Habitat fragmentation is one of the most serious threats to plant diversity. In general, fragmentation negatively impacts the genetic variability of plant populations due to increased random genetic drift, inbreeding, and reductions in gene flow. To investigate the effect of habitat fragmentation on genetic diversity and the mating system of Tsoongiodendron odorum, in this study, we analyzed genetic diversity and the mating system in hierarchical levels at the population, stands, and the individual scales in a fragmented T. odorum population. We sampled and mapped 61 adult individuals from the population. Using eight microsatellite loci, we genotyped a total of 780 seeds from 15 maternal trees for the mating system analysis. The results revealed moderate levels of genetic diversity in both adults (H_E = 0.522) and seeds (H_E = 0.499) with no significant differences between the two ontogenic stages. In addition, we did not observe a significant increase in the seeds inbreeding coefficient. Results from the multilocus mating system analysis indicated that T. odorum was an outbreeding species with a multilocus outcrossing rate (t_m) of 1.000. A small number of biparental inbreeding and correlated mating events were detected in this fragmented population. We found a small number of effective pollen donors (N_ep is between 3.7 and 5.4), which seems to be a common character of insect-pollinated canopy trees. Minor differences in outcrossing rates were detected among stands, and more pollen donors were found in smaller stands. However, outcrossing rate was significantly different among individuals, and a few selfing events were detected in some seed trees. These results may provide fundamental information required to establish long term conservation strategies for this endangered tree which is endemic to China.

Pollination has been recognized to be an important ecological process to maintain and promote biodiversity on earth. The order Hymenoptera, especially the superfamily Apoidea, is one of the most abundant and diverse groups among natural pollinators. However, studies on the ecological functions of wild bee pollinators have been underappreciated in China. More professional taxonomists are urgently needed to sur-vey fauna, to provide accurate identifications and to do comprehensive studies. Our understanding of nesting biology or pollination biology is not comprehensive for most wild bees. On the other side, much more atten-tion should be paid to the associations between bees and plants for their application to various projects on pollination biology, biodiversity or conservation biology in the future. We review current research on the status of wild bees, and we anticipate further applications to practices in agriculture and forestry with infor-mation from taxonomy, nesting biology and pollination biology of wild bees.

Complete understanding of floral function requires the recognition of floral traits at two aspects: floral design and floral display. Floral display, the fundamental unit of plant mating, refers to the number, type and arrangement of the open flowers on the plant in a certain period. Interactions between the flowers on a plant could influence pollinator behaviors on the plant and consequently may govern the mating outcomes. Pollinators prefer large floral displays, which often receive more visitations than smaller displays and thus could facilitate geitonogamous pollination. However, visitations for each flower do not increase with display size directly, suggesting that the variation of display size would balance the benefit of attractiveness against the cost of self-pollination. Flower morphs, such as unisexual or perfect flowers on one plant, may have dif-ferent gender function varying with pollinator preference and when or how much reward is available. Flow-ers on one plant also vary on size, color and reward to affect pollinator’s visitation behavior and minimize the potential self-pollination. Floral display often contains a three-dimensional floral arrangement that could in-fluence pollinator attraction and forage energy at a large scale, and is dynamic in response to the complex pollination and physical environments. Here we briefly review the progress of studies on floral display. It is clear that further studies on the interactions between floral traits and various pollination environments would provide insights into ecological functions of diverse floral displays in angiosperms.

The majority of flowering plants and crops rely in whole or part on animals for pollination. The mutualism between plants and pollinators has attracted ecologists and evolutionists to use this type of interspecific interaction as a model system to study species adaptation and diversification since Charles Darwin. Recent debate on the nature of pollination systems call for studies of this interaction at different levels, ranging from single species to entire communities in a given area. At the species level, detailed studies suggest that floral traits are under selection from mutualists and antagonists as well as the physical environment. In contrast, studies at community-level are rare, but recent analyses indicate considerable spatial and temporal variation in both generalized and specialized pollination systems. This special issue of Biodiversity Science focuses on plant-pollinator interaction, presenting current research status in this area from China. Papers include floral traits and pollinator behaviors addressed by phenotypic manipulation, estimates of pollen removal and receipt, anatomy of flowers, histochemistry analysis and spatial and temporal comparison. The taxa being investigated include wild orchid and cultivated legume, endemic, endangered and invasive species with diverse sexual systems. These thirteen experimental studies and three reviews show the development of pollination biology in China and expose how to facilitate our understanding of the critical ecological proc-esses underlying interspecific interaction in both natural and agricultural ecosystems.

Foraging behavior of pollinating insects can directly influence plant–pollinator interactions in many aspects, thus studies on pollinator behavior are important for understanding plant diversity and ecological processes of plant reproduction. In this paper, we describe the characteristics of major pollinating insects and discuss the methods for studying foraging behavior of pollinating insects and factors potentially influencing pollinator behaviors. We also suggest some practical methods for studying the roles of nocturnal pollinators and how to conduct studies of insect foraging behaviors at community level. The purpose of this introduction is to provide some useful information for future studies on this multidisciplinary research in China, particularly for botanists and entomologists.

Orchids are a flagship group in plant conservation. There are about 1,247 species and about 171 genera in China. The Chinese orchid flora is distinguished by having rich diversity in geographical types, especially by having a broad subtropical area lying between the Qinling Mountains and the Tropic of Cancer. This paper generally reviews the present situation of research and conservation of Chinese orchids. The publication of three volumes of flora of China (Orchidaceae), and some local flora of orchids, such as those for Hainan, Yunnan, Sichuan, and Guangxi Provinces, has laid down a profound foundation not only for correctly identifying and naming Chinese orchids but also for properly making conservation policy and strategies. Compared with relatively rich taxonomic knowledge, however, there is a wide gap in our knowledge of orchid ecology and many other aspects of orchid biology related to conservation. Five main priorities for education and scientific research for the long and short terms are suggested. First, strengthen and improve basic research such as ecology, biology and floristics. Second, identify and categorize Chinese orchid diversity hot spots at the country level. Third, promote study of ex situ conservation techniques, particularly developing artificial propagation methods for commercially important species, and starting reintroduction of selected endangered species. Fourth, develop public education programs through school lessons, the press, TV and radio. Finally, establish a scientific information pool at national level.

Flowering date along with interactions of other co-flowering species may be factors that determine pollination and reproductive success of a focal plant species. We investigated an alpine bumblebee-pollinated herb Pedicularis gruina to assess the effects of flowering date and co-flowering species on pollination, reproduction, and predation over a three year period (2003-2005). We analyzed pollen load of bumblebees to determine floral constancy, and calculated pollination rates of flowers, fruit set, seed production per capsule and the percentage of seed predation. We found that fruit sets did not differ significantly among years and between different flowering dates when there were no co-flowering species, with 34-38% of flowers producing fruits. However, the presence of co-flowering species could either increase or decrease the pollination and reproduction in P. gruina, which depends on pollinator type and flowering date of the co-flowering species P. densispica. Pedicularis densispica enhanced pollination and reproduction through increased floral resource diversity, whearas P. siphonantha enhanced pollination and reproduction through increased attractiveness with larger displays. Flowering date did not affect fruit set, but significantly affected seed production and seed predation. Seed production decreased significantly at later flowering dates given the same pollination condition(i.e., no co-flowering species, pollinator competition, or facilitative pollination). However, fruits produced in early flowering periods suffered more severe seed predation compared to those in later flowering periods.

Floral color and scent are two important components of floral design. Although previous studies on floral design have been mostly descriptive. The exploration on their quantitative studies can provide key evidence to learn pollination mechanisms. Our study focuses on recently developed methods for the quantitative measurement of floral color and floral scent. We introduce three floral color measurement and plotting methods, including color chart, colorimeter and spectrometer. Dynamic headspace collection-adsorption, adsorption-solvent desorption and solid phase micro-extraction are the three frequently-used floral scent collection methods and gas chromatography-mass spectrometry and zNose are usually used in floral scent analysis. Lastly, gas chromatography with electroantennographic detector, Y-tube olfactometer and flight cage experiment can help us to evaluate insect behavior.

Weather patterns in desert environments (e.g., dry heat, strong winds, dust storms, and temperature fluctuations) can be highly unpredictable and can have deleterious effects on sexual reproduction of flowering plants. The herb Limonium otolepis is a dominant species in desert areas of Xinjiang, northwest China; however, reproductive strategies in unfavorable environments remain to be explored. Overall, we addressed flowering pattern and pollination modes, flowering process, floral traits, pollinators, and fruiting characteristics of L. otolepsis in Junggar Basin region. We employed scanning electron microscopy to examine the morphology of pistil stigmas and pollen grains. The flowering period of L. otolepis was from early June to mid-July, with extended flowering due to unsynchronized flowering phenology among individuals in the same population. During a given day, it showed a concentrated flowering and pollen release, with anthesis lasting for about 6-8 h. Petals unfolded at about 8:00 and reached the peak of flowering at about 9:00, with no more flowers opening after 11:00. One flower produced 752 pollen grains with a small amount of nectar. The pollen exine ornamentation was reticulate. Insects, including bees and hoverflies, were primary pollinators. Highest visiting frequencies occurred between 9:00 and 14:00. L. otolepsis was mainly outcrossing given that automatic selfing yielded a few seeds. Overall, fruit sets of open-pollinated flowers were 36.1%. Within inflorescences, fruit sets on basal flowers (45.7%) were significantly higher than those on neighboring flowers (3.3%). Mast flowering in L. otolepis reported here may facilitate pollination in unpredictable weather conditions. The reproductive strategy may also dispense reproductive risk by maintaining flowering asynchrony to prolong overall flowering period.

Hermaphroditic plants that simultaneously display multiple flowers may attract more pollinators and create more chances for mating. However, this may also lead to geitonogamy, which may have major impacts on evolution of sex allocation, floral characteristics, and dioecy. To explore adaptive significance of mass-flowering in Hedychium coccineum, we conducted studies on pollination biology of H. coccineum, via hand-pollination and manipulated experiments, pollinator observation, and investigations on population density and natural fruit sets. Overall, inflorescences of H. coccineum were composed of terminal spikes containing 57.33 ± 1.68 (n = 30) cincinnus, and each cincinnus possessed 3.8 ± 0.15 (n = 30) flowers. Flowers in each cincinnus opened in turn, and the same round flowers among each cincinnus within inflorescences opened in successive synchronous cycles. This provided a spike that kept a mass of flowers blooming simultaneously throughout the flowering period. H. coccineum displayed a large floral display at the inflorescence level. Additionally, H. coccineum was self-compatible, and spontaneous self-pollination did not occur. Fruit production of H. coccineum in natural populations was severely pollinator-limited. Meanwhile, H. coccineum was also resource-limited in fruit production. Three butterflies, Papilio memnon, Dercas lycorias, and Appias indra aristoxemus were effective pollinators of H. coccineum. Visiting frequencies of these butterfly species to three inflorescences with different floral display size were obviously different and visitation frequencies increased as the inflorescence increased in floral display size. Pollen grains and ovules of H. coccineum varied significantly among flowers of different rounds in the cincinnus. Ovule production decreased significantly from the first to the fourth round, but pollen grains increased significantly from the first round to the third round. The decrease in ovules and resource limitation in fruit production indicated that mass-flowering in H. coccineum was not for increasing fruit production. The increase in pollen grains from the first round to the third round benefits pollen export and increases male fitness.

The diversification of floral traits is the outcome of evolution by natural selection, and the variation in floral traits between species has a certain correlation with the plant habitats. The northern tropical karst seasonal rain forests have a great habitat heterogeneity, a complex structure of plant community, and abundant endemic components. Analyzing the relationship between the variation of plant traits and their habitats in those forests is helpful to understand species coexistence, coevolution and their adaptations to habitats, and to provide clues for revealing the ecological adaptability of plants and the maintenance mechanisms of biodiversity in karst forests. We conducted a preliminary study at a 15 ha plot of the northern tropical karst seasonal rain forest in Nonggang to examine the differences in floral traits among species and how these traits varied with habitats, when a relatively large number of woody plants bloomed synchronously, with a total of 21 species being found flowering. We divided the 21 species into three preferred habitat types (peak, slope and valley) according to the spatial distribution of species and their habitat associations, and analysed the differentiations among their floral traits. We also divided 21 species into three clustering groups of floral traits, and had a comparison between the habitat types and the clustering groups. The results showed that species dominance had a significantly negative correlation with flower size and flower color vividness, which indicated that it should be a favored trait for trees to have small flower size or unattractive flower color in the community. Only the flower color lightness was significantly different among the three types, while other flower traits were not significantly different. However, there was weak consistency between the three types and the three main groups, with 57.14% species overlapped , which showed that habitats had something to do with flower traits differentiations. In conclusion, we believe that the floral traits are closely related to the dominance of species and habitats in the northern tropical karst forest. Moreover, floral traits might be more deeply affected by habitats than pollination in this region.

The highly specific mutualism between fig trees and their obligate pollinating fig wasps is usually exploited by non-agaonid wasps, and some of these wasps can enter and pollinate the figs just like the obligate pollinating wasps. Therefore, the agaonid and non-agaonid wasps have convergently evolved in their morphological characteristics and phenology. However, there are few data about the convergence of host recognition behaviors among these wasps. In Ficus curtipes, there are three internally ovipositing wasps, i.e. one obligate pollinating wasp, Eupristina sp., and two inquiline wasps (Diaziella yangi and Lipothymus sp.), which can also pollinate the figs if they enter the figs. In this study, we carried out several behavioral experiments with a Y-tube olfactometer to test the hypothesis of convergence of host recognition behaviors among these wasps. We observed and recorded the wasps’ behavior of choice among figs at different developmental phases and among 6-methyl-5-hepten-2-ol, 6-methyl-5-hepten-2-one and the mixture of the two chemicals. Our results showed that all three of the wasps were significantly attracted by the receptive F. curtipes figs when presented with choices between receptive figs and figs at other developmental phases and were significantly repelled by the male phase figs of F. curtipes when presented with choices between male phase figs and figs at other developmental phases. In addition, all of them also preferred to the compound 6-methyl-5-hepten-2-ol at the dose level of 1µL. These results provide evidence for the hypothesis of convergence of host recognition behavior among obligate pollinating fig wasps and non-pollinating fig wasps. The role of behavioral convergence in the evolution of non-obligate pollinating wasps into obligate ones is also discussed.

Mutualism, or a mutually beneficial interaction between two organisms, are ubiquitous in ecological systems. However, some “empty flowers”, which offer pollinators no any kinds of rewards, design different strategies to attract pollinators without providing rewards to the pollinators. These pollination mechanisms are called deceptive pollination. The family Orchidaceae, representing one of the largest groups in angiosperms, is distinguished by high floral diversity and intricate adaptations to pollinators. Darwin described and identified most of the functional floral morphology and biomechanics in orchid pollination. However, he never recognized that many of the flowers that he examined lacked food rewards for pollinators. Floral evolution in the Orchidaceae appears to be dominated by modes of deceptive pollination, and more than one third of orchid species are thought to be pollinated by deceit. Deceptive pollination is thought to be one of key roles which has lead to relatively high species diversity within the Orchidaceae. Deceptive orchids frequently exploit the food foraging, sexual, oviposition and sleep/warmth behaviors of insects. The most common deception mechanisms include generalized food deception, Batesian floral mimicry, sexual deception, brood-site imitation and shelter imitation. Additionally, floral color, morphology and fragrance play key roles to cheat target pollinators. Relationships between deceptive orchids and their pollinators possibly involve asynchronous evolution; therefore orchids track the diversification of their pollinators. However, deception has negative impacts on the pollinators, which may exert selection on the pollinators. Because duped pollinators tend to avoid rewardless flowers, deceptive orchids suffer low visitation rates and fruit set, various environment factors can affect the reproductive success of these orchids. Deceptive orchids depend largely on insect pollinators for reproduction, and the proclivity of these species to use deceptive pollination strategies puts many of these species at a relatively higher rate of extinction. Therefore, pollination biology studies are needed to provide a scientific basis for proper conservation of orchids. At the community level, co-occurring mimic, non-mimic, and mimic signal providing plant species affect the reproductive fitness of orchids. Therefore, it is necessary to further study the co-evolution webs of deceptive orchids and pollinators along with other related biological and ecological factors.

Variable insect behaviors may result in different flower visiting frequencies, which may result in diverse impacts on plant pollination. Although flies, hoverflies, and Italian bees (Apis mellifera ligustica) were the most common visitors of Chimonanthus praecox, these insect groups had disparate visiting behaviors. Flies traveled quickly, and crawled or jumped between flowers. Italian bees also flew quickly or occasionally crawled between flowers. In contrast, hoverflies generally spent a majority of time foraging on a given flower. Based on the average foraging time and inter-floral flight time, one Italian bee could visit 4.57 flowers per minute, whereas individual bees and hoverflies could visit 2.65 and 0.53 flowers on average, respectively. After incorporating the total number of individuals for each insect group, we extrapolated the visiting rates of Italian bees, flies and hoverflies as 498.19, 1,089.74 and 99.78 flowers per minute, respectively. Flies and Italian bees had higher pollination effectiveness (number of pollinated flowers by the same number of visitors in a given time interval under the same condition), with estimates of 93% and 100% (n = 30), respectively, while only 13% (n = 30) for hoverflies. If we consider the pollen load ability and pollination effectiveness for Italian bees and flies, visiting rate may be the main limiting factor for their potential in pollinating flowers. Thus following these conclusions, flies are the major pollinators of C. praecox, followed by Italian bees (pollination role was limited by their low number) and hoverflies (limited by their foraging and hovering behaviors and lower pollination effectiveness).

Thermopsis lanceolata is sand-fixing plant that plays an important role as a spring nectar source in northeastern China; however, little is known about its reproductive characteristics. To ascertain the major pollinators of T. lanceolata, we identified foraging behaviors, pollination process, activity rhythm of floral visitors and the effect of nectar robbers on the seed setting were investigated in Mu Us Sandland, Inner Mongolia. We used photo, video and visual measurement to observe floral visitors in a 10 m × 10 m quadrat. We recorded the species, pollinating behaviors and visiting frequency of all pollinators from 06:00 hours until 19:00 hours in sunny days. Based on visitation frequencies and pollen amounts carried on the body, two bee species, Megachile japonica and Hoplitis princeps were identified as the primary pollinators of T. lanceolata., but visiting frequencies between the two species were significantly different. Daily activity of M. japonica was bimodal with diurnal foraging activities during 9:00-13:00 and 16:00-18:00 hours, whereas daily activity of H. princeps was of unimodal type from 11:30 to 16:30 hours. We also noted nectar robbing behaviors by M. japonica due to the presence of nectar acquisition holes at the base of the flower of T. lanceolata, but found that these behaviors had no effect on seed setting. Pollinator activity peaks indicate a complementary relationship between M. japonica and H. princeps. Our observations of visiting frequencies suggest that M. japonica was a dominant pollinator for T. lanceolata.

Classic size-dependent sex allocation models for animal-pollinated plants predict that female allocation increases with plant size. However, some studies have shown a male biased size-dependent sex allocation at flowering stage. We investigated the size-dependent sex allocation strategies within flower and inflorescence in gynomonoecious Aster ageratoides within two naturally-occurring populations in which hermaphroditic and female flowers occur in the same individual. At the flowering stage, male biased size-dependent sex allocation has been observed at both flower and inflorescence level. The pollen/ovule ratio ranged from 750 to 5,250, with a mean of 2,757.14± 206.635, and increased significantly with plant height in one of the populations. At the inflorescence level, the male allocation, which was measured by the proportion of hermaphroditic flowers, increased with plant height in both populations. These results are consistent with the expectation of male biased size-dependent sex allocation at flowering stage, through which A. ageratoides may maximize fitness by enhancing pollen dispersal.

The relationship between sexual reproduction and resource allocation plays an important role in plant life history and in the evolution of breeding systems. Tulipa sinkiangensis is an early spring perennial ephemeral species endemic to the desert zone of the northern piedmont of the Tianshan Mountains, Xinjiang, China. This species produces offspring only by sexual reproduction, and produces 1-8 flowers per individual in natural populations. The relationships between sexual reproduction allocation and vegetative growth and plant size, respectively, and resource allocation among flowers and fruits at different positions within the inflorescence were studied in T. sinkiangensis. Our aims were to explore the effect of vegetative growth, plant size and flowering order on sexual reproduction allocation in this species. There was a negative correlation between the resources that plants allocate to both vegetative organs (bulb and aboveground vegetative organs) and sexual reproductive organs at flowering stage and fruit maturation stage (P<0.01), suggesting that resource allocation between vegetative growth and sexual reproduction is a trade-off. Production of multiple flowers is a stable character in this species―the total number and biomass of flowers, total fruit biomass, and total number of seeds per individual were positively correlated with plant biomass (P<0.01), indicating that sexual reproduction allocation is size-dependent. In individuals with 2-5 flowers, the biomass of flowers, pollen and ovule production, fruit-set, fruit biomass, seed number, seed-set, and 100-seed weight declined successively with flowering order within the inflorescence, indicating that resource competition for floral allocation was significant and that resource limitation for each flower or fruit is related to flowering order. Plants ensure their reproductive success by reducing resource allocation to late flowers or fruits and increasing the resource allocation to early flowers or fruits in this species.

Scented nectar secreted by flowers is an interesting natural phenomenon in angiosperms that is often overlooked by most researchers. Scented nectar is an evolutionary feature coupling the behavior between flower-visitors and their rewards. The effect of scented nectar on pollinator or nectar-thief is a scientific problem which deserves further exploration. We evaluated Mucuna sempervirens flowering dynamics, floral nectar volatiles, behaviors of Dremomys pernyi and Callosciurus erythraeus to inflorescence opening, nectar’s attraction effect to Apis cerana cerana, and the toxic effect of M. sempervirens nectar to Tapinoma sp. Aliphatic compounds (87.2%) were the main volatiles in nectar of M. sempervirens, whereas ketones accounted for 56.1% of the total volatiles. Lastly, volatile components with sulfur elements were absent from nectar samples. Previous studies of bat pollinated plant species have found that plants in the genus Mucuna always release sulfur compounds. Additionally, we found that the nectar of M. sempervirens was poisonous to Tapinoma sp. ant species, whereas it attracted A. cerana cerana using the Y olfactory bioassay. Squirrel species (D. pernyi and C. erythraeus) were effective pollinators of M. sempervirens, so we suspected that M. sempervirens might not be pollinated by bats only. This study provides data of pollination mechanisms in Mucuna and provides a new perspective of composition and function study of nectar in other plant groups in future.

The theory of sex allocation can be used to predict the optimal allocation of reproductive resources and considered as individual fitness between male and female function. Variations in resource allocation between sexes and among different positions in the flower inflorescence have great significance for understanding how choice of reproductive strategy affects reproductive success. Changes in floral sex allocation among different positions, pollinator movement and reproductive success within sequentially flowering inflorescences of Ammopiptanthus mongolicus were studied at the Turpan Eremophytes Botanical Garden of the Chinese Academy of Sciences. Our major results are as follows: (1) Floral longevity shortened significantly, stamen mass/(stamen mass + pistil mass), pollen number and pollen/ovule ratio all increased among positions along with the turn of intermediate earlier developing flowers→lower flowers→upper later developing flowers. Corolla diameter, corolla mass and nectar production, on the other hand, all diminished and ovule number showed no differences. Upper flowers showed male-biased sex allocation; (2) Within an inflorescence, Apis mellifera and Lasiglossum sp.1 landed first on the intermediate flowers, moved among different positions in the inflorescence, and finally flew away from the upper flower. Intermediate flowers had the highest first visiting rate but lowest last visiting rate while upper flowers had the lowest last visiting rate but highest first visiting rate; (3) During the two years of our study, the seed set rate and seed weight increased after supplying outcross pollen to the upper flowers; the fruit set rate, seed set rate and seed weight all increased significantly after simultaneously supplying outcross pollen to the upper flowers and removing the intermediate and lower flowers, and these measures did not differ between the two treatments. These results suggested that lower fruit set, seed set and seed weight of upper flowers under natural conditions can be attributed to the pollinators’ directionality within inflorescence and a lack of outcrossing pollen in upper flowers rather than resource limitation in A. mongolicus. Increasing investment in male functions on the upper fowers may be an adaptive strategy to sustain pollination success for the species.

Heterostyly is a genetically controlled floral polymorphism, which includes both distyly and tristyly. We investigated morph ratio, floral and pollen morphology, and self-incompatibility of Hedyotis pulcherrima. Overall, the natural population of H. pulcherrima was isoplethic, containing long-styled and short-styled morphs with an equilibrium 1:1 ratio. Long-styled and short-styled morph exhibited a precise reciprocal herkogamy, which was significantly correlated with corolla length. Stigma-lobe length, pollen size, and starch content in pollen grains were dimorphic in the two morphs of H. pulcherrima, whereas pollen germination and pollen tube growth in vitro were not significantly different between the two morphs. Artificial pollination revealed that pollen tube shape was normal in both morphs where pollen tubes reached the ovary 24 h after pollination. However, pollen tube growth was arrested in the stigma with the accumulation of callose in the swollen tips in two morphs with self and intramorph pollination, indicating strict heteromorphic self-incompatibility in H. pulcherrima. No fruit was produced in emasculated netted flowers, suggesting the absence of apomixis. Artificial intermorph pollination resulted in 100% of fruit set, significantly higher than those with open pollination. Our results indicate that H. pulcherrima is a typically distylous species with heteromorphic self-incompatibility.

Interactions among co-evolved species has been assumed to operate in a symmetrical manner, resulting in stable equilibrium or evolutionary stable strategies for the observed species. However, recent observational and experimental data highlight the existence of asymmetrical interactions, which may lead to meta-populations or non-equilibrium states (e.g. chaos) between co-evolved species. Here, we explore the asymmetrical interaction in the classic ‘fig-fig wasp’ co-evolution model system and show how such asymmetric interactions may produce meta-populations. It is well known that there are two different kinds of fig wasps inhabiting fig fruits (i.e. pollinator wasps and non-pollinating exploiter wasps). Exploiter wasps have a fitness advantage over pollinator wasps because they do not pay the cooperative cost. However, figs can effectively restrict exploiter wasps, and reward the pollinator wasps, resulting in complex asymmetric interactions among species. Specifically, the discriminative restriction of paretic wasps by fig trees will cause drastic population decreases or even local extinction of the exploiter species in some fig trees, syconia, or associated habitats. In patches where populations of exploiters are low or extinct, pollinator wasps will immigrate leading to concomitant population size increases due to the high reward of fig trees. The prosperity of pollinator wasps will then attract more exploiters, and population increases of exploiter wasps will lead to the sanction by fig trees again. Over the long term, populations of different wasps will chaotically oscillate either temporally or evolutionary through asymmetric interactions.

We measured floral traits of Trollius ranunculoides at different flowering time, observed the taxa of pollinators and pollinator visiting frequencies, and collected the seeds in the end of growing season to estimate pollinator preferences and the overall phenotypic selection gradient. We estimated phenotypic selection on floral traits at different flowering time using female fitness (seed number). We found significant differences of floral traits among different flowering time, with the corresponding preference, taxa, and visiting frequencies of pollinators also different. During early flowering time, visiting frequencies of bees were higher, flower calxy and petal size were wider, and total stalk length was shorter on flowers that bees preferred; however, there was no significant flower preference by different fly species. Furthermore, our estimates of phenotypic selection through female fitness differed at variable flowering time. The results of our study indicate that divergent flowering time and corresponding variation in pollinators might lead to different phenotypes, ultimately resulting in the divergence of floral traits. However, pollinator preference was dissimilar from phenotypic selection estimated by female fitness due to reasons not sufficiently evaluated in this study.

One of the main purposes of scientific research is to form questions and collect data to find patterns that best explain a particular process. In plant reproductive ecology, scientists have found important patterns including overwhelming occurrence of co-sexual flowers, separate flowering and fruiting periods, low seed-set ratio, simultaneous mass blooming, high occurrence of outcrossing in harsh habitats, interactions of mating systems and life-history, flower differentiations on the same inflorescence, and heteranthery (stamen differentiations in same flower). The underlying mechanisms of these patterns are in debate and are still in need of further examination. In this paper, we emphasize resource allocation at different levels, i.e. individual, inflorescence, and flower, and try to explain above patterns to reconcile some contradictory results obtained in studies based on different foci. Resource and sexual allocations directly determine the sexual role of a flower and floral traits, and as a consequence, plant mating systems are determined. In the future, studies considering both ecological factors and hierarchical levels will become more important and it is essential to focus more effort to study resource allocation to gain greater understanding and accurate predictions of floral adaptations and evolution.

Crocus alatavicus, a geophytic-geocarpic plant, is an early spring ephemeral species that grows in subalpine areas of the western Tianshan Mountains. To understand the ecological significance of geocarpy in plants, we studied the breeding system and pollination biology of C. alatavicus, with special reference to its adaptive strategies for a subalpine existence. We found that C. alatavicus is hysteranthous and its flowering pattern was explosive at the population level. It began to flower in early- to mid-April, and the white flowers had neither nectar nor scent. The inferior ovary was below ground at anthesis, while the other floral parts were aboveground. The flowers were open during the day and closed at night, and individual flower duration was 6-9 days. Pollen viability was 75.39±5.69% at the end of anthesis, and stigma receptivity lasted eight days. Results of artificial pollination experiments suggested that the breeding system ofC. alatavicusis facultative xenogamy with the ability to self-pollinate spontaneously. Crocus alatavicushas a generalist pollination system, with Bombus lucorum, Anthophora senilis and Andrena capillosa being effective pollinators. These insects initiated pollination while foraging for pollen, and their visitation frequencies were 0.50±0.27, 0.18±0.08 and 0.13±0.05 per flower per hour, respectively. Thus,C. alatavicusnot only has evolved unique flowering pattern and generalist pollination system, but also utilizes an otherwise vacant niche in the early spring to improve the effectiveness of pollination. Furthermore, characteristics of its breeding system, such as self-compatibility and spontaneous self-pollination, ensure reproductive success even when pollinators are scarce and pollinator visitation frequencies are low in early spring.

The orchid family is renowned for its enormous diversity of pollination mechanisms and unusually high occurrence of non-rewarding flowers compared to other plant families. We investigated the pollination biol-ogy of Cymbidium lancifolium during 2005 and 2006 in the Yachang Nature Reserve, Guangxi Province, SW China. Our observations showed that Apis cerana cerana (Hymenopterous, Apidae) was the only pollinator. The bee directly landed on the mid-lobe, then adjusted its direction and entered into the flower. When it found no rewards in the flower, it would exit from the flower with the hind legs forcibly grasping the mid-lobe. The pollinaria together with the anther cap adhered to the thorax by the viscidium during the re-treating process of the bees. The pollinator stayed for 8–71 s in the flower with an average of 18.3 s (N = 11). It is likely that C. lancifolium attracted the bee by the purplish chestnut spots on the labellum (false nectar guides) exploiting the foraging preference of bees. The anther cap that remained on the back of the pollinator blocked the pollinaria from being received by the subsequently visited flower. The height of the anther cap (0.154 ± 0.032 cm) (N = 10) together with thorax of the pollinator (2005: 0.37 ± 0.03 cm (N = 10), 2006: 0.35 ± 0.04 cm (N = 7)) was higher than the entrance of the flower (0.29 ± 0.04 cm) (N = 21), suggesting a possi-bility of anther cap retention for geitonogamy. The reproductive success in the population was respectively 21.13% and 21.28% in 2005 and 2006. Breeding experiments showed that C. lancifolium was self-compatible and there was no significant difference in reproductive success between self-pollination and cross-pollination. There was neither apomixes nor spontaneous autogamy. The pollination success of this species was depend-ent on pollinators. The viability of seed by TTC was estimated as 85.78% (N = 11), and it was not considered to be the key limiting factor for seed germination. Therefore, it is suggested that other factors, such as the density of the pollinators and their pollination frequency, influence the fruit set and maintenance of the popu-lation of C. lancifolium.

Primula merrilliana, a rare and endangered National Grade Ⅲ protected plant, is endemic to Anhui Province, China. We studied its reproductive characteristics from the aspects of phenology, breeding system structure and pollination. Our observations showed that: (1) P. merrilliana is typically distylous in both mor-phology and function, and it is a biennial plant that grows from September in the first year and lasts until late June or early July of the next year. In addition, the long-styled plants tend to flower 3 to 5 days earlier than the short-styled morphs in each population; (2) The pantoporate pollen of P. merrilliana shows that the spe-cies has vital taxonomical and evolutionary significance in Primula. The short-styled pollen diameter is 1.78 times as wide as the long-styled, but the pollen production of a short-styled flower is less than that of a long-styled flower; (3) The long-styled pistil is about 1.64 times as long as that of the short-styled morph with shorter and more frequent papillae; (4) The pollen-ovule ratio (P/O) is different in the three studied populations, and we noted larger variation in pollen than ovule production, which may be related to habitats; (5) P. merrilliana is an outcrossing plant pollinated by thrips, and the seed-set of legitimate pollination is higher than that of illegitimate; and (6) The seed morphology of P. merrilliana usually appears like a non-equilateral heptahedron, a few seeds resembles an irregular polyhedron, but there is no obvious difference between the two morphs, the long-styled and the short-styled. Based on this study, biased morph frequencies in small populations of P. merrilliana, destruction of habitats and disappearance of accompanying species are possibly the main reasons for the endangerment of this species.

During May to July, 2006 and April to May, 2007, we studied pollination biology in Ipomoea cairica, an invasive weed in Guangdong Province, China. Ipomoea cairica is a perennial creeping or climbing herbaceous vine, blooming all year round in Guangdong. The flowers gathered in cymes, with a purple or bluish purple bell-formed corolla. The petals unfolded at about 4:30–5:20 and closed at 17:40, lasting for about 12 hours. The proximity of the stigma and anthers during flowering period facilitated the transfer of self-pollen. The mean volume of nectar per flower was 5.82 ± 2.47 μL (n = 10), with a sugar concentration of 41.03 ± 6.52% and an amino acid concentration of ca. 0.042 mg/mL. During anthesis, both pollen viability and stigma receptivity remained at a high level (>90%). The mean P/O ratio was 4,778.18 ± 264.58 (n = 10). Carpenter Bees (Xylocopa sp.) were considered as effective pollinators, while flies and butterflies were occa-sional visitors. Ipomoea cairica is self-incompatible as determined by an absence of fruit set in the flowers following controlled hand self-pollination, producing fruits and viable seeds only after cross-pollination.

Nectar guides are special floral structures that are assumed to attract and/or direct pollinators towards a concealed nectar reward. Nectar guides are commonly found in many angiosperm plants, especially entomophilous flowers, but their functional significance and impacts on the reproductive success of plants are poorly understood. For example, Delphinium caeruleum of the family Ranunculaceae possesses two barbate staminodes, between which is the nectar entrance. To test the effects of these barbs on male and female fitness, we experimentally investigated the effects of barb removal on seed production and pollen exportation of D. caeruleum at the Haibei Station of Qinghai Province. Our results showed that flowers of D. caeruleum were dichogamous and herkogamous, which prevented autonomous selfing in this species. The mating system of D. caeruleum belongs to obligate xenogamy and thus pollinators are necessary for the successful pollination. Compared to control flowers, the number of pollen grains exported was reduced significantly after removing the barbs, but fruit set and seed number was not affected. Our primary results indicate that the barbs on the staminode act as nectar guides, which improve the male fitness of D. caeruleum, but have little effect on female fitness. Our results suggest that selection through male fitness might have played an important role in the floral evolution of D. caeruleum at our study site. These results need to be further investigated at a larger number of sites to evaluate selection mechanisms on floral traits of D. caeruleum.

Self-incompatibility occurs widely in the Solanaceae, but a few species in this family, or a few intraspecific populations or a few individuals within the populations are highly self-compatible and their breeding systems have evolved from outcrossing to selfing. We studied pollination biology of Anisodus tanguticus from this family, a perennial endemic to the Qinghai-Tibetan Plateau. This species flowers in the early summer and probably suffers from the serious limitations of the outcrossing pollinator. We aimed to test whether the breeding systems of a few individuals in this species have shifted from outcrossing to selfing. Our experiments and observations were conducted in two populations with different altitudes. Most flowers of this species were incompletely protogynous, and the mean distance between the stigma and anthers decreased gradually with floral longevity. Both stigma and anthers did not come into touch when the corolla wilted in most of the monitored flowers and such herkogamy favours outcrossing. However, the complete contact of stigma and anthers in 4.9% of the measured flowers suggested potential occurrence of “autonomous selfing” in these flowers. Most of the examined individuals in both populations are self-incompatible and have to depend on insects for pollination. However, a few of them were found to be self-compatible. Flies were the effective outcrossing pollinators in the high altitude population and “autonomous selfing” was detected in a few individuals of this population. In contrast, the effective outcrossing pollinators were rarely observed in the low altitude population and the dominant visitors or pollinators, ants in this population, transferred pollen within the single flower, which finally resulted in “facilitated selfing”. Pollination limita-tions were obvious in both populations due to the lack of enough outcrossing pollinators. Two different self-ing mechanisms, uncovered here, may provide a partial compensation for the natural reproduction of this al-pine species in the arid alpine habitats when flowering at the early summer.

Flexistyly is a unique sexual dimorphic system found in Amomum and Alpinia species of the ginger family (Zingiberaceae). The populations of flexistylous species have two phenotypes, named an anaflexistylous morph and a cataflexistylous morph, and all individuals of both morphs separate their male and female functions spatio-temporally. We conducted manipulated pollinations and pollen tube growth experiments on Alpinia blepharocalyx and A. galanga to detect the manner of separation of male and female functions within the individual and its adaptive significance. The results showed that the outcrossing rates of manipulated and natural pollination in the cata-morph did not differ significantly (P>0.05). However, the number of seeds per fruit of manipulated cata-morphs was significantly lower (P<0.01) than that of control individuals, perhaps due to the inbreeding depression caused by ovule discounting. Pollen tube growth experiments showed that, when stigma were located at the receptive position (ana-morph in AM, cata-morph in PM), stigma provided appropriate conditions (had stigmatic secretion) for pollen grain germination, and pollen tubes penetrated into the style within 2 hours after pollination, regardless treatment of selfing or outcrossing. However, when stigma were beyond the anther (ana-morph in PM, cata-morph in AM, without stigmatic secretion), it usually took 6–10 hours for pollen germination and pollen tube penetration. Pollen tubes, however, could reach the ovary within 24 hours under both treatments. Hand-pollination also showed that pollen grains of anaflexistylous flowers have matured before the dehiscence of pollen sacs. Our research suggests that flexistyly is a floral dimorphism comprising reciprocal mobile herkogamy and heterodichog-amy. Heterodichogamy encourages outcrossing, meanwhile reciprocal curvatures of stigmas play a role of reducing interference between male and female functions.

Investigation of intraspecific morphological variation is critical for the understanding of evolutionary and ecological processes, and may provide insights for uncovering the roles of natural selection, gene flow, and genetic drift in the distribution and abundance of species diversity. In the present study, we analyzed statistically the patterns of variation of 13 morphological traits among 12 populations in three localities of an endemic orchid, Changnienia amoena. The results showed that, among the three localities, the mean value of all measured morphological traits is the highest in the Lushan populations. There is abundant variation at both the species and population levels. Coefficient of variation (CV) varied between 0.02 and 0.30 at the species level, and the CV values of reproductive organs were lower than those of vegetative structures. ANOVA analysis showed that there was significant difference in most morphological traits among the three localities. UP-GMA analysis showed that the populations in three localities were clustered into three main clades, the populations from Xinning and Lushan formed one clade, while the Shennongjia populations formed another clade. It is noteworthy that the Shennongjia populations were separated into two subclades corresponding to the populations at Longmenhe and Guanmenshan, respectively. This morphological differentiation is most likely a result of pollinator-mediated selection because this species is pollinated by different bumblebee species at these two sites. Pearson correlation analysis showed that high correlation occurred within the vegetative organs, reproductive organs, and be-tween them. These results have important implications for our understanding of the evolutionary processes and mechanism of C. amoena and might provide useful information for the conservation and management of this endangered orchid.

The “Malpighiaceae route” is proposed based on the distribution pattern of the family Malpighiaceae to explain plant inter-continent disjunctions and long-distance dispersal during historical periods. The route involves seven inter-continent dispersals from New World (America) to Old World (Africa and Asia) in the Eocene (~ 65 Ma). Malpighiaceae has about 1,300 species, most of which are endemic to the New World and are characterized by “floral conservatism”. Floral conservatism in this family refers to its stereotyped yellow or pink flowers with paired oil-glands on each sepal, a result from co-evolution with specialized oil-collecting Anthophoridae bees. These bees, however, are absent from the Old World. In the African Malpighiaceae, floral conservatism disappeared as a result of adaptations to local pollen-collecting bees. Their flowers became white and radially symmetric, without sepal gland. Furthermore, the floral sex changed to be morphological androdioecy but functional dioecy. These results indicated that African Malpighiaceae had shifted from floral conservatism to generalized floral syndromes. In Asian Malpighiaceae, some studies reported in Aspidopterys and Ryssopterys generalized evolutionary adaptations in floral traits and pollination systems similar to African taxa. But a recent study found mirror-image flowers, a highly-specialized pollination system, in the Asia-endemic genus Hiptage. Mirror-image flowers in Hiptage show a sexual polymorphism in which the style deflects either to the left or the right side of the floral axis, which is a highly specialized mechanism promoting cross-pollination between left- and right-styled flowers via pollinators touching two sexual organs respectively with their left and right side of abdomens. Hiptage is also notable for its heteromorphic stamens, bilaterally-symmetric corolla, and extremely-reflected petals. Its main pollinators are the pollen-collecting honeybees such as Apis dorsata. These results indicate that the Asian Malpighiaceae could shift from floral conservatism to specialized pollination systems adapted to pollen-gathering honeybees. Hiptage is also distinctive for its position in a clade with complete migration history of Malpighiaceae route, having endemic genera respectively in South and Central America, Africa, and Asia. Further experimental studies on this genus and other Asia-endemic genera are needed to fully understand the Malpighiaceae route and its associated evolutionary adaptations, which will be helpful for studies on plant long-distance dispersal and inter-continent disjunctions.

Ex-situ conservation is an effective measure to prevent extinction of species. In order to learn the plant-pollinator relationship in ex situ habitat and shed insights into the ex situ conservation, we studied the pollination system and reproductive strategy of two orchids Ludisia discolor from Wutong Mountains, Shenzhen and Calanthe argenteo-striata from Malipo, Yunnan in the Botanical Garden of the National Orchid Conservation Center (NOCC), Shenzhen, southeastern China in 2007 and 2008. Both species have inflorescences with many white flowers which have spur with a narrow entrance. The butterfly, Pieris rapae, was their solely legitimate pollinators. The pollinaria of L. discolor were carried on the legs of butterflies, while that of C. argenteo-striata on the proboscis. The pollination efficiency in both orchids was high, and the ratio of pollinaria removal to massulae deposition was 1.14:1 and 0.74:1 in L. discolor and C. argenteo-striata, respectively. In 2007 and 2008, fruit set of open pollination in L. discolor was higher than that in C. argenteo-striata, which could be explained by that the former provides rewards to its pollinators while the latter is deceptive, as well as by that the flowering period of L. discolor was synchronized with the abundance of butterflies, which favored the orchid to exploit its pollinators, whereas the flowering period of C. argenteo-striata was lagged behind the outbreak period of butterflies, which made the flowers suffer insufficient pollinator visitation. Both species were self-compatible but there were no spontaneous autogamy or apomixes. The fruit sets of cross-pollination and self-pollination showed no significant difference. Ludisia discolor and C. argenteo-striata grow well after being transplanted to NOCC, and new seedlings were found near parent plants, indicating the population tended to expand, and the ex situ conservation was feasible. It could be predicted that these two orchids have the potential to establish stable populations in NOCC in the future.

Mutualism is one of the most important ecological interactions, with strong influences on almost all levels of biological systems. Their long-term persistence raises many challenging evolutionary questions, especially those involving high-level coevolution and coadaptation. Figs and their pollinating wasps are among the most tightly integrated mutualisms known, and provide a model system for developing and testing theories of coevolution. Initial studies suggested specific coevolution between them, described as the famous rule of one fig one wasp. However, more and more exceptions have been revealed by recent studies, and cryptic species in pollinating wasps and host switching were found common in some regions and within some Ficus groups, inducing debates on the levels of species specificity and coevolution. A broad-sense coevolution model to describe the relationship of the related groups of figs and their pollinating wasps was proposed recently. The diverse relationships between figs and their pollinating wasps indicated coexistence of both specific and diffuse coevolution in this mutualism system, producing different species-specificity level. However, which model is the dominant one in this system is still keeping open. The species specificity could be tight or loose in different regions and fig groups involved. Consequently, the frequencies and mechanisms of breakdowns of the one-to-one rule within different fig groups as well as in different regions are essential for the understanding of the relative importance of the competing finer-scale cospeciation or broad-sense coevolution models.

Interaction between animal-pollinated plants and their pollinators is thought to be an important selecting force shaping the evolution of flowers. Successful pollination of a plant relies on its attractiveness to pollinators and the ability of the pollinators to deposit enough compatible pollen on the flower’s stigma. To assess the attractiveness, we manipulated the flower of Aconitum kusnezoffii by removing the sepals and pet-als, respectively. We then evaluated the mating system based on pollinator behavior and flower biology of A. kusnezoffii. The bumblebee, Bombus ignites, was the effective pollinator of A. kusnezoffii. Bumblebee be-havior differed on the manipulated flowers. The visiting frequency was significantly reduced by se-pal-removal treatment, but remained unchanged by nectar-removal treatment. Bumblebees were attracted by the flower exterior morphology formed by five sepals, not by the nectary formed by the two petals. The two lower sepals provided a landing platform, and the uppermost sepal functioned as a nectar guide, ensuring bumblebee find the nectar precisely and rapidly. The two lateral sepals restricted the direction in which a bumblebee can enter the flower, ensuring pollination efficiency in zygomorphic flowers. Nectar was the main reward for bumblebees, with a sugar concentration of 39.23% and histidine concentration of 0.25 μg/μL. The flower of A. kusnezoffii was large, protandrous and herkogamous, suggesting an outcrossing mating system based on the outcrossing index. In Aconitum inflorescences the oldest flowers were functionally female at the bottom of acropetal racemes, while younger flowers were functionally male at the upper positions. Bombus ignites generally flew from a younger functionally male flower near the top of an A. kusnezoffii inflorescence to an older functionally female flower near the bottom of another inflorescence, and then moved progres-sively upward. This behavior promoted the outcrossing of A. kusnezoffii. Aconitum kusnezoffii flowers were strongly protandrous, with the male phase lasting 2–4 days, during which the anthers gradually dehisced. Aconitum packaged pollen within each flower as individual anthers dehisced over a protracted period. Such pollen packaging necessarily limited pollen removal during single visits and increased male fitness by pro-moting pollen dispersal.

Cryptic species are morphologically undetected yet genetically divergent species. The presence of cryptic pollinating-fig wasp species has potentially important implications for understanding fig/fig-wasp mutualisms and the coexistence of cryptic species. It is essential to identify cryptic species quickly and accurately using molecular markers since the species are morphologically undistinguishable. We used PCR-RFLP to analyze the mtDNA COI gene in cryptic species of Ceratosolen emarginatusandC. gravelyi. The PCR-RFLP was useful for determining the composition of cryptic pollinating-fig wasp species. Ceratosolen emarginatus included two cryptic species (A and B) with an XhoI and a BssSI digesting site in the COIgene, respectively. Ceratosolen gravelyi included two cryptic species with a BmrI and an AvaI digesting site, respectively. Genomic DNA samples were extracted from a mixture of cryptic species A and B withinC. emarginatus, and XhoI andBssSI endonucleases were used to digest the amplified COIfragments. XhoI-specific fragments were found when cryptic species A was present, and BssSI-specific fragments were found when cryptic species B was present. The digestion of COI gene with BmrI andAvaI in two cryptic species ofC. gravelyi showed similar results. Taken together, PCR-RFLP, based on PCR and DNA restriction enzyme digestion, can quickly and accurately identify cryptic pollinating-fig wasp species.

In order to understand the nature and intensity of interactions between non-pollinating fig wasps and fig-pollinator mutualist wasps, and consequently shed light on why they coexist, we studied interactions between Walkerella sp. and the Ficus curtipes fig-pollinator mutualism in Xishuangbanna Botanical Garden in 2008. The reproductive behaviour of Walkerella sp. was studied by means of observation, and a controlled experiment was applied to study interspecies relationships. Figs only with pollinator, figs only with Walkerella sp., figs with both pollinator and Walkerella sp. and natural figs were compared in this research. Walkerella sp. was observed ovipositing on the outside of the fig. One Walkerella sp. can oviposit on many figs. ANOVA showed that the number of pollinator in figs with both pollinator and Walkerella sp. is significantly lower than that in figs only with pollinator (P < 0.01), while the number of seed did not differ between figs with only pollinators and figs with both Walkerella sp. and pollinator (P = 0.33). The number of Walkerella sp. did not differ among figs with only Walkerella sp., figs with both pollinator and Walkerella sp. and natural figs. Our results support the hypothesis that Walkerella sp. affects pollinator number, but not seed number and there is no inquiline or parasitoid affecting the number of Walkerella sp. The requirement of male pollinators to dig a hole in the fig wall may limit the number of Walkerella sp. eggs in the figs.