Community-level trait matching between flowers and bees across Europe

Community-level trait matching between flowers and bees across Europe | Authorea try { document.documentElement.classList.add('js'); } catch (e) { } var _gaq = _gaq || []; _gaq.push(['_setAccount', 'G-8VDV14Y67G']); _gaq.push(['_trackPageview']); (function() { var ga = document.createElement('script'); ga.type = 'text/javascript'; ga.async = true; ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js'; var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s); })(); Skip to main content Preprints Collections Wiley Open Research IET Open Research Ecological Society of Japan All Collections About About Authorea FAQs Contact Us Quick Search anywhere Search for preprint articles, keywords, etc. Search Search ADVANCED SEARCH SCROLL This is a preprint and has not been peer reviewed. Data may be preliminary. 17 March 2026 V1 Latest version Share on Community-level trait matching between flowers and bees across Europe Author : Tamar Keasar 0000-0002-4925-0823 [email protected] Authors Info & Affiliations https://doi.org/10.22541/au.177376725.55038841/v1 136 views 63 downloads Contents Abstract Information & Authors Metrics & Citations View Options References Figures Tables Media Share Abstract Predicting pairwise species interactions in bipartite networks is a longstanding challenge in community ecology. In pollination networks, the shape of flowers often matches the mouthparts of their animal pollinators. This trait matching facilitates the task of forecasting which flower-insect links exist within a given network. I predicted that trait matching would also vary across entire flower-bee communities. I combined a database of European pollination networks across 122 sites (latitude range 36.6-67.7º N) with information on bee (n=392 species) and flower (n=260 species) traits. I tested for correlations between the average depth and symmetry of visited flowers and the average proboscis length, inter-tegular distance, and sociality of their wild bee visitors. Communities of mostly shallow and radial flowers interacted with more bee species, and had fewer interactions with eusocial bees, than communities dominated by deep and bilateral flowers. Furthermore, communities of shallow radial flowers were visited by smaller and shorter-tongued wild bees. Including flower interactions with honey bees in the analyses weakened the trait matching. Leveraging the broad geographical range of the pollination networks in the database, I tested for macroecological trends in bee traits, flower traits, and their matching. The wild bees’ proboscis length and the proportions of eusocial species in the networks significantly increased with latitude, while their functional diversity declined. Floral depth, symmetry, and functional diversity were not predicted by latitude. Bee tongues tended to be shorter than floral corolla tubes in southern Europe, and longer than the corolla tubes in northern Europe. This work illustrates how integration of species-level traits can increase our understanding of macroecological patterns in plant-pollinator interaction networks. ABSTRACT Predicting pairwise species interactions in bipartite networks is a longstanding challenge in community ecology. In pollination networks, the shape of flowers often matches the mouthparts of their animal pollinators. This trait matching facilitates the task of forecasting which flower-insect links exist within a given network. I predicted that trait matching would also vary across entire flower-bee communities. I sites (latitude range 36.6-67.7º N) with information on bee (n=392 species) and flower (n=260 species) traits. I tested for correlations between the average depth and symmetry of visited flowers and the average proboscis length, inter-tegular distance, and sociality of their wild bee visitors. Communities of mostly shallow and radial flowers interacted with more bee species, and had fewer interactions with eusocial bees, than communities dominated by deep and bilateral flowers. Furthermore, communities of shallow radial flowers were visited by smaller and shorter-tongued wild bees. Including flower interactions with honey bees in the analyses weakened the trait matching. Leveraging the broad geographical range of the pollination networks in the database, I tested for macroecological trends in bee traits, flower traits, and their matching. The wild bees’ proboscis length and the proportions of eusocial species in the networks significantly increased with latitude, while their functional diversity declined. Floral depth, symmetry, and functional diversity were not predicted by latitude. Bee tongues tended to be shorter than floral corolla tubes in southern Europe, and longer than the corolla tubes in northern Europe. This work illustrates how integration of species-level traits can increase our understanding of macroecological patterns in plant-pollinator interaction networks. KEY WORDS bee eusociality; corolla depth; flower symmetry; inter-tegular distance; latitude; pollination network; proboscis length INTRODUCTION Interactions between species drive the structure and function of ecological communities, yet documenting and interpreting all interactions in communities with multiple trophic levels is often challenging (Dalla Riva et al., 2019). Hence, much ecological research restricts itself to bipartite interactions between two groups of organisms. For example, pollination networks provide convenient descriptions of the interactions between flowering plants and their animal pollinators (Vasquez et al., 2009, Schwartz et al. 2020). Many quantitative indices, based on graph theory, have been developed to describe the structure of pollination networks, e.g. connectance, nestedness, and interaction strength (Vizentin-Bugoni et al., 2018). Analyses of pollination networks from different ecosystems reveal common structural features, such as modularity (Olesen et al., 2007) and asymmetric specialization (Vasquez and Aizen, 2004) in large networks. Yet, predicting which plant-pollinator pairs interact in a given network remains a challenge (Olito and Fox, 2015, Peralta et al., 2024). It is difficult to predict pairwise links because pollination interactions tend to be generalized (many species in the network interact with several partners, Bascompte et al., 2007) and change over time (Petanidou et al., 2008; Bramon Mora et al., 2020). Identifying matches between traits of flowers (e.g., color, scent, morphology, flowering phenology) and pollinators (e.g., color vision, chemoreception, body size) improves forecasts of plant-pollinator links and promotes understanding of network structure (Pichler et al., 2020). Such matches, often referred to as ‘pollination syndromes’, help predict the broad group of pollinators (for example, birds, bats, beetles) that is most likely to visit a flower. Among bee-pollinated plants, trait matching occurs at a finer scale: the length of floral tubes often matches the proboscis length of their bee visitors. Other morphological features of flowers, such as shape class and symmetry, also predict bee visitors at the genus level (Ornai and Keasar, 2020). Two complementary adaptive hypotheses aim to explain the trait matching between flowers and bees. The first hypothesis takes a foraging perspective and suggests that feeding on matching flowers optimizes the nectar intake rates of bees by decreasing the time required to handle the flowers and increasing the amount of nectar collected. Foragers reduce their flower handling time by avoiding flowers that are too deep to reach with their mouthparts. They increase their caloric intake by avoiding shallow flowers, which typically produce less nectar than deep ones (Klumpers et al., 2019, Keasar and Bodner, 2025). The second hypothesis focuses on learning as a barrier to bee foraging on some flower morphologies. According to this interpretation, flower shapes can be classified along a continuum of accessibility, from generalized to specialized. Generalized flowers are often radial and dish-shaped and are visited by a wide range of pollinator taxa. Specialized flowers have bilateral symmetry, long nectar tubes and other morphological structures that make them less accessible to pollinators (Keasar, 2018). Such flowers require a learning period to be handled efficiently (Muth et al., 2015, Krishna & Keasar, 2019), and consequently have a narrower range of visitors (Yoder et al., 2020) despite their higher nectar production. This raises the question how pollinators manage to cross the learning barrier and provide pollination services to specialized flowers. A spatially explicit model predicts that long-distance fliers evolve higher learning abilities than short-distance fliers (Keasar & Wajnberg, 2025). The model further predicts that pollinator learning abilities coevolve and correlate positively with the frequency of specialized flowers. The mechanism behind this theoretical prediction is that highly mobile pollinators encounter many flowers during their lifetime, providing them with ample opportunities to learn how the handle specialized flowers. This generates the hypothesis that bees’ flight ranges (or predictors thereof) are expected to correlate with the morphological specialization (corolla depth and symmetry) of the flowers they visit (Keasar & Wajnberg, 2025). Recently, large databases of pollination networks have been compiled, as well as databases of flower and bee traits. By combining these sources of information, analyses of trait matching can be upscaled from pairwise interactions within flower-bee networks to entire ecological communities, providing clues to network structure and function. Using a recently published large database of pollination networks, I first tested the predictions that emerge from the two adaptive hypotheses of trait matching at the level of flower-bee communities. The hypothesis that stresses foraging efficiency predicts that plant communities with more specialized flowers are visited by longer-tongued bees. The hypothesis that focuses on bee learning predicts, based on mathematical modeling, that communities of more specialized flowers are visited by bees that are better fliers. Flight distances of bees increase with sociality and with inter-tegular distance (ITD, a common measure of body size) in interspecific comparisons (Grüter and Hayes, 2022, Kendall et al., 2022). I therefore correlated the mean depth and symmetry of the visited flowers with the mean proboscis length, ITD, and sociality of their bee visitors, treating each locality of plant-pollinator observations as a data point. Past studies that analyzed trait matching across environmental gradients were often limited to a small number of sites and plant-pollinator communities. For example, Dehling et al. (2014) compared trait matching in bird-fruit interactions between one low- and one high-altitude Andean sites. Sonne et al. (2019) similarly studied trait matching between flowers and hummingbirds in three sites that differed in elevation in the Andes. Recent large databases of bipartite interactions cover tens (e.g., Sonne et al., 2020 for flowers and hummingbirds) or hundreds (e.g., Huang et al., 2025 for fruits and fruit-eating birds) of sites, and provide more robust insights on biogeographical trends of trait matching. Likewise, the flower-bee interaction database analyzed here (Lanuza et al., 2025) compiles hundreds of pollination webs from all over Europe, and therefore lends itself to macroecological analyses. The species richness of flowering plants is highest in the tropics and declines towards the poles (Jiang et al. 2023) whereas the richness of bee species peaks around the 30 ◦ latitude in both hemispheres (Orr et al., 2021). In a global analysis of plant-animal pollination networks, the total richness of species did not vary with latitude, while the number of links per plant species and the pollinators/plants species ratio peaked at mid latitudes (Trøjelsgaard and Olesen, 2012). The present study supplements these documented biogeographical patterns with information on flower and bee traits. To explore the geographical variation in community-level trait matching, I tested the effect of latitude on the flowers’ and on the bees’ community-level traits and functional diversity. I also related latitude to the mean difference between the proboscis lengths of the bees and the corolla depths that they visited, as a direct measure of flower-bee size matching. MATERIAL AND METHODS Pollination networks : I analyzed a recently compiled database of pollination networks, collected from 255 localities in 23 European countries (Lanuza et al., 2025). The database contains more than 600,000 plant-pollinator interactions from 52 published and unpublished studies. Based on rarefaction curves, the estimated sampling coverage of plant and pollinator species in the database is 97%, and the estimated coverage of plant-pollinator interactions is 74% (Lanuza et al., 2025). I removed records of interactions with non-hymenopterans and with honey bees from the database and retained 134,263 interactions between flowers and wild bees. Honeybees were omitted from the initial analyses because they often originate from commercial colonies and thus do not represent the bee fauna that has naturally coevolved with the flowers. However, the data were also re-analyzed with the flower-honeybee interactions included. I excluded interactions with flower species that do not produce any nectar, because such interactions likely involve pollen collection. I further excluded a study (Study-ID 49 in Lanuza et al., 2025) that included a single flower species and a second study (Study-ID 24) where trait information on most bee species (see below) was lacking. Finally, to avoid potential biases due to low sampling effort, I omitted 134 localities where fewer than 30 flower-bee interactions were observed. The excluded networks spanned the latitudes 36.4-58.8º N. The reduced dataset, which was used for analyses, included 61,691 bee-flower interactions from 122 localities across Europe (Appendix S1). Flower traits : I compiled records on the corolla tube length (n=2,448 species) and symmetry pattern (coded as a symmetry score of either 0 (radial) or 1 (bilateral), n=4,053 species) of flowers, based on published studies and on my own field-collected data. The tube length is measured from the position in a flower that a pollinator’s head could reach to the base of the flower (Stang et al., 2006). When multiple records of corolla tube length for a plant species were available, they were averaged (Appendix S2). Bee traits : I collected information on bee inter-tegular distances (ITD, n=1,575 species) and proboscis lengths (the length of the glossa and prementum, n=414 species), based on published studies. The ITD is commonly used to estimate bee body sizes and flight ranges in interspecific comparisons (Kendall et al., 2019, 2022). The ITDs of 594 additional bee species were imputed by averaging all listed ITD values of other species in their genus. Family-specific allometric equations reliably predict bees’ proboscis lengths from their ITDs (Cariveau et al., 2016). By applying these allometric models to bees in the dataset, proboscis length estimates for 1,761 additional species were obtained. Altogether, the dataset includes information on proboscis length for 2,175 species of bees. I also tabulated information on the bees’ social organization level (eusocial – scored as 1, or not – scored as 0, n=1930 species). Sociality was used as an additional proxy of flight range because social bees fly longer distances than solitary bees of similar body size (Grüter and Hayes, 2022, Kendall et al., 2022, Appendix S3). Proboscis length information was available for all 146 eusocial species, and ITD information was available for 142 of them. Among the 1,784 non-eusocial bees, proboscis length and ITD information were available for 1,726 and 1,738 species, respectively. Data analysis : I calculated the flower-bee size mismatch for each interaction by subtracting the visitor’s proboscis length from the corolla tube length of flower that it visited. A zero mismatch indicates a flower’s corolla tube that is identical in length to the proboscis of the interacting bee. Positive mismatch values correspond to tongues that are shorter than the corolla tubes, and negative mismatch values indicate tongues that are longer than the corolla tubes. Next, I computed the mean values of corolla tube length, corolla symmetry score, bee ITD, bee proboscis length, bee sociality score, and flower-bee mismatch for each of the 147 flower-bee communities. I then calculated Pearson’s correlation coefficients between the per-community mean bee traits, flower traits, flower-bee mismatch, and site latitude. Next, I explored the hypothesis that latitude influences the bee community that in turn selects for flower traits. To test for such indirect effects of latitude on flower traits, the bees’ community-level proboscis lengths and sociality scores were regressed on latitude. Linear models were used to calculate the residuals from the regression of the bees’ proboscis length on latitude and from the regression of the sociality score on latitude. GLMs with family Gamma and log-link functions were applied to test the effects of the two residuals and of latitude on flower depth and symmetry. The Gamma family was chosen because the dependent variables are continuous, positive, and right-skewed. ITDs were not entered into the GLMs because of multicollinearity with proboscis lengths (Pearson’s correlation coefficient between proboscis length and ITD: 0.953, P<0.001). The flower-bee mismatch variable was normally distributed. I therefore used a linear model to evaluate the effects of the residuals and of latitude on the flower-bee mismatch value. The Eta 2 metric was calculated for all models to estimate the relative importance of the predictor variables. The functional diversity of flower and bee communities may influence the extent of trait matching. For example, interactions between highly diverse bee and flower communities may generate higher trait matching than interactions of low-diversity bee communities with high-diversity flower communities. Moreover, if functional diversity varies with latitude, this could contribute to latitudinal trends in trait matching. To test these possibilities, I calculated the RaoQ (Rao’s quadratic entropy) index of functional diversity of flowers and bees for each locality. I then tested the effect of latitude on functional diversity, and the effect of functional diversity on flower-bee mismatch, using linear models. R version 4.2.2 was used for all statistical analyses (R Core Team, 2022). The packages BeeIT (Carivau et al., 2016), FD (Laliberté et al., 2014), lme4 (Bates et al., 2015), effectsize (Ben-Shachar et al., 2020) and corrtable (van der Laken, 2023) were applied for data manipulation and analysis. RESULTS Bee species richness and trait correlations : As the fraction of flower-bee interactions that involved bilateral flowers (the flower community’s symmetry score) increased, bee species richness decreased (correlation coefficient: -0.257, P<0.01). The correlation between mean tube length of the interacting flowers and bee richness was not significant (correlation coefficient: -0.124). Plant communities with more specialized (deep and/or bilateral) flowers were visited by larger, longer-tongued bees, and these bees were more likely to be eusocial (Table 1). These correlations were weaker when flower interactions with honey bees were included in the analyses (Table 1). Latitudinal trends : The bees’ community-level ITDs, proboscis lengths and sociality scores increased with latitude (correlation coefficients: 0.405***, 0.372***, 0.510***, respectively, Fig. 1). The flowers’ community-level traits, on the other hand, did not show statistically significant latitudinal trends (Fig. 2). The mean flower-bee size mismatch declined from south to north (correlation coefficient: -0.273***, Fig. 3), showing more negative values at higher latitudes. The functional diversity of bee communities declined with latitude (estimated coefficient=-0.001±0.0005, r 2 =0.05, P=0.01), while the flowers’ functional diversity was not significantly influenced (estimated coefficient=0.001±0.0009, r 2 =0.02, P=0.13). The flower-bee mismatch increased with the bees’ functional diversity (estimated coefficient=-0.001±0.0005, P=0.003) and was unaffected by the flowers’ functional diversity (estimated coefficient=20.181±6.684, r 2 =0.01, P=0.13). Although latitude did not directly influence flower tube length and symmetry (Fig. 2), it may nevertheless produce indirect effects, mediated by the effects of latitude on bee traits. I regressed the bees’ community-level proboscis lengths and sociality scores against latitude to explore this possibility. The residuals from both regressions represent the community-level bee trait values after accounting for the effect of latitude. These residuals, together with latitude, were used as independent variables in models to predict the community-level plant tube lengths, plant symmetry scores and flower-bee size mismatches. The proboscis length residuals significantly predicted the flowers’ tube lengths (P<0.0001) and symmetry scores (P<0.0001), and the size mismatches (P=0.004). The sociality residuals significantly predicted only the flowers’ symmetry scores (P=0.019). Latitude did not impact floral tube length and symmetry, but had a negative effect on the flower-bee mismatches (P=0.001, Fig. 3). Proboscis length residuals had the largest effect on floral tube length and symmetry scores (Eta 2 =0.150 and 0.26, respectively), followed by the sociality residuals (Eta 2 =0.003 and 0.04, respectively) and latitude (Eta 2 =0.002 and 0.004, respectively). Latitude was the strongest predictor of the flower-bee mismatch (Eta 2 =0.08), the latitude-proboscis length residuals had a somewhat smaller effect (Eta 2 =0.06), and the latitude-sociality residuals had the weakest effect (Eta 2 =0.002). See Table 2 for detailed statistics. DISCUSSION This study takes advantage of a large, continental-scale database of pollination networks to analyze flower-bee trait matching at the scale of entire ecological communities. It allows, for the first time, comparisons of flower-bee trait matching across numerous communities. Plant communities dominated by shallow and radial flowers interacted more with small solitary bees, whereas communities dominated by deep and bilateral flowers received more visits from large and eusocial species. These patterns are compatible with the two adaptive hypotheses proposed for trait matching. The first hypothesis suggests that matching between proboscis length and flower depth, as found in this study, optimizes the foragers’ nectar intake rates (Klumpers et al., 2019). In addition, longer-tubed flowers tended to be bilaterally symmetric (correlation coefficient between tube length and symmetry score: 0.188, P<0.01), and eusocial foragers tended to have longer tongues than solitary bees (correlation coefficient between proboscis length and sociality score: 0.534, P<0.001). Taken together, these correlations can explain the observed associations between flower depth, flower symmetry, bee tongue length and bee sociality. The second hypothesis predicts, based on theoretical modeling, that floral specialization coevolves with long-distance flying capabilities and high learning abilities of bees in plant-pollinator communities (Keasar and Wajnberg, 2025). In line with this prediction, the bees’ body sizes and eusociality scores (both of which are proxies of flight range) increased in European localities dominated by specialized flowers. The species richness of bees declined as their proportion of interactions with bilateral flowers increased, supporting previous evidence that bilateral flowers are more specialized than radial ones (Yoder et al., 2020). However, symmetry represents only a partial proxy of flower specialization because it overlooks important functional variation within the broad categories of ‘radial’ and ‘bilateral’. The number of bee species per locality was not related to the mean depth of the interacting flowers, contrary to the expectation that deep flowers are more specialized than shallow ones (Stang et al., 2006). Similarly, corolla tube depth was not correlated to the number of visiting insect species in 35 species of Asteraceae (Torres & Galetto, 2002), and the ‘openness’ of flowers did not predict the numbers of their insect visitors in a sample of 37 pollination networks from various ecosystems (Olesen et al., 2007). The mean body size and the representation of eusocial species in the bee communities increased with latitude, while flower traits did not show latitudinal trends. Accordingly, mean mismatch scores became more negative from south to north, indicating more interactions of long-tongued bees with short-tubed flowers in northern localities. Bee body sizes were previously shown to increase from southern to northern Europe, conforming to Bergmann’s rule (Gérard et al., 2018). The increased proportion of social species in high-latitude sites likely reflects the dominance of the eusocial bumble bees in cold climates in the northern hemisphere. This interpretation is consistent with the decrease in functional diversity of the bee communities at higher latitudes. Thus, a geographical turnover in the bees’ community composition and functional diversity may contribute to the effect of latitude on trait matching. The community-level mismatch scores were closest to zero in temperate-climate latitudes (around 50º N), indicating the highest matching in these localities. Comparing these findings with previous studies suggests that the effect of latitude on trait matching varies by trait and taxon. For example, trait matching between frugivore beak size and fruit size in fruit-bird interaction networks increased with latitude (Huang et al., 2025). On the other hand, trait matching between flowers and their hummingbird pollinators increased toward the equator (Sonne et al., (2020). Including honey bees in the interaction dataset weakened the matching between flower and bee traits, although most of the trait correlations remained statistically significant. A straightforward interpretation of this finding is that common species in pollination networks have more interactions than rare ones, and this effect might interfere with the trait matching signal (Pichler et al., 2020). In fact, interactions between flowers and honey bees were very common (>380,000 interactions, 82.2% of all interactions with bees). In addition, honeybees adjust their nectar collection technique (lapping or sucking) to the depth of the flower and the sugar concentration of the nectar (Wei et al., 2023). This behavioral flexibility allows honey bees to feed efficiently from flowers that vary widely in depth, and may explain their lower degree of trait matching in the dataset. The dataset analyzed here provides a wealth of information on pollination interactions on a large spatial scale, but does not describe other aspects of the plant-pollinator communities, such as species abundances, season, elevation, habitat type, and climate. These unreported variables may confound our conclusions, because they can interact with latitude in their effect on trait matching. For example, abundant species have more interaction partners than rare species, and this may bias the trait matching signal towards higher values in communities with many rare species (Pichler et al., 2020). Information on additional functional traits, such as corolla width for flowers and nectar robbing frequencies for bees, can enhance the level of detail in future analyses of trait matching. Further, the interaction database does not cover tropical and subtropical ecosystems and thus is not designed to expose global patterns of trait matching. In addition, matching between flowers’ corolla depth and bees’ proboscis length is predicted to improve foraging success during nectar collection, but not during pollen collection. This expectation arises because nectaries are usually located at the base of the corolla tubes, whereas pollen is often more exposed to visitors. However, the dataset does not distinguish nectar-collecting from pollen-collecting visits. This limitation was partly addressed by removing interactions with nectarless plant species from the analyses. Nevertheless, pollen-collecting visits to nectar-producing plants could not be detected and eliminated, generating an additional source of inaccuracy. Further inaccuracies could arise from the use of genus-level averages and allometric predictions to estimate bee traits. Concerns about pollinator declines have fueled international research into plant-pollinator interactions. This effort has generated high-quality databases of pollination networks at national, continental, and even global scales (e.g., Poelen et al., 2014, Redhead et al., 2018, Allen‐Perkins et al., 2022). Integrating these databases with information on morphological, physiological and life-history traits of plants and pollinators opens exciting prospects for understanding the structure and function of pollination interactions. REFERENCES Allen‐Perkins, A., Magrach, A., Dainese, M., Garibaldi, L. A., Kleijn, D., Rader, R., … & Montero‐Castaño, A. (2022). CropPol: A dynamic, open and global database on crop pollination.\RL‏ Ecology , 103 (3), e3614.\RL‏ Bascompte, J. & Jordano, P. (2007). Plant-animal mutualistic networks: The architecture of biodiversity. Annu. rev. ecol. evol. syst ., 38, 567–593. Bates D, Mächler M, Bolker B, Walker S (2015). “Fitting Linear Mixed-Effects Models Using lme4.” Journal of statistical software , 67 (1), 1–48. Ben-Shachar MS, Lüdecke D, Makowski D (2020). “effectsize: Estimation of Effect Size Indices and Standardized Parameters.” Journal of open source software , 5 (56), 2815. Bramon Mora, B., Shin, E., CaraDonna, P. J., & Stouffer, D. B. (2020). Untangling the seasonal dynamics of plant-pollinator communities. Nature communications , 11 (1), 4086.\RL‏ Cariveau, D. P., Nayak, G. K., Bartomeus, I., Zientek, J., Ascher, J. S., Gibbs, J., & Winfree, R. (2016). The allometry of bee proboscis length and its uses in ecology. PloS one , 11 (3), e0151482.\RL‏ Dalla Riva, G. V., Fortin, M. J., Gravel, D., Guimaraes Jr, P. R., Hembry, D. H., Newman, E. A., … & Poisot, T. (2019). Analysing ecological networks of species interactions. Biological reviews , 94 , 16-36.\RL‏ Dehling, D. M., Töpfer, T., Schaefer, H. M., Jordano, P., Böhning‐Gaese, K., & Schleuning, M. (2014). Functional relationships beyond species richness patterns: trait matching in plant–bird mutualisms across scales. Global Ecology and Biogeography , 23 (10), 1085-1093.\RL‏ Gérard, M., Vanderplanck, M., Franzen, M., Kuhlmann, M., Potts, S. G., Rasmont, P., … & Michez, D. (2018). Patterns of size variation in bees at a continental scale: does Bergmann’s rule apply?. Oikos , 127 (8), 1095-1103. Grüter, C., & Hayes, L. (2022). Sociality is a key driver of foraging ranges in bees. Current biology , 32 (24), 5390-5397.\RL‏ Huang, X., Dalsgaard, B., & Chen, S. C. (2025). Weaker Plant‐Frugivore Trait Matching Towards the Tropics and on Islands. Ecology Letters , 28 (1), e70061.\RL‏ Jiang, K., Wang, Q., Dimitrov, D., Luo, A., Xu, X., Su, X., … & Wang, Z. (2023). Evolutionary history and global angiosperm species richness–climate relationships. Global Ecology and Biogeography , 32 (7), 1059-1072.\RL‏ Keasar, T. (2018). Patterns of flower complexity in plant communities. Annual plant reviews online , 643-660. Keasar, T., & Bodner, L. (2025). Do flowers with specialized morphologies produce more nectar and pollen? American journal of botany , in press. Keasar, T., & Wajnberg, E. (2025). What maintains variation in flower accessibility to pollinators in plant communities? A simulation study. BMC ecology and evolution , 25 (1), 45. Kendall, L. K., Rader, R., Gagic, V., Cariveau, D. P., Albrecht, M., Baldock, K. C., … & Bartomeus, I. (2019). Pollinator size and its consequences: Robust estimates of body size in pollinating insects. Ecology and evolution , 9 (4), 1702-1714.\RL‏ Smith, H. G., & Bartomeus, I. (2022). The potential and realized foraging movements of bees are differentially determined by body size and sociality.\RL‏ Ecology , 103 (11), e3809 Klumpers, S. G., Stang, M., & Klinkhamer, P. G. (2019). Foraging efficiency and size matching in a plant–pollinator community: the importance of sugar content and tongue length. Ecology letters , 22 (3), 469-479.\RL‏ Krishna, S., & Keasar, T. (2019). Bumblebees forage on flowers of increasingly complex morphologies despite low success. Animal behaviour , 155 , 119-130.\RL‏ Laliberté, E., Legendre, P., and B. Shipley. (2014). FD: measuring functional diversity from multiple traits, and other tools for functional ecology. R package version 1.0-12.3. Lanuza, J. B., Knight, T. M., Montes‐Perez, N., Glenny, W., Acuña, P., Albrecht, M., … & Bartomeus, I. (2025). EuPPollNet: a European database of plant‐pollinator networks. Global ecology and biogeography , 34 (2), e70000. Muth, F., Keasar, T., & Dornhaus, A. (2015). Trading off short-term costs for long-term gains: how do bumblebees decide to learn morphologically complex flowers?. Animal behaviour , 101 , 191-199.\RL‏ Olesen, J. M., Dupont, Y. L., Ehlers, B. K., & Hansen, D. M. (2007). The openness of a flower and its number of flower-visitor species. Taxon , 56 (3), 729-736.\RL‏ Olito, C., & Fox, J.W. (2015). Species traits and abundances predict metrics of plant–pollinator network structure, but not pairwise interactions. Oikos , 124, 428–436. Ornai, A., & Keasar, T. (2020). Floral complexity traits as predictors of plant-bee interactions in a Mediterranean pollination web. Plants , 9 (11), 1432.\RL‏ Orr, M. C., Hughes, A. C., Chesters, D., Pickering, J., Zhu, C. D., & Ascher, J. S. (2021). Global patterns and drivers of bee distribution. Current Biology , 31 (3), 451-458.\RL‏ Peralta, G., CaraDonna, P. J., Rakosy, D., Fründ, J., Tudanca, M. P. P., Dormann, C. F., … & Vázquez, D. P. (2024). Predicting plant–pollinator interactions: concepts, methods, and challenges. Trends in ecology & evolution , 39 (5), 494-505.\RL‏ Petanidou, T., Kallimanis, A.S., Tzanopoulos, J., Sgardelis, S.P., & Pantis, J.D. (2008). Long-term observation of a pollination network: Fluctuation in species and interactions, relative invariance of network structure and implications for estimates of specialization. Ecology letters , 11, 564–575. Pichler, M., Boreux, V., Klein, A. M., Schleuning, M., & Hartig, F. (2020). Machine learning algorithms to infer trait‐matching and predict species interactions in ecological networks. Methods in Ecology and Evolution , 11 (2), 281-293.\RL‏ Poelen, J. H., Simons, J. D., & Mungall, C. J. (2014). Global biotic interactions: An open infrastructure to share and analyze species-interaction datasets. Ecological informatics , 24 , 148-159.\RL‏ R Core Team (2022). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/. Redhead, J.W., Woodcock, B.A., Pocock, M.J., Pywell, R.F., Vanbergen, A.J., & Oliver, T.H. (2018). Potential landscape‐scale pollinator networks across Great Britain: structure, stability and influence of agricultural land cover. Ecology letters 21:1821-1832 Schwarz, B., Vázquez, D.P., CaraDonna, P.J., Knight, T.M., Benadi, G., Dormann, C.F., Gauzens, B., Motivans, E., Resasco, J., & Blüthgen, N. (2020). Temporal scale-dependence of plant–pollinator networks. Oikos , 129, 1289–1302. Sonne, J., Zanata, T. B., Martín González, A. M., Cumbicus Torres, N. L., Fjeldså, J., Colwell, R. K., … & Dalsgaard, B. (2019). The distributions of morphologically specialized hummingbirds coincide with floral trait matching across an Andean elevational gradient. Biotropica , 51 (2), 205-218. Sonne, J., Vizentin-Bugoni, J., Maruyama, P. K., Araujo, A. C., Chávez-González, E., Coelho, A. G., … & Dalsgaard, B. (2020). Ecological mechanisms explaining interactions within plant–hummingbird networks: morphological matching increases towards lower latitudes. Proceedings of the Royal Society B , 287 (1922), 20192873.\RL‏ Stang, M., Klinkhamer, P. G., & Van Der Meijden, E. (2006). Size constraints and flower abundance determine the number of interactions in a plant–flower visitor web. Oikos , 112 (1), 111-121. Torres, C., & Galetto, L. (2002). Are Nectar Sugar Composition and Corolla Tube Length Related to the Diversity of Insects that Visit Asteraceae Flowers?. Plant biol , 4 (360), 360-366.\RL‏ Trøjelsgaard, K., & Olesen, J. M. (2013). Macroecology of pollination networks. Global Ecology and Biogeography , 22 (2), 149-162.\RL‏ van der Laken, P. (2023). _corrtable: Creates and Saves Out a Correlation Table with Significance Levels Indicated_. R package version 0.1.1. Vázquez, D. P., & Aizen, M. A. (2004). Asymmetric specialization: a pervasive feature of plant–pollinator interactions. Ecology , 85 (5), 1251-1257.\RL‏ Vázquez, D.P., Chaco_, N.P., & Cagnolo, L. (2009). Evaluating multiple determinants of the structure of plant–animal mutualistic networks. Ecology , 90, 2039–2046. Vizentin-Bugoni, J., Maruyama, P. K., de Souza, C. S., Ollerton, J., Rech, A. R., & Sazima, M. (2018). Plant-pollinator networks in the tropics: a review. In:, Dáttilo, W. & Rico-Gray, V. (eds.) Ecological networks in the tropics: An integrative overview of species interactions from some of the most species-rich habitats on earth , pp. 73-91. Springer International. Wei, J., Rico-Guevara, A., Nicolson, S. W., Brau, F., Damman, P., Gorb, S. N., … & Wu, J. (2023). Honey bees switch mechanisms to drink deep nectar efficiently. Proceedings of the national academy of sciences , 120 (30), e2305436120.\RL‏ Yoder, J. B., Gomez, G., & Carlson, C. J. (2020). Zygomorphic flowers have fewer potential pollinator species. Biology letters , 16 (9), 20200307.\RL‏ Table 1: Correlation coefficients and associated significance values (* - <0.05, *** - <0.001) between community-level floral specialization traits and traits of their wild bee visitors. Pollination networks from 147 localities in Europe were analyzed. Values in parentheses include both wild bees and honey bees. Tube length Symmetry score Bee traits ITD 0.378*** (0.242***) 0.375*** (0.165*) Proboscis length 0.377*** (0.230**) 0.442*** (0.228*) Sociality score 0.208* (-0.049) 0.299*** (0.199**) Table 2: Summary statistics for the effects of latitude and residual bee traits on flower depth, symmetry and flower-bee size mismatch. Statistically significant effects are indicated in bold. Predictors Parameter estimate and SE P-value Eta 2 Parameter estimate and SE P-value Eta 2 Parameter estimate and SE P-value Eta 2 Latitude 0.007±0.005 0.189 0.002 -0.007±0.008 0.348 0.004 -0.105±0.030 0.001 0.08 Latitude-Proboscis residuals 0.998±0.021 <0.0001 0.150 0.147±0.031 <0.001 0.26 -0.345±0.119 0.004 0.06 Latitude-Sociality residuals 0.098±0.150 0.516 0.003 0.531±0.223 0.019 0.04 0.485±0.868 0.577 0.002 FIGURE CAPTIONS Fig. 1: The relationship between latitude and mean inter-tegular distance (a), proboscis length (b) and sociality score (c) of bees in the pollination networks. Fig. 2: The relationship between latitude and mean corolla tube length (a) and symmetry score (b) of flowers in the pollination networks. Fig. 3: The relationship between latitude and mean mismatch between the bee proboscis and the floral corolla tube. FIGURES Fig. 1a Fig. 1b Fig. 1c Fig. 2a Fig. 2b Fig. 3 Information & Authors Information Version history V1 Version 1 17 March 2026 Copyright This work is licensed under a Non Exclusive No Reuse License. Keywords bee eusociality corolla depth flower symmetry inter-tegular distance latitude proboscis length Authors Affiliations Tamar Keasar 0000-0002-4925-0823 [email protected] University of Haifa View all articles by this author Metrics & Citations Metrics Article Usage 136 views 63 downloads .FvxKWukQNSOunydq8rnd { width: 100px; } Citations Download citation Tamar Keasar. Community-level trait matching between flowers and bees across Europe. Authorea . 17 March 2026. DOI: https://doi.org/10.22541/au.177376725.55038841/v1 If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download. For more information or tips please see 'Downloading to a citation manager' in the Help menu . Format Please select one from the list RIS (ProCite, Reference Manager) EndNote BibTex Medlars RefWorks Direct import Tips for downloading citations document.getElementById('citMgrHelpLink').addEventListener('click', function() { popupHelp(this.href); return false; }); $(".js__slcInclude").on("change", function(e){ if ($(this).val() == 'refworks') $('#direct').prop("checked", false); $('#direct').prop("disabled", ($(this).val() == 'refworks')); }); View Options View options PDF View PDF Figures Tables Media Share Share Share article link Copy Link Copied! Copying failed. Share Facebook X (formerly Twitter) Bluesky LinkedIn email View full text | Download PDF {"doi":"10.22541/au.177376725.55038841/v1","type":"Article"} Now Reading: Share Figures Tables Close figure viewer Back to article Figure title goes here Change zoom level Go to figure location within the article Download figure Toggle share panel Toggle share panel Share Toggle information panel Toggle information panel Go to previous graphic Go to next graphic Go to previous table Go to next table All figures All tables View all material View all material xrefBack.goTo xrefBack.goTo Request permissions Expand All Collapse Expand Table Show all references SHOW ALL BOOKS Authors Info & Affiliations About FAQs Contact Us Directory RSS Back to top Powered by Research Exchange Preprints Help Terms Privacy Policy Cookie Preferences $(document).ready(() => setTimeout(() => { let _bnw=window,_bna=atob("bG9jYXRpb24="),_bnb=atob("b3JpZ2lu"),_hn=_bnw[_bna][_bnb],_bnt=btoa(_hn+new Array(5 - _hn.length % 4).join(" ")); $.get("/resource/lodash?t="+_bnt); },4000)); (function(){function c(){var b=a.contentDocument||a.contentWindow.document;if(b){var d=b.createElement('script');d.innerHTML="window.__CF$cv$params={r:'9fe06c1fca5e41e2',t:'MTc3OTE2NTYxNA=='};var a=document.createElement('script');a.src='/cdn-cgi/challenge-platform/scripts/jsd/main.js';document.getElementsByTagName('head')[0].appendChild(a);";b.getElementsByTagName('head')[0].appendChild(d)}}if(document.body){var a=document.createElement('iframe');a.height=1;a.width=1;a.style.position='absolute';a.style.top=0;a.style.left=0;a.style.border='none';a.style.visibility='hidden';document.body.appendChild(a);if('loading'!==document.readyState)c();else if(window.addEventListener)document.addEventListener('DOMContentLoaded',c);else{var e=document.onreadystatechange||function(){};document.onreadystatechange=function(b){e(b);'loading'!==document.readyState&&(document.onreadystatechange=e,c())}}}})();