A heterogeneous landscape does not guarantee high crop pollination

The expansion of pollinator-dependent crops, especially in the developing world, together with reports of worldwide pollinator declines, raises concern of possible yield gaps. Farmers directly reliant on pollination services for food supply often live in regions where our knowledge of pollination services is poor. In a manipulative experiment replicated at 23 sites across an Ethiopian agricultural landscape, we found poor pollination services and severe pollen limitation in a common oil crop. With supplementary pollination, the yield increased on average by 91%. Despite the heterogeneous agricultural matrix, we found a low bee abundance, which may explain poor pollination services. The variation in pollen limitation was unrelated to surrounding forest cover, local bee richness and bee abundance. While practices that commonly increase pollinators (restricted pesticide use, flower strips) are an integral part of the landscape, these elements are apparently insufficient. Management to increase pollination services is therefore in need of urgent investigation.     

Effects of grazing intensity on pollinator abundance and diversity,and on pollination services

1. Pollinating insects provide important ecosystem services and are influenced by the intensity of grazing. Based on the Intermediate Disturbance Hypothesis (IDH), pollinator diversity is expected to peak at intermediate grazing intensities. However, this hump‐shaped relationship is rarely found. 2. The effect of grazing intensity was tested on flower cover, on the abundance and richness of bees, hoverflies and bee flies, and on pollination services to early‐flowering bee‐pollinated Asphodelus ramosus L. For that, we used data on 11 plant–pollinator phryganic communities from Lesvos Island (Greece) widely differing in grazing intensities. 3. Flower abundance and richness showed hump‐shaped relationships with grazing intensity. Grazing affected the abundance and richness of bees and hoverflies directly and also indirectly, through changes in the flower community. Grazing influenced directly the richness but not the abundance of bee flies. Overall, pollinator abundance and richness showed hump‐shaped relationships with grazing intensity, but variations in strength (hoverfly abundance) and direction (bee community) of the effect appeared along the season. Early in the season, grazing increased bee abundance but decreased richness, resulting in increased pollen limitation in A. ramosus. 4. The effects of grazing on pollinators vary with the intensity of the disturbance, generally supporting the IDH, and the timing of land‐use activities may influence pollination services. Management strategies should include moderate grazing levels to preserve overall diversity in this area, however, the conservation of particular early bee or bee‐pollinated species may benefit from reduced grazing in early spring.     

Restoration of Plant–Pollinator Interactions: Pollination Neighborhood and Asymmetric Pollen Flow Between Restored Habitats in a Beetle‐Pollinated Aroid

Two processes globally threatening natural ecosystems are changes in land use and deforestation. Two methods used to restore threatened ecosystems are: (1) unassisted forest regeneration, which promotes the establishment of plants and fauna arriving from surrounding habitats and (2) assisted restoration, which involves the reconstruction of forests by planting native or exotic trees. Functional attributes, such as plant–pollinator interactions, are essential for ecosystem recovery. Unfortunately, information regarding the effect of restoration on pollination systems is limited. Forty years ago, a tropical cloud forest in Colombia was restored through unassisted forest regeneration, as well as by establishing monospecific plantations of the exotic Chinese ash Fraxinus chinensis. The understories of both restoration strategies were colonized by the beetle‐pollinated aroid Xanthosoma daguense. Using isolation‐by‐distance and multi‐strata mark‐recapture models, I estimated in each restoration strategy two fundamental pollination processes: (1) the magnitude of X. daguense pollination neighborhood and (2) the directionality of pollen flow among plants colonizing both restoration strategies. In addition, I recorded pollinator visits and fruit production for X. daguense in each restored habitats. The pollination neighborhood of X. daguense in the ash plantation is two times larger than its pollination neighborhood in natural regeneration. Inflorescences in the ash plantation donated 10 times more pollen to inflorescences in natural regeneration. Plants in natural regeneration produced two times more infructescences and more fruits than plants in ash plantation. Results show that the selection of different restoration strategies can alter two major components of plant–pollinator interactions in plants colonizing restored habitats, pollination neighborhoods, and pollen flow within the population.     

Density of insect‐pollinated grassland plants decreases with increasing surrounding land‐use intensity

Pollinator declines have raised concerns about the persistence of plant species that depend on insect pollination, in particular by bees, for their reproduction. The impact of pollinator declines remains unknown for species‐rich plant communities found in temperate seminatural grasslands. We investigated effects of land‐use intensity in the surrounding landscape on the distribution of plant traits related to insect pollination in 239 European seminatural grasslands. Increasing arable land use in the surrounding landscape consistently reduced the density of plants depending on bee and insect pollination. Similarly, the relative abundance of bee‐pollination‐dependent plants increased with higher proportions of non‐arable agricultural land (e.g. permanent grassland). This was paralleled by an overall increase in bee abundance and diversity. By isolating the impact of the surrounding landscape from effects of local habitat quality, we show for the first time that grassland plants dependent on insect pollination are particularly susceptible to increasing land‐use intensity in the landscape.     

Florally rich habitats reduce insect pollination and the reproductive success of isolated plants

Landscape heterogeneity in floral communities has the potential to modify pollinator behavior. Pollinator foraging varies with the diversity, abundance, and spatial configuration of floral resources. However, the implications of this variation for pollen transfer and ultimately the reproductive success of insect pollinated plants remains unclear, especially for species which are rare or isolated in the landscape. We used a landscape‐scale experiment, coupled with microsatellite genotyping, to explore how the floral richness of habitats affected pollinator behavior and pollination effectiveness. Small arrays of the partially self‐compatible plant Californian poppy (Eschscholzia californica) were introduced across a landscape gradient to simulate rare, spatially isolated populations. The effects on pollinator activity, outcrossing, and plant reproduction were measured. In florally rich habitats, we found reduced pollen movement between plants, leading to fewer long‐distance pollination events, lower plant outcrossing, and a higher incidence of pollen limitation. This pattern indicates a potential reduction in per capita pollinator visitation, as suggested by the lower activity densities and richness of pollinators observed within florally rich habitats. In addition, seed production reduced by a factor of 1.8 in plants within florally rich habitats and progeny germination reduced by a factor of 1.2. We show this to be a consequence of self‐fertilization within the partially self‐compatible plant, E. californica. These findings indicate that locally rare plants are at a competitive disadvantage within florally rich habitats because neighboring plant species disrupt conspecific mating by co‐opting pollinators. Ultimately, this Allee effect may play an important role in determining the long‐term persistence of rarer plants in the landscape, both in terms of seed production and viability. Community context therefore requires consideration when designing and implementing conservation management for plants which are comparatively rare in the landscape.     

Does the morphological fit between flowers and pollinators affect pollen deposition? An experimental test in a buzz‐pollinated species with anther dimorphism

Some pollination systems, such as buzz‐pollination, are associated with floral morphologies that require a close physical interaction between floral sexual organs and insect visitors. In these systems, a pollinator's size relative to the flower may be an important feature determining whether the visitor touches both male and female sexual organs and thus transfers pollen between plants efficiently. To date, few studies have addressed whether in fact the “fit” between flower and pollinator influences pollen transfer, particularly among buzz‐pollinated species. Here we use Solanum rostratum, a buzz‐pollinated plant with dimorphic anthers and mirror‐image flowers, to investigate whether the morphological fit between the pollinator's body and floral morphology influences pollen deposition. We hypothesized that when the size of the pollinator matches the separation between the sexual organs in a flower, more pollen should be transferred to the stigma than when the visitor is either too small or too big relative to the flower. To test this hypothesis, we exposed flowers of S. rostratum with varying levels of separation between sexual organs, to bumblebees (Bombus terrestris) of different sizes. We recorded the number of visits received, pollen deposition, and fruit and seed production. We found higher pollen deposition when bees were the same size or bigger than the separation between anther and stigma within a flower. We found a similar, but not statistically significant pattern for fruit set. In contrast, seed set was more likely to occur when the size of the flower exceeded the size of the bee, suggesting that other postpollination processes may be important in translating pollen receipt to seed set. Our results suggest that the fit between flower and pollinator significantly influences pollen deposition in this buzz‐pollinated species. We speculate that in buzz‐pollinated species where floral morphology and pollinators interact closely, variation in the visitor's size may determine whether it acts mainly as a pollinator or as a pollen thief (i.e., removing pollen rewards but contributing little to pollen deposition and fertilization).     

Floral size variation in Campanula rotundifolia (Campanulaceae) along altitudinal gradients: patterns and possible selective mechanisms

We investigated patterns of flower‐size variation along altitudinal gradients in the bee‐pollinated perennial Campanula rotundifolia (Campanulaceae) by examining 22 Norwegian populations at altitudes between 240 and 1100 m a.s.l. We explored potential mechanisms for the underlying pattern by quantifying pollinator–faunal composition, pollinator‐visitation rates and pollen limitation of seed set in subsets of the study populations. Despite a decrease in plant size, several measures of flower size increased with elevation. Bumble bees were the main pollinators at both alpine and lowland sites in the study area. However, species composition of the pollinator fauna differed, and pollinators were larger in higher‐elevation than in lower‐elevation sites. Pollinator visitation rates were lower at higher‐elevations than at lower elevations. Pollen limitation of seed set did not vary significantly with altitude. Our results are consistent with differences in bumble‐bee size and visitation rates as causal mechanisms for the relatively larger flowers at higher elevations, in three non‐mutually exclusive ways: 1) Larger flowers reflect selection for increased attractiveness where pollinators are rare. 2) Larger and fewer flowers represent a risk avoidance strategy where the probability of pollination is low on any given day. 3) Flower size variation reflects selection to improve the fit of pollinators with fertile structures by matching flower size to pollinator size across sites.     

Restoration of Nontarget Species: Bee Communities and Pollination Function in Riparian Forests

Nontarget species such as pollinators may be of great importance to the restoration process and the long‐term functioning of restored habitats, but little is known about how such groups respond to habitat restoration. I surveyed bee communities at five equal‐aged restored sites, paired with five reference sites (riparian remnants) along the Sacramento River, California, United States. Flower availability and bee visitation patterns were also measured to examine the restoration of pollination function. Restoration of structural vegetation allowed diverse and abundant native bee communities to establish at the restoration sites; however, the composition of these important pollinator communities was distinct from that in the remnant riparian sites. Differences did not arise primarily from differences in the composition of the flowering‐plant community; rather there must be other physical characteristics of the restored sites or differences in nesting site availability that led to the different pollinator communities. Because sites were spatially paired, the differences are unlikely to be driven by landscape context. Bee life‐history and other biological traits may partially explain the differences between bee communities at restored and remnant sites. Patterns of visitation to native plant species suggest that pollination function is restored along with pollinator abundance and richness; however, function may be less robust in restored habitats. An examination of interaction networks between bees and plant species found at both restored and remnant riparian sites showed less redundancy of pollinators visiting some plants at restored habitats.     

Effects of landscape metrics on Apis and non-Apis pollinators and seed set in common buckwheat

Crop pollination is an essential ecosystem service, which, when provided by wild pollinator species, can be negatively affected by human activities at a landscape scale. Various pollinator species can be affected by landscape factors and/or act at different spatial scales. We investigated whether different landscape factors at different spatial scales are correlated with pollinator abundances and pollination success. We examined common buckwheat, which is a self-incompatible crop species, pollinated by managed honeybees, native honeybees, and wild non-honeybee insects. To determine the effects of landscape factors at different spatial scales, we sampled flower-visiting insects and recorded seed sets. Results from the sampled flower visitors indicated that the abundances of managed honeybees, native honeybees, and wild non-honeybee insects were significantly related to the distance of managed hives, the area of forest cover within a 3-km radius, and the area of forest and grassland cover within a 100-m radius, respectively. Additionally, the latter two influential factors of the landscape, but not the distance of managed hives, showed significant positive correlations with the seed set. These results suggest that considering landscape factors at different spatial scales associated with the characteristics of managed and wild pollinators would help to ensure crop yields.     

Autonomous self‐pollination in the nectarless orchid Pogonia minor

The pollination biology of the nectarless orchid Pogonia minor was investigated in central Japan. The investigation revealed that the solitary flowers failed to attract pollinators, while high rates of fruit set were observed in the natural population. Comparable levels of fruit set were obtained in bagged, artificial self‐pollinated and artificial cross‐pollinated plants, indicating that the species is not pollinator‐limited for fruit set under natural conditions. Autonomous self‐pollination in P. minor resulted from a reduced rostellum, which allowed contact between the pollinia and the stigma. Self‐pollination is thought to be an adaptive response that provides reproductive assurance under conditions of pollinator limitation. Since pollen limitation is generally known to be frequent among deceptive orchids, strong pollen limitation is probably a driving force in the autonomous self‐pollination mechanism in the nectarless orchid P. minor.     

Effects of local density on pollination and reproduction in Delphinium nuttallianum and Aconitum columbianum (Ranunculaceae)

Plant populations vary in density both naturally and as a consequence of anthropogenic impacts. Density in turn can influence pollination by animals. For example, plants in dense populations might enjoy more frequent visitation if pollinators forage most efficiently in such populations. We explored effects of plant density on pollination and seed set in the larkspur Delphinium nuttallianum and monkshood Aconitum columbianum. At our site in the Colorado Rocky Mountains, flowers of D. nuttallianum are pollinated primarily by queen bumble bees, solitary bees, and hummingbirds, whereas those of A. columbianum are pollinated primarily by queen and worker bumble bees. We found that the quantity of pollination service to both species (pollinator visitation rate and pollen deposition) was at best weakly related to density. In contrast, seed set declined by approximately one-third in sparse populations relative to nearby dense populations. This decline may stem from the receipt of low-quality pollen, for example, inbred pollen. Alternatively, sparsity may indicate poor environmental conditions that lower seed set for reasons unrelated to pollination. Our results demonstrate the value of simultaneously exploring pollinator behavior, pollen receipt, and seed set in attempting to understand how the population context influences plant reproductive success.     

Proximity to forest edge does not affect crop production despite pollen limitation

A decline in pollination function has been linked to agriculture expansion and intensification. In northwest Argentina, pollinator visits to grapefruit, a self-compatible but pollinator-dependent crop, decline by approximately 50% at 1km from forest edges. We evaluated whether this decrease in visitation also reduces the pollination service in this crop. We analysed the quantity and quality of pollen deposited on stigmas, and associated limitation of fruit production at increasing distances (edge: 10, 100, 500 and 1000m) from the remnants of Yungas forest. We also examined the quantitative and qualitative efficiency of honeybees as pollen vectors. Pollen receipt and pollen tubes in styles decreased with increasing distance from forest edge; however, this decline did not affect fruit production. Supplementation of natural pollen with self- and cross-pollen revealed that both pollen quantity and quality limited fruit production. Despite pollen limitation, honeybees cannot raise fruit production because they often do not deposit sufficient high-quality pollen per visit to elicit fruit development. However, declines in visitation frequency well below seven visits during a flower's lifespan could decrease production beyond current yields. In this context, the preservation of forest remnants, which act as pollinator sources, could contribute to resilience in crop production. Like wild plants, pollen limitation of the yield among animal-pollinated crops may be common and indicative not only of pollinator scarcity, but also of poor pollination quality, whereby pollinator efficiency, rather than just abundance, can play a broader role than previously appreciated.     

Rapid Shift in Pollinator Communities Following Invasive Species Removal

Ecological restoration is increasingly used to reverse degradation of rare ecosystems and maintain biological diversity. Pollinator communities are critical to maintenance of plant diversity and, in light of recent pollinator loss, we tested whether removal of invasive glossy buckthorn (Frangula alnus L.) from portions of a prairie fen wetland altered plant and pollinator communities. We compared herbaceous plant, bee, and butterfly abundance, diversity, and species composition in buckthorn invaded, buckthorn removal, and uninvaded reference plots. Following restoration, we found striking differences in plant and pollinator abundance and species composition between restored, unrestored, and reference plots. Within 2 years of F. alnus removal, plant species diversity and composition in restored plots were significantly different than invaded plots, but also remained significantly lower than reference plots. In contrast, in the first growing season following restoration, bee and butterfly abundance, diversity, and composition were similar in restored and reference plots and distinct from invaded plots. Our findings indicate that a diverse community of mobile generalist pollinators rapidly re‐colonizes restored areas of prairie fen, while the plant community may take longer to fully recover. This work implies that, in areas with intact pollinator metapopulations, restoration efforts will likely prevent further loss of mobile generalist pollinators and maintain pollination services. On the other hand, targeted restoration efforts will likely be required to restore populations of rare plants and specialist pollinators for which local and regional species pools may be lacking.     

Effects of pollination and pollen source on the seed set ofPedicularis palustris

Habitat alteration can deteriorate plant-pollinator interactions and thereby increase the risk of population extinction. As part of a larger study on the effects of changes in land use on fen grassland vegetation, factors influencing the seed set of a short-lived, endangered wetland plant,Pedicularis palustris, were studied. We conducted field pollination experiments in one large and one small population. To investigate the effect of pollen source on seed set, individual flowers of caged plants were left unpollinated or were pollinated with pollen from the same flower, the same population or another population. To study pollen limitation and flower display, whole plants were subjected to pollinator exclosure, hand pollination or natural pollination. Self-compatibility was high, but differed between populations (61% and 97% of seed set after cross-pollination within populations). Cross-pollination between populations did not significantly alter seed number per capsule. Pollinator exclosure resulted in a very low seed set (<15% of natural seed set), despite high self-compatibility. The most likely explanations for high self-compatibility in combination with low autofertility are geitonogamy as reproductive assurance, selective neutrality of self-compatibility and phylogenetic constraints. Because of low autofertility, the seed set inP. palustris depends on pollinators. In the study populations, natural pollination was clearly sufficient for maximum seed production per plant, but seed set per capsule was significantly pollen-limited in the smaller population. Plants in this population also had a higher maximum percentage of simultaneously open flowers than those of the large population (31% vs. 13%), while flower longevity was generally extended without pollination. It is concluded thatP. palustris may influence pollinator behaviour and therefore the risk of pollen limitation by flower display.P. palustris showed a flexible reaction to differing pollination regimes without losses in overall seed set in the study populations.     

The challenge of accurately documenting bee species richness in agroecosystems: bee diversity in eastern apple orchards

Bees are important pollinators of agricultural crops, and bee diversity has been shown to be closely associated with pollination, a valuable ecosystem service. Higher functional diversity and species richness of bees have been shown to lead to higher crop yield. Bees simultaneously represent a mega‐diverse taxon that is extremely challenging to sample thoroughly and an important group to understand because of pollination services. We sampled bees visiting apple blossoms in 28 orchards over 6 years. We used species rarefaction analyses to test for the completeness of sampling and the relationship between species richness and sampling effort, orchard size, and percent agriculture in the surrounding landscape. We performed more than 190 h of sampling, collecting 11,219 specimens representing 104 species. Despite the sampling intensity, we captured <75% of expected species richness at more than half of the sites. For most of these, the variation in bee community composition between years was greater than among sites. Species richness was influenced by percent agriculture, orchard size, and sampling effort, but we found no factors explaining the difference between observed and expected species richness. Competition between honeybees and wild bees did not appear to be a factor, as we found no correlation between honeybee and wild bee abundance. Our study shows that the pollinator fauna of agroecosystems can be diverse and challenging to thoroughly sample. We demonstrate that there is high temporal variation in community composition and that sites vary widely in the sampling effort required to fully describe their diversity. In order to maximize pollination services provided by wild bee species, we must first accurately estimate species richness. For researchers interested in providing this estimate, we recommend multiyear studies and rarefaction analyses to quantify the gap between observed and expected species richness.     

Field of Dreams: Restitution of Pollinator Services in Restored Bird‐Pollinated Plant Populations

Ecosystem functionality is an increasingly important objective of ecological restoration. Despite this, a few studies have rigorously assessed reproductive functionality within restored plant populations, and it is largely assumed that pollinators follow restoration of plant communities—“build it and they will come.” Here, we applied an ecological genetic approach to determine the impact of spatial separation on mating in Banksia menziesii (Proteaceae), a dominant bird‐pollinated species of Banksia woodlands of Western Australia. All plants at three post‐mining restored sites (n = 72 [13 years old], n = 21 [8 years old], and n = 20 [9 years old]), as well as a sample from an adjacent natural reference site (n = 42), were genotyped at nine microsatellite loci. Seed set, mating system parameters, realized pollen dispersal through the assignment of paternity to seed, and avian pollinator species composition, abundance and behavior, were assessed. All patches displayed equivalent heterozygosity (H_e = 0.53–0.59) and very weak genetic divergence (F_ST ≤ 0.01). Seed of plants within restored sites showed complete outcrossing and relatively high seed set, 26% of which were sired by pollen donors located beyond the local patch. Similar abundance and movement of nectar‐feeding birds was observed in restored and natural sites, despite lower bird species diversity in the restored site, where a smaller, less aggressive species was dominant. Our results demonstrate the restitution of wide outcrossing in these restored Banksia patches within an active mine‐site, and suggest that restored bird‐pollinated Banksia populations are resilient to human impacts, due largely to their generalist pollinator requirements and highly‐mobile avian pollinators.     

Enhancement of fruit quality in Capsicum annum through pollination by Hypotrigona gribodoi in Kakamega,Western Kenya

A study was carried out in Kakamega forest, in the western region of Kenya, to evaluate the effectiveness of the stingless bee Hypotrigona gribodoi (Magretti, 1884) on the pollination of green pepper. Three treatments were applied and consisted of self‐pollination, pollination by feral pollinators in the open field and pollination by H. gribodoi in a net cage. The differences in fruit yield and seed quality were compared among treatments. Flowers pollinated by H. gribodoi produced the heaviest fruits with the highest seed numbers followed by feral pollinators and lastly self‐pollinated flowers. Moreover, seeds were significantly bigger in size in fruits resulting from flowers pollinated by H. gribodoi compared to fruits obtained from self‐pollinated flowers or flowers pollinated by feral insects. We, therefore, conclude that H. gribodoi is an efficient pollinator of green pepper in the tropical region of East Africa.     

Proximity to wildflower strips did not boost crop pollination on small,diversified farms harboring diverse wild bees

The yield of many agricultural crops depends on pollination services provided by wild and managed bees, many of which are experiencing declines due to factors such as reductions in floral resources. Thus, improving pollinator habitat on farmlands using management strategies like planting wildflower strips is vital for wild bee conservation and sustainable crop pollination. Yet, few studies have examined whether and at what spatial scales wildflower strips enhance crop pollination and yields, and most research has been conducted in large-scale commercial agriculture. Therefore, we investigated the effects of wildflower strips on crop pollination on small, diversified farms (i.e., those growing a variety of crop species) where wild bee diversity and abundance is predicted to be comparatively high. Over three years, on four diversified farms in Montana USA, we tested the hypothesis that distance (20, 60, and 180 m) of crops from native perennial wildflower strips planted alongside crop fields affected wild bee visitation, pollination, and yields of squash and sunflower crop plants. We found that distance to wildflower strips did not affect bee visitation or pollination in crops. Squash yield was pollen-limited in the growing season prior to wildflower strip establishment, and in one of the two years after wildflower strip establishment, but proximity to wildflower strips did not influence the magnitude of pollen limitation. Sunflower seed production was not pollen-limited in any year. Our findings demonstrate that even on diverse farms with wildflower strips and a demonstrated high diversity of bees, some crops do not necessarily receive maximum pollination, regardless of distance from the wildflower strips. However, the value of wildflower strips for supporting wild bee diversity, and other ecological or economic benefits, needs consideration for a full understanding of this pollinator habitat management strategy.     

Quantitative and qualitative consequences of reduced pollen loads in a mixed‐mating plant

Greater pollination intensity can enhance maternal plant fitness by increasing seed set and seed quality as a result of more intense pollen competition or enhanced genetic sampling. We tested experimentally these effects by varying the pollen load from a single pollen donor on stigmas of female flowers of Dalechampia scandens (Euphorbiaceae) and measuring the effects on seed number and seed mass. Seed set increased rapidly with pollen number at low to moderate pollen loads, and a maximum set of three seeds occurred with a mean pollen load of 19 pollen grains. We did not detect a trade‐off between the number of seeds and seed mass within a fruit. Seed mass increased with increasing pollen load, supporting the hypothesis of enhanced seed quality via increased pollen‐competition intensity or genetic sampling. These results suggest that maternal fitness increases with larger pollen loads, even when the fertilization success is already high. Our results further highlight the importance of high rates of pollen arrival onto stigmas, as mediated by reliable pollinators. Comparing the pollen‐to‐seed response curve obtained in this experiment with those observed in natural populations suggests that pollen limitation may be more severe in natural populations than predicted from greenhouse studies. These results also indicate that declines in pollinator abundance may decrease plant fitness through lowered seed quality before an effect on seed set is detected.     

Relative impact of mate versus pollinator availability on pollen limitation and outcrossing rates in a mass‐flowering species

Plant mating systems are driven by several pre‐pollination factors, including pollinator availability, mate availability and reproductive traits. We investigated the relative contributions of these factors to pollination and to realized outcrossing rates in the patchily distributed mass‐flowering shrub Rhododendron ferrugineum. We jointly monitored pollen limitation (comparing seed set from intact and pollen‐supplemented flowers), reproductive traits (herkogamy, flower size and autofertility) and mating patterns (progeny array analysis) in 28 natural patches varying in the level of pollinator availability (flower visitation rates) and of mate availability (patch floral display estimated as the total number of inflorescences per patch). Our results showed that patch floral display was the strongest determinant of pollination and of the realized outcrossing rates in this mass‐flowering species. We found an increase in pollen limitation and in outcrossing rates with increasing patch floral display. Reproductive traits were not significantly related to patch floral display, while autofertility was negatively correlated to outcrossing rates. These findings suggest that mate limitation, arising from high flower visitation rates in small plant patches, resulted in low pollen limitation and high selfing rates, while pollinator limitation, arising from low flower visitation rates in large plant patches, resulted in higher pollen limitation and outcrossing rates. Pollinator‐mediated selfing and geitonogamy likely alleviates pollen limitation in the case of reduced mate availability, while reduced pollinator availability (intraspecific competition for pollinator services) may result in the maintenance of high outcrossing rates despite reduced seed production.