Journal of Plant Ecology Volume 8, Number 2, Pages 213–223 April 2015 doi:10.1093/jpe/rtv021 available online at www.jpe.oxfordjournals.org High ratio of illegitimate visitation by small bees severely weakens the potential function of heteranthery Jiao-Kun Li1,2, Yun-Peng Song3, Hui Xu1, Yan-Wen Zhang4, Jian-Yu Zhu1 and Lu-Lu Tang3,* 1 School of Minerals Processing and Bioengineering, Central South University, No. 932, South lushan Road, Changsha 410083, China 2 School of Metallurgy and Environment, Central South University, No. 932, South lushan Road, Changsha 410083, China 3 School of Life Sciences, Central South University, No. 172, Tongzipo Road, Changsha 410013, China 4 Department of Biology, Changchun Normal University, No. 3291, North Changji Road, Changchun 130032, China *Correspondence address. School of Life Sciences, Central South University, No. 172, Tongzipo Road, Changsha 410013, China. Tel: +86-13755137002; Fax: +86-731-82650230; E-mail: [email protected] Abstract Aims In heterantherous plants, ‘division of labor’ among structurally different stamens, i.e. pollinating and feeding functions, has been thought to reduce the evolutionary conflict of relying on pollen both as the carrier of male gametes and as the food for pollinators. The key to successful division of labor among different sets of stamens is the size match between stamens and legitimate pollinators, which results in the precise deposition of pollen onto specific locations on pollinator’s body and facilitates cross pollination. However, the potential impact of small illegitimate insects that are ubiquitous during the pollination process on the plant reproduction in heterantherous species has been largely neglected in previous studies and never been demonstrated experimentally. Methods Here, we investigated the functions of three different types of stamens in Commelina communis. The pollinator visitation, pollen removal and deposition were compared among flowers with different types of anthers emasculated at two natural populations. Moreover, the mating systems of C. communis in wild populations were estimated using microsatellite markers. Important Findings Our data showed that the main floral visitors for C. communis at the two studied populations were small illegitimate bees rather than legitimate pollinators, accounting for 77.5 and 92.2% of total flower visits, respectively. Flower manipulations in C. communis demonstrated that the two types of brightly yellow stamens separately functioned as ‘deceptive attraction’ and ‘feeding’ functions. Although the brown inconspicuous stamens of C. communis with the largest amount of fertile pollen had the potential function in offering pollen for cross pollination, the high ratio of illegitimate visitation by small bees significantly affected the dispersal and deposition of pollen from the pollinating anthers, and subsequently decreased the levels of outcrossing (tm = 0.23–0.32) in wild populations. Our work further confirmed that the size match between pollinators and the floral morphology is the prerequisite to successfully fulfill the functional differentiation among different sets of stamens in heterantherous plants. Local high ratio of illegitimate visitation by size unmatched insects could significantly weaken the potential functions of heteranthery, affecting the dispersal and deposition of functional pollen in heterantherous plants and further the whole mating systems. Keywords: Commelina communis, division of labor, heteranthery, illegitimate visitation, mating system Received: 14 January 2015, Accepted: 24 January 2015 Introduction In flowering plants, there are approximately 20 000 plant species that offer pollen as the only reward for pollinators (Buchmann 1983). However, if the pollen grains, the vehicle for the transport of male gametes, are consumed as food by pollinators, it will inevitably reduce the number of male gametes for cross fertilization and subsequently reduce the plant © The Author 2015. Published by Oxford University Press on behalf of the Institute of Botany, Chinese Academy of Sciences and the Botanical Society of China. All rights reserved. For permissions, please email: [email protected] 214 fitness. Thus, it is an evolutionary dilemma for these plants that they need to protect their precious pollen from the pollinators while maintaining enough attractiveness to pollinators. Heteranthery, the occurrence of two or more distinct types of stamens in the same flower, has been thought to reconcile the evolutionary conflict by allowing structurally different stamens to specialize in ‘pollinating’ and ‘feeding’ functions, respectively (‘division of labor’ hypothesis) (Darwin 1899; Müller 1881, 1882, 1883). The ‘feeding’ stamens mainly satisfy the pollinator’s demand for pollen as food, whereas the ‘pollinating’ stamens are directly involved in cross pollination and satisfy the plant’s need for safe transport. Heteranthery is commonly associated with bee-pollinated, nectarless flowers (Buchmann 1983; Endress 1994; Jesson and Barrett 2003; Vallejo-Marín et al. 2010; Vogel 1978). The stamens of heterantherous species usually differ in position, color, size and/or shape within a flower, with two types being most common (Barrett 2010; Vallejo-Marín et al. 2010). The first type of stamens (‘feeding’ stamens) is located at the center in the flower and composed of a few brightly colored anthers that are easily accessible to pollen-collecting visitors. The other stamen (s) (‘pollinating’ stamens) with cryptically colored anther (s) usually displaced away from the central axis of the flower and is primarily involved in promoting cross pollination. In addition, some species also possess a third type of stamens in the floral center, which are slightly larger or staminodes (Luo et al. 2009; Whalen 1979). Despite the ‘division of labor’ hypothesis has been substantiated in some species with dimorphic stamens and considered as the most reasonable explanation for the evolutionary significance of heteranthery (Graham and Barrett 1995; Jesson and Barrett 2003; Luo et al. 2008; Marazzi et al. 2007; Tang and Huang 2007; Vallejo-Marín et al. 2009), few empirical tests have been performed in species with trimorphic stamens. The function and evolutionary significance for the third type of stamens is still unclear. In addition, compared with other floral rewards, such as nectar (Cresswell 2001; Klinkhamer et al. 2001; Mitchell and Waser 1992 ) or heat (Li and Huang 2009; Seymour et al. 2003), it is much more costly and risky to use pollen as the only reward in heterantherous plants. If the flower visitors are legitimate pollinators that consume pollen but also transport it to stigmas of other plants, pollen consumption represents payment for services rendered and may not actually reduce the pollination efficiency. By contrast, if the pollen is mainly collected by illegitimate visitors without pollinating, pollen consumption can significantly affect the plant reproduction by reducing siring opportunity directly, or causing pollen limitation at the population level (Hargreaves et al. 2009, 2010). Unlike some flowers in Lecythidaceae (Mori et al. 1980), Fabaceae (Westerkamp 1997; Yeo 1993) and Polygalaceae (Westerkamp 1999) which can use hood-like petals or other protective structures to prevent access by small visitors, there is no special mechanism in flowers of heterantherous plants to avoid the illegitimate insects to access to the pollinating Journal of Plant Ecology anthers. In heterantherous species studied before, some small illegitimate bees has been found to land directly on pollinating anthers (Gross and Kukuk 2001), buzz only individual anthers and so do not contact stigmas (Renner 1983). Trigona bees have been found to chew poricidal anthers to access pollen, acting as pollen robbers in many Melastomataceae (Gross 1993; Renner 1983; Snow and Roubik 1987). In contrast to the extensive studies on the mutually beneficial relationship between legitimate pollinators and heterantherous plants (Luo et al. 2008; Tang and Huang 2007; Vallejo-Marín et al. 2009), the potential impact from illegitimate visitors on plant reproduction have been largely neglected and seldom demonstrated in heterantherous species. Here, we test the ‘division of labor’ hypothesis in Commelina communis, a species with trimorphic stamens, including two long stamens (L-anther), one medium length stamen (M-anther) and three short stamens (S-anther, staminodes) (Fig. 1a; Ushimaru et al. 2003a, 2003b, 2007). The flowers of C. communis do not produce nectar, and pollen is the only reward for floral visitors (Faden 1992). Unlike buzz-pollinated species that require vibration by bees to release pollen from poricidal anthers, the anthers of C. communis are longitudinal dehiscence, and therefore generalist insects, involving bees and flies have been reported as the main pollinators for C. communis (Faden 1992, 2000; Ushimaru et al. 2007). In experimental arrays, we systematically removed different types of stamens in flowers of C. communis and quantified the attractive roles of the three types of stamens through documenting pollinator responses in our experimental arrays. The subsequent pollen receipt and removal in flowers with different treatments were measured to test the effects of pollinator visits on male and female reproductive success. In addition, the variation in pollen viability among structurally different anthers of C. communis was investigated in our study. Taking advantage of recently developed microsatellite markers for C. communis (Table S1, see online supplementary material), we further estimated the outcrossing rates of C. communis in two wild populations based on multilocus genotypes in progeny arrays (Ritland 2002) and addressed the following questions: (i) Whether morphological differentiation among the three types of stamens in C. communis is accompanied by division of function? (ii) Whether illegitimate visitor exists during the pollination process and if so, how does it affect the mating system of C. communis? MATERIALS AND METHODS Study species and sites Commelina communis L. (Commelinaceae) is an annual, andromonoecious weed that often occurs around rice fields and along roads in temperate northeast Asia. Each plant usually has many inflorescences that contain both perfect and staminate flowers. The number of perfect and staminate flowers is equal and the flowers of both types open at sunrise each day and last only part of a day before fading (therefore their common Li et al. | Small bees weaken the function of heteranthery215 Figure 1: artificially manipulated flowers exposed to natural visitors. The abbreviated letter S, M and L in (a) indicate the three different stamens in Commelina communis, representing short, medium and long anther, respectively. name ‘dayflowers’) (Morita and Nigorikawa 1999; Ushimaru et al. 2003a, 2003b). The flowers of C. communis show bilateral symmetry and contain three types of stamens: two long, one medium and three short stamens. Long and medium anthers (L-and M-anther) produce fertile pollen, whereas short anthers (S-anther) produce only a small amount of sterile pollen (Hardy et al. 2009; Ushimaru et al. 2003a, 2003b). Meanwhile, C. communis is self-compatible and thought to exhibit delayed autogamy if the entomophilous pollination is not success at the end of anthesis (Morita and Nigorikawa 1999). Field observations and experiments were carried out from 10 to 20 July in 2011 at ‘Caizhuan’ population located in Changsha, Hunan Province, China (28°12′N, 112°55′E) and from 15 to 25 August in 2012 at ‘Shenghua’ population, about 30 km apart away from ‘Caizhuan’ population. Floral characteristics and sex allocation To explore the degree of size matching between pollinators and floral organs, 20 perfect flowers of C. communis were randomly collected from each population and the lengths of S-, M- and L-filaments were measured using digital calipers. Filament length was recorded as the distance between the pedicel–receptacle junction and the anther–filament junction (Ushimaru and Nakata 2002); style length was determined in a similar manner, from the pedicel–receptacle junction to the stigma. Meanwhile, the linear distances between stigmas and the three types of anthers were also measured, respectively (Fig. 1a). To examine the floral sex allocation, 25 individuals of C. communis were randomly selected in the ‘Caizhuan’ population. Perfect and staminate flowers in C. communis produce similar amounts of pollen (Morita and Nigorikawa 1999), and thus only perfect flowers were studied in this study. One perfect flower for each individual was collected before anther dehiscence in the early morning and stored in FAA solution (formal + acetic acid + 70% ethanol, at 5: 5: 90 v/v) until examination in the laboratory. The numbers of ovules and pollen grains per anther type were calculated. Each one type of anther per flower was split and pollen grains were suspended in 1 ml of water. The number of pollen grains in 0.05 ml pollen suspension sample was counted and nine repeat counts were performed under a microscope. For each flower, the total number of pollen grains per anther equaled the average number of pollen grains in 0.05 ml suspension sample multiplied by 20. Ovaries from all collected flowers were dissected to determine ovule number per flower with the aid of a stereomicroscope. Pollen viability of different types of stamens To compare the viability of pollen from different types of stamens in C. communis, we measured the pollen germination rate in vivo rather than in vitro and the pollen tube length after 2 h postpollination. Preliminary studies showed that the short anthers of C. communis produced only few malformed sterile pollen grains that are poor contributors for the siring of seeds (Hardy et al. 2009; Morita and Nigorikawa 1999; Ushimaru 216 et al. 2007). Therefore, only the pollen grains from M- and L-anthers were tested for this study. Both mature pollen grains of M- and L-anthers from 20 randomly chosen perfect flowers in ‘Shenghua’ population were separately hand pollinated onto the receptive stigmas of female flowers, which were emasculated all stamens and bagged before hand pollination. Pollinated stigmas were cut after 2 h and fixed in FAA solution for later microscopic observation. Pollen tubes were measured following the aniline blue method as described by Dafni et al. (2005). Pollinator preferences for heteranthery To explore the function of three different types of stamens in pollinator attraction in C. communis, we carried out experiments in two populations with > 100 individuals in 2011 and 2012, respectively. Four different types of flowers of C. communis were prepared artificially and exposed to natural visitors (Fig. 1): intact flowers as controls; S-removed flowers (flowers without S-anthers); M-removed flowers; L-removed flowers. Treatments were performed on five sunny days when C. communis was at peak flowering. Two arrays were established each day at c. 10-m intervals. In each array, eight perfect flowers were randomly subjected to the four treatments. All floral treatments were carried out in the early morning before anther dehiscence, and then all treated flowers were bagged until insects became active at c. 06:00 h. We tested responses of visitors to these different types of flowers by: (i) recording the flower type visited; (ii) recording the visitor behavior and the number of total visitor landings for each type of flower. Special attention was paid to whether visitors contacted the stigmas. Observations were continuously monitored for about 4 h every day and ended c.10:00 h, when no insect came to visit. All visitors were photographed and sample specimens were captured for identification. To examine the effect of pollinator visits on male and female reproductive success, pollen removal from brown L-anthers and yellow M-anthers and pollen receipt per stigma in flowers with different treatments were measured. After visitor activity ceased for 2 h, we collected the flowers that had been subjected to the above treatments. Brown L-anthers and yellow M-anthers were stored separately in FAA solution in centrifuge tubes and brought back to the laboratory. The number of pollen removed was equal to the mean number of pollen grains produced by L- and M-anthers (see ‘floral sex allocation’) minus the corresponding remaining pollen number. The pistil of each flower with S-, M- and L-removed was separately fixed in FAA solution and stained with aniline blue. Pollen grains deposited on the stigma were counted under a fluorescence microscope. Mating system Sampling design The fruit of C. communis is a dehiscent, ellipsoid capsule with two locules each containing two seeds (Hong and DeFillipps 2000).The mature seeds easily fell down to the ground from Journal of Plant Ecology the dehiscent fruits and were difficult to collect all seeds from per individual, therefore, we tried to sample four to seven mature seeds per plant from up to 20 randomly chosen individuals per population, with at least 10 m apart from each other. A total of 194 progeny plants and 40 maternal plans were included in the experiment. All seeds belonging to a mother plant were grouped as a family. Seeds from the maternal plants were geminated in soil in the glasshouse at Central South University. Fresh leaves sampled from maternal plants in the field and from young seedlings (2–3 weeks post germination) in the glasshouse were dried in silica gel. Total genomic DNA was extracted using a modified CTAB protocol (Doyle and Doyle 1990). Simple sequence repeat analysis For mating system analysis, 40 polymorphic microsatellite loci were developed as described by Zane et al. (2002). Of those 40 microsatellites, six were polymorphic and selected to use in this study (Table S1, see online supplementary material). PCR amplification conditions for all loci were as follows: A denaturing period of 5 min at 94°C, followed by 35 cycles of 30 s at 94°C, 45 s at 53–57°C (the melting temperature was specific for different markers), 45 s at 72°C, then 12 min at 72°C for final extension. The PCR reactions were performed in 20 μl of reaction volume containing 0.25 mM of dNTPs, 2 μl of 10 × Taq buffer (10 mM Tris–HCl (pH 8.3), 1.5 mM MgCl2 and 50 mM KCl), 0.2 μM of each primer, 2 units of Taq polymerase (TaKaRa Inc.) and 50 ng of genomic DNA. We checked PCR products with a 3% agarose × TBE electrophoresis, and sent them to DNA sequencing and services (Beijing Bio-ulab Co., Ltd, China) for fragment analysis in an ABI 3730xl capillary sequencer with a GeneScan 500 LIZ internal size standard (Applied Biosystem). Data analysis To analyze the effect of different types of stamens on pollinator visitation, pollen removal and pollen receipt, one-way analysis of variance (ANOVA) was performed. The differences in pollen viability and pollen tube length between the two types of stamens were compared using Student’s t-tests. All statistical analyses were performed using IBM SPSS Statistics Version 21.0 (IBM Corp., Armonk, NY, USA). Throughout this report, all means are accompanied by their standard deviation unless otherwise indicated. MICRO-CHECKER version 2.2.3 (van Oosterhout et al. 2004) was used to estimate the frequencies of null alleles in the microsatellite markers. The number of alleles per locus (Na), observed heterozygosity (Ho) and expected heterozygosity (He) were estimated in GENEPOP version 4.0.10 (Rousset 2008). The mating system was analyzed according to the mixed mating model of Ritland and Jain (1981) using MLTR software (Ritland 2002). The following parameters at population level were calculated: (i) multilocus outcrossing rates (tm); (ii) single-locus outcrossing rates (ts); (iii) the difference between multilocus and single-locus estimates of outcrossing Li et al. | Small bees weaken the function of heteranthery217 (tm − ts), that is often used to characterize the level of biparental inbreeding (the outcrossing resulted between closely related individuals) rates (Ritland 2002); and (iv) the correlation of outcrossed paternity within progeny (rp). In addition, the number of pollen donors contributing to each family was estimated as 1/rp (Ritland 1989). The inbreeding coefficient of maternal parents (Fm) was also calculated using the MLTR software (Ritland 2002). The maximum likelihood was used to estimate all the parameters with 1000 bootstraps. RESULTS Floral traits and sex allocation In C. communis, the filament of L-stamen was nearly as long as the style and was longer than that of the central S- and M-stamens (Fig. 1a; Table S2, see online supplementary material). Compared with the two types of yellow anthers, the lateral L-anther was much closer to the stigma (Fig. 1a; Table S2 , see online supplementary material). There was no significant difference on the number of pollen grains produced by per L- (1416.4 ± 256.2) and M-anther (1291.2 ± 288.2). However, the S-anther contained significantly less pollen grains (33.8 ± 22.6) than did both L-and M-anther (F2, 72 = 293.88, P < 0.0001). The number of ovules was four in each flower, and the pollen/ovule ratio therefore was 1031.0 ± 194.1. In vivo pollen germination rate and pollen tube growth The pollen grains from both M- and L-anthers have germinated on the receptive stigmas in C. communis, even only 2 h after pollination (Fig. 2a and b). The mean rates of germination for pollen grains of M- and L-anthers were 74.1 ± 23.1 and 74.9 ± 12.9%, respectively. There was no significant difference in germination rate between these two types of pollen grains (t = −0.14, P = 0.89; Fig. 2c). In addition, the mean length of pollen tubes at 2 h after pollination was 87.5 ± 15.9 μm for M-anthers and 86.0 ± 26.0 μm for L-anthers, respectively, and no significant difference was observed between these two figures (t = 0.22, P = 0.83; Fig. 2d). Pollinator behavior on different types of anthers Three types of insects, including Lasioglossum sp., Halictus sp. and syrphid flies were observed to visit the flowers of Figure 2: in vivo pollen germination rate and pollen tube growth on stigmas of Commelina communis after 2 h hand pollination. (a) pollen grains from M-anthers germinated on the receptive stigmas (M-stigma). (b) pollen grains from L-anthers germinated on the receptive stigmas (L-stigma). Values (mean ± SD) with the same superscript letters in (c) and (d) indicate that differences are not significant (P > 0.05). Abbreviations: pg (pollen grain) and pt (pollen tube). Scale bars 100 μm. 218 C. communis at both sites. Lasioglossum sp. was the most frequent visitor in our experiment, accounting for 77.5% (310) of total flower visits (400) in ‘Caizhuan’ and 92.2% (368) of total flower visits (399) in ‘Shenghua’. Flower visits by Halictus sp. and syrphid flies were much less frequent compared with Lasioglossum sp., accounting for 16.0 and 6.5% in ‘Caizhuan’ and for 3.5 and 4.3% in ‘Shenghua’. The straight body length (from the head to the tail) of Lasioglossum sp. was 6.27 ± 0.11 mm (n = 15), which was shorter than the distance between S- or M-anther and stigma (Table S2, see online supplementary material). Although Lasioglossum sp. usually alighted on the filaments of M-stamens and collected pollen grains from M-anthers, they rarely touched the stigmas of C. communis due to their small body size (Fig. 1b). Meanwhile, Lasioglossum sp. was also observed to directly collect pollen grains from the L-anthers (Fig. 1c). The visitation rates were similar in ‘Caizhuan’ and ‘Shenghua’ during 2011 and 2012 (Fig. 3a and b). Compared with intact flowers (Control), both S- and M- removed flowers received significantly fewer visits. By contrast, L-removed flowers received as many visits and were as attractive to pollinators as intact flowers (Fig. 3a and b; F3, 76 = 10.85, P < 0.01 for ‘Caizhuan’; F3, 76 = 3.42, P = 0.02 for ‘Shenghua’). There was no significant difference on pollinator preference between S- and M-removed flowers at both sites (Fig. 3a and b). Pollen deposition and removal In both populations, flowers with L-anther removed received significantly fewer pollen grains than did intact flowers, while removal of S- and M-anther had no effect on pollen receipt (Fig. 3c and d; F3, 76 = 7.11, P < 0.01 for ‘Caizhuan’; F3, 76 = 3.24, P < 0.01 for ‘Shenghua’). Moreover, almost all of pollen grains in both M- and L-anthers were removed by pollinators in two populations (Fig. 3e and f). The number of pollen grains removed from L-anthers was not significantly different between intact flowers and M- or S-anther removed flowers in two populations (Fig. 3e and f; F2, 57 = 0.54, P = 0.58 for ‘Caizhuan’; F2, 57 = 1.92, P = 0.16 for ‘Shenghua’). Similarly, there was no significant difference between the intact flowers and flowers with L- or S-removed in pollen removal from M-anthers in population of ‘Caizhuan’ (F2, 57 = 1.00, P = 0.37) and ‘Shenghua’ (F2, 57 = 1.90, P = 0.16). Outcrossing rate and mating system parameters The populations of C. communis studied here showed relatively low outcrossing rates (Table 1). The average multilocus outcrossing rate (tm) was 0.281 ± 0.066, ranging from 0.234 ± 0.065 in ‘Shenghua’ to 0.318 ± 0.112 in ‘Caizhuan’, indicating a mixed mating system with a predominance of self-fertilization in C. communis (Table 1). The estimates of tm − ts values were positive for the species (0.165 ± 0.039) and populations (0.141 ± 0.041 and 0.179 ± 0.060, respectively, for ‘Shenghua’ and ‘Caizhan’), indicating that biparental inbreeding contributed to the selfing rate of the C. communis populations (Table 1). The maternal inbreeding coefficient (Fm) Journal of Plant Ecology agreed with the other estimates indicating the presence of a relatively high inbreeding level in the species (0.365 ± 0.028) (Table 1). In addition, the correlation of outcrossed paternity within progeny (rp) in the two populations was very high (0.999 ± 0.738 and 0.837 ± 0.340, respectively for ‘Shenghua’ and ‘Caizhuan’), indicating the populations studied here possessed a low number of pollen donator parents (Table 1). Discussion Our data indicated that the three different types of stamens in C. communis were responsible for different functions and played different roles during the pollination process. The brightly yellow S-anthers with little sterile pollen mainly functioned as ‘deceptive attraction’ for pollinators (Hardy et al. 2009; Hrycan and Davis 2005) and the M-anthers with a large number of high viability pollen grains were considered as feeding anthers, providing pollen as ‘food’ for pollinators (Ushimaru et al. 2007). Although the brown inconspicuous L-anthers produced the largest amount of fertile pollen and had the potential role in offering pollen for cross pollination, the high ratio of illegitimate visitation by small bees could significantly affect the dispersal and deposition of pollen from the L-anthers in C. communis, decreasing the levels of outcrossing (tm = 0.23–0.32) in wild populations and severely weakening the potential functional significance of heteranthery. Functional differences among the three types of anthers The similar pollinator visiting frequencies between intact flowers and flowers without L-anthers in C. communis at two populations demonstrated that the brown L-anthers did not play a direct role in attracting the pollinators (Fig. 3a and b). By contrast, the brightly yellow anthers (S- or M-anthers) were the main factor to affect the pollinator visitation and functioned as flower attraction for pollinators (Fig. 3a and b). Although staminodes have been shown to have nutritive function in some species, offering food for pollinators (Ronse Decrane and Smets 2001; Vogel 1978; Walker-Larsen and Harder 2000), the total number of pollen grains produced by three S-anthers was very small in C. communis, only about one-thirteenth of that produced by one M-anther and no insects were observed to forage pollen from the S-anthers in our study. Because of the shorter filaments and nonviable pollen (Hardy et al. 2009; Morita and Nigorikawa 1999; Ushimaru et al. 2007), the S-anthers of C. communis were also unlikely to play a role in the pollination (Kaul and Koul 2008, 2012).Therefore, the nonrewarding S-anthers that were striking against the blue background of the corolla (Fig. 1) might mainly serve as ‘deceptive attraction’ to mimic large quantities of pollen and lead potential pollinators to the central anther where the food reward was (Hrycan and Davis 2005; Hardy et al. 2009). For Commelina flowers, lack of nectar is one of the most outstanding features, which significantly affects their reproductive Li et al. | Small bees weaken the function of heteranthery219 Figure 3: effects of three types of stamens on visitor responses (a, b), consequent pollen receipt (c, d) and pollen removal (e, f) in the two field populations of Commelina communis. The black and blank bars in (e) and (f) represent the number of pollen removal from L-and M-anthers, respectively. Values (mean ± SD) with different superscript letters indicate that differences are significant (P < 0.05) in pollinator visits, pollen receipt and pollen removal using a Tukey post hoc comparison. success (Morita and Nigorikawa 1999). Thus, the pollen must serve two functions, pollinating and rewarding the pollinators. Stamen placement within a flower determines whether pollen will be fed upon or used for pollination (Lloyd 2000; Luo et al. 2012). McCollum et al. (1984) called the central anther of C. erecta a feeding anther and proposed it to be analogous to nectar, serving only to provide a reward for pollinators and having very little to do with fertilization outside this role. Like other Commelina species (Faden 1992; Hrycan and Davis 2005; Oziegbe et al. 2013), the central M-anther occupied a similar position with S-anthers in the flower of C. communis (Fig. 1a). Meanwhile, the brightly yellow M-anthers that were attractive for pollen-collecting visitors (Fig. 3a and b) occupied a large number of fertile pollen grains (Fig. 2c and d), indicating their potential role in providing a nutritional reward for pollinators. There was no significant difference among the total number of pollen produced by M-anthers and that of removed pollen from M-anthers in intact flowers (t = 1.08, P = 0.30 for ‘Caizhuan’; t = −0.12, P = 0.90 for ‘Shenghua’), indicating that nearly all pollen in M-anthers was depredated by visitors (Fig. 1b). Thus, once pollinators were lured into a flower of C. communis, the medium stamens might primarily function as ‘feeding’ anthers to reward the pollinators (Ushimaru et al. 2007). 220 Journal of Plant Ecology Table 1 : mating system parameters in two Commelina communis populations: multilocus outcrossing rate (tm), single-locus outcrossing rate (ts), outcrossing rate between related individuals (tm − ts), correlation of paternity (rp), inbreeding coefficient of maternal parents (Fm) Parameters Number of families Caizhuan Shenghua Species 20 20 40 Number of individuals 100 94 194 tm 0.318 (0.112) 0.234 (0.065) 0.281 (0.066) ts 0.140 (0.056) 0.093 (0.027) 0.116 (0.030) tm − ts 0.179 (0.060) 0.141 (0.041) 0.165 (0.039) rp 0.837 (0.340) 0.999 (0.738) 0.956 (0.208) 1/rp Fm 1.19 0.392 (0.039) 1.00 0.341 (0.041) 1.05 0.365 (0.028) Data are means with standard errors in parentheses. The anther color is thought as another attractant signal for pollen-collecting insects by a virtue of being of a similar color to pollen and reflecting comparable UV patterns (Lunau 2000). Similar with other heterantherous species, the color of L-anthers in C. communis is more inconspicuous and cryptic than that of both S- and M-anthers, and hence less attractive for pollinators (Fig. 3a and b). For ‘division of labor’ hypothesis, the pollen from the less conspicuously colored stamens is not intended as a reward for pollinators and while legitimate pollinators are foraging in the central stamens, their abdomen will frequently come in contact with both the pollen-laden anthers and the stigma, facilitating outcrossing (Luo et al. 2008; Vallejo-Marín et al. 2009; Tang and Huang 2007). However, the main visitor of C. communis in our study is Lasioglossum sp. (accounting for 77.5 and 92.2% of total flower visits in ‘Caizhuan’ and ‘Shenghua’), which is a relative small bees to the flower and could not contact L-anthers or the receptive stigma when they fed on the central M-anthers in the flower at the same time (Fig, 1b). Thus, removal of S- or M-anthers that sharply declined the insect visitation did not significantly affect the pollen receipt in flowers of C. communis in our study (Fig. 3c and d). In addition, Lasioglossum sp. was also observed to collect the pollen from L-anthers (Fig. 1b) and there was no significant difference on the number of pollen removal from L-anthers between intact flowers and S- or M- removed flowers (Fig. 3e and f), indicating that Lasioglossum sp. was likely a pollen thief rather than a legitimate pollinator. Thus, although the inconspicuous brown L-anthers with the largest amount of fertile pollen had the potential pollinating function, local high ratio of illegitimate visitation by size unmatched bees severely impaired the dispatch and receipt of functional pollen in C. communis. The influence of illegitimate visitation on outcrossing rate in C. communis The ratio of pollen per ovule (1031.0 ± 194.1) in perfect flowers of C. communis was between the ranges set by Cruden (1997) for facultative xenogamy and xenogamy (796.6 and 5859.2, respectively), suggesting that pollinator mediated outcrossing was important for this species. Although we found that C. communis could be described as mixed-mating (0.20 ≤ t ≥ 0.80) (Schemske and Lande 1985), the average outcrossing rates in the two wild populations as measured by six simple sequence repeat loci were relatively low (tm = 0.23–0.32) and fell towards the low threshold of the mixed mating spectrum, suggesting self-pollination predominated the reproduction of this species here. The difference between multilocus and single locus outcrossing rates (tm − ts) represents an estimate of the biparental inbreeding degree, as in its presence ts will be smaller than tm because outcrossing events that are not detected at a single locus have a higher probability of being detected as more loci are examined (Ritland 1996).This difference was significant in both populations, being 0.141 ± 0.041 and 0.179 ± 0.060, respectively for ‘Shenghua’ and ‘Caizhan’, which indicates that of the fraction of multilocus outcrossing, around 14 and 18% has occurred between related individuals. The correlation of paternity gives the probability that a randomly chosen pair of progenies from the same array were full sibs (rp). This probability was very high for both populations of C. communis (0.999 ± 0.738 and 0.837 ± 0.340, respectively for ‘Shenghua’ and ‘Caizhuan’), indicating that most progeny derived from outcrossing within a family were full-siblings that shared the same male parents. Even though the progenies were derived from 20 maternal plants of C. communis in the two populations, only about 1–2 plants (1/rp) contributed as pollen donors (Table 1). Especially, the number of pollen donors contributing to each family (1/rp) was 1 for ‘Shenghua’ population, suggesting that absolute autogamy existed in this population, with the seed parent donating both male and female gametes (ovule and pollen grains). As observed by the maternal inbreeding coefficient and by the multilocus and single-locus rates, the mating systems C. communis were significantly affected by the limited functional pollen dispersion, which might be due to the high ratio of visitations by smaller illegitimate visitors. The possible reasons leading to self-pollination in C. communis During the legitimate visits of medium to large bees, heteranthery results in the deposition of pollen on different parts of the Li et al. | Small bees weaken the function of heteranthery221 pollinator’s body and therefore pollinator-assisted cross pollination firstly happens (Bowers 1975; Jesson and Barrett 2005; Lloyd and Schoen 1992; Vallejo-Marín et al. 2009). However, autonomous selfing may occur in C. communis when cross pollination fails. Many members of Commelinaceae, including C. communis have been reported to be self-compatible and a similar mechanism of style coiling causing autogamy was seen at the end of anthesis (Hrycan and Davis 2005; Oziegbe et al. 2013). In flowers of C. communis, the stigma and L-anthers are in closest proximity (Fig. 1; Table S2, see online supplementary material). Thus, the significantly reduced number of pollen on the stigmas in L-removed flowers may be due to the failure of contact between L-anther and stigma (Fig. 3c and d). In addition, the central anthers also have been considered to be responsible for autogamy in some C. species. The styles of perfect flowers of Commelina dianthifolia coil, brushing the stigma against the fertile central stamen and effecting pollination (Hrycan and Davis 2005). This mechanism also has been observed in C. erecta, C. lagosensis and C. diffusa (Oziegbe et al. 2013). The high viability of pollen from central stamens in C. communis, with equal viability as functional pollen, suggested that the central stamens might be also responsible for autogamy rather than just feeding pollinators. Furthermore, self-pollination within flower may occur when the smaller Lasioglossum sp. and other opportunistic insects illegitimately manipulate the flower (Renner 1989). on the plant reproduction in heterantherous species. Our data suggested that local high ratio of illegitimate visitation by smaller bees could significantly affect the dispersal and deposition of functional pollen and further the whole mating systems in C. communis, significantly weakening the potential functions of heteranthery. CONCLUSION Barrett SCH (2010) Darwin’s legacy: the forms, function and sexual diversity of flowers. Philos Trans R Soc Lond B Biol Sci 365:351–68. Functional division of different types of stamens into attraction and fertilization functions is one of the important ways to reconcile the dual function of pollen in heterantherous species. Evidently, the brightly colored stamens can be employed by plants to draw the attention of pollinators away from the inconspicuous and cryptic stamens, protecting the functional pollen from being consumed. Heteranthery with various forms occurs in diverse taxonomic groups and is phylogenetically associated with buzz pollination, enantiostyly and zygomorphy, indicating that it has probably evolved as a result of pollinator-mediated selection (Endress 1994; Graham and Barrett 1995; Jesson and Barrett 2003; Vallejo-Marín et al. 2010). The size match between pollinators and the floral morphology may be required to assure the precise pollen placement on the pollinator’s body, facilitating crossing pollination. However, pollen that provides a rich source of protein can also attract a variety of illegitimate insects, which may act as pollen thieves and cannot be excluded by the heterantherous plant itself. In this scenario, excess pollen consumption may be detrimental to plant fitness and thus there is no selection for anther specialization and dimorphism. In our study, the absence of legitimate pollinators in natural populations of C. communis that may be due to changes of pollination environment, provides us an opportunity to explore the potential influence from illegitimate visitors SUPPLEMENTARY MATERIAL Supplementary material is available at Journal of Plant Ecology online. Funding Postdoctoral Science Foundation of Central South University (502080023); National Science Foundation of China (31000106, 31100173). Acknowledgements We thank RL Guo, Y Gao, Y Zhang and L Fen for field assistance and YB Gong for data analysis, and SQ Huang for valuable comments on and improvements to the manuscript. Conflict of interest statement. None declared. References Bowers KAW (1975) The pollination ecology of Solanum rostratum (Solanaceae). Am J Bot 62:633–38. Buchmann SL (1983) Buzz pollination in angiosperms. In Jonesn CE, Little RJ (eds). Handbook of Experimental Pollination Biology. New York, NY: Scientific and Academic Editions, 73–113. Cresswell JE (2001) The influence of nectar and pollen availability on pollen transfer by individual flowers of oil-seed rape (Brassica napus) when pollinated by bumblebees (Bombus lapidaries). J Ecol 87:670–77. Cruden RW (1977) Pollen-ovule rations: a conservative indicator of breeding systems in flowering plants. Evolution 31:32–46. Dafni A, Kevan PG, Husband BC (2005) Practical Pollination Biology. Cambridge, UK: Enviroquest Ltd. Darwin F (1899) The botanical work of Darwin. Ann Bot 13:ix–xix. Doyle JJ, Doyle JL (1990) Isolation of plant DNA from fresh tissue. Focus 12:13–5. Endress PK (1994) Diversity and Evolutionary Biology of Tropical Flowers. Cambridge, UK: Cambridge University Press. Faden RB (1992) Floral attraction and floral hairs in the Commelinaceae. Ann Mo Bot Gard 79:46–52. Faden RB (2000) Floral biology of Commelinacea. In Wilson KL, Morrison DA (eds). Monocots: Systematics and Evolution. Melbourne, Australia: CSIRO Publishing, 309–17. Graham SW, Barrett SCH (1995) Phylogenetic systematics of Pontederiales: implications for breeding-system evolution. In 222 Journal of Plant Ecology Gross CL, Kukuk PF (2001) Foraging strategies of Amegilla anomala at the flowers of Melastoma affine – no evidence for separate feeding and pollinating anthers. Acta Hortic 561:171–78. McCollum TM, Estes JR, Sullivan JR (1984) Reproductive biology of Commelina erecta (Commelinaceae). In Horner NV (ed). Festschrift for Walter W. Dalquist in Horhour of His Sixty-Sixth Birthday. Wichita Falls, TX: Midwestern State University, Department of Biology, 57–66. Gross CL (1993) The breeding system and pollinators of Melastoma affine (Melastomataceae): a pioneer shrub in tropical Australia. Biotropica 25:468–74. Mitchell RJ, Waser NM (1992) Adaptive significance of Ipomopsis aggregate nectar production: pollination success of single flower. Ecology 73:633–38. Hardy CR, Sloat LL, Faden RB (2009) Floral organogenesis and the developmental basis for pollinator deception in the asiatic dayflower, Commelina communis (Commelinaceae). Am J Bot 96:1236–44. Mori SA, Orchard JE, Prance GT (1980) Intrafloral pollen differentiation in the new world lecythidaceae, subfamily lecythidoideae. Science 209:400–3. Rudall PJ, Cribb PJ (eds). Monocotyledons: Systematic and Evolution. Kew, UK: Royal Botanical Gardens, 415–41. Hargreaves AL, Harder LD, Johnson SD (2009) Consumptive emasculation: the ecological and evolutionary consequences of pollen theft. Biol Rev 84:259–76. Hargreaves AL, Harder LD, Johnson SD. (2010) Native pollen thieves reduce the reproductive success of a hermaphroditic plant, Aloe maculata. Ecology 91:1693–703. Morita T, Nigorikawa T (1999) Phenotypic plasticity of floral sex. In Ohara M (ed). Natural History of Flowers. Sapporo, Japan: Hokkaido University Press, 227–42. Müller H (1881) Two kinds of stamens with different functions in the same flower. Nature 24:307–8. Müller H (1882) Two kinds of stamens with different functions in the same flower. Nature 26:30. Hong DY, DeFillipps RA (2000) Commelina communis. In Wu ZY, Raven PH (eds). Flora of China. Beijing, China: Science Press, 36. Müller H (1883) Two kinds of stamens with different functions in the same flower. Nature 27:364–65. Hrycan WC, Davis AR (2005) Comparative structure and pollen production of the stamens and pollinator-deceptive staminodes of Commelina coelestis and C. dianthifolia (Commelinaceae). Ann Bot 95:1113–30. Oziegbe M, Akinlua IO, Olalekan AA (2013) Comparative pollination role of stamens and breeding system in three species of Commelina (Commelinaceae) in Ile-Ife, Nigeria. Acta Bot Bras 27:543–50. Jesson LK, Barrett SCH (2003) The comparative biology of mirrorimage flowers. Int J Plant Sci 164:S237–S249. Renner SS (1983) The widespread occurrence of anther destruction by Trigona bees in Melastomataceae. Biotropica 15:257–67. Jesson LK, Barrett SCH (2005) Experimental tests of the function of mirror-image flowers. Biol J Linn Soc Lond 85:167–79. Renner SS (1989) A survey of reproductive biology in Neotropical Melastomataceae and Memecylaceae. Ann Mo Bot Gard 76:496–518. Kaul V, Koul AK (2008) Floral phenology in relation to pollination and reproductive output in Commelina caroliniana (Commelinaceae). Aust J Bot 56:59–66. Kaul V, Koul AK (2012) Staminal variation and its possible significance in Commelina benghalensis L. and Commelina caroliniana Walter. Curr Sci 105:419–26. Klinkhamer PGL, de Jong TJ, Linnebank LA (2001) Small-scale spatial patterns determine ecological relationships: an experimental example using nectar production rates. Ecol Lett 4:559–67. Li JK, Huang SQ (2009) Effective pollinators of Asian sacred lotus (Nelumbo nucifera): contemporary pollinators may not reflect the historical pollination syndrome. Ann Bot 104:845–51. Lloyd DG, Schoen DJ (1992) Self-fertilization and cross-fertilization in plants. I. Functional dimensions. Int J Plant Sci 153:358–69. Lloyd DG (2000) The selection of social actions in families. III. Reproductively disabled individuals and organs. Evol Ecol Res 2:29–40. Lunau TDK (2000) The ecology and evolution of visual pollen signals. Plant Syst Evol 222:89–111. Luo ZL, Chen S, Zhang DX (2012) Floral reward presentation favored the expression of male function in the pollen-only flower Melastoma malabathricum. J Syst Evol 50:488–95. Ritland K, Jain S (1981) A model for the estimation of outcrossing rate and gene frequencies using n indepentdent loci. Heredity 47:35–52. Ritland K (1989) Correlated matings in the partial selfer, Mimulus guttatus. Evolution 43:848–59. Ritland K (1996) Estimators for pairwise relatedness and individual inbreeding coefficients. Genet Res 67:175–85. Ritland K (2002) Extension of models for the estimation of mating systems using n independent loci. Heredity 88:221–28. Ronse Decrane LP, Smets EF (2001) Staminodes: their morphological and evolutionary significance. Bot Rev 67:351–402. Rousset F (2008) GENEPOP’007: a complete re-implementation of the genepop software for Windows and Linux. Mol Ecol Resour 8:103–6. Schemske DW, Lande R (1985) The evolution of self-fertilization and inbreeding depression in plants. II. Empirical observations. Evolution 39:41–52. Seymour RS, White CR, Gibernau M (2003) Heat reward for insect pollinators. Nature 426:243–44. Snow AA, Roubik DW (1987) Pollen deposition and removal by bees visiting two tree species in Panama. Biotropica 19:57–63. Luo ZL, Gu L, Zhang DX (2009) Intrafloral differentiation of stamens in heterantherous flowers. J Syst Evol 47:43–56. Luo ZL, Zhang DX, Renner SS (2008) Why two kinds of stamens in buzz-pollinated flowers? Experimental support for Darwin’s division-of-labour hypothesis. Funct Ecol 22:794–800. Tang LL, Huang SQ (2007) Evidence for reductions in floral attractants with increased selfing rates in two heterandrous species. New Phytol 175:588–95. Marazzi B, Conti E, Endress PK (2007) Diversity in anthers and stigmas in the buzz-pollinated genus Senna (Leguminosae, Cassiinae). Int J Plant Sci 168:371–91. Ushimaru A, Itagaki T, Ishii HS (2003a) Variation in floral organ size depends on function: a test with Commelina communis, an andromonoecious species. Evol Ecol Res 5:615–22. Ushimaru A, Nakata K (2002) The evolution of flower allometry in selfing species. Evol Ecol Res 4:1217–27. Li et al. | Small bees weaken the function of heteranthery223 Ushimaru A, Itagaki T, Ishii HS (2003b) Floral correlations in an andromonoecious species, Commelina communis (Commelinaceae). Plant Spec Biol 18:103–6. Vogel S (1978) Evolutionary shifts from reward to deception in pollen flowers. In Richards AJ (ed). The Pollination of Flowers by Insects. London, UK: Academic Press, 89–96. Ushimaru A, Watanabe T, Nakata K (2007) Colored floral organs influence pollinator behavior and pollen transfer in Commelina communis (Commelinaceae). Am J Bot 94:249–58. Walker-Larsen J, Harder LD (2000) The evolution of staminodes in angiosperms: patterns of stamen reduction, loss, and functional reinvention. Am J Bot 87:1367–84. Vallejo-Marín M, Manson JS, Thomson JD, et al. (2009) Division of labour within flowers: heteranthery, a floral strategy to reconcile contrasting pollen fates. J Evol Biol 22:828–39. Westerkamp C (1997) Keel blossoms: bee flowers with adaptations against bees. Flora 192:125–32. Vallejo-Marín M, Da Silva EM, Sargent RD, et al. (2010) Trait correlates and functional significance of heteranthery in flowering plants. New Phytol 188:418–25. van Oosterhout C, Hutchinson WF, Wills DPM, et al. (2004) MICROCHECKER: software for identifying and correcting genotyping errors in microsatellite data. Mol Ecol Notes 4:535–53. Westerkamp C (1999) Keel flowers of the Polygalaceae and Fabaceae: a functional comparison. Bot J Linn Soc 129:207–21. Whalen MD (1979) Taxonomy of Solanum section Androceras. Gentes Herb 11:359–426. Yeo PF (1993) Secondary Pollen Presentation. New York, NY: Springer. Zane L, Bargelloni L, Patarnello T (2002) Strategies for microsatellite isolation: a review. Mol Ecol 11:1–16.
© Copyright 2026 Paperzz