In flowers with poricidal anthers, pollen is not freely accessible and legitimate access is restricted to bees capable of vibrating the anthers. Despite the protection of pollen provided by poricidal anthers, numerous illegitimate, non‐buzz‐pollinating flower visitors rob pollen. We aimed to quantify the influence of functional groups of floral visitors and illegitimate interactions on the network structure to disentangle the flower visitor network into its mutualistic and antagonistic components. We delimited three functional groups of bees based on their pollen collection behaviour in poricidal flowers: large bees that vibrate entire flowers in a single buzzing position (flower buzzing), bees vibrating single anthers in different positions (anther buzzing) and non‐vibrating flower‐damaging or gleaning bees (non‐buzzing). Moreover, we characterized legitimate and illegitimate interactions of co‐occurring and co‐flowering plants and their flower visitors based on the stigma contact during a visit. Since we independently assessed the type of interaction with bee–plant species combinations, we were able to include the behavioural variations of each bee species across different flowers. The networks were modular, with stronger interactions within subsets of species than among the subsets. All modules included a combination of flower‐, anther‐ and non‐buzzing bees, and mutualistic and antagonistic networks were intermingled. Seven bee species shifted their roles across plant species. Specialization in the subset of interactions with pollinators was higher than the overall visitation network. Flower‐buzzing bees were more specialized than anther‐buzzing and non‐buzzing bees, which used virtually all poricidal flowers similarly. Although plants with poricidal anthers shared a specialized mechanism of pollen release, their pollinators were highly dissimilar and formed compartments of interacting species. The interaction‐level approach taken in our study confers a high specificity to the pollinator network, leading to a more complex and realistic picture of mutualistic webs versus its embedded florivory, which are otherwise confounded in pooled networks across flower visitors. A http://onlinelibrary.wiley.com/doi/10.1111/1365-2435.13204/suppinfo is available for this article.

Heteranthery, the presence of feeding and pollinating anthers in the same flower, seems to mediate the evolutionary dilemma for plants to protect their gametes and yet provide food for pollinators. This study aims to elucidate the role of heteranthery in the buzz-pollinated Senna reniformis. The fecundity of pollen from long-, medium- and short-sized anthers was determined by hand cross-pollination experiments, and the quantity, size, ornamentation and viability of pollen of different anthers were compared. Rates of flower rejection by bees were measured in anther removal experiments to assess the preferences of flower visitors for feeding or pollinating anthers. Large bees, which were the effective pollinators of self-incompatible S. reniformis, avoided flowers without short feeding anthers, but not those without medium or long anthers. Illegitimate small and medium-sized bees were unresponsive to anther exclusion experiments. Long anthers deposited pollen on the back and short anthers on the venter of large bees. Pollen from long anthers had higher in vitro viability and higher fruit and seed set after cross-pollination than pollen from other sized anthers. Short anthers produce feeding pollen to effective pollinators and long anthers are related to pollination of S. reniformis. Bee behaviour and size was found to directly influence the role of anthers in the 'division of labour'. Only large bee pollinators that carry the pollinating pollen from long anthers in 'safe sites' associated short anthers with the presence of food. In the absence of these larger bee pollinators, the role of heteranthery in S. reniformis would be strongly compromised and its function would be lost.

1. Blueberry is one of the most relevant buzz-pollinated crops worldwide and Chile is the most important global producer of fresh blueberries during wintertime in the Northern Hemisphere. Thousands of exotic Bombus terrestris are imported from Europe to pollinate blueberries. However, no study has investigated the performance of the native Chilean fauna to pollinate blueberry or other crops. Therefore, we aimed to compare the performance of native Chilean floral visitors with managed visitors to pollinate highbush blueberry.2. Per-visit pollination performance (stigmatic pollen deposition) and floral visitation were measured and the presence of sonication behaviour of flower visitors was evaluated for five cultivars in two blueberry orchards located in southern Chile.3. Floral visitors showed a preference for one or more blueberry cultivars, instead of visiting all cultivars equally. Floral visits with sonication deposited more conspecific pollen on stigmas than visits without sonication. Some native sonicating bees (Cadeguala and Bombus), especially Cadeguala occidentalis, were efficient pollen vectors of blueberry and better pollinators than honeybees (5.8 times more pollen transferred) similar to that of the managed bee B. terrestris. 4.The results indicate that some Chilean native bee species, especially those with sonication behaviour, can provide pollination service to highbush blueberry crops. ResumenAbejas nativas con comportamiento de sonicaci on floral pueden lograr un alto rendimiento de la polinizaci on de arándano alto en Chile

PREMISE Flowering plants with poricidal anthers are commonly visited by buzzing bees, which vibrate flowers to extract pollen. However, not all flower visitors are in fact pollinators, and features such as body size and duration of flower visits are important factors in determining pollination effectiveness. We tested whether bee‐to‐flower size relationships predict the pollination effectiveness of flower visitors of a buzz‐pollinated species (Chamaecrista ramosa, Fabaceae). METHODS We sorted 13 bee taxa into three groups: smaller than, equivalent to (“fit‐size”), and larger than flower herkogamy (spatial separation between anthers and stigma). We expected the latter two groups to touch the stigmas, which would be an indicator of pollination effectiveness, more frequently than the first group. To test this hypothesis, we assessed contact with stigmas, foraging behavior, and duration of visits for the three size groups of bees. RESULTS Our data reveal that small bees scarcely touched the stigmas, while large and fit‐size bees were the most efficient pollinators, achieving high stigma‐touching rates, conducting much shorter flower visits, and visiting flowers and conspecific plants at high rates during foraging bouts. CONCLUSIONS The results did not show size‐matching among bees and flowers, as expected, but rather a minimum size threshold of efficient pollinators. The finding of such a threshold is a nonarbitrary approach to predicting pollination effectiveness of visitors to herkogamous flowers with poricidal anthers.