Introduction
Roughly 90% of plant species are pollinated by animals and insects play an especially important role in pollination (Bawa, Reference Bawa1990; Linder, Reference Linder1998; Ollerton et al., Reference Ollerton, Winfree and Tarrant2011). For example, most Apocynaceae are insect-pollinated, with only a few records of bird pollination (Ollerton, Reference Ollerton2018). Among pollinating insects, bees play a dominant role in most ecosystems and many species have been identified as the most-effective pollinators for various kinds of plants (Crane and Walker, Reference Crane and Walker1984; Pernal and Currie, Reference Pernal and Currie2001; Winfree et al., Reference Winfree, Williams, Dushoff and Kremen2007). Further, 90% of the world's 107 most-important crops depend on bees for pollination (Potts et al., Reference Potts, Imperatriz-Fonseca, Ngo, Aizen, Biesmeijer, Breeze, Dicks, Garibaldi, Hill, Settele and Vanbergen2016). For example, the managed honey bee Apis mellifera is famous for the portable pollination services they provide, more easily enhancing the yields of the target plants while improving the quality of their fruits (Morse and Calderone, Reference Morse and Calderone2000; Klein et al., Reference Klein, Vaissiere, Cane, Steffan-Dewenter, Cunningham, Kremen and Tscharntke2007).
In recent years, native bees have become a prominent focus due to honey bee declines and potential ecological consequences of their high abundance (Roubik et al., Reference Roubik, Moreno, Vergara and Wittmann1986; Steffan-Dewenter and Tscharntke, Reference Steffan-Dewenter and Tscharntke2000; Paini, Reference Paini2004; Aizen et al., Reference Aizen, Garibaldi, Cunningham and Klein2008; Allsopp et al., Reference Allsopp, de Lange and Veldtman2008; Rader et al., Reference Rader, Bradley, Saul, David, Linda, Melanie and Edwards2009; Suzuki et al., Reference Suzuki, Kawaguchi, Munidasa and Toquenaga2009; Kleijn et al., Reference Kleijn, Winfree, Bartomeus, Carvalheiro, Henry, Isaacs, Klein, Kremen, M'Gonigle, Rader, Ricketts, Williams, Adamson, Ascher, Báldi, Batáry, Benjamin, Biesmeijer, Blitzer, Bommarco, Brand, Bretagnolle, Button, Cariveau, Chifflet, JF, Danforth, Elle, Garratt, Herzog, Holzschuh, Howlett, Jauker, Jha, Knop, KM, Féon, Mandelik, May, MG, Pisanty, Reemer, Riedinger, Rollin, Rundlöf, Sardin˜as, Scheper, AR, Smith, Steffan-Dewenter, Thorp, Tscharntke, Verhulst, Viana, Vaissie're, Veldtman, Westphal and Potts2015; Frier et al., Reference Frier, Somers and Sheffield2016). However, more research is needed to understand the dynamics of pollination throughout most cropping systems in China and monitoring data are also lacking generally for wild pollinators (Liu et al., Reference Liu, Douglas, Wu, Zhou and Zhu2018). To date, the majority of studies have focused on managed honey bees (Teichroew et al., Reference Teichroew, Xu, Ahrends, Huang, Tan and Xie2017). For example, Apis cerana was among 18 flower-visiting insect taxa observed in cabbage fields and was the focus of the study (Johnson et al., Reference Johnson, Khushboo, Avupati and JaiPrakash2017). Of these, the order Hymenoptera constituted 59.40% and 40.70% in honey bee introduced and control fields, respectively. It is crucial that we investigate the specific visitors in systems like this because species differ not only in their effectiveness but also their specialization (or oligolecty), such that many may only visit a small number of plants (Gauld and Bolton, Reference Gauld and Bolton1988; Cane and Sipes, Reference Cane and Sipes2006).
The case of the economically-important tea oil plant, Camellia oleifera (Theaceae) is especially interesting because it requires cross-pollination but seemingly discourages pollinators (Zhuang, Reference Zhuang1998; Huang et al., Reference Huang, He, Gu, Su and Zhu2017). Not only does this species bloom during winter in China when few pollinators are available, it is also chemically defended, causing losses to honey bee colonies used to pollinate it (Zhang and Chen, Reference Zhang and Chen2013). These factors may contribute to pollen limitation, leading to the observed phenomenon of ‘with a thousand flowers and only one fruit’ observed by He et al. (Reference He, Kang and He2002).
Given these challenges, few species are thought to be effective pollinators of C. oleifera (Lin and Li, Reference Lin and Li1991; Qiu et al., Reference Qiu, Zhang and Liu2018). Among them, Andrena camellia Wu (Hymenoptera, Apoidea, Andrenidae) is an oligolectic species and a primary pollinator of C. oleifera based on their unusual activity period, peaking during the bloom of C. oleifera and extensive observations of floral visitation and nesting density among Camellia fields (Wu, Reference Wu1977; Ding et al., Reference Ding, Huang, Zhang, Huang, Li and Zhu2007; Huang et al., Reference Huang, Ding, Zhang, Hao and Zhu2008; Xie et al., Reference Xie, Chen and Qiu2013; Qiu et al., Reference Qiu, Zhang and Liu2018). Shortages of this crucial pollinator could result in limited visitation and, therefore, pollen limitation causes diminished fruit production of C. oleifera. Conversely, the fruit setting rate could be improved by increasing the abundance and density of bees that pollinate C. oleifera such as Andrena camellia, Colletes gigas Cockerell, Andrena striata Wu, and Andrena hunanensis Wu (Lin and Li, Reference Lin and Li1991; Kunin, Reference Kunin1997; Zhuang, Reference Zhuang1998). Recent studies reported the nesting biology, the biological features and behaviors of A. camellia (Ding et al., Reference Ding, Huang, Zhang, Huang, Li and Zhu2007; Huang et al., Reference Huang, Ding, Zhang, Hao and Zhu2008). However, up to the present, its flower foraging habits remain poorly known and unquantified, preventing us from building a framework to evaluate and better manage the pollination services that A. camellia and other species provide for this vital crop.
Quantitative pollination ecology has gained attention in recent years, based mainly on a need to increase agricultural sustainability for production (Lonsdorf et al., Reference Lonsdorf, Kremen, Ricketts, Winfree, Williams and Greenleaf2009). Although many studies have taken top-down approaches to estimate ecosystem services, seeking to determine total pollination based on correlations with landscape or other factors, bottom-up methods that calculate service provision species-by-species are a powerful way to incorporate interspecies variation in effectiveness (Joseph et al., Reference Joseph, Santibáñez, Laguna, Abramson, Kuperman and Garibaldi2020). As an example, the essential numbers of Osmia cornifrons and Osmia excavata per unit area were estimated using both flower-foraging habits and foraging trip duration, but the results were quite different between these metrics (Wei et al., Reference Wei, Yuan and Zhang2000). This may be due to various factors such as debilitating weather or low density of plantings, both of which can greatly reduce the amount of time that an insect is actually pollinating within a given foraging period. Studies must also incorporate measures of pollination efficiency, the comparative ability of individuals to provide pollination services (Traveset, Reference Traveset, Pugnaire and Valladares1999; Young, Reference Young2002). Foraging frequency is also an important index of pollination efficiency and some studies have found that the effective foraging frequency is lower than that of total foraging frequency. Some studies have tried to control for this, for instance, only if the time spent on a flower exceeds 5 s and there is nectar- or powder-gathering behavior, will an effective visit be counted (Luo et al., Reference Luo, Chang, Chen and Zhang2019). Consequently, it is vital that we include the amount of pollen transported, as it does not matter how long a bee spends on a flower if it is not carrying any suitable pollen or contacting floral reproductive structures.
Here, we report new advances in the incorporation of these factors into the calculation of the essential number of pollinators for C. oleifera. To this end, we report the foraging habits of Andrena camellia and provide two methods to determine their essential number per unit forest area, incorporating both foraging time and the number of pollen grains carried as metrics. This represents an important first step to building a holistic model of pollination service provision for C. oleifera and other crops that require animal pollination.
Material and methods
Study species
Camellia oleifera Abel (family Theaceae is an important woody oil plant used in China for both human consumption and industries such as textile manufacture) having originated in China, is now extensively planted in 16 provinces in southern China as well as other countries such as Vietnam (Zhuang, Reference Zhuang1998). Currently, the total planted area of C. oleifera in China is about 1 million hectares (http://www.forestry.gov.cn/). The flowering period of C. oleifera ranges mainly from the middle of October to the end of December.
Study area
Field investigations were carried out in two C. oleifera sample plots located in Jiangxi province, China, from October 2015 to December 2018. The sample plot was located in Xinyu city (115°04′ E, 28°04′ N). The C. oleifera have been growing for 10 years and have consistently high yields. The numbers of C. oleifera per hectare to determine the planting density and the numbers of flowers on each target plant were counted.
Sampling programme
A cubic net measuring 4 m × 2.5 m × 2.5 m (Devkota and Thapa, Reference Devkota and Thapa2005) was used to collect 60 A. camellia individuals. After each foraging trip, the numbers of pollen grains carried on the body was counted. One foraging period was measured for the 60 collected females every 2 h from 8:00 am to 18:00 pm under natural conditions. For each individual, the total foraging time in 1 day was measured using a cup placed over the nest (fig. 1a); In the morning, we take the lid off, let the bees out of the nest, count the time, and close it when it is time to go back to its nest, so that it stays around its nest and cannot get in. This has been observed in our previous research, so we designed the device (Huang et al., Reference Huang, Ding, Zhang, Hao and Zhu2008). The data on 60 female individuals were recorded every day for 10 days. Finally, 60 homing females were collected returning to their nests and the number of pollen grains carried by each was counted under a microscope (100×).
Figure 1 (a) Cup trap for measuring the total foraging time in 1 day. (b) A female Andrena camellia on a flower of Camellia oleifera.
Based on nest excavations, 100 fresh pollen balls made by A. camellia were collected and the number of pollen grains contained in each pollen ball was counted out under a microscope (100×). One hundred newly-hatched individuals collected from nests using covering cups were examined under a microscope (100×) for whether any residual pollen grain existed on their body surfaces.
For each relevant treatment, to collect the pollen grains from females, their bodies were washed with 75% (v/v) alcohol solution repeatedly. From these samples, representative 0.01 ml samples were slide mounted and the pollen grains were counted under a microscope (100×), with total amounts per body then calculated based on the total volume related to this subsample. The criteria for determining sex were based on a previous study, drones are small, with a body length of 8.501 ± 0.244 mm, and with no pollen picking organs (corbiculate). The female bee has a large body length of 10.233 ± 0.365 mm, has a special pollen-collecting organ (corbiculate) to collect pollen and suck nectar without harming flowers, and its body surface is hairy and easy to get pollen (Huang et al., Reference Huang, Ding, Zhang, Hao and Zhu2008).
Data analysis
The essential number of bees per unit area was calculated according to the number of flowers in the Camellia oleifera plantation located in Xinyu city. The plants in this area were planted with a density of 1800 ha−1. The following six parameters were used for calculating the essential number of bees: speed of processing (time spent per flower), duration of a single flight period, mean number of pollen grains carried in one foraging instance (during the time spent on a single flower), mean number of pollen grains carried in a single total foraging period, mean number of pollen grains contained in one pollen ball, and mean number of eggs laid by one female over their lifetime.
Results
Essential number of pollinators based on foraging time
The observation of 60 nests of A. camellia showed that the foraging frequency of a female individual on a sunny day ranged from one to four trips, averaging 2.98 trips ± 0 (mean ± SD, n = 60). The time that 60 homing females spent in a single period (SFT) was 2644 s ± 120 (44.06 mins ± 2) (mean ± SD, n = 60), and the mean number of pollen grains collected from one foraging periods (NPSF) was 90,692.5 ± 80 (mean ± SD, n = 60). The mean number of time spent per flower for 100 female visits (SP) was 79.26 s ± 7 (mean ± SD, n = 60). A female bee needs about 51 flights to collect sufficient pollen grains during its lifetime, considering average fecundity and the pollen balls required for its offspring (Table 1). The number of plants foraged on in one foraging period (NFOF) was conservatively set to one in this formula. In this case, about 1998 A. camellia female individuals were essential to the efficient pollination of 1800 (NPH) C. oleifera plants in 1 ha of garden.
Table 1. Data for the essential individual numbers of Andrena camellia per unit area of C. oleifera
NPL, mean number of pollen grains in a big pollen ball; NPS, mean number of pollen grains in a small pollen ball; NPOF, mean number of pollen grains collected from one flower; NPSF, mean number of pollen grains collected from one foraging period; SP, mean time spent per flower (S) (Speed of Processing); SFT, mean time of single flight period (min); NE, mean number of eggs laid by a female; FFD, frequency of flight per day; NFP, mean number of flowers per plant.
Based on foraging time:
$$\lpar {{\rm Essential}\;{\rm number}\;{\rm of}\;{\rm pollinators}} \rpar \;{\rm ENP} = \displaystyle{{{\rm NPH} \times {\rm NFP} \times {\rm NFOF} \times {\rm SP} \times {\rm NPSF}} \over {{\rm SFT} \times {\rm NPFL}}}$$where ENP is the essential number of A. camellia per hectare (essential number of pollinators), NPH is the mean number of plants per hectare; NFP is the mean number of flowers per plant, NFOF is the mean number of flowers foraged on in one foraging period, SP is the mean time spent per flower (S) (Speed of Processing), NPSF is the mean number of pollen grains collected from one foraging period, SFT is the mean time of single flight period (min), and NPFL is the mean number of pollen grains collected by a female throughout lifetime
$${\rm NPFL} = \left( {\displaystyle{1 \over 4}} \right) \times ({\rm NE)} \times ({\rm NPS}) + \left( {\displaystyle{3 \over 4}} \right) \times ({\rm NE}) \times ({\rm NPL}).$$Essential number of pollinators based on quantity of pollen carried
The mean number of pollen grains collected by a female individual from one flower (NPOF) was 2876 ± 28 (mean ± SD, n = 60). The mean number of pollen grains contained in big (NPL) and small pollen balls (NPS) were 384,900 ± 15,800 (mean ± SD, n = 60) and 282,440 ± 8700 (mean ± SD, n = 60), respectively. The mean number of eggs laid by a female in its lifetime (NE) was 17, with each pollen ball sustaining one offspring (Table 1). The sex ratio of female-to-male among these eggs was approximately 3:1. In order to estimate the essential number of pollinators, it is assumed that there are only A. camellia pollinators in 1 ha of C. oleifera garden with 1800 6-year old plants (NPH) and each flower only needs one individual for successful pollination. The number of plants foraged on in one foraging period (NFOF) was conservatively set to one in this formula. Pollen carried in all parts of the body should be considered effective pollen of the belt as shown in fig. 1b (Qiu et al., Reference Qiu, Zhang and Liu2018). Under this circumstance, atleast 2107 female individuals are estimated to be required to complete the whole pollination task.
Based on pollen grains’ numbers:
$$\left( {{\rm Essential}\;{\rm number}\;{\rm of}\;{\rm pollinators}} \right)\;{\rm ENP} = \displaystyle{{{\rm NPH} \times {\rm NFP} \times {\rm NFOF} \times {\rm NPOF}} \over {{\rm (1}/{\rm 4)}\,{\rm (NE} \times {\rm NPS)} + {\rm (3}/{\rm 4)}\,{\rm (NE} \times {\rm NPL)}}}$$where ENP is the essential number of A. camellia per hectare (essential number of pollinators), NPH is the mean number of plants per hectare, NFOF is the mean number of plants foraged on in one foraging period, NFP is the mean number of flowers per plant, NPOF is the mean number of pollen grains collected from one flower, NE is the mean number of eggs laid by a female, NPS is the mean number of pollen grains in a small pollen ball, and NPL is the mean number of pollen grains in a big pollen ball.
Discussion
At present, research on quantification of pollinator populations remains incipient. Crucially, there has too often been a disconnect between the calculation of pollination services at either the field level, which is relevant to farmers, or at the individual level, where important information about specific pollinators and their effectiveness can be more readily detected (Perrot et al., Reference Perrot, Gaba, Roncoroni, Gautier, Saintilan and Bretagnolle2019; Jauker et al., Reference Jauker, Bondarenko, Becker and Steffan-Dewenter2012). By scaling up from the individual to the field level, however, we can satisfy practitioners while still incorporating variation in efficiency both within and across different pollinator species.
In quantifying the essential number of pollinators, prior studies have relied primarily on flight period data alone (Wei et al., Reference Wei, Yuan and Zhang2000). However, this is problematic because it does not account for time spent foraging for nectar vs pollen, nor does it account for time spent off flowers. The number of bees visiting flowers on a daily basis can be affected by many climatic factors, such as temperature. In addition, the time spent by bees on any single flower depends on many factors, including the size of flower and the amount of nectar or pollen present, the latter two being directly related to time of day and how many resources other visitors have already removed (Joshi and Joshi, Reference Joshi and Joshi2010; Field et al., Reference Field, Paxton, Soro, Craze and Bridge2012). Given that pollen carrying capacity is an important characteristic value of statistical pollination efficiency, determining fitness outcomes for both the pollinator and the plant, this experiment has improved this method on the basis of selecting stable flowering stage and estimated the effective pollinator quantity based on pollen quantity.
The critical issue with using total foraging period time is that it tends to overestimate the time spent pollinating, thereby underestimating the essential number of pollinators. Here, flight period-based analyses suggest that only 1988 females are required, but the integration of pollen quantity instead suggests that at least 2107 females are required. If underestimations based on flight period were followed, insufficient pollination services may result, impacting grower profits and, on larger scales, this could potentially lead to food or product shortages. At present, however, these estimates may be high or low, given that the exact amount of pollen needed for successful pollination was not measured, nor was the rate of deposition; it may be that only a small amount of pollen is necessary, but we do not know how much is successfully transferred. Future work should also more directly incorporate the amount of pollen required for optimal fruit production on the individual level, as well as single visit deposition trials to confirm that each individual visit by A. camellia is sufficient for fruit set (King et al., Reference King, Ballantyne and Willmer2013).
Summary and future perspectives
The number of pollinating insects present during a given crop's flowering period plays a central role in improving fruit setting rate, fruit shape index and fruit quality overall, and efforts to enhance their numbers are necessary to ensure food security in developing countries such as China. Quantifying the insects present and their value as ecosystem service providers is a crucial first step to this type of management. Our study significantly improves upon previous approaches, demonstrating the importance of methods which directly incorporate flower visiting activity into essential number calculation. In addition, this study provides foundational data for the rearing and management of A. camellia, a crucial pollinator of C. oleifera. In the future, by extending this method to additional pollinators in this or other cropping systems, it will finally be possible to build an integrative, holistic scheme for managing wild pollinators in crops to ensure optimal pollination.
Acknowledgements
The authors appreciate valuable advice from Prof. Yan-Ru Wu and other contributing team members in the Institute of Zoology. They also thank Mr Bo He for his assistance with experiments in the field. This study was supported by the National Natural Science Foundation of China (31970484); National Natural Science Foundation of Chongqing (cstc2018jcyjAX0382); National Science & Technology Fundamental Resources Investigation Program of China (2018FY100405).
