Logo: to the web site of Uppsala University

Flowering plants have many modes of sexual reproduction, notably varying from selfing to outcrossing, and from bisexual flowers to individuals with separate sexes (dioecy). These reproductive modes are associated with floral and life-history traits that have evolved together, making it difficult to interpret correlations between traits.

We analysed variation in 21 traits related to flowers, pollination, mating, sexual systems and life history from 361 species representative of flowering plant diversity.

Outcrossing was mainly found among long-lived, large-stature plants, but hermaphroditic outcrossers and dioecious species appeared as largely nonoverlapping strategies in the trait space. Level of floral investment was the main difference between these strategies, with dioecious species having generally smaller, less rewarding flowers, a pattern that also occurred in biotically pollinated species.

This multi-trait study shows that pollination can be achieved in many, often contrasting, ways. Despite extensive variation in reproductive traits, dioecy stands out as being linked to floral traits primarily, while correlations with lifespan and dispersal traits appear spurious. We provide a conceptual framework based on lifespan, floral investment and sexual separation that can be used to integrate pollination, reproduction and plant growth in future research on plant evolution and ecology.

Floral handling can be energetically costly for bees, yet these costs are rarely measured. We provide the first direct quantification of the metabolic cost of floral buzzing in bumblebees and evaluate its ecological significance. Using flow-through respirometry synchronized with laser vibrometry, we measured carbon dioxide production during floral buzzing by Bombus terrestris and compared it with flight take-off, which is powered by the same thoracic muscles. Floral buzzing required high muscular effort compared with other behaviours, with approximately 0.10 J per event and mass-specific power approximately 293 W/kg. Overall costs were comparable to take-off because floral buzzing bouts are longer despite a lower metabolic rate. Metabolic rate scaled with body mass, whereas intertegular span did not, implying that transient load rather than structural size better explains energetic demand in short, high-intensity behaviours. Metabolic traits were highly repeatable within individuals, but variable between individuals, and colony identity also explained additional variance. Converting costs to nectar equivalents showed that floral buzzing required slightly more nectar than take-off and that requirements rose as nectar sugar concentration declined. We conclude that floral buzzing is a major, previously unquantified component of bee energy budgets that is likely to shape nectar demands, flower visitation patterns and generally plant-pollinator interactions.

1. The rate at which flowers dispense pollen across successive pollinator visits is a key functional trait influencing male reproductive success. Pollen presentation theory (PPT) predicts that when pollinators are abundant and actively groom their bodies to collect pollen, plants should evolve gradual pollen release. Despite its intuitive appeal and generality, PPT has rarely been tested empirically using real bee visits, especially in buzz-pollinated plants where pollen dispensing is controlled mechanically.
2. Buzz-pollinated plants are ideal to study PPT because they are often nectarless and rely on pollen-grooming bees, which must apply vibrations to specialised tube-like or poricidal anthers to release pollen. The vibrations generated by some bees directly determine pollen release and thus yield a direct mapping between bee behaviour and floral function.
3. Here, we provide the first quantitative assessment of pollen dispensing in a buzz-pollinated plant, Solanum rostratum, using real bumblebee (Bombus terrestris) visits.
4. We conducted 80 trials with individual bees and measured over 2000 pollination buzzes to estimate the relationship between total buzzing time and pollen dosing. Individual pollination buzzes had a fundamental frequency of 340 ± 26 Hz (mean ± SD) and were relatively short in duration (0.748 ± 0.604 s).
5. We found that S. rostratum shows a strong pattern of gradual pollen release with flowers dispensing on average 2%–4% of the available pollen grains for each buzzing second.
6. Our study links an ecologically relevant pollinator behaviour to a key floral functional trait that affects male fitness, offering empirical support for PPT in a buzz-pollinated system.

Grooming by pollen vectors and the positioning of floral reproductive organs jointly influence pollen transfer by shaping where pollen grains are deposited on, redistributed on, and removed from the vector's body, thereby affecting plant mating patterns. Linking these processes to reproductive outcomes requires tracking the fate of individual pollen grains, which remains challenging. We used a spatially explicit agent-based model to simulate sequential floral visits by pollen vectors with varying grooming intensities to flowers differing in anther and stigma positions. From these simulations, we generated quantitative mating matrices and analyzed mating patterns at both the population and individual levels. Grooming reduced the proportion of realized mating interactions relative to all possible plant pairings ("connectance"), while increasing partner exclusivity ("specialization") and clustering ("modularity"). These changes in mating network structure concentrated pollen exchange within smaller plant groups, reducing mating partner diversity and substantially weakening the influence of floral morphological variation. Without grooming, by contrast, variation in stigma position increased mating diversity, and visit order intensified first-male and last-female advantages. Overall, grooming seems to constrain the diversity of effective mating interactions and diminishes the functional impact of stigma position, promoting more uniform and selective mating patterns. In bee-pollinated plants, this may favor stabilizing selection on stigma positions, whereas in low-grooming systems, greater variation in stigma positions may be tolerated or even advantageous.

During the Paleogene, bees diversified extensively after expanding from tropical regions into northern forests surrounding the future Baltic Sea, evolving alongside diverse angiosperms. Most fossil bee diversity is preserved in amber, which captures morphology but rarely behavior or ecology. Using synchrotron X-ray imaging, we describe a new leaf-cutter bee species, †Protolithurgus acarophorus sp. nov., from two specimens in Baltic amber. This is only the second known record of the extinct tribe †Protolithurgiini, which displays both ancestral and derived megachilid traits. The holotype preserves exceptional anatomical detail and reveals ecological insights, including compacted floral rewards in the hind-leg scopa, a pollen-carrying strategy rare in living Megachilidae. We also document the first fossil bee-mite association, suggesting unspecialized kleptoparasitic interactions, and reconstruct thoracic musculature dimensions. Together, these findings provide rare evidence of the foraging ecology, parasite relationships, and biomechanics of the earliest leaf-cutter bees.

Background and Aims: Intra-individual variation in the quantity and quality of floral rewards can affect pollinator foraging and plant fitness. Few studies have examined how this variation affects pollinator behaviour, particularly in pollen-rewarding plants.

Methods: We quantified intra-individual variation in pollen production in two nectarless species (Solanum sisymbriifolium and S. rostratum), in which pollen is concealed inside poricidal anthers. Then, for one of these species (S. rostratum), we manipulated pollen availability to create plants with increased intra-individual variation (three flowers with high, natural levels and no pollen) or plants with natural levels of variation (three flowers with natural pollen levels). We exposed both treatments simultaneously to bumblebees (Bombus terrestris), recording visitation and buzzing behaviour.

Key Results: Natural intra-individual variation in pollen production was high, with 46-58 % of the variation attributable to differences among flowers of the same plant, and coefficients of variation of 14-23 %. Analysis of foraging patterns capturing >2500 visits and 9700 buzzing events showed that bumblebees performed fewer and shorter buzzing visits on pollen-less flowers, whereas they were more likely to remain on a plant after visiting higher-reward flowers, potentially reducing outcrossing opportunities. These flower-level responses did not scale up to the plant level. Contrary to our expectations and to previous work in nectar-rewarding systems, plants in both variation treatments received visits in similar quantity and quality.

Conclusions: These findings suggest that B. terrestris can detect and respond to differences among flowers in pollen quantity despite high intra-individual variation, but that there is no plant-level effect. Given that bees respond to individual flowers rather than integrating information at the plant level, the behavioural patterns documented here might help plants to buffer fitness consequences of intra-individual variation in pollen rewards, which we show is substantial even at the level of pollen production and is likely to be exacerbated following visitation.

Bees use vibrations across behavioural contexts, including a specialised foraging routine in which they vibrate flowers to release pollen (floral vibrations). Floral vibrations often have complex spectral properties, combining fundamental frequencies and their harmonics, yet their functional consequences remain unknown. We experimentally replicated bee buzzes with identical fundamental frequency (300 Hz) but different harmonic content (0, 2 or 4 harmonics) to determine their effect on pollen release, a key functional outcome for both the bee and the flower. We normalised signals to the same average power by holding root mean squared (RMS) acceleration constant and applied three RMS levels to test for interactions between power and spectral composition. We applied vibrations to flowers and quantified both the input signal (at the shaker) and the vibration transmitted at a distal anther, and measured pollen release. Harmonic content did not explain pollen release at any RMS level. In contrast, RMS acceleration strongly predicted pollen removal. The RMS measured at the distal anther was only weakly correlated with the input RMS, suggesting that the flower strongly modifies the vibration signal. Both input RMS and distal anther RMS explained pollen release, irrespective of spectral composition. Our results suggest that pollen release is determined by vibration amplitude (RMS acceleration) rather than spectral composition and is jointly governed by the bee-generated input and the mechanical filtering properties of the flower.

Introduced flowering plants often encounter novel pollinator assemblages in their non-native ranges, potentially reshaping plant-pollinator interactions and driving evolutionary change. We tested whether variation in pollinator assemblages mediates local adaptation of introduced Mimulus guttatus (Phrymaceae). Using a reciprocal transplant experiment between the southern and northern extremes of the British Isles, we combined pollinator censuses with measurements of floral traits and fruit production to assess pollinator assemblages, phenotypic differentiation and selection. Pollinator assemblages differed between sites, with more abundant bumblebees in the northern garden. Pollinator visitation rate was positively associated with fruit set, indicating that pollinators remain essential for sexual reproduction in the introduced range. Plants from northern populations received more visits in the northern than in the southern garden, and more visits than southern populations in the northern garden. This pattern paralleled a previously reported local and home advantage in fruit production. Corolla width was under positive and stabilizing selection mediated by bumblebees in the north, suggesting that pollinator-driven selection contributes to floral differentiation between regions. Our findings demonstrate that M. guttatus has undergone pollinator-mediated adaptive differentiation within two centuries in the British Isles, highlighting how strong biotic pressures can generate rapid ecological diversification during plant introductions.

The evolution of tube-like floral morphologies that control pollen release via small openings (functionally poricidal flowers) represents a taxonomically and geographically widespread instance of repeated and independent evolution of a functionally similar morphology. Poricidal flowers are also often closely associated with buzz pollination by bees. Yet we lack an updated angiosperm-wide survey of their phylogenetic distribution. We identify all known angiosperm genera containing poricidal flowers via a literature survey. We determined their phylogenetic distribution and minimum number of independent gains and losses via a species-level angiosperm-wide phylogeny. We estimated if evolution of poricidal flowers is associated with changes in speciation/extinction via diversification rate analyses. Poricidal flowers occur across 87 angiosperm families and 639 genera containing > 28,000 species. At the species level, an average of 205 independent gains and 215 losses of poricidal flowers occurred. Angiosperm-wide analyses suggest an early burst in poricidal evolution, but no differences in net diversification (origination-extinction) between non-poricidal and poricidal taxa. Analyses for two focal families however indicate strong context-dependent effects of poricidal flowers on diversification. Poricidal evolution thus represents a large-scale example of convergent evolution in floral form, but effects on diversification appear to be strongly contingent on phylogenetic and ecological background.

Flowers display remarkable diversity, much of which is shaped by interactions with pollinators. While various floral traits are considered adaptations to pollination, the function of many morphological features remains untested. In buzz-pollinated species, where bees use vibrations to extract pollen, anther morphology is diverse and can influence pollen removal and transfer. Poricidal anthers, common in these species, sometimes bear spur-like projections. These spurs are hypothesized to aid pollen release, but this has not been experimentally tested. Here, we investigated the function of anther spurs in Vaccinium myrtillus (Ericaceae). Using a laboratory setup, we conducted two complementary experiments: one where Bombus terrestris bumblebees foraged on flowers, and another applying 1 s artificial vibrations to mimic bee buzzes. We compared pollen release between flowers with intact anther spurs and those where spurs had been removed. Bumblebees produced vibrations during 90% of visits, irrespective of spur presence. Buzzing visits were shorter (21.70 ± 24 s; mean ± SD) and removed more pollen (60 ± 29%) than non-buzzing visits, which were three times longer (63.79 ± 70.45 s) but removed 23 ± 43% of pollen. Artificial 1 s buzzes removed an average of 23% of pollen. Spur removal increased pollen release in both the bee (54% to 60%) and the artificial buzzing experiment (18% to 28%). Our results suggest that anther spurs in V. myrtillus potentially act as a pollen-dispensing mechanism, but this effect depends on bee behaviour and visit duration. Further studies could explore other roles the spur has in interactions with non-buzzing visitors, such as regulating pollen release when visitors collect nectar.