Complementary roles of hummingbirds and bees: Pollen pickup, pollen deposition, and fruit production in the distylous Palicourea rigida.

PREMISE: Distyly is a condition in which individual plants in a population express two floral morphs, L- and S-morph, characterized by reciprocal placements of anthers and stigmas between morphs. The function of distyly requires that pollinators collect pollen from L- and S-morphs on different parts along their bodies to then deposit it on the stigmas of the opposite morph, known as legitimate pollination. However, different pollinator groups might differ in the ability to transfer pollen legitimately. METHODS: We investigated patterns of pollen pickup along the body of different functional groups (hummingbirds and bees) using preserved specimens to analyze their role in the reproductive success of Palicourea rigida. We measured pollen deposition on the body of pollinators, on stigmas, and fruit production after a single visit. RESULTS: Pollen from L- and S-flowers appeared segregated on different body parts of the hummingbird and bee used in the study. S-pollen was deposited primarily on the proximal regions (near the head), and L-pollen was placed in the distal regions (tip of the proboscis and bill). Hummingbirds were more efficient at legitimate pollination than bees, particularly to S-stigmas. However, fruit formation after single visits by both pollinators was similar. CONCLUSIONS: The morphology of distylous flowers allows the segregated placement of L-and S-pollen on different body parts of the animal specimens used, facilitating the promotion of legitimate pollen transfer, an observation consistent between the two functional pollinator groups. Also, the results show that full fruit set requires more than one visit.

Patterns of variation in distylous traits and reproductive consequences in Erythroxylum species and populations.

PREMISE: Distylous species possess two floral morphs with reciprocal positioning of stigmas and anthers that is hypothesized to promote disassortative pollination. Theoretical models predict equal morph frequencies, but many populations depart from the expected 1:1 ratio, a pattern that often correlates with asymmetric mating between morphs and/or presence of a weak incompatibility system. Variation in reciprocity can also affect the likelihood of disassortative pollination and, hence, reproductive fitness. METHODS: We described variation in incompatibility systems and morph ratio in four Erythroxylum species to test if greater deviations from 1:1 ratios occur in populations of self-compatible species. Using adaptive inaccuracy, we described upper and lower organ reciprocity in species and populations and assessed the relationship of reciprocity to population means and coefficients of variation for fruit set to test if reciprocity could predict female reproductive success. RESULTS: Morphs occurred in 1:1 ratios in most populations of three Erythroxylum species with distylous self-incompatibility. In self-compatible E. campestre populations showed an excess of the long-styled morph, the short-styled morph, or were monomorphic for the short-styled morph. We detected deviations from reciprocity, with total inaccuracy ranging between 9.39% and 42.94%, and inaccuracy values were lowest in low organs. Across populations, we found a positive relationship between inaccuracy and the coefficient of variation of fruit set. CONCLUSIONS: Erythroxylum species showed variation in the distylous syndrome, with changes in the incompatibility system that corresponded with deviations from 1:1 morph ratio, and variation in reciprocity that correlated with variation in female reproductive fitness.

Evolutionary history of the buildup and breakdown of the heterostylous syndrome in Plumbaginaceae.

The evolutionary pathways leading to the heterostylous syndrome are not well understood, and models concerning the origins of distyly differ in the order in which reciprocal herkogamy and self-incompatibility evolve. We investigated the evolution and breakdown of distyly in Plumbaginaceae, a family with considerable diversity of floral traits and reproductive systems. Using Bayesian Markov chain Monte Carlo analyses and stochastic character mapping, we examined the evolutionary assembly and breakdown of the heterostylous syndrome based on a well-resolved phylogeny of 121 species of Plumbaginaceae and six outgroup taxa using five nuclear and plastid gene regions. We used the distribution of reproductive traits and reconstructed ancestral characters across phylogenies to evaluate competing models for the evolution of distyly. The most likely common ancestor of Plumbaginaceae was self-incompatible and monomorphic for sex-organ arrangement and pollen-stigma characters. Character state reconstructions indicated that reciprocal herkogamy evolved at least three times and that shifts to selfing and apomixis occurred on multiple occasions. Our results provide comparative support for the early ideas of H. G. Baker on evolutionary pathways in Plumbaginaceae, and the more recent selfing avoidance model by D. & B. Charlesworth in which distyly evolves from self-incompatible ancestors.

Reciprocal herkogamy promotes disassortative mating in a distylous species with intramorph compatibility.

Mating patterns in heterostylous species with intramorph compatibility have the potential to deviate from symmetrical disassortative mating owing to ecological and reproductive factors influencing pollen dispersal. Here, we investigate potential and realized patterns of mating in distylous Luculia pinceana (Rubiaceae), a species with intramorph compatibility. Our analysis provides an opportunity to test Darwin's hypothesis that reciprocal herkogamy promotes disassortative pollen transfer. We combined measurements of sex-organ reciprocity and pollen production to predict potential pollen transfer and mating patterns in a population from SW China. Marker-based paternity analysis was then used to estimate realized patterns of disassortative and assortative mating at the individual and floral morph levels. Both potential and realized mating patterns indicated a significant component of disassortative mating, satisfying theoretical conditions for the maintenance of floral dimorphism. Levels of assortative mating (37.7%) were significantly lower than disassortative mating (62.3%), but numerous offspring resulting from intramorph mating were detected in the majority of maternal seed families in both floral morphs. Our results provide empirical support for Darwin's cross-promotion hypothesis on the function of reciprocal herkogamy, but indicate that in most heterostylous species strong diallelic incompatibility may be a general requirement for complete disassortative mating.

Floral morph variation mediated by clonal growth and pollinator functional groups of Limonium otolepis in a heterostylous fragmented population.

Abstract. Heterostyly, a genetic style polymorphism, is linked to symmetric pollen transfer, vital for its maintenance. Clonal growth typically impacts sexual reproduction by influencing pollen transfer. However, the floral morph variation remains poorly understood under the combined effects of pollinators and clonal growth in heterostyly characterized by negative frequency-dependent selection and disassortative mating. We estimated morph ratios, ramets per genet and heterostylous syndrome and quantified legitimate pollen transfer via clonal growth, pollinators and reciprocal herkogamy between floral morphs in Limonium otolepis, a fragmented population composed of five subpopulations in the desert environment of northwestern China, with small flower and large floral morph variation. All subpopulations but one exhibited pollen-stigma morphology dimorphism. The compatibility between mating types with different pollen-stigma morphologies remained consistent regardless of reciprocal herkogamy. Biased ratios and ramets per genet of the two mating types with distinct pollen-stigma morphologies caused asymmetric pollen flow and varying fruit sets in all subpopulations. Short-tongued insects were the primary pollinators due to small flower sizes. However, pollen-feeding Syrphidae sp. triggered asymmetry in pollen flow between high and low sex organs, with short-styled morphs having lower stigma pollen depositions and greater variation. Clonal growth amplified this variation by reducing intermorph pollen transfer. All in all, pollinators and clonal growth jointly drive floral morph variation. H-morphs with the same stigma-anther position and self-incompatibility, which mitigate the disadvantages of sunken low sex organs with differing from the classical homostyly, might arise from long- and short-styled morphs through a 'relaxed selection'. This study is the first to uncover the occurrence of the H-morph and its associated influencing factors in a distylous plant featuring clonal growth, small flowers and a fragmented population.

Floral polymorphism in Chamaecrista flexuosa (Fabaceae-Caesalpinioideae): a possible case of atypical enantiostyly?

BACKGROUND AND AIMS: Reciprocal herkogamy, including enantiostyly and heterostyly, involves reciprocity in the relative positions of the sexual elements within the flower. Such systems result in morphologically and, since pollen is deposited on and captured from different parts of the pollinator, functionally distinct floral forms. Deviations from the basic pattern may modify the functionality of these mechanisms. For heterostylous species, such deviations are generally related to environmental disturbances, pollination services and/or reduced numbers of one floral morph. Deviations for enantiostylous species have not yet been reported. This study aims to investigate enantiostyly in Chamaecrista flexuosa, in particular the presence of deviations from the standard form, in an area of coastal vegetation in north-east Brazil. METHODS: Observations and investigations of floral biology, the reproductive system, pollinator behaviour, floral morphology and morphometry were performed. KEY RESULTS: In C. flexuosa flowers, anthers of different size but similar function are grouped. The flowers were self-compatible and set fruits after every treatment, except in the spontaneous self-pollination experiment, thereby indicating their dependence on pollen vectors. The flowers were pollinated by bees, especially Xylocopa cearensis and X. grisencens. Pollen is deposited and captured from the ventral portion of the pollinator's body. Variations in the spatial arrangement of floral elements allowed for the identification of floral morphs based on both morphological and functional criteria. Using morphological criteria, morphologically right (MR) and morphologically left (ML) floral morphs were identified. Three floral morphs were identified using functional criteria: functionally right (FR), functionally central (FC) and functionally left (FL). Combinations of morphologically and functionally defined morphs did not occur in equal proportions. There was a reduced frequency of the MR-FR combination. CONCLUSIONS: The results indicate the occurrence of an atypical enantiostyly in C. flexuosa. This seems to improve reproductive success by increasing the efficiency of pollen deposition and capture.

Comparative intra- and interspecific sexual organ reciprocity in four distylous Primula species in the Himalaya-Hengduan Mountains.

Distyly is a mechanism promoting cross-pollination within a balanced polymorphism. Numerous studies show that the degree of inter-morph sexual organ reciprocity (SOR) within species relates to its pollen-mediated gene flow. Similarly, a lower interspecific SOR should promote interspecific isolation when congeners are sympatric, co-blooming and share pollinators. In this comparative study, we address the significance of SOR at both intra- and interspecific levels. Seventeen allopatric and eight sympatric populations representing four Primula species (P. anisodora, P. beesiana, P. bulleyana and P. poissonii) native to the Himalaya-Hengduan Mountains were measured for eight floral traits in both long- and short-styled morphs. GLMM and spatial overlap methods were used to compare intra- and interspecific SOR. While floral morphology differed among four Primula species, SOR within species was generally higher than between species, but in species pairs P. poissonii/P. anisodora and P. beesiana/P. bulleyana, the SOR was high at both intra- and interspecific levels. We did not detect a significant variation in intraspecific SOR or interspecific SOR when comparing allopatric versus sympatric populations for all species studied. As intraspecific SOR increased, disassortative mating may be promoted. As interspecific SOR decreased, interspecific isolation between co-flowering species pairs also may increase. Hybridisation between congeners occurred when interspecific SOR increased in sympatric populations, as confirmed in two species pairs, P. poissonii/P. anisodora and P. beesiana/P. bulleyana.

Variation in the degree of reciprocal herkogamy affects the degree of legitimate pollination in a distylous species.

Distyly is a widespread floral polymorphism characterized by the flowers within a population showing reciprocal placement of the anthers and stigma. Darwin hypothesizes that distyly evolves to promote precise pollen transfer between morphs. Primula chungensis exhibits two types of anther heights, and these two types of anthers show pollen of two different size classes. To understand whether the stigma could capture more pollen grains from the anthers of the pollen donor as the separation between the stigma of pollen receiver and the anther of pollen donor decreased, the present research assessed the source of the pollen load in a series of open-pollinated flowers with continuous variation of style lengths. Individuals with continuous variation of style length were tagged, and the selected flowers in the tagged plants were emasculated the day before dehiscence. The stigma of the emasculated flowers was fixed in fuchsin gel at the end of blooming. We assessed the pollen sources on each stigma by taking photos under a microscope and measured the diameter of each conspecific pollen grain with ImageJ. We found that a shorter distance from the stigmas to the anthers of a pollen donor gave the flower a higher capacity to receive pollen from those anthers. Our result provides a new evidence that distyly could promote the pollen transfer between morphs, which is consistent with Darwin's hypothesis of disassortative pollination. An alternative hypothesis for the evolution of distyly (e.g. selfing avoidance) might also be true, but less likely, because self-incompatibility would greatly avoid self-fertilization for many distylous species.

Both morph- and species-dependent asymmetries affect reproductive barriers between heterostylous species.

The interaction between floral traits and reproductive isolation is crucial to explaining the extraordinary diversity of angiosperms. Heterostyly, a complex floral polymorphism that optimizes outcrossing, evolved repeatedly and has been shown to accelerate diversification in primroses, yet its potential influence on isolating mechanisms remains unexplored. Furthermore, the relative contribution of pre- versus postmating barriers to reproductive isolation is still debated. No experimental study has yet evaluated the possible effects of heterostyly on pre- and postmating reproductive mechanisms. We quantify multiple reproductive barriers between the heterostylous Primula elatior (oxlip) and P. vulgaris (primrose), which readily hybridize when co-occurring, and test whether traits of heterostyly contribute to reproductive barriers in unique ways. We find that premating isolation is key for both species, while postmating isolation is considerable only for P. vulgaris; ecogeographic isolation is crucial for both species, while phenological, seed developmental, and hybrid sterility barriers are also important in P. vulgaris, implicating sympatrically higher gene flow into P. elatior. We document for the first time that, in addition to the aforementioned species-dependent asymmetries, morph-dependent asymmetries affect reproductive barriers between heterostylous species. Indeed, the interspecific decrease of reciprocity between high sexual organs of complementary floral morphs limits interspecific pollen transfer from anthers of short-styled flowers to stigmas of long-styled flowers, while higher reciprocity between low sexual organs favors introgression over isolation from anthers of long-styled flowers to stigmas of short-styled flowers. Finally, intramorph incompatibility persists across species boundaries, but is weakened in long-styled flowers of P. elatior, opening a possible backdoor to gene flow through intramorph pollen transfer between species. Therefore, patterns of gene flow across species boundaries are likely affected by floral morph composition of adjacent populations. To summarize, our study highlights the general importance of premating isolation and newly illustrates that both morph- and species-dependent asymmetries shape boundaries between heterostylous species.

EXPERIMENTAL STUDIES ON THE FUNCTIONAL SIGNIFICANCE OF HETEROSTYLY.

Heterostyly has been viewed as both an antiselfing device and a mechanism that increases the proficiency of pollen transfer between plants. We used experimental manipulation of the morph structure of garden populations of self-compatible, tristylous Eichhornia paniculata to investigate the function of floral polymorphism. Outcrossing rates (t), levels of intermorph mating (d), and morph-specific male and female reproductive success were compared in replicate trimorphic and monomorphic populations. In trimorphic populations, t and d averaged 0.81 (2 SE = 0.03) and 0.77 (2 SE = 0.03) respectively, with no difference in either parameter among morphs. Ninety-five percent of outcrossed seeds were therefore the result of intermorph fertilizations. Male reproductive success of the long-styled morph was low, especially in comparison with plants of the short-styled morph. Outcrossing rates for each morph were higher in trimorphic than monomorphic populations where t averaged 0.71 (2 SE = 0.01), 0.30 (2 SE = 0.04) and 0.43 (2 SE = 0.1) for the long-, mid-, and short-styled morphs, respectively. Seed set was lower in monomorphic populations, particularly those composed of the L morph, reflecting reduced pollen deposition. Floral polymorphism therefore increased both outcrossing rate and fecundity but the magnitude of the differences varied among morphs. If the ancestral condition in heterostylous groups resembled the L morph, as has been suggested, data from this study suggests that the selective basis for the establishment of floral polymorphism could have been increased pollen transfer rather than higher levels of outcrossing.