Pollination determines plant fitness, genetic diversity, and the overall functioning and long-term stability of the world’s ecosystems (Biesmeijer et al., 2006; Potts et al., 2010; Neuschulz et al., 2016). In agricultural production systems, insects or wind either transfer pollen grains naturally or human-assisted pollination can be implemented (Strange, 2015; Toni et al., 2021); these pollination processes are worth 9.5% of the total value of food production globally (Gallai et al., 2009; Breeze et al., 2011; Hanley et al., 2015). Insect-mediated pollination in particular ensures the sexual reproduction and primary productivity of multiple fruits, vegetables, oil crops, cereal grains and forages (Eilers et al., 2011; Jauker et al., 2012; Strange, 2015). More than 1,500 crops are pollinated by insects and these entomophilous plants produce 35% of the world’s food items (Klein et al., 2007; Aizen et al., 2009). Moreover, insect pollination routinely improves fruit quality and increases farmer incomes (Eilers et al., 2011; Chaplin-Kramer et al., 2014). Pollinating insects such as bees, flies, beetles, moths and butterflies thus generate economic dividends worth US$ 780.8 billion in 2016 globally (Rader et al., 2020). In different production systems e.g., almond or fruit orchards, or greenhouse vegetables, farmers fortify insect-mediated pollination through the establishment of rented or purchased hives of European honeybee (Apis mellifera L.) or bumblebees Bombus spp. (Roubik, 2002; Velthuis and Van Doorn, 2006; Potts et al., 2016).

In many farming systems, beneficial insects also provide a second ecosystem service i.e., biological pest control (Oerke, 2006; Power, 2010) which is valued at US$ 4.5 billion per year in the U.S. alone (Losey and Vaughan, 2006). In greenhouse settings, crop productivity is regularly affected by pestiferous herbivores such as aphids, thrips, whiteflies, and mites (Malais and Ravensberg, 2003; Maleknia et al., 2016). Aphids are common pests of greenhouse crops, where they reduce yields through direct (phloem) feeding, excretion of sugar-rich liquids that impede photosynthesis and the vectoring of debilitating viruses. Though chemical insecticides are routinely used for aphid control, these products cause resistance development, negatively affect resident pollinator populations, pollute the environment and pose important food safety hazards (Devine and Denholm, 2009; Garibaldi et al., 2018; Wyckhuys et al., 2020). Biological control is an environmentally friendly alternative to insecticide-based approaches, and has been continually used and refined for nearly 2000 years (Bale et al., 2008; Heimpel and Mills, 2017). In greenhouse production systems across the globe, biological control has replaced chemical pest control over the past decades (Pilkington et al., 2010).

In China and across the globe, beneficial insects are increasingly deployed for pest, disease and weed management in protected and open-field farming systems alike. However, the joint deployment of insect pollinators and biological control agents can either lead to positive (synergies) or negative (trade-offs) effects on crop yield (Lundin et al., 2013; Sutter and Albrecht, 2016; Garibaldi et al., 2018). For example, insect pollinators and pest control act synergistically on red clover (Trifolium pratense L.) to produce higher seed yields (Lundin et al., 2013). Similarly, managed pollinators such as bumblebees often vector insect-killing or antagonistic fungi and thus simultaneously deliver pollination and pest or disease control services (Shafir et al., 2006; Kapongo et al., 2008). On the other hand, high pest loads i.e., low levels of biological control enhanced pollinator benefits for seed yield in oilseed rape (Brassica napus L.) (Bartomeus et al., 2015).

Given their variable impact on crop productivity and the additional costs associated with a simultaneous use of pollinators and natural enemies, this practice has only received marginal attention in greenhouse agriculture. Science however can unlock opportunities to tap the synergistic interactions between pollinators and natural enemies, as demonstrated by the successful use of bumblebee pollinators for thrips biological control in Canadian tomato and sweet pepper crops (Kapongo et al., 2008).

Syrphid flies (Diptera: Syrphidae) commonly forage in agricultural crops, where they simultaneously provide pollination and biological control services (Ssymank et al., 2008; Omkar, 2016; Dunn et al., 2020). Hoverfly adults feed on pollen and nectar, making them the world’s second most important pollinators after bees (Rotheray and Gilbert, 2011; Rader et al., 2016; Klecka et al., 2018; Wotton et al., 2019; Rader et al., 2020). Hoverfly-mediated pollination has been valued at US$ 300 billion per year (Food and Agriculture Organisation of the United Nations, 2017). Members of the Syrphinae subfamily have carnivorous larvae that prey upon aphids, lepidopteran larvae and other soft-bodied herbivores (Nelson et al., 2012; Ramsden et al., 2017; Bellefeuille et al., 2021; Li et al., 2021; Li and Wu, 2022). The vagrant hoverfly Eupeodes corollae is a widely distributed hoverfly species that naturally occurs in Central Europe, North Asia and North Africa (Schweiger et al., 2007; Djellab et al., 2019). Adults of E. corollae consume pollen and floral nectar of cultivated crops and deposit eggs on aphid-infected plants (Jauker and Wolters, 2008; Pekas et al., 2020; Moerkens et al., 2021; Jiang et al., 2022). This hoverfly species is easily reared under laboratory conditions, creating lucrative opportunities for augmentation biological control (Van Lenteren et al., 2018). Untilrecently, few scientific studies have assessed how E. corollae augmentative releases affect fruit set, aphid pest control or yield in greenhouse crops. Pioneering work in Europe however has revealed 88% higher yields on sweet pepper crops that were subject to E. corollae releases (Pekas et al., 2020).

In this study, we investigated the relative contribution of E. corollae to pollination and aphid biological control in three entomophilous crops in China i.e., tomato Solanum lycopersicum L. (Solanaceae), melon Cucumis melo L. (Cucurbitaceae), and strawberry Fragaria ananassa Duch. (Rosaceae). In local greenhouses, these crops are regularly affected by the cotton aphid Aphis gossypii Glover (Rondon et al., 2005). For varying hoverfly release rates, we assessed fruit set, A. gossypii population dynamics, and the number of E. corollae eggs and larvae on melon and strawberry plants. Our work defines and validates crop-specific augmentative release schemes for E. corollae, opening new vistas for biodiversity-driven pollination and pest management in multiple greenhouse crops.

Tomato pollination Tomato fruit set was higher in plots that received hoverfly releases than in bagged plots. With increasing E. corollae release rate, fruit set initially increased but then remained constant during both spring and autumn ( Figure 1 ). During spring, the highest fruit set (97.8%) was recorded in plots with releases of 80 E. corollae adults. Fruit set under this treatment was comparable to hormone-treated plants (98.9%), and markedly higher than for the bagging treatment (60.2%) (F _6,210 = 36.73, P=0.000). During autumn, fruit set in plots with releases of 60 hoverfly adults (95.7%) did not differ from that of hormone-treated plants (96.3%) (F _1,60 = 0.44, P=0.73). Similarly, the weight of harvested tomatoes was highest in plots that were subject to hoverfly releases. With increasing release rates, tomato weight initially increased and then remained unchanged during either season ( Figure 2 ). During spring, a release rate of 60 hoverflies yielded a fruit weight of 1794.0 ± 92.4 g, which was identical to that of hormone-treated plants (1871.3 ± 71.1 g), but higher than for the bagging treatment (765.4 ± 37.2 g) (F _6,207 = 34.1, P=0.000). For different hoverfly release rates i.e., 60, 80, 120 adults and the hormone treatment, fruit weight during spring was higher than during autumn (60: F _1,53 = 0.52, P=0.032; 80: F _1,53 = 0.1, P=0.005; 120: F _1,48 = 0.14, P=0.018; and hormone: F _1,55 = 1.1, P=0.008 respectively).

Melon pollination and biological control. Melon fruit set was higher in plots that received hoverfly releases than in bagged plots. With increasing E. corollae release rate, fruit set initially increased and then remained unchanged during spring and autumn ( Figure 3 ). Upon release of 20 E. corollae pairs, 100% fruit set was attained during either season; a rate identical to that of hormone-treated plants and substantially higher than for the bagging treatment (spring: F _6,157 = 675.8, P=0.000; autumn: F _6,157 = 415.6, P=0.000). For a given release rate, fruit set did not differ between spring and autumn season ( Figure 3 ). During spring, aphid infestation levels declined at higher hoverfly release rates with the extent of A. gossypii population decline ranging from 65.5-99.3% ( Figure 4A ). Once mesh bags were removed (i.e., day #13), aphid population growth rate declined in the hoverfly release treatments and peak population numbers were reached on day #19. By the end of the experiment (i.e., day #31), aphid infestation levels were reduced by 65.5 ± 12.3%, 86.7 ± 8.8%, 89.7 ± 7.8%, 97.3 ± 4.2% and 99.3 ± 2.2% under the different hoverfly release treatments i.e., 5, 10, 20, 40, 80 pairs as compared to the 0 pairs (Hormone group) ( Figure 4A ). Similar results were obtained during autumn, with the extent of aphid population reduction directly proportional to hoverfly release rate ( Figure 4B ). However, the degree of population reduction was smaller than during spring, especially at low hoverfly release rates. On day #31, aphid population levels were reduced by 53.5 ± 12.9%, 66.5 ± 12.2%, 80.4 ± 10.2%, 95.0 ± 5.6% and 97.8 ± 3.8% as compared to the hormone group ( Figure 4B ). During either season, the number of E. corollae eggs on melon plants increased with the hoverfly release rate ( Figures 4C, D ). The number of hoverfly eggs on a single melon plant was highest on day #19. Under a given release treatment, egg deposition rates were higher during spring than autumn (5 pairs: F _1,46 = 2.349, P=0.034; 10 pairs: F _1,46 = 0.026, P=0.000; 20 pairs: F _1,39 = 4.205, P=0.001; 40 pairs: F _1,45 = 0.393, P=0.002; and 80 pairs: F _1,40 = 1.452, P= 0.007).

Similarly, the number of hoverfly larvae increased with the hoverfly release rate during either season ( Figures 4E, F ). Larval numbers were highest on day #25 in all treatments except for the lowest hoverfly release rate i.e., 5 pairs ( Figures 4E, F ). Under a given release treatment, hoverfly larvae were more abundant during spring than autumn (20 pairs: F _1,39 = 0.001, P=0.005; 40 pairs: F _1,45 = 0.424, P=0.007; and 80 pairs: F _1,40 = 3.146, P=0.000 respectively). Lastly, during both seasons, melon weight increased proportionally with hoverfly release rate ( Figure 5 ). Under a given release treatment, fruit weight was higher during spring than in autumn ( Figure 5 ) and the highest melon weights were recorded for a release rate of 80 E. corollae pairs. The latter values i.e., 878.2 ± 39.7 g and 560.9 ± 24.2 g during spring or autumn, respectively were comparable to those of the hormone treatment i.e., 832.1 ± 11.6 g and 519.6 ± 30.7 g ( Figure 5 ).

Strawberry pollination and biological control. Strawberry fruit set was higher in plots that received hoverfly releases than in bagging plots, and increased with release rate. Fruit set under release treatments ranged from 76.6-87.8%, and was markedly higher than for bagged plants i.e., 67.3% (F _4,148 = 13.7, P=0.000) ( Figure 6 ). Aphid populations were suppressed under all hoverfly release schemes and the lowest A. gossypii abundance was recorded for release rates of 80 pairs ( Figure 7A ). As compared to the bagged treatment where A. gossypii attained peak abundance of 1061.3 ± 40.9 individuals per plant on day # 21, aphid population growth slowed markedly following hoverfly release. Under release rates of 10, 20, 40 or 80 E. corollae pairs, aphid population levels were reduced by a respective 57.9 ± 12.7%, 63.2 ± 12.5%, 50.0 ± 12.9% and 70.2 ± 11.8% as compared to the bagging group. Under release rates of 40 and 80 E. corollae pairs, high numbers of eggs were deposited on strawberry plants during two weeks following hoverfly release ( Figure 7B ). Similarly, the number of hoverfly larvae was highest on day #16 following release, attaining respective maxima of 1.07 ± 0.2, 1.5 ± 0.27, 3.3 ± 1.3 and 3.58 ± 0.73 larvae per plant ( Figure 7C ). Lastly, the highest strawberry weight (78.5 ± 5.5 g/plant) was recorded for a release rate of 80 pairs; this weight surpassed that for other release rates or for the bagging treatment i.e., 31.8 ± 2.9 g ( Figure 8 ; F _4,148 = 14.485, P=0.000).

Syrphid flies, such as E. corollae, provide globally important ecosystem services such as pollination and biological control in both natural and agricultural ecosystems. Aside from being a widely-distributed pollinator, E. corollae larvae consume aphid pests and lepidopteran larvae in various agricultural crops and agro-ecosystems (Hopper et al., 2011; Li et al., 2021). Pioneering work in Europe has shown how laboratory-reared E. corollae effectively deliver non-chemical pest control and enhance fruit set in sweet pepper (Pekas et al., 2020). Building upon these initial findings, we demonstrate how E. corollae augmentative releases slow aphid pest build-up and raise yield levels in other greenhouse crops i.e., tomato, muskmelon and strawberry. Even at small release rates, considerable increments in fruit set, fruit weight and aphid biological control rates were observed for all three crops. Considering how multiple hoverfly species can be reared under laboratory conditions (Iwai et al., 2007; Iwai et al., 2009), our work unlocks opportunities for augmentative hoverfly releases in (greenhouse) cropping systems in China and abroad.

Though we investigated hoverfly-mediated pollination and aphid pest control on muskmelon and strawberry, we were unable to assess the latter service on tomato crops. Upon inoculation of tomato plants with different aphid species (A. gossypii, Myzus persicae (Sulzer), Aphis craccivora Koch), these failed to colonize the plants. This can possibly be ascribed to the effect of glandular trichomes on tomato leaves (Blanco-Sánchez et al., 2021), which inhibit aphid colonization and feeding. Tomatoes are a self-pollinating plants that produce many flowers and exhibit a long flowering period. Wind and insect vectors promote pollen deposition, fertilization and fruit set through at a low rate. In greenhouses, hormones and honeybees or bumblebees are often used to promote tomato fruit set (Sabara and Winston, 2003; Higo et al., 2004). Pollination is thus a crucial factor in commercial tomato production, and hoverfly releases improve crop output. Also, as E. corollae larvae do not consume common tomato pests such as thrips and whiteflies (Bleeker et al., 2009; Gupta et al., 2018), hoverfly adults did not forage or oviposit on (flowerless) tomato plants in our trials. Hence, in the sole presence of tomato, E. corollae populations likely cannot sustain themselves in greenhouse systems and may require recurrent adult releases.

As compared to the European or North American horticultural sector, Chinese producers commonly use low-tech plastic (vs. glass) greenhouses with limited or no climate control. As our study mimicked these conditions, temperatures occasionally reached peaks of 38-39°CC i.e., as recorded in the upper half of the closed structure. Such elevated temperatures may have negatively affected behavior, development and overall fitness of the study organisms i.e., aphids and hoverflies. Those unfavorable climatic conditions possibly explain some of the differences in hoverfly or aphid population dynamics or fruit yield between spring and autumn seasons i.e., in muskmelon and tomato. Notably, hoverfly egg and larval densities on muskmelon were substantially lower during autumn. This however may also be attributed to shortened daylength or cooler night-time temperature. Overall, it is challenging to fully ascertain the magnitude of those effects as laboratory-derived temperature sensitivity measures do not fully account for hoverfly behavioral adaptations under field conditions (e.g., Hassall et al., 2017). Yet, lower E. corollae reproductive output during autumn is plausibly mirrored in dampened aphid suppression at intermediate release rates, and such needs to be taken into account in future biological control endeavors.

In order to optimally gauge E. corollae effects on strawberry pollination, we opted not to employ fruit thinning. As a result, we attained up to 88% fruit set and 10-15 fruits per plant at release rates of 80 hoverfly pairs. This practice however contrasts with commercial strawberry cultivation in which thinning is routinely used to improve fruit quality. Given that E. corollae releases concurrently enhance fruit set and fruit weight, this practice potentially results in higher overall yields at lower thinning rates and could be of major interest to producers. Our novel experimental set-up further allowed to distinguish hoverfly impacts on pollination vs. aphid pest control, and thus provides advantages over a simple determination of yield (Hodgkiss et al., 2018). Aside from providing direct aphid control, pollinators such as E. corollae also selectively favor direct or indirect plant defence in a strawberry crop (Egan et al., 2021) and thereby enhance further opportunities for sustainable pest management (Wyckhuys et al., 2022).

Aphids often co-occur with other greenhouse pests such as whiteflies, thrips, mites. As E. corollae does not necessarily prey upon all of these other pests, hoverfly releases can easily be combined with other (augmentative, conservation) biological control measures or preventative pest management. Living and UV-reflective mulches can be deployed against thrips and whitefly pests, while natural enemies such as Phytoseiulus persimilis Athias-Henriot and Diglyphus isaea (Walker) have been effectively used against mites and agromyzid leafminers (Nyoike and Liburd, 2010; Park and Lee, 2021). Aphids and thrips can be simultaneously controlled with coordinated releases of Aphidius colemani Viereck and Neoseiulus cucumeris Oudemans, while a combination of hoverfly releases with bee-vectored microbials carries ample potential for multi-target pest or pathogen control (Kapongo et al., 2008; Messelink et al., 2013; Gonzalez et al., 2016). Yet, when combining hoverflies with other (invertebrate, microbial) natural enemies under integrated control programs, hoverfly release rates and multi-species augmentation schemes will need to be defined based upon ecological interactions and overall cost-effectiveness. Especially when tailored extension services are lacking, smallholders may consider it impractical to use multiple natural enemies and might still resort to pesticide applications. Several of the insecticides that are commonly used against whitefly or thrips however cause important lethal and sublethal effects on hoverflies (Colignon et al., 2003; Moens et al., 2011). Hence, an effective integration of hoverfly-mediated services with chemical pest control is only possible for a narrow set of selective chemical compounds (Jansen et al., 2011; Moens et al., 2011). A judicious use of chemical insecticides and their proper timing, dosage and placement is thus of paramount importance to integrated pest management (IPM) success.

Aside from its poor compatibility with chemical control, economic aspects may also hamper a further diffusion of hoverfly augmentative releases especially in low-value crops. In order to incentivize and accelerate its adoption, comprehensive benefit-cost analyses, fine-tuned release schemes and multi-stakeholder communication is crucial. This can comprise a full accounting of the (human, environmental) health benefits of foregone insecticide applications (Wyckhuys et al., 2020). Habitat management tactics can also improve economic viability. Costly, repetitive releases of laboratory-reared hoverflies possibly can be avoided through the use of banker plants, as demonstrated for the hoverfly Eupeodus americanus (Wiedemann) in Canadian sweet pepper (Miller and Rebek, 2018; Bellefeuille et al., 2021). For example, legume companion plants can provide alternative aphid prey and sustain viable hoverfly populations within greenhouse settings (Tscharntke et al., 2005; Winqvist et al., 2011). In order to enhance hoverfly fitness and stimulate oviposition, plants that provide alternative, non-pest prey can be paired with ones that bear copious amount of pollen or (floral, extra-floral) nectar (Campbell et al., 2012; Laubertie et al., 2012; Martínez-Uña et al., 2013; Gurr et al., 2017). To pinpoint the most appropriate species of nectar-producing plants, banker plants and non-target aphids that can accompany E. corollae augmentative releases, further-study is imperative.

In conclusion, our work uncovers how E. corollae hoverflies assume a dual role as pollinators and biological control agents in three different greenhouse crops. Our research shows that augmentative hoverfly releases augment fruit set and crop yield while securing insecticide-free aphid pest control. To fully exploit the benefits of hoverflies for (greenhouse) horticulture in China and abroad, follow-up research is required to define (economically sound) release schemes, assess habitat management tactics and explore a further integration with other (insect-vectored) natural enemies.

The original contributions presented in this study are included in the article/supplementary material. Further inquiries can be directed to the corresponding author.

HL: Investigation, Formal analysis, Writing-original draft. KAW: Writing - review and editing. KW: Funding acquisition, Conceptualization, Supervision, Resources, Writing - review and editing. All authors contributed to the article and approved the submitted version.