correlated characters.     

Floral morphology and pollination in Gastrodia elata,a mycoheterotrophic orchid

Pollination biology in Gastrodia elata was investigated on Rebun Island, Japan. The flowers have an urceolate perianth chamber whose opening is constricted by the column and labellum. This structure plays a crucial role in the selection of small‐bodied bees from flower visitors of various body sizes. In addition, the furrow on the ventral side of the column possibly contributes to the admission of sweat bee pollinators, whose thoracic widths match the width of the furrow. Sweat bees receive the pollinarium on their thorax while escaping head first from the chamber. No pollinator rewards, such as previously reported harvestable spongy tissue, are likely to be produced by the orchid. The flowers, however, have a glossy area reminiscent of nectar secreted into the chamber that likely acts as a deceptive signal for pollinators. Fruit‐set ratios were found to vary spatiotemporally; the values were higher under bright light conditions, probably reflecting the presence of sweat bees for sunny places.     

Multiple aspects of the selfing syndrome of the morning glory Ipomoea lacunosa evolved in response to selection: A Qst‐Fst comparison

The frequent transition from outcrossing to selfing in flowering plants is often accompanied by changes in multiple aspects of floral morphology, termed the “selfing syndrome.” While the repeated evolution of these changes suggests a role for natural selection, genetic drift may also be responsible. To determine whether selection or drift shaped different aspects of the pollination syndrome and mating system in the highly selfing morning glory Ipomoea lacunosa, we performed multivariate and univariate Qst‐Fst comparisons using a wide sample of populations of I. lacunosa and its mixed‐mating sister species Ipomoea cordatotriloba. The two species differ in early growth, floral display, inflorescence traits, corolla size, nectar, and pollen number. Our analyses support a role for natural selection driving trait divergence, specifically in corolla size and nectar traits, but not in early growth, display size, inflorescence length, or pollen traits. We also find evidence of selection for reduced herkogamy in I. lacunosa, consistent with selection driving both the transition in mating system and the correlated floral changes. Our research demonstrates that while some aspects of the selfing syndrome evolved in response to selection, others likely evolved due to drift or correlated selection, and the balance between these forces may vary across selfing species.     

Phenotypic selection on flowering phenology and size in two dioecious plant species with different pollen vectors

Dioecious plants may be pollinated biotically by animals or abiotically via wind or water currents. It has been hypothesized that these two types of pollen vectors might impose contrasting selective pressures on plant flowering phenology. In the present study we describe the flowering phenology of two sympatric dioecious species with contrasting pollination modes: Mercurialis perennis (wind‐pollinated) and Tamus communis (insect‐pollinated). We estimated selection differentials and gradients for flowering time and flowering synchrony. As flowering time might depend on the accumulation of enough internal resources, we also estimated direct and indirect selection on plant size. Both species have male‐biased sexual ratios, and males are bigger and produce larger flower displays than females, but only in T. communis do males bloom earlier and for longer than females. Selection gradients suggest that selection tends to favor early‐flowering females of T. communis. There is no evidence of direct current selection on the flowering phenology of M. perennis. Intersexual differences in phenology fit with sex allocation and sexual selection theories. As we hypothesized, phenology of the animal‐pollinated species is under stronger selection than that of the wind‐pollinated species and we discuss the potential role of pollen vectors in shaping the flowering phenologies of the study species.     

Floral scent evolution in the section Pseudophrys: pollinator‐mediated selection or phylogenetic constraints?

• Sexually deceptive orchid species from the Mediterranean genus Ophrys usually interact with one or a few pollinator species by means of specific floral scents. In this study, we investigated the respective role of pollinator‐mediated selection and phylogenetic constraints in the evolution of floral scents in the section Pseudophrys.
• We built a phylogenetic tree of 19 Pseudophrys species based on three nuclear loci; we gathered a dataset on their pollination interactions from the literature and from our own field data; and we extracted and analysed their floral scents using solid phase microextraction and gas chromatography‐mass spectrometry. We then quantified the phylogenetic signal carried by floral scents and investigated the link between plant–pollinator interactions and floral scent composition using phylogenetic comparative methods.
• We confirmed the monophyly of the section Pseudophrys and demonstrated the existence of three main clades within this section. We found that floral scent composition is affected by both phylogenetic relationships among Ophrys species and pollination interactions, with some compounds (especially fatty acid esters) carrying a significant phylogenetic signal and some (especially alkenes and alkadienes) generating dissimilarities between closely related Pseudophrys pollinated by different insects.
• Our results show that in the section Pseudophrys, floral scents are shaped both by pollinator‐mediated selection and by phylogenetic constraints, but that the relative importance of these two evolutionary forces differ among compound classes, probably reflecting distinct selective pressures imposed upon behaviourally active and non‐active compounds.

     

毛茛状金莲花不同花期的花特征和访花昆虫的变化及表型选择

为了研究植物生长季内开花时间对花特征表型选择的影响,我们以青藏高原东缘高寒草地的毛茛状金莲花Trollius ranunculoides)为实验材料,在生长季内不同开花时间(花前期、花末期)测定花特征,观察访花昆虫的类群和访花频率,生长季结束后收集种子.根据昆虫访花的喜好和季节内类群与访花频率的变化,分析了不同开花时间毛茛状金莲花的花特征与昆虫的选择;并用种子产量表示雌性适合度,估计了毛茛状金莲花的花特征在不同开花时间所受的表型选择.结果表明:不同花期植物的花特征有显著差异,相应的访花昆虫的类群和频率也存在差异,不同类群昆虫访花喜好也不一样.蜂喜好花瓣和花萼较宽、花茎短和花茎数少的个体,这正符合花前期的特征,因而蜂的访花频率在花前期较高;蝇对花特征没有明显的偏好.而通过雌性适合度估计毛茛状金莲花花特征所受的表型选择则是:花前期,花茎较长和花茎数多的植株适合度大;花末期,花茎数多的植株适合度大.我们的研究表明:在植物生长季,花期的分化伴随着传粉昆虫活动的变化.不同花期,访花昆虫的变化可能对植物花特征的分化起了至关重要的作用.但是访花昆虫对花特征的选择与通过雌性适合度估计植物受到的选择不尽相同,这可能是由于其他因素造成的.     

海拔对全缘叶绿绒蒿植株性状和花特征的表型选择分析

为了研究海拔差异对植株性状、花特征表型选择的影响,以青藏高原高寒草甸的全缘叶绿绒蒿（Meconopsis integrifolia）为研究材料,于盛花期内,测定不同海拔（4 452、4 081和3 681 m）种群中个体植株性状、花特征、单果结实数并进行统计分析,采用线性回归模型估计不同海拔种群间植株性状、花特征所受的表型选择（选择差与选择梯度）。结果表明：（1）随着海拔升高,全缘叶绿绒蒿植株性状、花特征及单果结实数显著降低,海拔越高的种群中株高越矮、叶面积越小、花数越少、花越小、单果结实数越低。（2）不同海拔种群中各性状的表型选择存在差异,较低海拔（3 681 m）种群中花数、花大小具有显著的选择差和选择梯度,表现为花越多、花越大的个体雌性适合度越高;海拔较高（4 081 m）的种群中株高、叶面积及花数更容易受到选择,表现为植株越高、叶面积越大、花越多的个体雌性适合度越高;海拔最高（4 452 m）的种群中叶面积与花数的选择梯度接近显著。（3）植物性状分化伴随着海拔的变化而呈现出差异,较低海拔种群中花特征容易受到选择,而较高海拔种群中可能由于传粉者稀少、资源限制等因素使得株高、叶面积更容易受到选择。     

Phenotypic selection in hawkmoth-pollinated Platanthera bifolia: targets and fitness surfaces

Abstract.— The present study explored phenotypic selection on phenological and morphological reproductive traits in hawkmoth-pollinated Platanthera bifolia (Orchidaceae), a Eurasian perennial herb displaying bisexual, long-spurred flowers. The work was carried out during three flowering seasons (1993–1995) in a Swedish population. Fitness was estimated as the number of pollinia removed (male fitness) and fruits produced (female fitness). Targets and patterns of selection were compared between years and sex functions by the use of multiple linear regression (including correlational selection estimates, i.e., of combination of traits), analysis of covariance, and projection pursuit regression (PPR). Results from the nonparametric surface-fitting-method PPR showed that selection was mostly linear, thus justifying the use of the parametric methods. In all study years, male and female fitness were highest in plants with many flowers. This reflects that flower number sets an upper limit to fitness and that a large inflorescence attracts more pollinators. In 1994, the summer was dry and the average spur length of P. bifolia was shorter than in the other years. In this year, male and female fitness were positively related to spur length, apparently because the spur of short-spurred plants was somewhat too short relative to the tongue length of the local pollinator for optimal pollen export and import. Additionally, the dry weather in 1994 caused a tendency for correlational selection, which was not found in the other years of study. Among small individuals (apparently more sensitive to drought than large ones), early-flowering plants had higher male and female fitness. The results show that patterns of selection may vary both between years and between sex functions in perennial hermaphroditic plants. The present study is one of the first to consider correlational selection in plants, which probably is of common occurrence and deserves to be investigated more.     

Floral traits of mammal‐pollinated Mucuna macrocarpa (Fabaceae): Implications for generalist‐like pollination systems

Floral traits are adapted by plants to attract pollinators. Some of those plants that have different pollinators in different regions adapt to each pollinator in each region to maximize their pollination success. Mucuna macrocarpa (Fabaceae) limits the pollinators using its floral structure and is pollinated by different mammals in different regions. Here, we examine the relationships between floral traits of M. macrocarpa and the external morphology of mammalian pollinators in different regions of its distribution. Field surveys were conducted on Kyushu and Okinawajima Island in Japan, and in Taiwan, where the main pollinators are the Japanese macaque Macaca fuscata, Ryukyu flying fox Pteropus dasymallus, and red‐bellied squirrel Callosciurus erythraeus, respectively. We measured the floral shapes, nectar secretion patterns, sugar components, and external morphology of the pollinators. Results showed that floral shape was slightly different among regions and that flower sizes were not correlated with the external morphology of the pollinators. Volume and sugar rate of nectar were not significantly different among the three regions and did not change throughout the day in any of the regions. However, nectar concentration was higher in Kyushu than in the other two regions. These results suggest that the floral traits of M. macrocarpa are not adapted to each pollinator in each region. Although this plant limits the number of pollinators using its flower structure, it has not adapted to specific mammals and may attract several species of mammals. Such generalist‐like pollination system might have evolved in the Old World.     

Floral volatile alleles can contribute to pollinator‐mediated reproductive isolation in monkeyflowers (Mimulus)

Pollinator‐mediated reproductive isolation is a major factor in driving the diversification of flowering plants. Studies of floral traits involved in reproductive isolation have focused nearly exclusively on visual signals, such as flower color. The role of less obvious signals, such as floral scent, has been studied only recently. In particular, the genetics of floral volatiles involved in mediating differential pollinator visitation remains unknown. The bumblebee‐pollinated Mimulus lewisii and hummingbird‐pollinated Mimulus cardinalis are a model system for studying reproductive isolation via pollinator preference. We have shown that these two species differ in three floral terpenoid volatiles – d ‐limonene, β‐myrcene, and E‐β‐ocimene – that are attractive to bumblebee pollinators. By genetic mapping and in vitro analysis of enzyme activity we demonstrate that these interspecific differences are consistent with allelic variation at two loci, LIMONENE‐MYRCENE SYNTHASE (LMS) and OCIMENE SYNTHASE (OS). Mimulus lewisii LMS (MlLMS) and OS (MlOS) are expressed most strongly in floral tissue in the last stages of floral development. Mimulus cardinalis LMS (McLMS) is weakly expressed and has a nonsense mutation in exon 3. Mimulus cardinalis OS (McOS) is expressed similarly to MlOS, but the encoded McOS enzyme produces no E‐β‐ocimene. Recapitulating the M. cardinalis phenotype by reducing the expression of MlLMS by RNA interference in transgenic M. lewisii produces no behavioral difference in pollinating bumblebees; however, reducing MlOS expression produces a 6% decrease in visitation. Allelic variation at the OCIMENE SYNTHASE locus is likely to contribute to differential pollinator visitation, and thus promote reproductive isolation between M. lewisii and M. cardinalis. OCIMENE SYNTHASE joins a growing list of ‘speciation genes’ (‘barrier genes’) in flowering plants.     

Pollen colour morphs take different paths to fitness

Colour phenotypes are often involved in communication and are thus under selection by species interactions. However, selection may also act on colour through correlated traits or alternative functions of biochemical pigments. Such forms of selection are instrumental in maintaining petal colour diversity in plants. Pollen colour also varies markedly, but the maintenance of this variation is little understood. In Campanula americana, pollen ranges from white to dark purple, with darker morphs garnering more pollinator visits and exhibiting elevated pollen performance under heat stress. Here, we generate an F2 population segregating for pollen colour and measure correlations with floral traits, pollen attributes and plant‐level traits related to fitness. We determine the pigment biochemistry of colour variants and evaluate maternal and paternal fitness of light and dark morphs by crossing within and between morphs. Pollen colour was largely uncorrelated with floral traits (petal colour, size, nectar traits) suggesting it can evolve independently. Darker pollen grains were larger and had higher anthocyanin content (cyanidin and peonidin) which may explain why they outperform light pollen under heat stress. Overall, pollen‐related fitness metrics were greater for dark pollen, and dark pollen sires generated seeds with higher germination potential. Conversely, light pollen plants produce 61% more flowers than dark, and 18% more seeds per fruit, suggesting a seed production advantage. Results indicate that light and dark morphs may achieve fitness through different means—dark morphs appear to have a pollen advantage whereas light morphs have an ovule advantage—helping to explain the maintenance of pollen colour variation.     

Climate drives among‐year variation in natural selection on flowering time

To predict long‐term responses to climate change, we need to understand how changes in temperature and precipitation elicit both immediate phenotypic responses and changes in natural selection. We used 22 years of data for the perennial herb Lathyrus vernus to examine how climate influences flowering phenology and phenotypic selection on phenology. Plants flowered earlier in springs with higher temperatures and higher precipitation. Early flowering was associated with a higher fitness in nearly all years, but selection for early flowering was significantly stronger in springs with higher temperatures and lower precipitation. Climate influenced selection through trait distributions, mean fitness and trait?fitness relationships, the latter accounting for most of the among‐year variation in selection. Our results show that climate both induces phenotypic responses and alters natural selection, and that the change in the optimal phenotype might be either weaker, as for spring temperature, or stronger, as for precipitation, than the optimal response.     

Morph‐specific selection on floral traits in a polymorphic plant

Correlations between phenotypic traits are common in many organisms, but the relative importance of nonadaptive mechanisms and selection for the evolution and maintenance of such correlations are poorly understood. In polymorphic species, morphs may evolve quantitative differences in additional characters as a result of morph‐specific selection. The perennial rosette herb Primula farinosa is polymorphic for scape length. The short‐scaped morph is less damaged by grazers and seed predators but is more strongly pollen limited than the long‐scaped morph. We examined whether morph‐specific differences in biotic interactions are associated with differences in selection on two other traits affecting floral display (number of flowers and petal size) and on one trait likely to affect pollination efficiency (corolla tube width) in three P. farinosa populations. Differences in selection between morphs were detected in one population. In this population, selection for more flowers and larger petals was stronger in the short‐scaped than in the long‐scaped morph, and although there was selection for narrower corolla tubes in the short‐scaped morph, no statistically significant selection on corolla tube width could be detected in the long‐scaped morph. In the study populations, the short‐scaped morph produced more and larger flowers and wider corolla tubes. Current morph‐specific selection was thus only partly consistent with trait differences between morphs. The results provide evidence of morph‐specific selection on traits associated with floral display and pollination efficiency, respectively.     

The search for Pleiades in trait constellations: functional integration and phenotypic selection in the complex flowers of Morrenia brachystephana (Apocynaceae)

Pollinator‐mediated natural selection on single traits, such as corolla tube or spur length, has been well documented. However, flower phenotypes are usually complex, and selection is expected to act on several traits that functionally interact rather than on a single isolated trait. Despite the fact that selection on complex phenotypes is expectedly widespread, multivariate selection modelling on such phenotypes still remains under‐explored in plants. Species of the subfamily Asclepiadoideae (Apocynaceae) provide an opportunity to study such complex flower contrivances integrated by fine‐scaled organs from disparate developmental origin. We studied the correlation structure among linear floral traits (i) by testing a priori morphological, functional or developmental hypotheses among traits and (ii) by exploring the organization of flower covariation, considering alternative expectations of modular organization or whole flower integration through conditional dependence analysis (CDA) and integration matrices. The phenotypic selection approach was applied to determine whether floral traits involved in the functioning of the pollination mechanism were affected by natural selection. Floral integration was low, suggesting that flowers are organized in more than just one correlation pleiad; our hypothetical functional correlation matrix was significantly correlated with the empirical matrix, and the CDA revealed three putative modules. Analyses of phenotypic selection showed significant linear and correlational gradients, lending support to expectations of functional interactions between floral traits. Significant correlational selection gradients found involved traits of different floral whorls, providing evidence for the existence of functional integration across developmental domains.     

Pollinator‐mediated selection in a specialized hummingbird–Heliconia system in the Eastern Caribbean

Phenotypic matches between plants and their pollinators often are interpreted as examples of reciprocal selection and adaptation. For the two co‐occurring plant species, Heliconia bihai and H. caribaea in the Eastern Caribbean, we evaluated for five populations over 2 years the strength and direction of natural selection on corolla length and number of bracts per inflorescence. These plant traits correspond closely to the bill lengths and body masses of their primary pollinators, female or male purple‐throated carib hummingbirds (Eulampis jugularis). In H. bihai, directional selection for longer corollas was always significant with the exception of one population in 1 year, whereas selection on bract numbers was rare and found only in one population in 1 year. In contrast, significant directional selection for more bracts per inflorescence occurred in all three populations of the yellow morph and in two populations of the red morph of H. caribaea, whereas significant directional selection on corolla length occurred in only one population of the red morph and one population of the yellow morph. Selection for longer corollas in H. bihai may result from better mechanical fit, and hence pollination, by the long bills of female E. jugularis, their sole pollinator. In contrast, competition between males of E. jugularis for territories may drive selection for more bracts in H. caribaea. Competitive exclusion of female E. jugularis by territorial males also implicates pollinator competition as a possible ecological mechanism for trait diversification in these plants.     

No geographic variation in thermoregulatory colour plasticity and limited variation in heat‐avoidance behaviour in Battus philenor caterpillars

Phenotypic plasticity can help organisms cope with variation in their current environment, including temperature variation, but not all environments are equally variable. In the least variable or extreme environments, plasticity may no longer be used. In this case, the plasticity could be lost altogether, or it could persist with either the same or an altered reaction norm, depending on factors such as the plasticity's costs. In the pipevine swallowtail caterpillar (Battus philenor), I tested for changes in two forms of heat‐avoidance plasticity, colour change and refuge‐seeking behaviour, across the species’ range in the United states, including the cooler eastern parts of its range where colour change has not been observed and is unlikely to be needed. I found that both heat‐avoidance behaviour and colour change persisted in all surveyed populations. Indeed, the reaction norm for colour change remained nearly unaltered, whereas the threshold for refuge‐seeking only changed slightly across populations. These results suggest that the costs of these plastic traits are low enough for them to be maintained by whatever minimal gene flow the population receives. I show that plasticity can be maintained unaltered in populations where it is not used and discuss the potential consequences of this persistence for both the ecology and evolution of plasticity.     

Selection of trait combinations through bee and fly visitation to flowers of Polemonium foliosissimum

Pollinators are known to exert natural selection on floral traits, but the extent to which combinations of floral traits are subject to correlational selection (nonadditive effects of two traits on fitness) is not well understood. Over two years, we used phenotypic manipulations of plant traits to test for effects of flower colour, flower shape and their interaction on rates of pollinator visitation to Polemonium foliosissimum. We also tested for correlational selection based on weighting visitation by the amount of conspecific pollen delivered per visit by each category of insect visitor. Although bumblebees were the presumed pollinators, solitary bees and flies contributed substantially (42%) to pollination. In manipulations of one trait at a time, insects visited flowers presenting the natural colour and shape over flowers manipulated to present artificial mutants with either paler colour or a more open or more tubular flower. When both colour and shape were manipulated in combination, selection on both traits arose, with bumblebees responding mainly to colour and flies responding mainly to shape. Despite selection on both floral traits, in a year with many bumblebees, we saw no evidence for correlational selection of these traits. In a year when flies predominated, fly visitation showed a pattern of correlational selection, but not favouring the natural phenotype, and correlational selection was still not detected for expected pollen receipt. These results show that flower colour and shape are subject to pollinator‐mediated selection and that correlational selection can be generated based on pollinator visitation alone, but provide no evidence for correlational selection specifically for the current phenotype.