Pollination of Blumenbachia amana (Loasaceae): flower morphology and partitioned pollen presentation guarantee a private reward to a specialist pollinator

Pollination of Blumenbachia amana (Loasaceae): flower morphology and partitioned pollen... Abstract Blumenbachia amana possesses the characteristic ‘tilt-revolver flowers’ and the pollen partitioning mechanism found exclusively in Loasoideae. Stamens are sheltered in naviculate petals that alternate with five nectar scales, each one associated with two free staminodes. In the staminate phase, stamens move successively towards the centre of the flower and present pollen. Bees of oligolectic Actenosigynes mantiqueirensis (Colletidae) are the unique effective pollinators. Clinching the nectar scale–staminode complex, they bend the scales outward one by one with their head to take up nectar, thereby triggering the movement of a stamen, which arrives in the centre of the flower ~3 min after the stimulus. Females returned to previously visited flowers after short time intervals by establishing pollen foraging routes in flower patches that were maintained for days. In flowers experimentally hand-stimulated in the nectar scales to simulate pollinator visits, twice as many stamens moved, and the longevity of bee-visited flowers decreased by shortening the staminate phase. Partitioned pollen presentation in B. amana leads to a unique pollen foraging behaviour characterized by absolute flower fidelity by this bee in this specialized plant–pollinator relationship. foraging behaviour, Loasaceae, oligolectic bee, tilt-revolver flowers INTRODUCTION Pollen presentation in flowering plants occurs between two extremes: simultaneously, when all anthers of simultaneously opening flowers open, or gradually, when flowers open during the course of the day and anthers open one by one during anthesis (Percival, 1955). The latter occurs by either packaging or dispensing mechanisms (Harder & Thomson, 1989), which are considered to have the function of making pollen transport to stigmas via pollinators more efficient by limiting the amount of pollen that can be removed during a single visit, forcing pollen foragers to increase their movement among flowers (Lloyd & Yates, 1982; Harder & Thomson, 1989; Harder & Wilson, 1994). Accentuated gradual pollen release within single flowers occurs in the subfamily Loasoideae (Loasaceae), where flowers are protandrous, and stamens ripen successively one by one. During the staminate phase, stamens initially hidden within naviculate petals move successively towards the centre of the flower after anther dehiscence. In addition to fertile antepetal stamens, the androecium contains a concave nectar scale in front of each sepal, which is interiorly opposed by two free staminodes (Urban, 1886, 1892; Brown & Kaul, 1981). Weigend & Gottschling (2006) denominated this melittophilous type of Loasoideae flower as a ‘tilt-revolver flower’, because the nectar scale–staminode complex forms a circle with five units, which must be probed separately by nectar-seeking bees. Oligolectic bees, which feed their larvae pollen of the same plant genus or family (Robertson, 1925), were the sole pollinators of Caiophora arechavaletae (Urb.) Urb. & Gilg (Schlindwein & Wittmann, 1997a) and Aosa rupestris (Gardner) Weigend (Leite, Nadia & Machado, 2016). While searching for nectar in flowers of C. arechavaletae, females of Bicolletes pampeana (Colletidae) elicit stamen movements by pressing the nectar scales outwards with their head. They establish micro-foraging routes and return to previously visited flowers in short intervals, similar to the time delay between stimulation and arrival of the stamen in the centre of the flower, to collect the presented pollen package (Schlindwein & Wittmann, 1997a). Delayed pollen presentation in response to mechanical stimulation of the nectar scale–staminode complex was later demonstrated in greenhouse-cultivated plants of numerous representative species of the genera Caiophora, Loasa, Nasa and Presliophytum, all in the subfamily Loasoideae (Henning & Weigend, 2012, 2013). We studied pollination of the recently described loasoid Blumenbachia amana Henning & Weigend, endemic to the Serra da Mantiqueira mountain range in southeastern Brazil, and its relationship with Actenosigynes mantiqueirensis Silveira (Colletidae). We focused on the following issues: What are the characteristics of pollen presentation in B. amana? Do flower visitors influence the pollen release pattern? Does the pollen presentation regime shape the foraging behaviour of the effective pollinator? To answer these questions, we characterized the floral biology and pollination of B. amana and analysed, in the field, the flower-handling behaviour and foraging flights of its bee pollinators. MATERIAL AND METHODS Study Area The study was conducted from November 2013 to February 2015 in the municipalities of Camanducaia and Gonçalves (22°41.275′S, 45°54.048′W and 22°42.695′S, 45°57.082′W), situated in the Serra da Mantiqueira mountain range, Minas Gerais, Brazil. The study sites are located between 1500 and 1800 m elevation. The climate is humid throughout the year, with mean annual rainfall ranging from 1600 to 1800 mm, and mean annual temperatures ranging between 14 and 19 °C (DER/MG, 1998). The surrounding vegetation is dominated by mixed Araucaria forest (Araucaria angustifolia) within the high-altitude domain of the Atlantic Forest. Studied Species The South American genus Blumenbachia Schrad. (Loasaceae) comprises 11 species, six of which occur in Brazil (Weigend, 1997; Henning et al., 2015). The recently described B. amana (Henning et al., 2015) is an ascending annual herb that occurs on humid organic soils along streams and in forest clearings, and in humid depressions in trample-protected pastures for cattle and vegetable gardening areas. Thus far, the species is known only from the northern slope of the Mantiqueira mountain range in the two municipalities of Camanducaia and Gonçalves. Like other species of Loasoideae (Urban, 1886, 1892; Schlindwein & Wittmann, 1995; Henning & Weigend, 2012, 2013; Weigend, Ackermann & Henning, 2010), B. amana shows a highly differentiated androecium composed of five antesepalous nectar scales internally opposed by two free staminodes with filiform apices. Five antepetalous fascicles of fertile stamens alternate with the nectar scale–staminode complex (Fig. 1). The nectar scales are concave, formed by three connate staminodes, and show three basal, abaxial, filiform appendices. The scale apex is recurved, brightly red merging into yellow and green at the base, and white laterally and near the apex. The numerous stamens are initially hidden within naviculate petals and move towards the flower centre throughout the staminate phase (Fig. 2A–C). Specimens of the studied species are deposited in Herbarium BHCB (holotype BHCB 162877, Belo Horizonte) and Herbarium Berolinense (isotypes B 10 0610559, Bonn). Figure 1. View largeDownload slide Floral diagram of Blumenbachia amana. fs, free staminodes; ns, nectar scale; o, ovary; p, petal; s, sepal; st, stamens. Figure 1. View largeDownload slide Floral diagram of Blumenbachia amana. fs, free staminodes; ns, nectar scale; o, ovary; p, petal; s, sepal; st, stamens. Figure 2. View largeDownload slide Flower of Blumenbachia amana. A, flower in the staminate phase; a stamen moves from the petal downwards into the centre of the flower. B, staminodial complex: 1, nectar scale; 2, free staminodes. C, longitudinal section through a flower in the staminate phase: 1, stamen bundle inside a petal; 2, immature style. D, flower in the pistillate phase: arrow, mature style. Scale bars represent 1 cm unless indicated otherwise. Figure 2. View largeDownload slide Flower of Blumenbachia amana. A, flower in the staminate phase; a stamen moves from the petal downwards into the centre of the flower. B, staminodial complex: 1, nectar scale; 2, free staminodes. C, longitudinal section through a flower in the staminate phase: 1, stamen bundle inside a petal; 2, immature style. D, flower in the pistillate phase: arrow, mature style. Scale bars represent 1 cm unless indicated otherwise. Floral Morphology and Longevity Given that flower size increases during anthesis, we measured variation in the size of petals, stamens and styles throughout the staminate and pistillate phases of anthesis. Forty-five flowers of 20 individual plants in pre-anthesis were marked in the field and divided into three sets of 15 flowers from different individuals. During three subsequent days, we collected 15 flowers per day and measured the length of the petals, stamens and styles (N = 45). The mean number of stamens per flower was obtained from 35 new flowers. We determined floral longevity and the duration of the staminate and pistillate phases. For flower longevity, we considered the number of daylight hours during which each flower remained open between 09.00 and 18.00 h. We compared flower longevity between non-visited bagged flowers (N = 20) and bee-visited flowers (N = 20). We considered the staminate phase to be the period from the beginning of flower opening until all the stamens had moved. The beginning of the pistillate phase was defined as the moment the stigma became visible in the centre of the flower and when all stamens had already moved (see Weigend et al., 2010; Henning & Weigend, 2012). From this moment onwards until flower senescence, flower visitors could contact the stigma. Stamen Movement We investigated stamen movement in terms of the duration of migration and the number of migrated stamens per hour. To evaluate whether flower visitors could induce stamen movements while taking up nectar, we experimentally applied mechanical stimuli to the nectar scales to simulate flower visits by bees. Stimuli consisted of the application of slight outward pressure on each of the five nectar scales using a toothpick. We then counted the number of migrated stamens and measured the time between stimuli and arrival of the stamens in the centre of the flower (N = 52 flowers). For 26 flowers, we applied stimuli, as described above, every 5 min for 1 h. This stimulation interval was similar to the mean interval of flower visits to B. amana. Another 26 flowers were bagged, to prevent access of bee visitors and to prevent stimulation of the flowers, as controls. In this paired experiment, all flowers were in the staminate phase. To certify that the flowers received the same level of stimulation, we stimulated all nectar scales of all flowers of the two treatments 30 min before the experiment. To examine whether stamens move overnight, at 18.00 h we gently removed the anthers of the already migrated stamens of 15 marked flowers. The next day, at 09.00 h, we inspected these flowers to count the stamens that had moved. Breeding System To determine whether flowers set fruits by spontaneous self-pollination, we compared the number of fruits and seeds produced in 20 flowers bagged throughout anthesis with 20 non-bagged control flowers that were visited by bees. Flower Visitors and Visiting Frequency Flower visitors of B. amana were collected with entomological nets and mounted with entomological pins. The specimens were identified and deposited in the Entomological Collection of UFMG (Centro de Coleções Taxonômicas da UFMG, Universidade Federal de Minas Gerais, Belo Horizonte, Brazil). Flower visiting frequency was determined throughout the day for four marked flowers per 30 min count in four 2 h time intervals (09.00–11.00, 11.00–13.00, 13.00–15.00 and 15.00–17.00 h; N = 158 flowers). We also compared the frequency of bee visits to flowers in the staminate and pistillate phases (N = 14 flowers per phase). Foraging Flights of Females of A. Mantiqueirensis Females of A. mantiqueirensis were captured with entomological nets and marked with individual colour codes on the mesoscutum and/or metasomal tergites using Revell ink (Revell, Germany). When the ink had dried, ~2 min after marking, the bees were liberated. This procedure did not exert any notable influence on the foraging behaviour of the marked females. All open flowers in the studied populations were numbered and the foraging flights of marked females accompanied. We measured the duration of partial and complete foraging flights of the individually marked bees. A complete flight included all flower visits of a marked female, from when she appeared in a foraging patch with empty scopae to when she left with pollen-filled scopae. When a marked female reappeared with empty scopae, we considered this interval as the time she needed to fly back to the nest, remove the pollen load and return to the foraging patch. During each foraging flight, we recorded duration, the total number of visits, the number of visits and revisits to individual flowers, and the duration of the intervals between revisits. For each floral visit, we recorded the number of nectar scales probed and whether pollen was collected. The mean flower-handling time per sequence of visits was calculated by dividing the duration of consecutive flower visits by the number of flowers visited by a marked female. Thus, the calculated handling time includes the duration of a flower visit plus the flight time to the next flower. The total time of observations of visiting sequences of individually marked females was ~22 h. To determine whether bees of A. mantiqueirensis visit other plants besides B. amana, samplings of floral visitors of other plants that occur near the foraging patches of B. amana were performed throughout the study. Statistical Analyses For all analyses, we first tested data for normality and homogeneity of variance using the Kolmogorov–Smirnov test and the Levene median test, respectively. We used a t-test to compare the following: (1) flower longevity and duration of staminate and pistillate phases of non-visited bagged flowers with bee-visited flowers; (2) number of moved stamens in experimentally stimulated flowers with non-stimulated flowers (data square root transformed); (3) seed set of visited flowers with non-visited flowers; and (4) frequency of visits by A. mantiqueirensis to flowers in the staminate and pistillate phases of anthesis. A non-parametric Kruskal–Wallis test was performed to compare the frequency of flower visits among the four established intervals throughout the day. All statistical analyses were conducted using SigmaStat 3.5 (Systat Software, Inc.) for Windows. RESULTS Floral Morphology and Anthesis The beginning of anthesis was not synchronized, so the flowers of an individual of B. amana opened at different times throughout the day between 09.00 and 18.00 h. New flowers were small, with flower size increasing during anthesis. Petals showed continuous growth throughout anthesis, stamens only until the second day, and the style doubled in size from the second to the third day, reaching about the length of the stamens at the end of anthesis (Fig. 3). Each of the five antepetalous fascicles of fertile stamens had on average 9.5 ± 1.7 stamens (N = 35 flowers) that had moved one by one during the staminate phase. After all stamens had moved, the pistillate phase commenced, and the style elongated and became visible in the centre of the flower (Fig. 2D). Figure 3. View largeDownload slide Mean length of petals, stamens and styles of Blumenbachia amana through anthesis (N = 45 flowers). Eorr bars indicate the standard deviation. Figure 3. View largeDownload slide Mean length of petals, stamens and styles of Blumenbachia amana through anthesis (N = 45 flowers). Eorr bars indicate the standard deviation. Stamen Movement Pollen partitioning in B. amana occurred by successive dehiscence of the anthers, accompanied by successive movements of the stamens. Stamen movement could be triggered by pushing the nectar scales slightly outwards, thus showing thigmonasty. After experimental stimuli, the stamens reached the centre of the flower after an average of 3.3 ± 1.1 min (N = 26). The comparison of flowers experimentally hand-stimulated in the nectar scales every 5 min for 1 h vs. non-stimulated bagged flowers showed that about twice as many stamens moved in hand-stimulated flowers, 6.3 ± 4.2 (N = 26), compared with the non-stimulated flowers, 3.2 ± 2.1 (N = 26) (t = 3.373; P = 0.001; N = 52; Fig. 4). Figure 4. View largeDownload slide Number of stamens moved per hour. Non-stimulated flowers (N = 26) and hand-stimulated flowers (N = 26). Values are the means ± SD. Different letters represent significant differences between means. Figure 4. View largeDownload slide Number of stamens moved per hour. Non-stimulated flowers (N = 26) and hand-stimulated flowers (N = 26). Values are the means ± SD. Different letters represent significant differences between means. In the 1 h stimulation experiments during the day, stamen movement occurred for 70% of the stimuli made. Stamen movements were almost completely restricted to daytime, with 15 marked flowers having only one stamen move at night (between 18.00 and 9.00 h). Pollen Presentation and Flower Longevity Non-visited flowers remained open for a longer time, on average 26.4 ± 4.8 h (N = 20), than flowers visited by bees, 21.0 ± 3.1 h (N = 20) (t = 4.285; P ≤ 0.001; N = 40). This difference was attributable to a 25% longer duration of the staminate phase in non-visited flowers. The duration of the staminate phase of visited flowers was on average 16.8 ± 3.3 h (N = 20) and of non-visited flowers 22.3 ± 3.9 h (N = 20) (t = 4.887; P ≤ 0.001; N = 40). There was no difference in the duration of the pistillate phase between visited and non-visited flowers, 4.2 ± 1.5 (N = 20) and 4.2 ± 2.3 h (N = 20), respectively (t = −0.059; P = 0.953; N = 40; Fig. 5). Figure 5. View largeDownload slide Flower longevity in Blumenbachia amana. Duration of staminate and pistillate phases of non-visited and bee-visited flowers. Non-visited flowers: staminate and pistillate phase (N = 20). Bee-visited flowers: staminate and pistillate phase (N = 20). Values are the means ± SD. Only the daylight hours of open flowers were considered. Different letters represent significant differences between means. Figure 5. View largeDownload slide Flower longevity in Blumenbachia amana. Duration of staminate and pistillate phases of non-visited and bee-visited flowers. Non-visited flowers: staminate and pistillate phase (N = 20). Bee-visited flowers: staminate and pistillate phase (N = 20). Values are the means ± SD. Only the daylight hours of open flowers were considered. Different letters represent significant differences between means. Breeding System The flowers of B. amana were self-compatible. All marked flowers set fruits, and there was no difference in seed set between non-visited flowers (spontaneous selfing), with 35.0 ± 11.5 seeds (N = 20), and flowers visited by pollinators, with 35.5 ± 14.0 seeds (N = 20) (t = −0.111; P = 0.912; N = 40). Flower Visitors and Visitation Frequency Females and males of A. mantiqueirensis were the almost exclusive flower visitors of B. amana, and no individual of this species was sampled on flowers of any other plant species in the vegetation surrounding B. amana plants. During a total of 70 days of observation (~500 h) of flowers, we recorded only sporadic visits of Plebeia saiqui (Friese, 1900), Plebeia droryana (Friese, 1900), Plebeia lucii Moure, 2004, Apis mellifera Linnaeus, 1758 (all Apidae), and Augochloropsis sp. (Halictidae). Bees of these species did not visit flowers of B. amana during the counts of flower visitation frequency. Given that these species occurred only sporadically (22 visits in 70 days of observation), did not visit flowers in the pistillate phase and also did not handle the flowers properly (bees of Apis and Augochloropsis slipped off the flower centre without gaining access to nectar, and those of Plebeia gleaned pollen from the petals), they were not considered in the present study. Bees of A. mantiqueirensis visited the flowers throughout the entire flowering period of B. amana (from November to February). There was no difference in the frequency of visits among the four established intervals throughout the day. We recorded on average 4.1–5.7 flower visits per 30 min (Kruskal–Wallis = 1.706; P = 0.636; N = 158). Females and males of A. mantiqueirensis also visited flowers in the staminate phase with the same frequency, 7.4 ± 4.0 visits (N = 14), as those in the pistillate phase, 7.9 ± 2.7 visits (N = 14) (t = −0.383; P = 0.705; N = 28). Flower Handling by A. Mantiqueirensis Flower visits by females of A. mantiqueirensis showed the following standard behavioural sequence. The bees approached the pendulous flowers bottom-up, with the head oriented towards the centre of the flower. Still in flight, the bees stretched their front legs towards the nectar scales and clinched the revolute apexes of the scales first with the hooked tarsal claws of these legs, then with the tarsal claws of the mid- and hindlegs (Fig. 6A, B). Figure 6. View largeDownload slide Handling of flowers of Blumenbachia amana by females of Actenosigynes mantiqueirensis. A, in search of nectar in a flower in the pistillate phase, a female inserts her head between the nectar scale and staminodes, pushing the scale outward, and contacts the long stigma with her ventral scopa filled with pollen (dotted arrow; continuous arrow see 6B). B, tarsal claws of fore-, mid- and hindlegs of a female A. mantiqueirensis clinching to the revolute apexes of two nectar scales of B. amana during nectar uptake (arrows indicate claws). C, female collecting pollen from a moved stamen in the centre of the flower, and another stamen moving to the centre. D, individually marked female pulling a stamen that has not moved with her mandibles. Scale bars represent 1 cm unless indicated otherwise. Figure 6. View largeDownload slide Handling of flowers of Blumenbachia amana by females of Actenosigynes mantiqueirensis. A, in search of nectar in a flower in the pistillate phase, a female inserts her head between the nectar scale and staminodes, pushing the scale outward, and contacts the long stigma with her ventral scopa filled with pollen (dotted arrow; continuous arrow see 6B). B, tarsal claws of fore-, mid- and hindlegs of a female A. mantiqueirensis clinching to the revolute apexes of two nectar scales of B. amana during nectar uptake (arrows indicate claws). C, female collecting pollen from a moved stamen in the centre of the flower, and another stamen moving to the centre. D, individually marked female pulling a stamen that has not moved with her mandibles. Scale bars represent 1 cm unless indicated otherwise. Immediately after landing, the bees looked for nectar in 99.8% (4894 visits) of the flower visits (N = 4905 total of visits). To do so, they inserted their head between the nectar scale apex and the two free staminodes interiorly opposed to it, pushing the scale outward in the process (Fig. 6A). The revolute collar-shaped apexes of the nectar scales provided a rigid supporting point to mortise the bees’ body forward until their short mouthparts reached the base of the scales. In each search for nectar in a nectar scale, the bees clinched the revolute apexes of the two neighbouring nectar scales with the tarsal claws of front and mid-legs, while with their hindlegs they grasped the opposite pair of scales. After nectar uptake in the scale, the bees turned to the neighbouring scale and repeated nectar uptake in the same relative position. While the bee removed its head from the nectar scale, the scale elastically bent back to the original position (Supporting Information, Video S1). In 47% (2281 visits) of the flower visits, the bees searched for nectar in all the five scales, pushing them in sequence in clockwise or anticlockwise rotation. Thereby, the bees always used the apical collar of the nectar scales as a foothold. Females and males showed no differences in nectar-searching behaviour, and males always left the flower immediately after nectar uptake. During nectar uptake in flowers in the pistillate phase, females and males continuously contacted the long stigma with the ventral face of their mesosoma and metasoma and thus transferred allochthonous pollen to the stigmatic surface. In females, the pollen-loaded ventral scopa was continuously in contact with the stigma in these flowers (Fig. 6A). After pushing the nectar scales, female bees actively collected pollen in 41.4% of the flower visits (2029 visits), showing two pollen-collection behaviours. The first behaviour is termed pollen brushing. Females brushed the pollen grains exposed on the anther surface of moved or moving stamens with scopal hairs of the metasoma and hindlegs. This behaviour was shown in 56.2% (1141 visits) of pollen collections in a flower. The second behaviour is termed stamen pulling. Females sought for non-moved stamens still hidden in the naviculate petals, moved to the petals, grasped a filament with the mandibles near its basal third and pulled the stamens to the flower centre (Fig. 6C, D). Using forelegs and mandibles, the females manipulated the filaments and advanced with the head in the direction to the anthers. For stamens with already dehisced anthers, females used pollen brushing to collect pollen. For pulled stamens with still closed anthers, the anther was torn with the mandibles and the exposed pollen grains were then transferred from the mandibles to the hindlegs using foreleg tarsal hairs. Next, the females directed the anther to the ventral scopa and rubbed it to remove more pollen grains. This behaviour lasted longer and was performed in 43.8% (888 visits) of pollen-collection visits. Foraging by A. Mantiqueirensis Monitoring of individually marked females of A. mantiqueirensis showed that they established foraging areas with limited numbers of B. amana flowers. These areas were maintained for up to 20 consecutive days. We recorded 46 foraging flights of 13 individually marked females. We accompanied 12 full foraging flights of five females, i.e. flights that included all flower visits from the moment a female appeared with empty scopae in an observation area until it left with scopae filled with pollen. These females reappeared in the observation patches with empty scopae after 12.5–24.1 min (15.9 ± 3.3 min; N = 12). The complete foraging flights lasted between 21.5 and 106.0 min (55.0 ± 26.7 min; N = 12). The individual foraging areas included 41–85 flowers (62.5 ± 13.4; N = 12), and the number of total flower visits per flight in a foraging area ranged from 93 to 449 (244.6 ± 112.8; N = 12). The mean handling time per flight sequence (flower handling plus duration of flight between flowers) of the different females varied between 9.0 and 23.0 s (14.1 ± 3.0 s; N = 46). During all the recorded flights, females repeatedly visited the same flowers (N = 3113 revisits). Revisit intervals to a flower were mostly short, and in 52.0% (1534 revisits) the females returned to the same flower within the first 6 min after a previous floral visit. Most revisit intervals were between 2 and 5 min (916 revisits) (Fig. 7). Figure 7. View largeDownload slide Revisit intervals for individually marked females of Actenosigynes mantiqueirensis during foraging flights to marked flowers of Blumenbachia amana (N = 3113 revisits). Revisit intervals between 41 and 97 min (0.83%) are not represented in the graph. Figure 7. View largeDownload slide Revisit intervals for individually marked females of Actenosigynes mantiqueirensis during foraging flights to marked flowers of Blumenbachia amana (N = 3113 revisits). Revisit intervals between 41 and 97 min (0.83%) are not represented in the graph. DISCUSSION The present study demonstrates a highly specialized interaction between B. amana and A. mantiqueirensis, its sole colletid bee pollinator. Both females and males of this bee species visit flowers in the staminate and pistillate phases with the same high frequency, deposit pollen on the stigma and continuously fly among flowers of conspecific individuals, showing absolute flower fidelity. These bees are, thus, highly effective pollinators of B. amana, guaranteeing the cross-flow of pollen. Our observations indicate that A. mantiqueirensis is narrowly oligolectic (in the sense of Robertson, 1925; Cane & Sipes, 2006): (1) the foraging behaviour of females is linked to the mechanism of continuous pollen presentation by the flowers; (2) whole-day monitoring of marked individuals showed that foraging is interrupted only when females need to fly back to their nest to unload pollen grains from their scopae, after which they return to their B. amana flower patches; and (3) the pollen loads of females contained pollen of B. amana exclusively, and the bees were never observed to visit flowers of other species. In the subfamily Neopasiphaeinae Cockerell, 1930 (sensuAlmeida & Danforth, 2009; Almeida et al., 2012; which were formerly included in Paracolletinae Cockerell, 1934 and Paracolletini Michener, 1989), pollen specialization seems to be common for Neotropical species (Schlindwein & Wittmann, 1995; Melo, 1996; Schlindwein & Wittmann, 1997a, b; Schlindwein, 1998, 2004; Gimenes, 2002; Michener, 2007; Roig-Alsina & Schlumpberger, 2008; Carvalho & Schlindwein, 2011). The continuous release of pollen, and probably nectar, in tiny portions throughout the day induces high frequencies of visitation to flowers by specialist visitors and makes the flowers of B. amana unattractive for polylectic bees, such as the highly eusocial species that efficiently recruit workers to collect synchronously offered resources in mass flowering plants (von Frisch, 1967; Roubik, 1989). Considering an almost constant rate of, on average, one visit per 5 min during ~8 h per day, a flower might receive 80–96 flower visits a day and ~250 visits during its lifetime by bees of A. mantiqueirensis. Given that a bee inserts its mouthparts in nectar scales on average 3.9 times per flower visit, each flower might receive ~1000 insertions of head and mouthparts during its lifetime. In the presence of the steadily visiting oligolectic pollinators, the flowers of B. amana are permanently empty of pollen and nectar, discouraging polylectic species from visiting their flowers. For the specialized A. mantiqueirensis, the positive consequence of its specialization is to guarantee a private resource, whereas the negative is its obvious inability to explore alternative host plants when flowers of B. amana are absent. The floral morphology of B. amana, and that of other representatives of melittophilous Loasoideae, seems to discourage visits of non-specialist bees. To access the pendulous flowers, bees must land ‘upside down’ and fix themselves with their tarsal claws at the revolute apical margin of the nectar scales. Only this exact position ensures support for pushing the nectar scales with the head to access nectar. Weigend & Gottschling (2006) previously considered this ‘recurved scale neck’, in addition to dorsal calli on the nectar scales, as morphological adaptations of melittophilous representatives of the genus Nasa for providing such a foothold for pollinating bees landing on the pendulous flowers. Only bees of A. mantiqueirensis handled the B. amana flowers adequately and used the revolute foothold to advance with the head and push the scales outward to access the nectar. The revolute nectar scale apex is therefore important for providing a rigid access point to obtain nectar and mortise the bees’ body in a position that guarantees contact with mature anthers and/or receptive stigmas. This seems to be a consistent feature in tilt-revolver flowers (Schlindwein & Wittmann, 1997a; Ackermann & Weigend, 2006; Weigend & Gottschling, 2006; Weigend et al., 2010; Leite et al., 2016; Strelin et al., 2016). Pollinators Influence the Pollen Release Pattern Pollen presentation in B. amana occurs in a manner similar to that demonstrated in other representatives of the subfamily Loasoideae. The anthers mature and open successively, followed by movement of the stamens (Urban, 1886, 1892; Brown & Kaul, 1981). When nectar scales are stimulated, the movement of stamens can be triggered, thereby showing thigmonasty (Schlindwein & Wittmann, 1997a; Weigend et al., 2010; Henning & Weigend, 2012, 2013). Phylogenetic analyses showed that thigmonasty is a synapomorphy of the ‘higher Loaseae’ within Loasoideae, the lineage to which B. amana belongs (see cladogram of Weigend et al., 2004). This species group includes both melittophilous tilt-revolver flowers and ornithophilous funnel-revolver flowers, with the former being the plesiomorphic condition (Weigend et al., 2004; Ackermann & Weigend, 2006; Weigend & Gottschling, 2006; Strelin et al., 2016). In particular, the Brazilian Loasoideae comprise a group of lowland species with rather uniform melittophilous flower morphology from the three genera Aosa, Blumenbachia and Caiophora, which seem to be pollinated mainly by solitary pollen-seeking bees. But thigmonasty probably did not evolve as an adaptation to specialized pollen foraging in colletid bees, as previously thought (Schlindwein & Wittmann, 1997a). In species where nectar rather than pollen is the central resource for pollinators, thigmonasty associated with protandry also strongly favours cross-pollination. This is the case in associations of representatives of ‘higher Loaseae’ in the Andes with floral nectar-seeking bees, such as the polylectic carpenter bee Neoxylocopa lachea Moure, 1951, which was identified as the effective pollinator of Nasa macrothyrsa (Loasoideae) (Weigend et al., 2010), and with hummingbirds and even rodents (Harter, Schlindwein & Wittmann, 1995; Cocucci & Sérsic, 1998; Ackermann & Weigend, 2006; Weigend & Gottschling, 2006). Compartmentalized pollen presentation in Loasoideae in the form of defined amounts of pollen was interpreted by Weigend et al. (2010) as a mechanism of tilt-revolver flowers that increases male fitness. The authors explained the complex morphology of Loasoideae flowers with the pollen presentation theory (PPT), which postulates that male fitness is favoured when the frequency of pollen release is adjusted to the activity patterns of pollinators (Lloyd & Yates, 1982; Harder & Thomson, 1989; Harder & Wilson, 1994). This is exactly what we observed as characteristic of the staminate phase of B. amana. We showed experimentally that continuous stimulation of nectar scales accelerates pollen donation. The flower-bagging experiment confirmed shortened flower longevity in flowers accessible to pollinators, but exclusively by abbreviating the staminate phase of flowering by ~25%. The already much shorter pistillate phase, important for pollen reception, is maintained without alteration. The flowers of B. amana, thus, show a complex mechanism of control of pollen release in the daytime hours, when the pollinators are active, but not during the night. Influence of Pollen Presentation Regime on Foraging Behaviour of the Effective Pollinator The possibility that females of A. mantiqueirensis trigger the release of a small amount of pollen through the movement of a stamen makes the moment of pollen presentation predictable to a few minutes after stimulation. We interpret the adoption of foraging routes of a restricted number of flowers as an adapted foraging behaviour for the return to previously visited flowers in intervals corresponding to the time between a stimulus and the arrival of the stamens in the centre of the flowers. However, a behavioural response of optimized collection of pollen depends on two unpredictable characteristics. First, not all stimuli applied to nectar scales result in a response of stamen movement. The present study found that ~30% of the stimuli in the staminate phase did not cause stamen movement. Second, other individual bees that forage in the same flower patch also stimulate nectar scales, harvest available pollen and, thus, interfere with the trigger/harvest mechanism. The interference of conspecific competitors should become greater when flowers of B. amana become scarce. Similar systems with narrowly oligolectic Neopasiphaeinae bees as highly effective unique pollinators of species of Loasoideae are known, such as for C. arechavaletae and Blumenbachia insignis Schrad. with Bicolletes pampeana (Wittmann & Schlindwein, 1995; Schlindwein & Wittmann, 1997a; Schlindwein, 2000); for Aosa rupestris (Gardner) Weigend with Bicolletes nordestina Urban (Leite et al., 2016); and probably also for Blumenbachia eichleri Urb. (cited as Caiophora eichleri) and Blumenbachia catharinensis Urb & Gilg with Actenosigynes fulvoniger (cited as Leioproctus fulvoniger) (Harter et al., 1995; Schlindwein, 2000). Caiophora arechavaletae and its oligolectic bee pollinator B. pampeana show a similar relationship to that found in the present study (Schlindwein & Wittmann, 1997a). In both cases, the bees continuously stimulate the nectar scales, triggering stamen movement, with the stamens arriving in the centre of the flower on average 2.4 min (C.arechavaletae) and 3.3 min (B. amana) after triggering. Females of both species adopt stable foraging areas for several days and return to previously visited flowers in short intervals, always searching for nectar and removing pollen from pollen-presenting anthers. Females of A. mantiqueirensis, however, include more flowers in their foraging patches, have longer handling times per flower, and their complete foraging flights last longer than those of B. pampeana. The two different pollen-collection behaviours of A. mantiqueirensis, however, seem to be the most striking difference from B. pampeana, which exhibits pollen brushing exclusively. Pollen harvest by stamen pulling from non-moved stamens might be advantageous for females of A. mantiqueirensis because they gain access to the pollen content of hidden anthers. A pollen-searching female, thus, would increase her competitive vigour over conspecifics because she ensures the collection of pollen before conspecifics do, who rely on ordinary pollen presentation of moved stamens. Nevertheless, females are not able to determine during a flower visit whether non-moved stamens show dehisced anthers, nor whether the flowers are in the staminate or pistillate phase. We frequently observed that females pull stamens with closed anthers that require more time for pollen grain removal or whose pollen is still not available. Given that the autogamous B. amana exhibits high fruit and seed set even in the absence of pollinators, as has been demonstrated for other annual Loasaceae (Schlindwein & Wittmann, 1997a; Henning & Weigend, 2013; Leite et al., 2016), only the narrowly oligolectic bee pollinator manifests reproductive dependence in this tight interaction. Nevertheless, the relationship between both partners is highly mutual, because in all the studied cases (Wittmann & Schlindwein, 1995; Schlindwein & Wittmann, 1997a; Leite et al., 2016), the sole colletid bee pollinators guarantee the important cross-pollen flow during their hundreds of uninterrupted flower visits to their specific host plants. It can be expected that A. mantiqueirensis bees would not survive in the absence of their host plant. Both interacting endemic species were described recently are known only from specific habitats within high-altitude remnants of the threatened Atlantic Forest, and are therefore at a high risk of extinction. SUPPORTING INFORMATION Additional Supporting Information may be found in the online version of this article at the publisher's web-site. Video S1. Nectar uptake. A female of Actenosigynes mantiqueirensis during nectar uptake in a flower of Blumenbachia amana. Clinching the nectar sacale-staminode complex, the bee bents the scales outwards one by one with it’s head. ACKNOWLEDGEMENTS We thank José Neiva, Leandro Sales, Nathália Falagan, Fernanda Figueiredo, Ana Luiza Cordeiro and Isabelle Cerceau for their help with fieldwork, and the residents of the villages of Cantagalo (Gonçalves) and Monte Azul (Camanducaia), especially Senhora Eurides, Suely, José Maria, Elizabeth, Pedro, Rosa, Maria and Altair, for logistic support. We thank three anonymous reviewers for their constructive comments, which improved the manuscript. We thank ICMBio for the collection license, Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq Universal – 481605-2011-8) for financial support and individual research grants from Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) to S.S.-O. and R.O., and from CNPq to C.S. REFERENCES Ackermann M, Weigend M. 2006. Nectar, floral morphology and pollination syndrome in Loasaceae subfam. Loasoideae (Cornales). Annals of Botany  98: 503– 514. Google Scholar CrossRef Search ADS PubMed  Almeida EA, Danforth BN. 2009. Phylogeny of colletid bees (Hymenoptera: Colletidae) inferred from four nuclear genes. Molecular Phylogenetics and Evolution  50: 290– 309. Google Scholar CrossRef Search ADS PubMed  Almeida EAB, Pie MR, Brady SG, Danforth BN. 2012. Biogeography and diversification of colletid bees (Hymenoptera: Colletidae): emerging patterns from the southern end of the world. Journal of Biogeography  39: 526– 544. Google Scholar CrossRef Search ADS   Brown DK, Kaul RB. 1981. Floral structure and mechanisms in Loasaceae. American Journal of Botany  68: 361– 72. Google Scholar CrossRef Search ADS   Cane JH, Sipes S. 2006. Characterizing floral specialization by bees: analytical methods and revised lexicon for oligolecty. In: Waser NM, Ollerton J, eds. Plant-pollinator interactions: from specialization to generalization . Chicago: The University of Chicago Press, 99– 121. Carvalho AT, Schlindwein C. 2011. Obligate association of an oligolectic bee and a seasonal aquatic herb in semi-arid north-eastern Brazil. Biological Journal of the Linnean Society  102: 355– 368. Google Scholar CrossRef Search ADS   Cocucci AA, Sérsic AN. 1998. Evidence of rodent pollination in Caiophora coronata (Loasaceae). Plant Systematics and Evolution  211: 113– 128. DER/MG. 1998. APA Fernão Dias. Plano de Gestão Ambiental , Vol. 1. Belo Horizonte, 225p. von Frisch K. 1967. The dance language and orientation of bees . Cambridge, Harvard University Press. Gimenes M. 2002. Interactions between bees and Ludwigia elegans (Camb.) Hara (Onagraceae) flowers at different altitudes in Sao Paulo, Brazil. Revista Brasileira de Zoologia  19: 681– 689. Google Scholar CrossRef Search ADS   Harder LD, Thomson JD. 1989. Evolutionary options for maximizing pollen dispersal of animal-pollinated plants. The American Naturalist  133: 323– 344. Google Scholar CrossRef Search ADS   Harder LD, Wilson WG. 1994. Floral evolution and male reproductive success: optimal dispensing schedules for pollen dispersal by animal-pollinated plants. Evolutionary Ecology  8: 542– 559. Google Scholar CrossRef Search ADS   Harter B, Schlindwein C, Wittmann D. 1995. Bienen und Kolibris als Bestäuber von Blüten der Gattung Cajophora (Loasaceae). Apidologie  26: 356– 357. Henning T, Oliveira S, Schlindwein C, Weigend M. 2015. A new, narrowly endemic species of Blumenbachia (Loasaceae subfam. Loasoideae) from Brazil. Phytotaxa  236: 196– 200. Google Scholar CrossRef Search ADS   Henning T, Weigend M. 2012. Total control – pollen presentation and floral longevity in Loasaceae (blazing star family) are modulated by light, temperature and pollinator visitation rates. PLoS ONE  7: e41121. Google Scholar CrossRef Search ADS PubMed  Henning T, Weigend M. 2013. Beautiful, complicated—and intelligent? Novel aspects of the thigmonastic stamen movement in Loasaceae. Plant Signaling & Behavior  8: e24605. Google Scholar CrossRef Search ADS PubMed  Leite AV, Nadia T, Machado IC. 2016. Pollination of Aosa rupestris (Hook.) Weigend (Loasaceae): are stamen movements induced by pollinators? Brazilian Journal of Botany  39: 559– 567. Google Scholar CrossRef Search ADS   Lloyd DG, Yates JM. 1982. Intrasexual selection and the segregation of pollen and stigmas in hermaphrodite plants, exemplified by Wahlenbergia albomarginata (Campanulaceae). Evolution  36: 903– 913. Google Scholar CrossRef Search ADS PubMed  Melo GAR. 1996. Two new Brazilian bees of the genus Protodiscelis (Hymenoptera, Apoidea). Revista Brasileira de Entomologia  40: 97– 100. Michener CD. 1989. Classification of American Colletinae (Hymenoptera, Apoidea). The University of Kansas Science Bulletin  53: 622– 703. Michener CD. 2007. The bees of the world . Baltimore, MD: Johns Hopkins University Press. Percival MS. 1955. The presentation of pollen in certain angiosperms and its collection by Apis mellifera. New Phytologist  54: 353– 368. Google Scholar CrossRef Search ADS   Robertson C. 1925. Heterotropic bees. Ecology  6: 412– 436. Google Scholar CrossRef Search ADS   Roig-Alsina A, Schlumpberger BO. 2008. The cactus-specialist bees of the genus Brachyglossula Hedicke (Hymenoptera: Colletidae): notes on host associations and description of a new species. Journal of the Kansas Entomological Society  81: 84– 91. Google Scholar CrossRef Search ADS   Roubik DW. 1989. Ecology and natural history of tropical bees . Cambridge, University Press. Schlindwein C. 1998. Frequent oligolecty characterizing a diverse bee–plant community in a xerophytic bushland of subtropical Brazil. Studies on Neotropical Fauna and Environment  33: 46– 59. Google Scholar CrossRef Search ADS   Schlindwein C. 2000. Verhaltensanpassungen oligolektischer Bienen an synchrone und an kontinuierliche Pollenpräsentation. In: Breckle SW, Schweizer B, Arndt U, eds. Ergebnisse weltweiter ökologischer Forschung . Stuttgart, Germany: Verlag Günter Heimbach, 235– 250. Schlindwein C. 2004. Are oligolectic bees always the most effective pollinators? In: Freitas BM, Pereira JOP, eds. Solitary bees. Conservation, rearing and management for pollinators . Fortaleza: Imprensa Universitária, 231– 240. Schlindwein C, Wittmann D. 1995. Specialized solitary bees as effective pollinators of South Brazilian species of Notocactus and Gymnocalycium (Cactaceae). Bradleya  13: 25– 34. Google Scholar CrossRef Search ADS   Schlindwein C, Wittmann D. 1997a. Micro-foraging routes of Bicolletes pampeana (Colletidae) and bee-induced pollen presentation in Cajophora arechavaletae. Plant Biology  110: 177– 183. Schlindwein C, Wittmann D. 1997b. Stamen movements in flowers of Opuntia (Cactaceae) favour oligolectic pollinators. Plant Systematics and Evolution  204: 179– 193. Google Scholar CrossRef Search ADS   Strelin MM, Benitez-Vieyra S, Ackermann M, Cocucci AA. 2016. Flower reshaping in transition to hummingbird pollination in Loasaceae subfam. Loasoideae despite absence of corolla tubes or spurs. Evolutionary Ecology  30: 401– 417. Google Scholar CrossRef Search ADS   Urban I. 1886. Die Bestäubungseinrichtungen der Loasaceen. Jahrbücher des Botanischen Gartens Berlin  4: 364– 388. Urban I. 1892. Blüten- und Fruchtbau der Loasaceen. Berichte der Deutschen Botanischen Gesellschaft  10: 259– 265. Weigend M. 1997. Loasoideae in Eastern South America and Hispaniola: names, types and a key. Sendnera  4: 207– 220. Weigend M, Ackermann M, Henning T. 2010. Reloading the revolver – male fitness as simple explanation for complex reward partitioning in Nasa macrothyrsa (Cornales, Loasaceae). Biological Journal of Linean Society  100: 124– 31. Google Scholar CrossRef Search ADS   Weigend M, Gottschling M. 2006. Evolution of funnel-revolver flowers and ornithophily in Nasa (loasaceae). Plant Biology  8: 120– 142. Google Scholar CrossRef Search ADS PubMed  Weigend M, Gottschling M, Hoot S, Ackermann M. 2004. A preliminary phylogeny of Loasaceae subfam. Loasoideae (Angiospermae: Cornales) based on trnL(UAA) sequence data and its relation to systematics and historical biogeography. Organisms, Diversity and Evolution  4: 73– 90. Google Scholar CrossRef Search ADS   Wittmann D, Schlindwein C. 1995. Melittophilous plants, their pollen and flower visiting bees in southern Brazil. 2. Loasaceae. Biociências  3: 19– 34. © 2018 The Linnean Society of London, Biological Journal of the Linnean Society This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Biological Journal of the Linnean Society Oxford University Press

Pollination of Blumenbachia amana (Loasaceae): flower morphology and partitioned pollen presentation guarantee a private reward to a specialist pollinator

Loading next page...
 
/lp/ou_press/pollination-of-blumenbachia-amana-loasaceae-flower-morphology-and-JWwfyiv5FW
Publisher
Oxford University Press
Copyright
© 2018 The Linnean Society of London, Biological Journal of the Linnean Society
ISSN
0024-4066
eISSN
1095-8312
D.O.I.
10.1093/biolinnean/bly061
Publisher site
See Article on Publisher Site

Abstract

Abstract Blumenbachia amana possesses the characteristic ‘tilt-revolver flowers’ and the pollen partitioning mechanism found exclusively in Loasoideae. Stamens are sheltered in naviculate petals that alternate with five nectar scales, each one associated with two free staminodes. In the staminate phase, stamens move successively towards the centre of the flower and present pollen. Bees of oligolectic Actenosigynes mantiqueirensis (Colletidae) are the unique effective pollinators. Clinching the nectar scale–staminode complex, they bend the scales outward one by one with their head to take up nectar, thereby triggering the movement of a stamen, which arrives in the centre of the flower ~3 min after the stimulus. Females returned to previously visited flowers after short time intervals by establishing pollen foraging routes in flower patches that were maintained for days. In flowers experimentally hand-stimulated in the nectar scales to simulate pollinator visits, twice as many stamens moved, and the longevity of bee-visited flowers decreased by shortening the staminate phase. Partitioned pollen presentation in B. amana leads to a unique pollen foraging behaviour characterized by absolute flower fidelity by this bee in this specialized plant–pollinator relationship. foraging behaviour, Loasaceae, oligolectic bee, tilt-revolver flowers INTRODUCTION Pollen presentation in flowering plants occurs between two extremes: simultaneously, when all anthers of simultaneously opening flowers open, or gradually, when flowers open during the course of the day and anthers open one by one during anthesis (Percival, 1955). The latter occurs by either packaging or dispensing mechanisms (Harder & Thomson, 1989), which are considered to have the function of making pollen transport to stigmas via pollinators more efficient by limiting the amount of pollen that can be removed during a single visit, forcing pollen foragers to increase their movement among flowers (Lloyd & Yates, 1982; Harder & Thomson, 1989; Harder & Wilson, 1994). Accentuated gradual pollen release within single flowers occurs in the subfamily Loasoideae (Loasaceae), where flowers are protandrous, and stamens ripen successively one by one. During the staminate phase, stamens initially hidden within naviculate petals move successively towards the centre of the flower after anther dehiscence. In addition to fertile antepetal stamens, the androecium contains a concave nectar scale in front of each sepal, which is interiorly opposed by two free staminodes (Urban, 1886, 1892; Brown & Kaul, 1981). Weigend & Gottschling (2006) denominated this melittophilous type of Loasoideae flower as a ‘tilt-revolver flower’, because the nectar scale–staminode complex forms a circle with five units, which must be probed separately by nectar-seeking bees. Oligolectic bees, which feed their larvae pollen of the same plant genus or family (Robertson, 1925), were the sole pollinators of Caiophora arechavaletae (Urb.) Urb. & Gilg (Schlindwein & Wittmann, 1997a) and Aosa rupestris (Gardner) Weigend (Leite, Nadia & Machado, 2016). While searching for nectar in flowers of C. arechavaletae, females of Bicolletes pampeana (Colletidae) elicit stamen movements by pressing the nectar scales outwards with their head. They establish micro-foraging routes and return to previously visited flowers in short intervals, similar to the time delay between stimulation and arrival of the stamen in the centre of the flower, to collect the presented pollen package (Schlindwein & Wittmann, 1997a). Delayed pollen presentation in response to mechanical stimulation of the nectar scale–staminode complex was later demonstrated in greenhouse-cultivated plants of numerous representative species of the genera Caiophora, Loasa, Nasa and Presliophytum, all in the subfamily Loasoideae (Henning & Weigend, 2012, 2013). We studied pollination of the recently described loasoid Blumenbachia amana Henning & Weigend, endemic to the Serra da Mantiqueira mountain range in southeastern Brazil, and its relationship with Actenosigynes mantiqueirensis Silveira (Colletidae). We focused on the following issues: What are the characteristics of pollen presentation in B. amana? Do flower visitors influence the pollen release pattern? Does the pollen presentation regime shape the foraging behaviour of the effective pollinator? To answer these questions, we characterized the floral biology and pollination of B. amana and analysed, in the field, the flower-handling behaviour and foraging flights of its bee pollinators. MATERIAL AND METHODS Study Area The study was conducted from November 2013 to February 2015 in the municipalities of Camanducaia and Gonçalves (22°41.275′S, 45°54.048′W and 22°42.695′S, 45°57.082′W), situated in the Serra da Mantiqueira mountain range, Minas Gerais, Brazil. The study sites are located between 1500 and 1800 m elevation. The climate is humid throughout the year, with mean annual rainfall ranging from 1600 to 1800 mm, and mean annual temperatures ranging between 14 and 19 °C (DER/MG, 1998). The surrounding vegetation is dominated by mixed Araucaria forest (Araucaria angustifolia) within the high-altitude domain of the Atlantic Forest. Studied Species The South American genus Blumenbachia Schrad. (Loasaceae) comprises 11 species, six of which occur in Brazil (Weigend, 1997; Henning et al., 2015). The recently described B. amana (Henning et al., 2015) is an ascending annual herb that occurs on humid organic soils along streams and in forest clearings, and in humid depressions in trample-protected pastures for cattle and vegetable gardening areas. Thus far, the species is known only from the northern slope of the Mantiqueira mountain range in the two municipalities of Camanducaia and Gonçalves. Like other species of Loasoideae (Urban, 1886, 1892; Schlindwein & Wittmann, 1995; Henning & Weigend, 2012, 2013; Weigend, Ackermann & Henning, 2010), B. amana shows a highly differentiated androecium composed of five antesepalous nectar scales internally opposed by two free staminodes with filiform apices. Five antepetalous fascicles of fertile stamens alternate with the nectar scale–staminode complex (Fig. 1). The nectar scales are concave, formed by three connate staminodes, and show three basal, abaxial, filiform appendices. The scale apex is recurved, brightly red merging into yellow and green at the base, and white laterally and near the apex. The numerous stamens are initially hidden within naviculate petals and move towards the flower centre throughout the staminate phase (Fig. 2A–C). Specimens of the studied species are deposited in Herbarium BHCB (holotype BHCB 162877, Belo Horizonte) and Herbarium Berolinense (isotypes B 10 0610559, Bonn). Figure 1. View largeDownload slide Floral diagram of Blumenbachia amana. fs, free staminodes; ns, nectar scale; o, ovary; p, petal; s, sepal; st, stamens. Figure 1. View largeDownload slide Floral diagram of Blumenbachia amana. fs, free staminodes; ns, nectar scale; o, ovary; p, petal; s, sepal; st, stamens. Figure 2. View largeDownload slide Flower of Blumenbachia amana. A, flower in the staminate phase; a stamen moves from the petal downwards into the centre of the flower. B, staminodial complex: 1, nectar scale; 2, free staminodes. C, longitudinal section through a flower in the staminate phase: 1, stamen bundle inside a petal; 2, immature style. D, flower in the pistillate phase: arrow, mature style. Scale bars represent 1 cm unless indicated otherwise. Figure 2. View largeDownload slide Flower of Blumenbachia amana. A, flower in the staminate phase; a stamen moves from the petal downwards into the centre of the flower. B, staminodial complex: 1, nectar scale; 2, free staminodes. C, longitudinal section through a flower in the staminate phase: 1, stamen bundle inside a petal; 2, immature style. D, flower in the pistillate phase: arrow, mature style. Scale bars represent 1 cm unless indicated otherwise. Floral Morphology and Longevity Given that flower size increases during anthesis, we measured variation in the size of petals, stamens and styles throughout the staminate and pistillate phases of anthesis. Forty-five flowers of 20 individual plants in pre-anthesis were marked in the field and divided into three sets of 15 flowers from different individuals. During three subsequent days, we collected 15 flowers per day and measured the length of the petals, stamens and styles (N = 45). The mean number of stamens per flower was obtained from 35 new flowers. We determined floral longevity and the duration of the staminate and pistillate phases. For flower longevity, we considered the number of daylight hours during which each flower remained open between 09.00 and 18.00 h. We compared flower longevity between non-visited bagged flowers (N = 20) and bee-visited flowers (N = 20). We considered the staminate phase to be the period from the beginning of flower opening until all the stamens had moved. The beginning of the pistillate phase was defined as the moment the stigma became visible in the centre of the flower and when all stamens had already moved (see Weigend et al., 2010; Henning & Weigend, 2012). From this moment onwards until flower senescence, flower visitors could contact the stigma. Stamen Movement We investigated stamen movement in terms of the duration of migration and the number of migrated stamens per hour. To evaluate whether flower visitors could induce stamen movements while taking up nectar, we experimentally applied mechanical stimuli to the nectar scales to simulate flower visits by bees. Stimuli consisted of the application of slight outward pressure on each of the five nectar scales using a toothpick. We then counted the number of migrated stamens and measured the time between stimuli and arrival of the stamens in the centre of the flower (N = 52 flowers). For 26 flowers, we applied stimuli, as described above, every 5 min for 1 h. This stimulation interval was similar to the mean interval of flower visits to B. amana. Another 26 flowers were bagged, to prevent access of bee visitors and to prevent stimulation of the flowers, as controls. In this paired experiment, all flowers were in the staminate phase. To certify that the flowers received the same level of stimulation, we stimulated all nectar scales of all flowers of the two treatments 30 min before the experiment. To examine whether stamens move overnight, at 18.00 h we gently removed the anthers of the already migrated stamens of 15 marked flowers. The next day, at 09.00 h, we inspected these flowers to count the stamens that had moved. Breeding System To determine whether flowers set fruits by spontaneous self-pollination, we compared the number of fruits and seeds produced in 20 flowers bagged throughout anthesis with 20 non-bagged control flowers that were visited by bees. Flower Visitors and Visiting Frequency Flower visitors of B. amana were collected with entomological nets and mounted with entomological pins. The specimens were identified and deposited in the Entomological Collection of UFMG (Centro de Coleções Taxonômicas da UFMG, Universidade Federal de Minas Gerais, Belo Horizonte, Brazil). Flower visiting frequency was determined throughout the day for four marked flowers per 30 min count in four 2 h time intervals (09.00–11.00, 11.00–13.00, 13.00–15.00 and 15.00–17.00 h; N = 158 flowers). We also compared the frequency of bee visits to flowers in the staminate and pistillate phases (N = 14 flowers per phase). Foraging Flights of Females of A. Mantiqueirensis Females of A. mantiqueirensis were captured with entomological nets and marked with individual colour codes on the mesoscutum and/or metasomal tergites using Revell ink (Revell, Germany). When the ink had dried, ~2 min after marking, the bees were liberated. This procedure did not exert any notable influence on the foraging behaviour of the marked females. All open flowers in the studied populations were numbered and the foraging flights of marked females accompanied. We measured the duration of partial and complete foraging flights of the individually marked bees. A complete flight included all flower visits of a marked female, from when she appeared in a foraging patch with empty scopae to when she left with pollen-filled scopae. When a marked female reappeared with empty scopae, we considered this interval as the time she needed to fly back to the nest, remove the pollen load and return to the foraging patch. During each foraging flight, we recorded duration, the total number of visits, the number of visits and revisits to individual flowers, and the duration of the intervals between revisits. For each floral visit, we recorded the number of nectar scales probed and whether pollen was collected. The mean flower-handling time per sequence of visits was calculated by dividing the duration of consecutive flower visits by the number of flowers visited by a marked female. Thus, the calculated handling time includes the duration of a flower visit plus the flight time to the next flower. The total time of observations of visiting sequences of individually marked females was ~22 h. To determine whether bees of A. mantiqueirensis visit other plants besides B. amana, samplings of floral visitors of other plants that occur near the foraging patches of B. amana were performed throughout the study. Statistical Analyses For all analyses, we first tested data for normality and homogeneity of variance using the Kolmogorov–Smirnov test and the Levene median test, respectively. We used a t-test to compare the following: (1) flower longevity and duration of staminate and pistillate phases of non-visited bagged flowers with bee-visited flowers; (2) number of moved stamens in experimentally stimulated flowers with non-stimulated flowers (data square root transformed); (3) seed set of visited flowers with non-visited flowers; and (4) frequency of visits by A. mantiqueirensis to flowers in the staminate and pistillate phases of anthesis. A non-parametric Kruskal–Wallis test was performed to compare the frequency of flower visits among the four established intervals throughout the day. All statistical analyses were conducted using SigmaStat 3.5 (Systat Software, Inc.) for Windows. RESULTS Floral Morphology and Anthesis The beginning of anthesis was not synchronized, so the flowers of an individual of B. amana opened at different times throughout the day between 09.00 and 18.00 h. New flowers were small, with flower size increasing during anthesis. Petals showed continuous growth throughout anthesis, stamens only until the second day, and the style doubled in size from the second to the third day, reaching about the length of the stamens at the end of anthesis (Fig. 3). Each of the five antepetalous fascicles of fertile stamens had on average 9.5 ± 1.7 stamens (N = 35 flowers) that had moved one by one during the staminate phase. After all stamens had moved, the pistillate phase commenced, and the style elongated and became visible in the centre of the flower (Fig. 2D). Figure 3. View largeDownload slide Mean length of petals, stamens and styles of Blumenbachia amana through anthesis (N = 45 flowers). Eorr bars indicate the standard deviation. Figure 3. View largeDownload slide Mean length of petals, stamens and styles of Blumenbachia amana through anthesis (N = 45 flowers). Eorr bars indicate the standard deviation. Stamen Movement Pollen partitioning in B. amana occurred by successive dehiscence of the anthers, accompanied by successive movements of the stamens. Stamen movement could be triggered by pushing the nectar scales slightly outwards, thus showing thigmonasty. After experimental stimuli, the stamens reached the centre of the flower after an average of 3.3 ± 1.1 min (N = 26). The comparison of flowers experimentally hand-stimulated in the nectar scales every 5 min for 1 h vs. non-stimulated bagged flowers showed that about twice as many stamens moved in hand-stimulated flowers, 6.3 ± 4.2 (N = 26), compared with the non-stimulated flowers, 3.2 ± 2.1 (N = 26) (t = 3.373; P = 0.001; N = 52; Fig. 4). Figure 4. View largeDownload slide Number of stamens moved per hour. Non-stimulated flowers (N = 26) and hand-stimulated flowers (N = 26). Values are the means ± SD. Different letters represent significant differences between means. Figure 4. View largeDownload slide Number of stamens moved per hour. Non-stimulated flowers (N = 26) and hand-stimulated flowers (N = 26). Values are the means ± SD. Different letters represent significant differences between means. In the 1 h stimulation experiments during the day, stamen movement occurred for 70% of the stimuli made. Stamen movements were almost completely restricted to daytime, with 15 marked flowers having only one stamen move at night (between 18.00 and 9.00 h). Pollen Presentation and Flower Longevity Non-visited flowers remained open for a longer time, on average 26.4 ± 4.8 h (N = 20), than flowers visited by bees, 21.0 ± 3.1 h (N = 20) (t = 4.285; P ≤ 0.001; N = 40). This difference was attributable to a 25% longer duration of the staminate phase in non-visited flowers. The duration of the staminate phase of visited flowers was on average 16.8 ± 3.3 h (N = 20) and of non-visited flowers 22.3 ± 3.9 h (N = 20) (t = 4.887; P ≤ 0.001; N = 40). There was no difference in the duration of the pistillate phase between visited and non-visited flowers, 4.2 ± 1.5 (N = 20) and 4.2 ± 2.3 h (N = 20), respectively (t = −0.059; P = 0.953; N = 40; Fig. 5). Figure 5. View largeDownload slide Flower longevity in Blumenbachia amana. Duration of staminate and pistillate phases of non-visited and bee-visited flowers. Non-visited flowers: staminate and pistillate phase (N = 20). Bee-visited flowers: staminate and pistillate phase (N = 20). Values are the means ± SD. Only the daylight hours of open flowers were considered. Different letters represent significant differences between means. Figure 5. View largeDownload slide Flower longevity in Blumenbachia amana. Duration of staminate and pistillate phases of non-visited and bee-visited flowers. Non-visited flowers: staminate and pistillate phase (N = 20). Bee-visited flowers: staminate and pistillate phase (N = 20). Values are the means ± SD. Only the daylight hours of open flowers were considered. Different letters represent significant differences between means. Breeding System The flowers of B. amana were self-compatible. All marked flowers set fruits, and there was no difference in seed set between non-visited flowers (spontaneous selfing), with 35.0 ± 11.5 seeds (N = 20), and flowers visited by pollinators, with 35.5 ± 14.0 seeds (N = 20) (t = −0.111; P = 0.912; N = 40). Flower Visitors and Visitation Frequency Females and males of A. mantiqueirensis were the almost exclusive flower visitors of B. amana, and no individual of this species was sampled on flowers of any other plant species in the vegetation surrounding B. amana plants. During a total of 70 days of observation (~500 h) of flowers, we recorded only sporadic visits of Plebeia saiqui (Friese, 1900), Plebeia droryana (Friese, 1900), Plebeia lucii Moure, 2004, Apis mellifera Linnaeus, 1758 (all Apidae), and Augochloropsis sp. (Halictidae). Bees of these species did not visit flowers of B. amana during the counts of flower visitation frequency. Given that these species occurred only sporadically (22 visits in 70 days of observation), did not visit flowers in the pistillate phase and also did not handle the flowers properly (bees of Apis and Augochloropsis slipped off the flower centre without gaining access to nectar, and those of Plebeia gleaned pollen from the petals), they were not considered in the present study. Bees of A. mantiqueirensis visited the flowers throughout the entire flowering period of B. amana (from November to February). There was no difference in the frequency of visits among the four established intervals throughout the day. We recorded on average 4.1–5.7 flower visits per 30 min (Kruskal–Wallis = 1.706; P = 0.636; N = 158). Females and males of A. mantiqueirensis also visited flowers in the staminate phase with the same frequency, 7.4 ± 4.0 visits (N = 14), as those in the pistillate phase, 7.9 ± 2.7 visits (N = 14) (t = −0.383; P = 0.705; N = 28). Flower Handling by A. Mantiqueirensis Flower visits by females of A. mantiqueirensis showed the following standard behavioural sequence. The bees approached the pendulous flowers bottom-up, with the head oriented towards the centre of the flower. Still in flight, the bees stretched their front legs towards the nectar scales and clinched the revolute apexes of the scales first with the hooked tarsal claws of these legs, then with the tarsal claws of the mid- and hindlegs (Fig. 6A, B). Figure 6. View largeDownload slide Handling of flowers of Blumenbachia amana by females of Actenosigynes mantiqueirensis. A, in search of nectar in a flower in the pistillate phase, a female inserts her head between the nectar scale and staminodes, pushing the scale outward, and contacts the long stigma with her ventral scopa filled with pollen (dotted arrow; continuous arrow see 6B). B, tarsal claws of fore-, mid- and hindlegs of a female A. mantiqueirensis clinching to the revolute apexes of two nectar scales of B. amana during nectar uptake (arrows indicate claws). C, female collecting pollen from a moved stamen in the centre of the flower, and another stamen moving to the centre. D, individually marked female pulling a stamen that has not moved with her mandibles. Scale bars represent 1 cm unless indicated otherwise. Figure 6. View largeDownload slide Handling of flowers of Blumenbachia amana by females of Actenosigynes mantiqueirensis. A, in search of nectar in a flower in the pistillate phase, a female inserts her head between the nectar scale and staminodes, pushing the scale outward, and contacts the long stigma with her ventral scopa filled with pollen (dotted arrow; continuous arrow see 6B). B, tarsal claws of fore-, mid- and hindlegs of a female A. mantiqueirensis clinching to the revolute apexes of two nectar scales of B. amana during nectar uptake (arrows indicate claws). C, female collecting pollen from a moved stamen in the centre of the flower, and another stamen moving to the centre. D, individually marked female pulling a stamen that has not moved with her mandibles. Scale bars represent 1 cm unless indicated otherwise. Immediately after landing, the bees looked for nectar in 99.8% (4894 visits) of the flower visits (N = 4905 total of visits). To do so, they inserted their head between the nectar scale apex and the two free staminodes interiorly opposed to it, pushing the scale outward in the process (Fig. 6A). The revolute collar-shaped apexes of the nectar scales provided a rigid supporting point to mortise the bees’ body forward until their short mouthparts reached the base of the scales. In each search for nectar in a nectar scale, the bees clinched the revolute apexes of the two neighbouring nectar scales with the tarsal claws of front and mid-legs, while with their hindlegs they grasped the opposite pair of scales. After nectar uptake in the scale, the bees turned to the neighbouring scale and repeated nectar uptake in the same relative position. While the bee removed its head from the nectar scale, the scale elastically bent back to the original position (Supporting Information, Video S1). In 47% (2281 visits) of the flower visits, the bees searched for nectar in all the five scales, pushing them in sequence in clockwise or anticlockwise rotation. Thereby, the bees always used the apical collar of the nectar scales as a foothold. Females and males showed no differences in nectar-searching behaviour, and males always left the flower immediately after nectar uptake. During nectar uptake in flowers in the pistillate phase, females and males continuously contacted the long stigma with the ventral face of their mesosoma and metasoma and thus transferred allochthonous pollen to the stigmatic surface. In females, the pollen-loaded ventral scopa was continuously in contact with the stigma in these flowers (Fig. 6A). After pushing the nectar scales, female bees actively collected pollen in 41.4% of the flower visits (2029 visits), showing two pollen-collection behaviours. The first behaviour is termed pollen brushing. Females brushed the pollen grains exposed on the anther surface of moved or moving stamens with scopal hairs of the metasoma and hindlegs. This behaviour was shown in 56.2% (1141 visits) of pollen collections in a flower. The second behaviour is termed stamen pulling. Females sought for non-moved stamens still hidden in the naviculate petals, moved to the petals, grasped a filament with the mandibles near its basal third and pulled the stamens to the flower centre (Fig. 6C, D). Using forelegs and mandibles, the females manipulated the filaments and advanced with the head in the direction to the anthers. For stamens with already dehisced anthers, females used pollen brushing to collect pollen. For pulled stamens with still closed anthers, the anther was torn with the mandibles and the exposed pollen grains were then transferred from the mandibles to the hindlegs using foreleg tarsal hairs. Next, the females directed the anther to the ventral scopa and rubbed it to remove more pollen grains. This behaviour lasted longer and was performed in 43.8% (888 visits) of pollen-collection visits. Foraging by A. Mantiqueirensis Monitoring of individually marked females of A. mantiqueirensis showed that they established foraging areas with limited numbers of B. amana flowers. These areas were maintained for up to 20 consecutive days. We recorded 46 foraging flights of 13 individually marked females. We accompanied 12 full foraging flights of five females, i.e. flights that included all flower visits from the moment a female appeared with empty scopae in an observation area until it left with scopae filled with pollen. These females reappeared in the observation patches with empty scopae after 12.5–24.1 min (15.9 ± 3.3 min; N = 12). The complete foraging flights lasted between 21.5 and 106.0 min (55.0 ± 26.7 min; N = 12). The individual foraging areas included 41–85 flowers (62.5 ± 13.4; N = 12), and the number of total flower visits per flight in a foraging area ranged from 93 to 449 (244.6 ± 112.8; N = 12). The mean handling time per flight sequence (flower handling plus duration of flight between flowers) of the different females varied between 9.0 and 23.0 s (14.1 ± 3.0 s; N = 46). During all the recorded flights, females repeatedly visited the same flowers (N = 3113 revisits). Revisit intervals to a flower were mostly short, and in 52.0% (1534 revisits) the females returned to the same flower within the first 6 min after a previous floral visit. Most revisit intervals were between 2 and 5 min (916 revisits) (Fig. 7). Figure 7. View largeDownload slide Revisit intervals for individually marked females of Actenosigynes mantiqueirensis during foraging flights to marked flowers of Blumenbachia amana (N = 3113 revisits). Revisit intervals between 41 and 97 min (0.83%) are not represented in the graph. Figure 7. View largeDownload slide Revisit intervals for individually marked females of Actenosigynes mantiqueirensis during foraging flights to marked flowers of Blumenbachia amana (N = 3113 revisits). Revisit intervals between 41 and 97 min (0.83%) are not represented in the graph. DISCUSSION The present study demonstrates a highly specialized interaction between B. amana and A. mantiqueirensis, its sole colletid bee pollinator. Both females and males of this bee species visit flowers in the staminate and pistillate phases with the same high frequency, deposit pollen on the stigma and continuously fly among flowers of conspecific individuals, showing absolute flower fidelity. These bees are, thus, highly effective pollinators of B. amana, guaranteeing the cross-flow of pollen. Our observations indicate that A. mantiqueirensis is narrowly oligolectic (in the sense of Robertson, 1925; Cane & Sipes, 2006): (1) the foraging behaviour of females is linked to the mechanism of continuous pollen presentation by the flowers; (2) whole-day monitoring of marked individuals showed that foraging is interrupted only when females need to fly back to their nest to unload pollen grains from their scopae, after which they return to their B. amana flower patches; and (3) the pollen loads of females contained pollen of B. amana exclusively, and the bees were never observed to visit flowers of other species. In the subfamily Neopasiphaeinae Cockerell, 1930 (sensuAlmeida & Danforth, 2009; Almeida et al., 2012; which were formerly included in Paracolletinae Cockerell, 1934 and Paracolletini Michener, 1989), pollen specialization seems to be common for Neotropical species (Schlindwein & Wittmann, 1995; Melo, 1996; Schlindwein & Wittmann, 1997a, b; Schlindwein, 1998, 2004; Gimenes, 2002; Michener, 2007; Roig-Alsina & Schlumpberger, 2008; Carvalho & Schlindwein, 2011). The continuous release of pollen, and probably nectar, in tiny portions throughout the day induces high frequencies of visitation to flowers by specialist visitors and makes the flowers of B. amana unattractive for polylectic bees, such as the highly eusocial species that efficiently recruit workers to collect synchronously offered resources in mass flowering plants (von Frisch, 1967; Roubik, 1989). Considering an almost constant rate of, on average, one visit per 5 min during ~8 h per day, a flower might receive 80–96 flower visits a day and ~250 visits during its lifetime by bees of A. mantiqueirensis. Given that a bee inserts its mouthparts in nectar scales on average 3.9 times per flower visit, each flower might receive ~1000 insertions of head and mouthparts during its lifetime. In the presence of the steadily visiting oligolectic pollinators, the flowers of B. amana are permanently empty of pollen and nectar, discouraging polylectic species from visiting their flowers. For the specialized A. mantiqueirensis, the positive consequence of its specialization is to guarantee a private resource, whereas the negative is its obvious inability to explore alternative host plants when flowers of B. amana are absent. The floral morphology of B. amana, and that of other representatives of melittophilous Loasoideae, seems to discourage visits of non-specialist bees. To access the pendulous flowers, bees must land ‘upside down’ and fix themselves with their tarsal claws at the revolute apical margin of the nectar scales. Only this exact position ensures support for pushing the nectar scales with the head to access nectar. Weigend & Gottschling (2006) previously considered this ‘recurved scale neck’, in addition to dorsal calli on the nectar scales, as morphological adaptations of melittophilous representatives of the genus Nasa for providing such a foothold for pollinating bees landing on the pendulous flowers. Only bees of A. mantiqueirensis handled the B. amana flowers adequately and used the revolute foothold to advance with the head and push the scales outward to access the nectar. The revolute nectar scale apex is therefore important for providing a rigid access point to obtain nectar and mortise the bees’ body in a position that guarantees contact with mature anthers and/or receptive stigmas. This seems to be a consistent feature in tilt-revolver flowers (Schlindwein & Wittmann, 1997a; Ackermann & Weigend, 2006; Weigend & Gottschling, 2006; Weigend et al., 2010; Leite et al., 2016; Strelin et al., 2016). Pollinators Influence the Pollen Release Pattern Pollen presentation in B. amana occurs in a manner similar to that demonstrated in other representatives of the subfamily Loasoideae. The anthers mature and open successively, followed by movement of the stamens (Urban, 1886, 1892; Brown & Kaul, 1981). When nectar scales are stimulated, the movement of stamens can be triggered, thereby showing thigmonasty (Schlindwein & Wittmann, 1997a; Weigend et al., 2010; Henning & Weigend, 2012, 2013). Phylogenetic analyses showed that thigmonasty is a synapomorphy of the ‘higher Loaseae’ within Loasoideae, the lineage to which B. amana belongs (see cladogram of Weigend et al., 2004). This species group includes both melittophilous tilt-revolver flowers and ornithophilous funnel-revolver flowers, with the former being the plesiomorphic condition (Weigend et al., 2004; Ackermann & Weigend, 2006; Weigend & Gottschling, 2006; Strelin et al., 2016). In particular, the Brazilian Loasoideae comprise a group of lowland species with rather uniform melittophilous flower morphology from the three genera Aosa, Blumenbachia and Caiophora, which seem to be pollinated mainly by solitary pollen-seeking bees. But thigmonasty probably did not evolve as an adaptation to specialized pollen foraging in colletid bees, as previously thought (Schlindwein & Wittmann, 1997a). In species where nectar rather than pollen is the central resource for pollinators, thigmonasty associated with protandry also strongly favours cross-pollination. This is the case in associations of representatives of ‘higher Loaseae’ in the Andes with floral nectar-seeking bees, such as the polylectic carpenter bee Neoxylocopa lachea Moure, 1951, which was identified as the effective pollinator of Nasa macrothyrsa (Loasoideae) (Weigend et al., 2010), and with hummingbirds and even rodents (Harter, Schlindwein & Wittmann, 1995; Cocucci & Sérsic, 1998; Ackermann & Weigend, 2006; Weigend & Gottschling, 2006). Compartmentalized pollen presentation in Loasoideae in the form of defined amounts of pollen was interpreted by Weigend et al. (2010) as a mechanism of tilt-revolver flowers that increases male fitness. The authors explained the complex morphology of Loasoideae flowers with the pollen presentation theory (PPT), which postulates that male fitness is favoured when the frequency of pollen release is adjusted to the activity patterns of pollinators (Lloyd & Yates, 1982; Harder & Thomson, 1989; Harder & Wilson, 1994). This is exactly what we observed as characteristic of the staminate phase of B. amana. We showed experimentally that continuous stimulation of nectar scales accelerates pollen donation. The flower-bagging experiment confirmed shortened flower longevity in flowers accessible to pollinators, but exclusively by abbreviating the staminate phase of flowering by ~25%. The already much shorter pistillate phase, important for pollen reception, is maintained without alteration. The flowers of B. amana, thus, show a complex mechanism of control of pollen release in the daytime hours, when the pollinators are active, but not during the night. Influence of Pollen Presentation Regime on Foraging Behaviour of the Effective Pollinator The possibility that females of A. mantiqueirensis trigger the release of a small amount of pollen through the movement of a stamen makes the moment of pollen presentation predictable to a few minutes after stimulation. We interpret the adoption of foraging routes of a restricted number of flowers as an adapted foraging behaviour for the return to previously visited flowers in intervals corresponding to the time between a stimulus and the arrival of the stamens in the centre of the flowers. However, a behavioural response of optimized collection of pollen depends on two unpredictable characteristics. First, not all stimuli applied to nectar scales result in a response of stamen movement. The present study found that ~30% of the stimuli in the staminate phase did not cause stamen movement. Second, other individual bees that forage in the same flower patch also stimulate nectar scales, harvest available pollen and, thus, interfere with the trigger/harvest mechanism. The interference of conspecific competitors should become greater when flowers of B. amana become scarce. Similar systems with narrowly oligolectic Neopasiphaeinae bees as highly effective unique pollinators of species of Loasoideae are known, such as for C. arechavaletae and Blumenbachia insignis Schrad. with Bicolletes pampeana (Wittmann & Schlindwein, 1995; Schlindwein & Wittmann, 1997a; Schlindwein, 2000); for Aosa rupestris (Gardner) Weigend with Bicolletes nordestina Urban (Leite et al., 2016); and probably also for Blumenbachia eichleri Urb. (cited as Caiophora eichleri) and Blumenbachia catharinensis Urb & Gilg with Actenosigynes fulvoniger (cited as Leioproctus fulvoniger) (Harter et al., 1995; Schlindwein, 2000). Caiophora arechavaletae and its oligolectic bee pollinator B. pampeana show a similar relationship to that found in the present study (Schlindwein & Wittmann, 1997a). In both cases, the bees continuously stimulate the nectar scales, triggering stamen movement, with the stamens arriving in the centre of the flower on average 2.4 min (C.arechavaletae) and 3.3 min (B. amana) after triggering. Females of both species adopt stable foraging areas for several days and return to previously visited flowers in short intervals, always searching for nectar and removing pollen from pollen-presenting anthers. Females of A. mantiqueirensis, however, include more flowers in their foraging patches, have longer handling times per flower, and their complete foraging flights last longer than those of B. pampeana. The two different pollen-collection behaviours of A. mantiqueirensis, however, seem to be the most striking difference from B. pampeana, which exhibits pollen brushing exclusively. Pollen harvest by stamen pulling from non-moved stamens might be advantageous for females of A. mantiqueirensis because they gain access to the pollen content of hidden anthers. A pollen-searching female, thus, would increase her competitive vigour over conspecifics because she ensures the collection of pollen before conspecifics do, who rely on ordinary pollen presentation of moved stamens. Nevertheless, females are not able to determine during a flower visit whether non-moved stamens show dehisced anthers, nor whether the flowers are in the staminate or pistillate phase. We frequently observed that females pull stamens with closed anthers that require more time for pollen grain removal or whose pollen is still not available. Given that the autogamous B. amana exhibits high fruit and seed set even in the absence of pollinators, as has been demonstrated for other annual Loasaceae (Schlindwein & Wittmann, 1997a; Henning & Weigend, 2013; Leite et al., 2016), only the narrowly oligolectic bee pollinator manifests reproductive dependence in this tight interaction. Nevertheless, the relationship between both partners is highly mutual, because in all the studied cases (Wittmann & Schlindwein, 1995; Schlindwein & Wittmann, 1997a; Leite et al., 2016), the sole colletid bee pollinators guarantee the important cross-pollen flow during their hundreds of uninterrupted flower visits to their specific host plants. It can be expected that A. mantiqueirensis bees would not survive in the absence of their host plant. Both interacting endemic species were described recently are known only from specific habitats within high-altitude remnants of the threatened Atlantic Forest, and are therefore at a high risk of extinction. SUPPORTING INFORMATION Additional Supporting Information may be found in the online version of this article at the publisher's web-site. Video S1. Nectar uptake. A female of Actenosigynes mantiqueirensis during nectar uptake in a flower of Blumenbachia amana. Clinching the nectar sacale-staminode complex, the bee bents the scales outwards one by one with it’s head. ACKNOWLEDGEMENTS We thank José Neiva, Leandro Sales, Nathália Falagan, Fernanda Figueiredo, Ana Luiza Cordeiro and Isabelle Cerceau for their help with fieldwork, and the residents of the villages of Cantagalo (Gonçalves) and Monte Azul (Camanducaia), especially Senhora Eurides, Suely, José Maria, Elizabeth, Pedro, Rosa, Maria and Altair, for logistic support. We thank three anonymous reviewers for their constructive comments, which improved the manuscript. We thank ICMBio for the collection license, Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq Universal – 481605-2011-8) for financial support and individual research grants from Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) to S.S.-O. and R.O., and from CNPq to C.S. REFERENCES Ackermann M, Weigend M. 2006. Nectar, floral morphology and pollination syndrome in Loasaceae subfam. Loasoideae (Cornales). Annals of Botany  98: 503– 514. Google Scholar CrossRef Search ADS PubMed  Almeida EA, Danforth BN. 2009. Phylogeny of colletid bees (Hymenoptera: Colletidae) inferred from four nuclear genes. Molecular Phylogenetics and Evolution  50: 290– 309. Google Scholar CrossRef Search ADS PubMed  Almeida EAB, Pie MR, Brady SG, Danforth BN. 2012. Biogeography and diversification of colletid bees (Hymenoptera: Colletidae): emerging patterns from the southern end of the world. Journal of Biogeography  39: 526– 544. Google Scholar CrossRef Search ADS   Brown DK, Kaul RB. 1981. Floral structure and mechanisms in Loasaceae. American Journal of Botany  68: 361– 72. Google Scholar CrossRef Search ADS   Cane JH, Sipes S. 2006. Characterizing floral specialization by bees: analytical methods and revised lexicon for oligolecty. In: Waser NM, Ollerton J, eds. Plant-pollinator interactions: from specialization to generalization . Chicago: The University of Chicago Press, 99– 121. Carvalho AT, Schlindwein C. 2011. Obligate association of an oligolectic bee and a seasonal aquatic herb in semi-arid north-eastern Brazil. Biological Journal of the Linnean Society  102: 355– 368. Google Scholar CrossRef Search ADS   Cocucci AA, Sérsic AN. 1998. Evidence of rodent pollination in Caiophora coronata (Loasaceae). Plant Systematics and Evolution  211: 113– 128. DER/MG. 1998. APA Fernão Dias. Plano de Gestão Ambiental , Vol. 1. Belo Horizonte, 225p. von Frisch K. 1967. The dance language and orientation of bees . Cambridge, Harvard University Press. Gimenes M. 2002. Interactions between bees and Ludwigia elegans (Camb.) Hara (Onagraceae) flowers at different altitudes in Sao Paulo, Brazil. Revista Brasileira de Zoologia  19: 681– 689. Google Scholar CrossRef Search ADS   Harder LD, Thomson JD. 1989. Evolutionary options for maximizing pollen dispersal of animal-pollinated plants. The American Naturalist  133: 323– 344. Google Scholar CrossRef Search ADS   Harder LD, Wilson WG. 1994. Floral evolution and male reproductive success: optimal dispensing schedules for pollen dispersal by animal-pollinated plants. Evolutionary Ecology  8: 542– 559. Google Scholar CrossRef Search ADS   Harter B, Schlindwein C, Wittmann D. 1995. Bienen und Kolibris als Bestäuber von Blüten der Gattung Cajophora (Loasaceae). Apidologie  26: 356– 357. Henning T, Oliveira S, Schlindwein C, Weigend M. 2015. A new, narrowly endemic species of Blumenbachia (Loasaceae subfam. Loasoideae) from Brazil. Phytotaxa  236: 196– 200. Google Scholar CrossRef Search ADS   Henning T, Weigend M. 2012. Total control – pollen presentation and floral longevity in Loasaceae (blazing star family) are modulated by light, temperature and pollinator visitation rates. PLoS ONE  7: e41121. Google Scholar CrossRef Search ADS PubMed  Henning T, Weigend M. 2013. Beautiful, complicated—and intelligent? Novel aspects of the thigmonastic stamen movement in Loasaceae. Plant Signaling & Behavior  8: e24605. Google Scholar CrossRef Search ADS PubMed  Leite AV, Nadia T, Machado IC. 2016. Pollination of Aosa rupestris (Hook.) Weigend (Loasaceae): are stamen movements induced by pollinators? Brazilian Journal of Botany  39: 559– 567. Google Scholar CrossRef Search ADS   Lloyd DG, Yates JM. 1982. Intrasexual selection and the segregation of pollen and stigmas in hermaphrodite plants, exemplified by Wahlenbergia albomarginata (Campanulaceae). Evolution  36: 903– 913. Google Scholar CrossRef Search ADS PubMed  Melo GAR. 1996. Two new Brazilian bees of the genus Protodiscelis (Hymenoptera, Apoidea). Revista Brasileira de Entomologia  40: 97– 100. Michener CD. 1989. Classification of American Colletinae (Hymenoptera, Apoidea). The University of Kansas Science Bulletin  53: 622– 703. Michener CD. 2007. The bees of the world . Baltimore, MD: Johns Hopkins University Press. Percival MS. 1955. The presentation of pollen in certain angiosperms and its collection by Apis mellifera. New Phytologist  54: 353– 368. Google Scholar CrossRef Search ADS   Robertson C. 1925. Heterotropic bees. Ecology  6: 412– 436. Google Scholar CrossRef Search ADS   Roig-Alsina A, Schlumpberger BO. 2008. The cactus-specialist bees of the genus Brachyglossula Hedicke (Hymenoptera: Colletidae): notes on host associations and description of a new species. Journal of the Kansas Entomological Society  81: 84– 91. Google Scholar CrossRef Search ADS   Roubik DW. 1989. Ecology and natural history of tropical bees . Cambridge, University Press. Schlindwein C. 1998. Frequent oligolecty characterizing a diverse bee–plant community in a xerophytic bushland of subtropical Brazil. Studies on Neotropical Fauna and Environment  33: 46– 59. Google Scholar CrossRef Search ADS   Schlindwein C. 2000. Verhaltensanpassungen oligolektischer Bienen an synchrone und an kontinuierliche Pollenpräsentation. In: Breckle SW, Schweizer B, Arndt U, eds. Ergebnisse weltweiter ökologischer Forschung . Stuttgart, Germany: Verlag Günter Heimbach, 235– 250. Schlindwein C. 2004. Are oligolectic bees always the most effective pollinators? In: Freitas BM, Pereira JOP, eds. Solitary bees. Conservation, rearing and management for pollinators . Fortaleza: Imprensa Universitária, 231– 240. Schlindwein C, Wittmann D. 1995. Specialized solitary bees as effective pollinators of South Brazilian species of Notocactus and Gymnocalycium (Cactaceae). Bradleya  13: 25– 34. Google Scholar CrossRef Search ADS   Schlindwein C, Wittmann D. 1997a. Micro-foraging routes of Bicolletes pampeana (Colletidae) and bee-induced pollen presentation in Cajophora arechavaletae. Plant Biology  110: 177– 183. Schlindwein C, Wittmann D. 1997b. Stamen movements in flowers of Opuntia (Cactaceae) favour oligolectic pollinators. Plant Systematics and Evolution  204: 179– 193. Google Scholar CrossRef Search ADS   Strelin MM, Benitez-Vieyra S, Ackermann M, Cocucci AA. 2016. Flower reshaping in transition to hummingbird pollination in Loasaceae subfam. Loasoideae despite absence of corolla tubes or spurs. Evolutionary Ecology  30: 401– 417. Google Scholar CrossRef Search ADS   Urban I. 1886. Die Bestäubungseinrichtungen der Loasaceen. Jahrbücher des Botanischen Gartens Berlin  4: 364– 388. Urban I. 1892. Blüten- und Fruchtbau der Loasaceen. Berichte der Deutschen Botanischen Gesellschaft  10: 259– 265. Weigend M. 1997. Loasoideae in Eastern South America and Hispaniola: names, types and a key. Sendnera  4: 207– 220. Weigend M, Ackermann M, Henning T. 2010. Reloading the revolver – male fitness as simple explanation for complex reward partitioning in Nasa macrothyrsa (Cornales, Loasaceae). Biological Journal of Linean Society  100: 124– 31. Google Scholar CrossRef Search ADS   Weigend M, Gottschling M. 2006. Evolution of funnel-revolver flowers and ornithophily in Nasa (loasaceae). Plant Biology  8: 120– 142. Google Scholar CrossRef Search ADS PubMed  Weigend M, Gottschling M, Hoot S, Ackermann M. 2004. A preliminary phylogeny of Loasaceae subfam. Loasoideae (Angiospermae: Cornales) based on trnL(UAA) sequence data and its relation to systematics and historical biogeography. Organisms, Diversity and Evolution  4: 73– 90. Google Scholar CrossRef Search ADS   Wittmann D, Schlindwein C. 1995. Melittophilous plants, their pollen and flower visiting bees in southern Brazil. 2. Loasaceae. Biociências  3: 19– 34. © 2018 The Linnean Society of London, Biological Journal of the Linnean Society This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Biological Journal of the Linnean SocietyOxford University Press

Published: May 24, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off