Ecosystem services provided by aculeate wasps

The aculeate wasps are one of the most diverse and speciose insect taxa; they are omnipresent across ecosystems and exhibit diverse co‐evolutionary and exploitative associations with other organisms. There is widespread conjecture that aculeate wasps are likely to perform essential ecological and economic services of importance to the health, well‐being and nutritional needs of our planet. However, the scope and nature of the ecosystem services they provide are not well understood relative to other insect groups (e.g. bees, butterflies, beetles); an appreciation of their value is further tarnished by their public reputation as pointless pests. Here, we conduct the first comprehensive review of how aculeate wasps contribute to the four main areas of ecosystem services: regulatory, provisioning, supporting and cultural services. Uniting data from a large but previously disconnected literature on solitary and social aculeate wasps, we provide a synthesis on how these insects perform important ecosystem services as parasites, predators, biological indicators, pollinators, decomposers and seed dispersers; and their additional services as a sustainable alternative to meat for human consumption, and medicinal potential as sources of research leads for anti‐microbials and cancer treatments. We highlight how aculeate wasps offer substantial, but largely overlooked, economic benefits through their roles in natural pest management and biological control programs. Accordingly, we provide data‐driven arguments for reasons to consider the ecosystem service value of aculeate wasps on a par with other ‘useful’ insects (e.g. bees). Finally, we provide a research roadmap identifying the key areas of research required to capitalise better on the services provided by these important insects.

"It really is unfortunate that one relatively small group of wasps should taint not only their own good work as predators, but also the work of thousands of solitary wasp species around the world." (Grissell, 2010)

I. INTRODUCTION
The natural world (from organisms to ecosystems) provides functions and goodsknown as ecosystem services (ESs)that are of value to society by supporting either directly or indirectly the quality of human life (Harrington et al., 2010). These processes are broadly categorised into four types of services: (i) regulating servicesi.e. regulation of ecosystems and ecosystem processes; (ii) provisioning servicesi.e. material or energy outputs which can be gained directly from ecosystems; (iii) supporting servicesi.e. helping maintain other ESs; and (iv) cultural servicesi.e. where ecosystems play important roles in education, recreation or conservation (MEA, 2005). Insects represent one of the largest and most biodiverse groups in the animal kingdom and are renowned for their contributions to all four categories of ESs (Losey & Vaughan, 2006;Noriega et al., 2018); for instance, 88% of angiosperm plants are pollinated by insects (Ollerton, Winfree & Tarrant, 2011), and the pollination services provided by insects are estimated to be worth over $250billion/year worldwide (IPBES, 2016). Recently, these services have been linked directly to the Sustainable Development Goals (Dangles & Casas, 2019), reaffirming their importance for human welfare, health and provision. Despite this, there is a severe taxon bias in research efforts to understand insect ESs (Donaldson et al., 2017): for example, almost 30% of all publications on insect ESs over the last 50 years have been on the Hymenoptera (bees, wasps, ants, and sawflies) (Noriega et al., 2018). This is unsurprising since the Hymenoptera are one of the most species-rich insect orders on earth, with over 150000 described species (Aguiar et al., 2013). At a finer taxonomic level within this group, however, further taxon biases are apparent: the majority of studies focus on ESs provided by bees (Noriega et al., 2018) and their role as pollinators (Klein et al., 2007;Klatt et al., 2014). Focussing on a specific ES overlooks taxonomic groups with more generalist life histories; e.g. insects that are both nectarivorous (thus providing pollination) and carnivorous (thus providing pest regulation). We lack an understanding of the functional roles for many of the other insects within ecosystems, how they contribute to ESs, and how anthropogenically driven changes in their populations will impact the services they provide (Noriega et al., 2018).
Here we provide the first comprehensive review on the scope and nature of ESs provided by one such understudied insect groupthe stinging (aculeate) wasps (Sumner, Law & Cini, 2018); this includes the non-social (solitary) aculeate wasps (excluding the relatively well-studied parasitoid wasps (non-aculeate) -previously known as the paraphyletic order Parasitica, and the fig wasps) as well as the social wasps. The potential ESs provided by this group remain only loosely defined (Elizalde et al., 2020), despite accounting for~20% of described hymenopteran species (Aguiar et al., 2013). A recent review of the ESs of social insects explained the importance of the services provided by social insects but highlighted the paucity of data for specific taxonomic groups, including social wasps (Elizalde et al., 2020). Consequently, the natural capital of aculeate wasps and their ESs are currently unclear and potentially under-valued. Further, assessing the ES contributions of understudied insects like aculeate wasps is especially important at a time of apparent global insect declines (Ollerton et al., 2014;Outhwaite et al., 2020;van Klink et al., 2020).
Wasps are a paraphyletic group (Branstetter et al., 2017) that account for approximately 103000 extant described species in the Hymenoptera; of these, 70% are parasitoid wasps (Aguiar et al., 2013), whose roles in the regulation of agricultural pests are well recognised and reviewed extensively elsewhere (e.g. Quicke, 2015;Wang et al., 2019). Here we focus on the other 33000 species of waspthe aculeate wasps (henceforth referred to as 'wasps' for simplicity) (Fig. 1); around 1000 of these species are social (belonging to the vespid subfamilies Polistinae, Stenogastrinae and Vespinae), meaning they live in colonies. These insects suffer from poor public image, and even scientists are reluctant to invest research effort in them (Lester, 2018;Sumner et al., 2018). As a result, their ES value remains poorly understood.
There are several reasons why it is important to review the provision of ESs by aculeate wasps. First, aculeate wasps are globally widespread, and exhibit high species richness ( Fig. 1), exceeding that of other well-studied ecosystem servicing insects such as the bees and ants combined (Aguiar et al., 2013). Second, the predatory value of aculeate wasps in regulating arthropod populations is largely anecdotal, lacking a synthesis of the empirical literature; and thus, the impact of aculeate wasps on natural and farmed ecosystems is currently unclear. This information is essential to motivate conservation programmes and for harnessing the ESs of these insects, as has been done very successfully for bees. Third, the ES provided by aculeate wasps through pollination of a variety of ecologically and economically important plants is almost entirely overlooked (Gess & Gess, 2010;Mello et al., 2011;Akhter, Khanday & Ahmad, 2016). Fourth, to date it is the disservicesas opposed to the beneficial services that predominate the ecological literature on aculeate wasps: they are an aggressive and destructive taxa outside of their native range, and serious pests as invasive species (Beggs et al., 2011;Lester, 2018). Further, culturally, humans have an antagonistic relationship with aculeate wasps due to their propensity to sting (Sumner et al., 2018;Jones, 2019). Finally, aculeate wasp populations are being affected by anthropogenic pressures in a similar way to bees (e.g. agricultural practices, habitat loss, climate change) (Dejean et al., 2011;Archer, 2015;Outhwaite et al., 2020).
In this review we adopt the broad-sense definition of ESs, namely, those services that support directly or indirectly the quality of human life (MEA, 2005). These include the provision of regulating services (e.g. pest control, pollination), provisioning services (e.g. human consumption, biomedical or pharmaceutical value); supporting services (e.g. decomposition, seed dispersal) and cultural services (e.g. bioindicators, recreation) (MEA, 2005;Noriega et al., 2018). One review of ESs by insects identified 42 specific roles (Noriega et al., 2018); another recent review of the ESs of social insects identified 10 roles (Elizalde et al., 2020). Here we use the same approach to identify 10 ways through which aculeate wasps (solitary and social) provide ESs ( Fig. 2; Table 1); we discuss the available evidence from extensive literature searches and provide a publicly available referenced database documenting this evidence (see online Supporting Information, Supplementary data sets S1-S6).  (Aguiar et al., 2013;Branstetter et al., 2017). For simplicity, this figure only includes 12 major aculeate wasp families (defined as those containing 200+ species); note that Formicidae and Anthophila are not aculeate wasps. Branch labels indicate the family name, common names (where available, taken from www.inaturalist.org), and the approximate number of species within each family. Colour of branch labels indicates the levels of social organisation found within each taxon: black, all species are solitary; blue, family includes both solitary and social species; pink, all species are social. Phylogenetic relationships are simplified from Branstetter et al. (2017)

II. REGULATING SERVICES
(1) Regulation of natural arthropod populations At least 30000 species of aculeate wasps are known to hunt a wide range of arthropods as predators or parasites (Goulet & Huber, 1993;Grissell, 2010), with impacts on arthropod populations in natural and farmed ecosystems. By regulating both carnivorous and phytophagous arthropod populations, wasps also indirectly deliver protection to invertebrate taxa and various plants lower down in food-chain networks. Limiting arthropod population growth is an essential function as arthropods can reproduce rapidly, reaching population sizes that can have detrimental effects on plant and other invertebrate taxa (Gaston & Lawton, 1988). The predatory impact of wasps is likely to be equal to or more effective than that of insectivorous birds, mammals and amphibians, as their short generation time allows them to match fluctuations in prey populations closely (Archer, 1985). Moreover, since wasps are central place foragers (delivering forage to one or more static nesting sites), they are likely to have considerable impact locally. It is timely and important to understand the predatorprey interactions of the solitary and social wasps. Until recently, the prey of social wasps were poorly studied; recent molecular analyses of the prey captured by social wasps returning to the nest are providing new insights into the range of insect populations they help regulate (see below). Conversely, there is a vast (but largely forgotten) literature on the prey of some solitary wasp subfamilies from the early 20th century when studying wasp hunting behaviour was a popular pastime of natural historians. However, in recent years solitary wasps have received little research attention, probably because the economic, ecological or societal importance of such research is regarded as low. Here we unite these disconnected data to generate extensive data sets on the predator-prey relationships of solitary and social aculeate wasps. Importantly, our data synthesis provides an evidence base for how solitary and social wasps occupy a mix of discrete and overlapping niches in their impact on arthropod populations, and highlights the ecological diversity of aculeate wasps. Our data sets, especially for the solitary wasps, are unlikely to be exhaustive as we limited our searches to the peer-reviewed literature; we acknowledge that there are likely to be many personal observations of wasp-prey interactions by researchers and amateur naturalists in the grey literature.

(a) Predatory solitary wasps as regulators of arthropods
Solitary wasps represent the vast majority (~97%) of aculeate wasp diversity (Fig. 1), and, with the exception of the Masarinae subfamily (pollen provisioners), all represent predators and parasites of insect and arachnid species, many of which are phytophagous (Goulet & Huber, 1993). Typically, a female wasp will capture a prey item in her mandibles, paralyse it with her sting and transport it to her nest where an egg is laid on the prey item (as opposed to inside the prey, as in parasitoid wasps); when the egg hatches, the larva feeds off the paralysed prey. Further prey may be added as the larva develops (progressive provisioning), or more typically, the brood is provisioned with several prey items soon after the egg is laid, providing sufficient food for the brood to develop fully to pupation (mass provisioning).
We surveyed the peer-reviewed literature to generate a database on the identity of prey captured by solitary  predatory wasps, including, where possible, representative members of each of the 22 families containing solitary aculeate wasps (see Fig. 1 for the major families). Using the Boolean terms "family name" AND "prey" OR "predat*" OR "parasit*" in Web of Knowledge (core databases; from 1900) and secondary searches using the paper-mining online tool, ConnectedPapers.com, we identified 2529 prey records for solitary wasp species from 15 families ( Fig. 3A; Supplementary data set S1). No records could be found for the families Bradynobaenidae, Chyphotidae, Mutillidae, Myrmosidae, Plumariidae, Sierolomorphidae or Thynnidae. Determining whether species are generalist (opportunistic) hunters or prey specialists is difficult to assess from the available data, as many studies sampled prey in nests but did not relate prey diversity to local prey availability (Federici et al., 2007;Polidori et al., 2010). As a functional group, however, solitary wasps are collectively generalists, implicating their importance in the maintenance of balanced ecosystems. Across the 15 families for which data could be collected (Fig. 3A), prey were hunted from across 14 different orders.
Solitary wasps are typically considered to be prey specialists; for example, the Ammophilinae (Sphecidae) and Eumeninae (Vespidae) mostly hunt Lepidoptera, with a minority hunting the larvae of beetles and sawflies; the Euparagiinae (Vespidae) hunt exclusively for weevil larvae (O'Neill, 2001) and the Pompilidae prey entirely on spiders. We found support for specialist hunting behaviours in our data set for 10 of the 15 families, each recorded hunting prey only from within a single order (see Fig. 3A). Analysis at a finer scale is beyond the scope of this review, but there is evidence that prey fidelity is not necessarily even a genus-level trait: some species of Cerceris appear to hunt just one family of beetle, whilst other species are generalist beetle hunters (Evans & Hook, 1986). Prey specialism may be due to taxon fidelity of the hunter, local prey ecology, and/or body size (Polidori et al., 2010). For example, the hunting habits of Sphex ichneumoneus on grasshoppers are thought to be dictated by prey size, as females appear to shift to new prey species when the species hunted earlier in the season becomes too large (Brockmann, 1985). We also identified families with significant levels of generalist hunting behaviour: the Crabronidae is a highly speciose assemblage of predatory species with massively diverse diets (O'Neill, 2001;Polidori et al., 2010Polidori et al., , 2011) that include prey from at least 12 arthropod orders Fig. 3A; Supplementary data set S1). Females may be hunting the most abundant arthropod species in the environment irrespective of ecology, size or other prey traits (Stubblefield et al., 1993). the node size represents the number of prey orders hunted by that specific wasp taxon, i.e. the extent to which each wasp taxon is a generalist or specialist. Blue nodes indicate the arthropod (prey) orders; the node size indicates the total number of wasp species (across all families/ subfamilies) recorded predating on that order. Tie strength (grey lines) indicate the total number of wasp species recorded as predators on each specific prey order. * highlights prey orders that have been recorded as prey of only solitary (A) or only social (B) wasps to illustrate how life history (solitary/social) influences the ecosystem services provided. Nodes are ordered to maximise clarity of visualisation rather than any ecological or evolutionary trait/relationship. Networks constructed using Cytoscape (Shannon et al., 2003).
Biological Reviews 96 (2021)  Unfortunately, quantitative data on the extent of arthropod regulation by solitary aculeate wasps is sorely lacking; but lifehistory, provisioning behaviour, and environment are likely to influence the impact that solitary wasps can have on prey populations. The biomass of prey taken will depend on the predator's brood size, the number of generations per year and the climate. In temperate zones, solitary wasps tend to be univoltine (one generation per year), with the pupae undergoing diapause through autumn and winter before completing development and emerging as an adult in late spring (Spradbery, 1973). In more tropical regions, solitary wasps may have several generations within a year and thus likely consume a larger amount of biomass per individual than solitary wasps with a univoltine life cycle (Spradbery, 1973).
In the absence of any quantitative measures of how much prey wasps consume, it is reasonable to make extrapolations based on the assumption that the total mass of prey provided for each larva must be (at least) approximately equal to the final size the wasp will attain upon completing development (Grissell, 2010). The life history of the wasp species will influence the number of individual prey items, and size of prey taken. For example, parasitic aculeate wasps often provide each of their larvae with a single large prey item [e.g. in the tarantula hawk wasps (Pompilidae)] whilst predatory species may provide each larva with several smaller items, such as in the Crabronidae and the Ammophilinae (subfamily: Sphecidae) (Polidori, 2011a). Within a Californian population of Symmorphus cristatus (Eumeninae; Vespidae), nests were found to contain a mean (SD) of 5 (1.5) occupied cells, each provisioned with 4.8 (1.5) larvae [equivalent to 84.6 (27.8) mg of prey biomass] of the beetle Chrysomela aeneicollis per cell, equating to approximately 423 mg of prey biomass per nest (Sears et al., 2001). Further, Smiley & Rank (1986) found Symmorphus to be a major cause of mortality of the leaf beetle C. aeneicollis in a willow forest in America. Among the Crabronidae, a single nest of Cerceris fumipennis, which hunts exclusively for wood-boring beetles (Buprestidae), may contain up to 83 prey beetles, representing approximately 4.8 g of prey biomass (Kurczewski & Miller, 1984).
The predator-prey relationships of solitary aculeate wasps are understudied relative to social wasps, given they outnumber social wasps by approximately 100:1. Our database provides an overview of how, collectively, solitary wasps are likely to contribute to regulating populations of Lepidoptera, Coleoptera, Diptera, Hymenoptera, Hemiptera and Araneae (Fig. 3A). Given their diversity, numerousness, and cosmopolitan distribution, the ES provided by solitary wasps as natural pest controllers has thus far been undervalued (Sumner et al., 2018).

(b) Predatory social wasps as regulators of arthropods
Although the social wasps account for less than 3% of all wasp species (Aguiar et al., 2013;Bell & Sumner, 2013), their ecological impact as predators is likely to be highly significant because of their large colony sizes, with hundreds to thousands of workers and brood accounting for a large amount of biomass in ecosystems (Thomas et al., 1990;Wilson, 1990;Elizalde et al., 2020). Adult wasps forage for water, wood pulp, nectar and, most importantly, protein (Spradbery, 1973;Richter, 2000). Unlike solitary preyhunting wasps, the larvae of all social wasps develop in an uncapped cell and are provisioned progressively with protein obtained from predation on various arthropod species (Spradbery, 1973).
There has been much more research interest (and success) in determining what social wasps hunt than for solitary wasps. Social wasps are easily located and observed due to their large numbers and central place foraging: by sitting at the nest, observers can collect prey from returning foragers. However, since foragers masticate prey items extensively before returning to the nest, it can be difficult to identify prey using morphology: success varies from 30% (Gambino, 1986), 39% (Sackmann et al., 2000), to 70% (Harris, 1991) and 95% (Harris & Oliver, 1993). These methods may not capture the range of species predated by aculeate wasps. More recently, molecular techniques have been used to identify masticated prey items to the family, genus, or species level (Kasper et al., 2004;Wilson, Mullen & Holway, 2009) and, given the cheaper sequencing technology available due to the genomics revolution (Schuster, 2008) and higher accuracy, these methods are likely to become the norm for prey identification in the future. Another approach is to sequence the gut contents of the larvae using the metabarcoding methods made possible by advances in the resolution of molecular technology and expanding online databases of species-level DNA barcodes (Ward & Ramón-Laca, 2013).
Here we collate data from across these different types of studies to compile a database of the prey of social wasps. The literature search was performed in the same way as for solitary wasps, searching for subfamilies within the family Vespidae that contain social species (i.e. Polistinae, Stenogastrinae and Vespinae). We identified 198 prey records across the three subfamilies of social wasps (Supplementary data set S2). Despite the vastly smaller data set compared to the solitary wasps, the data clearly indicate that social wasps (even at the subfamily level) collectively prey on a huge variety of arthropods; but moreover, in contrast to the solitary wasps, all three subfamilies of the social wasps appear to be highly opportunistic (and thus generalist) foragers (Fig. 3B). By far the largest proportion of data comes from the Polistinae [e.g. Agelaia vicina in Brazil was observed capturing at least 10 orders of insects: Lepidoptera, Coleoptera, Dermaptera, Hymenoptera, Heteroptera, Mantodea, Diptera, Neuroptera, Blattodae and Hemiptera (Oliveira, Noll & Wenzel, 2010); Mischocyttarus flavitarsis was observed predating on flies and caterpillars (Snelling, 1953); and Mischocyttarus drewseni reportedly preys on small spiders, ants and a hemipteran nymph (Jeanne, 1972)], and the Vespinae [e.g. Vespula sulphurea in California are reported foraging on flies, caterpillars and grasshoppers (Akre & Myhre, 1994 and season: the prey of both Vespula vulgaris and Vespula germanica differ depending on the abundance of arthropod families throughout the season, as well as the location of the nest (Harris & Oliver, 1993). Reports on prey types hunted by the third subfamily of social wasps -Stenogastrinaeare more scarce: Parischnogaster nigricans serrei have been reported returning to nests with Araneae, Blatttodea and Diptera, P. mellyi was observed capturing tiny midges in sunny patches of forest or plucking prey from spider webs, whilst Liostenogaster vechtii were observed predating on Blattodea and Hymenoptera (Turillazzi, 1983. The opportunistic foraging of social wasps is apparent from the reported variation in diet composition; for example, the diet of V. germanica varied dramatically with location and season (Harris, 1991;Sackmann et al., 2000); likewise, the diet of Vespa velutina in its invasive French range varied depending on habitat . The success of the social wasps beyond their native ranges (Wilson et al., 2009;MacIntyre & Hellstrom, 2015), where prey are novel, is likely due to their ability to prey opportunistically upon a huge range of arthropods. Irrespective of where social wasps live, there is a predominance of Lepidoptera, Diptera, Hymenoptera, Hemiptera and Coleoptera in their diets ( Fig. 3B; Supplementary data set S2); these orders include many economically and ecologically important species. Finally, it is important to note that although arthropod prey makes up the majority of social wasp diets, many also scavenge vertebrate and invertebrate carrion [suggesting further ecological roles as ecosystem recyclerssee Section IV.2 (Wilson et al., 2009;Wilson & Wolkovich, 2011)] and are even known to prey on live vertebrates, such as frog tadpoles (Warkentin, 2000) and hummingbird chicks (Grant, 1959).
Quantitative estimates of the ecological impact of social wasp predation can be extrapolated from their colony sizes and productivity. Over the course of the colony cycle, nests of the yellowjacket wasp V. vulgaris are estimated to produce an average of 9601 adult wasps, including 7274 workers, 1438 males and 889 queens (Archer, 2003). Assuming a 1:1 larva: prey mass ratio, and an average adult mass of~23, 163 and~85 mg for worker, queen, and male respectively (Archer, 1998), a single V. vulgaris nest may therefore require 0.43 kg of prey biomass during the colony cycle.
Ecological impact can also be measured by counting prey loads collected by foragers on arrival at their colony; such data are available for a few species, but heavily skewed to the ecologically dominant Vespula species in invasive regions like New Zealand. For example, Harris (1991) reported that V. vulgaris can capture up to 4.8 million prey loads per hectare in one season, and Thomas et al. (1990) estimated them to be responsible for the capture of 8.1 kg of prey per season/hectare. This biomass intake by wasps/hectare over just the summer and autumn months is comparable to that of the local insectivorous bird population throughout the whole year (Harris, 1991). In their native range, Vespula typically have an annual colony cycle. In warmer latitudes such as California (Gambino, 1986) and New Zealand (Plunkett, Moller & Hamilton, 1989), colonies can over-winter and as a result be up to 10 times larger than colonies in temperate regions, containing up to 3-4 million cells and producing over 115000 adults in some cases (Spradbery, 1973;Archer, 2003).
Nest density strongly influences the prey biomass removed from an ecosystem: in high-density areas (as found in invasive zones; Lester & Beggs, 2019), V. vulgaris and V. germanica are estimated to remove up to 11.28 kg of prey biomass/hectare/season, whilst in low-density areas, it could be as low as 0.01 kg (Harris & Oliver, 1993). In New Zealand, where V. vulgaris has become a problematic invasive species (MacIntyre & Hellstrom, 2015), Thomas et al. (1990) found a mean of 11.9 nests per hectare. Using the above estimation of 0.43 kg of prey biomass collected per colony, this amounts to at least 5.1 kg of prey biomass consumed per hectare by V. vulgaris alone. Two nests of V. germanica in New Zealand were estimated to have consumed 26.6 and 99 kg of prey biomass respectively over their two-year life cycles (Harris, 1996). However, in their native range overwintering colonies are very rare (accounting for just 4% of nests) and thus their overall predatory contribution was estimated to be 0.16 kg of prey per colony, per season, per hectare (Harris, 1996).
Few data are available on prey biomass taken by species other than Vespula species. Data exist for another invasive species, the Asian paper wasp Polistes chinensis (Polistinae) which has become established in New Zealand (Clapperton, Tilley & Pierce, 1996). Although nests of P. chinensis are far smaller than those of Vespula species, the number of P. chinensis nests was far higher per hectare, resulting in estimates of 0.96 kg of prey biomass/season/hectare (Clapperton, 1999), a similar value to that of the two invasive vespine wasps (Harris & Oliver, 1993). Thus, nest size alone is not indicative of impact on insect populations. Therefore, estimates of nest density as well as prey consumption and colony productivity are required for an estimate of quantitative impact. Further estimates exist for other Polistes species: single colonies of Polistes versicolor are estimated to capture around 4015 prey in a year (Prezoto et al., 2006).
The generalist, adaptable and opportunistic nature of predation by social wasps, together with their capacity to remove many kilograms of insect biomass from habitats each year suggest that the social wasps have a very important role in maintaining balanced phytophagous arthropod populations in natural ecosystems. Accurate quantification of their ecological impact on native and invasive arthropod populations (e.g. Losey & Vaughan, 2006) is essential to fully understand the extent to which they provide this ES, the economic and ecological value of this service and the consequences of changing populations of social wasps in natural ecosystems.
(2) Regulation of agricultural arthropod pests As well as their role as regulators of insect populations in natural ecosystems, aculeate wasps are also likely to be important as bio-control agents. Biological control of pests is likely to play an integral role in meeting the ever-growing Biological Reviews 96 (2021)  demand for food production in future decades (Carvalho, 2006). Unlike chemical control methods (e.g. pesticides), biological control is based on pre-existing predator-prey relationships, meaning that pest resistance is unlikely. The ES provided by biological control has an estimated value of US $417 billion per year (Costanza et al., 1998). In North America alone, the value of natural control provided by insects was estimated at US$4.5 billion annually (Losey & Vaughan, 2006). Most successful examples of biological control using invertebrates have involved parasitoid wasps (Greathead & Greathead, 1992) and their economic value in such programmes is well documented (e.g. Dean et al., 1979;Zeddies et al., 2001;Bokonon-Ganta, De Groote & Neuenschwander, 2002). The reputation of wasps as effective agents of biocontrol entirely refers to the non-aculeate parasitoids, which make up 48.7% of the 230 invertebrate species commercially available for augmentative biological control ( van Lenteren, 2012). The potential roles of aculeate wasps in such programmes have largely been ignored, with only a few notable successes; this is surprising given their ecosystem contributions as predators and parasites.
As specialist predators and parasites, the solitary aculeate wasps offer potential for biocontrol programs. Despite this, only four species of aculeate wasps are commercially available for augmentative biological control practices, all of which are solitary (van Lenteren, 2012): Cephalonomia stephanoderis and Prorops nasuta (both Bethylidae) are used to control coffee berry borer beetle (Hypothenemus hampei); Goniozus legneri (Bethylidae) is used to control lower naval orangeworm (Amyelois transitella) and Ampulex compressa (Ampulicidae) is used to control domestic cockroaches (see Supplementary data set S1). Unfortunately, the number of unsuccessful introductions of aculeate wasps for biological control far outweighs the number of successful introductions (e.g. Krombein, 1948;O'Neill, 2001). This is likely due to a lack of knowledge on life histories and poor management strategies. For example, whilst the introduction of Campsomeris marginella to Hawaii was a success, its introduction to Taiwan to control other sugarcane pests was unsuccessful (Cheng, 1991). Our review of solitary wasp predation (see Section II.2a) highlights that some of these wasp families have more generalist prey choice than commonly assumed, with many species attacking either multiple species within a genus or, in some cases, multiple genera; such lack of prey specificity would preclude their release into novel habitats (Supplementary data set S1). Ultimately, if the success rate of future solitary aculeate biological control programmes is to match that of the parasitoid wasps, then more effort is needed to understand wasp life history fully, prior to introductions.
Social wasps hold untapped promise as biocontrol agents Elizalde et al., 2020). Firstly, they are generalist predators. In a meta-analysis of (non-wasp) generalist arthropod predators, significant reductions in pest numbers were reported in 75% of studies (Symondson, Sunderland & Greenstone, 2002). The generalist predatory nature of social wasps (Fig. 3B) along with their high demand for arthropod protein makes them a potentially effective agent of pest control across agricultural ecosystems. Ironically, it is exactly because of the social wasps' generalist diet that they been largely ignored as a means of biological control: effective biocontrol agents are typically specialist predators or parasites. Secondly, social wasps hold great potential for biocontrol because Lepidoptera and Diptera make up a large part of their preferred diet (Fig. 3B); many lepidopterans and dipterans are significant crop pests. Predatory wasps are common in a range of agricultural environments, suggesting that such crops provide rewarding forage patches (Bommarco et al., 2011;Rodriguez et al., 2012;Souza et al., 2013;Jacques et al., 2015Jacques et al., , 2018Tomazella et al., 2018). Moreover, observational studies report social predatory wasps as one of the main natural predators of lepidopteran pests in agricultural settings (Rabb & Lawson, 1957;Picanço et al., 2006Picanço et al., , 2010Bommarco et al., 2011;Picanço, 2012;Pereira et al., 2018;Lourido et al., 2019) and that they can successfully locate such prey using olfactory cues (Pietrantuono, Moreyra & Lozada, 2018;Southon et al., 2019). A third reason why social wasps are promising biocontrol agents is due to the sheer numbers of hunters provided by each colony; a single V. vulgaris colony has up to 10000 workers and produces thousands of larvae; Vespula has been suggested as a viable candidate for development as a manageable pest controller (Donovan, 2003). Finally, social wasps are well equipped to deal with pest species outbreaks effectively. Many social wasp species appear to lack the efficient forager recruitment traits found in other social insects [e.g. ; although see Overmyer & Jeanne (1998) and Wilson Rankin (2014)]. However, wasp foragers in some species (e.g. Polistes) are known to return to sites where they have previously had success in hunting and to 'fixate' on a particular prey type (Richter, 2000) whilst others are known to be capable of learning from nest-mate workers how to extract caterpillar prey from silk retreats spun on host plants (Weiss, Wilson & Castellanos, 2004). Given the tendency for caterpillars to feed in aggregations (Greeney, Dyer & Smilanich, 2012), such hunting behaviour would cause workers to focus their hunting in areas where caterpillars aggregate. Therefore, social wasps have the potential to control lepidopterous pests directly through predation, and indirectly by breakingup beneficial feeding aggregations, exposing caterpillars to lessoptimal feeding conditions (Nakasuji, Yamanaka & Kiritani, 1976;Stamp & Bowers, 1988).
Despite these compelling traits of social wasps that make them promising biocontrol agents, the experimental research base for their efficacy is sparse, even though their potential was recognised over a century ago; e.g. Polistes were cited as the most important predatory enemies of cotton worm (Alabama argillacea) in several West Indian islands 90 years ago (Myers, 1931;Rabb & Lawson, 1957); Ballou (1915) attributed 10 successful crop years of pesticide-free cotton production on the island of St Vincent to the predatory impact of Polistes annularis, nesting in sheds erected in the fields by planters (Ballou, 1909). More recently, there have been observations of a diverse range of social wasp species predating on pest species in agricultural environments; e.g. Prezoto & Braga (2013)  the fig fly Zaprinus indianus, a pest to many fruit crops in Brazil; seven species of polistine wasps were observed naturally predating on the key tomato crop pest Tuta absoluta in Brazil (Picanço et al., 2011). Observational studies like these clearly show that Vespidae wasps preferentially predate the larger 2nd-5th instar larvae (Gonring et al., 2003;Picanço et al., 2010;Picanço, 2012), without quantifying the impact of this on the prey population. These studies are only illustrative of foraging behaviour, not the foraging capacity of the wasp species, and so do not help estimate of the value of wasps as biological control agents. More recently, an experimental study on the Brazilian species Polistes satan showed that their presence has a significant effect in reducing the population sizes and impact of the globally important Lepidopteran pest, the fall army worm (Spodoptera frugiperda) maize plants (Southon et al., 2019). This study, however, was performed under semi-controlled conditions in a screen house. Field-realistic data on the impact of social wasps as biocontrol agents are lacking.
Social wasps are unlikely to replace other forms of biocontrol (e.g. parasitoids), but the evidence suggests they hold a great deal of promise as part of an integrated pest management programme, especially in developing countries in tropical/sub-tropical climates where social wasps are abundant and relatively tolerant of anthropogenic landscapes, and where farming is small scale (Barbosa et al., 2018;Prezoto et al., 2019). Augmentative control using artificially inflated populations of social wasps to control pest outbreaks shows great promise. The generalist nature of social wasp diets would theoretically keep most pest populations under control, and thus unlike species-specific biological control agents (like parasitoids) they are unlikely to enable other pest species to become prominent (Pearson & Callaway, 2003).

(3) Pollination services
Over 75% of human-cultivated crops worldwide depend directly on (or show increased fruit or seed set as a result of) insect pollination (Klein et al., 2007). Insect pollination is estimated to be worth over $250billion/year worldwide (IPBES, 2016), amounting to at least 9.5% of the value of world agricultural production (Gallai et al., 2009). We collated records of plants that were visited by aculeate wasps using the Boolean terms "family name" AND "pollinat*" OR "flower" NOT "ficus" (to exclude the fig wasps, reviewed elsewhere) in Web of Knowledge. We identified 2288 unique flower-wasp visitation records for aculeate wasp species from seven families ( Fig. 4; Supplementary data sets S3 and S4). We categorised these records into those for plants that are known to rely solely on aculeate wasps for pollination (obligate pollination; Supplementary data set S3), and those where wasps have a non-specialist relationship with plants, visiting them for nectar opportunistically and thus having the potential to contribute to their pollination (facultative pollination; Supplementary data set S4). Our survey reveals relationships of 960 plant species with aculeate wasp pollinators.
(a) Obligate pollinationwhere plants rely solely on aculeate wasps for pollination We identified 164 plant species across six plant families and four orders that are reported to depend solely on aculeate wasps for pollination ( Fig. 4A; Supplementary data set S3). Wasp pollination of orchids is an example of coevolution between the plant and the wasps. Pollination by chemical deception is primarily restricted to the Orchidaceae (Gaskett, 2011), although such deception has been reported outside of this group (Ellis & Johnson, 2010;Vereecken et al., 2012). Within the orchids, sexual deception is one of the most specialised forms of chemical deception used to achieve pollination, with the Scoliidae and Thynnidae wasp families especially important in this relationship (Supplementary data set S3). Such deception involves the attraction of male wasps to the orchid through speciesspecific mimicry of the appearance and sex pheromones released by females (Mant et al., 2002;Ayasse et al., 2003). Upon attempting to copulate with the flower, pollinaria are attached to the male and are then transferred to the next flower the male attempts to copulate with, completing the pollination process (Stoutamire, 1975) (Fig. 4B). Such interactions tend to be highly species specific, ensuring that hybrids are not formed through cross-species pollination (Mant et al., 2002). Furthermore, due to orchid chemical profile changes after pseudocopulation, males tend not to revisit the same plant, enhancing the reproductive success of the orchid and ensuring that self-pollination does not occur (Schiestl & Ayasse, 2001).
Orchids also deceive wasps by falsely advertising the presence of food. At least two species of orchid, Epipactis helleborine and E. purpurata, employ food deception to achieve specialist pollination from the social wasps V. germanica and V. vulgaris (Brodmann et al., 2008). The orchids advertise the presence of caterpillar prey (a protein reward) through the floral release of green-leaf volatiles, mimicking the chemical profiles of plants under attack from phytophagous insects, to attract their social wasp pollinators (Brodmann et al., 2008). Other orchids release honey bee alarm pheromone to attract hornets for pollination (Brodmann et al., 2009). Given the location in which these orchids grow (dark forest understorey) and that their green-leaf volatiles fail to attract other potential pollinators, it is thought that both species are solely reliant on social wasps for pollination (Brodmann et al., 2008).
Besides orchids, plants that are solely reliant on wasps for pollination include representatives of the Apocynaceae, Asparagaceae, Euphorbiaceae and Iridaceae families, most of which are pollinated by spider wasps (Pompilidae) in the genera Hemipepsis and Pepsis (Supplementary data set S3). This guild of plants demonstrates highly convergent traits to attract their spider wasp pollinators, such as cryptic floral colouring and synchronised blooming periods when the wasps are at their most active (Shuttleworth & Johnson, 2012). Furthermore, to deter other non-pollinating insect species, certain plants within the guild produce bitter, unpalatable nectar which is preferentially Biological Reviews 96 (2021)   consumed by spider wasps (Shuttleworth & Johnson, 2006, 2009b. (b) Facultative pollinationwhere plants are opportunistically pollinated by aculeate wasps Whilst the diet of wasp larvae is heavily reliant on protein obtained from arthropods, adult diets consist largely of carbohydrates, most of which are obtained by feeding on nectar from flowers (Spradbery, 1973). During these visits to flowers, wasps may inadvertently transfer pollen between flowers (Fig. 4C). The efficacy of this putative facultative pollination is almost completely unstudied. The data we present here, therefore, are of plant-wasp relationships that could potentially constitute pollination services. We identified 798 plant species across 106 families and 35 orders which fit this category (Supplementary data set S4; Fig. 4A). Due to sparse reporting, the number of plant species we report is likely to be a vast underestimate of the true number that have the potential to benefit from wasps. One exception are the well-studied wasp-plant interactions reported for Southern Africa (Gess & Gess, 2010). Elsewhere, partial pollination networks have been reported for the paper wasps (Polistinae) of South America (Santos, Aguiar & Mello, 2010;Mello et al., 2011;Clemente et al., 2012), and the vespid wasps of the Crimea (Fateryga, 2010) and various localities across Asia (Kato, 2000;Corlett, 2001).
The Vespidae family are the most dominant generalist pollinators ( Fig. 4A; Supplementary data set S4): variability of flower species visited is extremely high, with wasps such as Brachygastra lecheguana pollinating at least 76 plant species across 29 families, whilst other wasps such as Celonites pulcher (which pollinates two species both belonging to the Scrophulariaceae) and Ceramius cerceriformis (which pollinates two species both belonging to the Aizoceae) seem restricted to the pollination of very few species (Supplementary data set S4). Across temperate to arid zones, the vespid subfamily Masarinae (pollen wasps) seem to be the most important wasps for generalist pollination (Gess & Gess, 2010; Supplementary data set S4). These wasps have reverted to a phytophagous lifestyle and exclusively provision their larvae with pollen instead of insect prey (Gess & Gess, 2010). It has been suggested that pollen wasps should be considered to perform the same functional pollination roles as bees in some ecosystems (Gess & Gess, 2004).
Social wasps (Polistinae and Vespinae) are likely to be especially important as facultative pollinators, across both temperate and tropical regions (Supplementary data set S4). Within plant-pollinator networks, social wasps show high degrees of connectivity; and are generalists in terms of the plants they visit . This is likely a result of their social nature, with multiple foraging workers utilising the most abundant species of plants within the foraging range of their nest for nectar collection . Thus, there are high levels of overlap between plants and their wasp pollinators, producing a plant-pollinator network that is highly robust to extinction on either side of the mutualism (Mello et al., 2011). Because of their lack of preference for specific plants, social wasps may be more abundant than bees in degraded or fragmented habitats and thus play an important role as 'backup' pollinators (Mello et al., 2011).
From an economic perspective, wasps have also been found to pollinate a number of important crops, providing similar 'Pollination Importance Values' to the key greenhouse pollinator, the bumble bee (Bombus spp.), in the pollination of mustard, apples, and pumpkins (Akhter et al., 2016). In some cases, wasps are considered to be more important for generalist pollination than bees. For example, the increased mobility of the sphecid wasps Bembix borrei and Sphex spp. allowed high levels of movement between patches of the shrub Allophylus serratus, ensuring rapid and efficient cross-pollination when compared to various bees, which tended to focus on a single plant (Aluri, Reddi & Das, 1998). Remarkably, a recent experiment showed that Polistes wasps were able to completely replace the pollination services of bumble bees in greenhouses, proving as efficient in pollen transfer as bumble bees (Hallett et al., 2017). This study alone brings into focus the extent to which the pollinating services of social wasps have been overlooked.
The paucity of data on wasps as pollinators, in both natural and farmed environments currently precludes putting a value on their services as pollinators. Given the importance of natural pollinators to human health, welfare and food security (Klein et al., 2007;Gallai et al., 2009;Potts et al., 2016), and current concerns over declining populations of well-recognised pollinators such as bees (Potts et al., 2010;Senapathi et al., 2015), it is timely to investigate in more detail the pollinating value of aculeate wasps. Their generalist interactions with plants, their potential for pollinating in degraded habitats, and the apparent resilience of some species to anthropogenic change (Jonsson et al., 2021) warrant more research effort to quantify their potential.

III. PROVISIONING SERVICES
(1) Aculeate wasps as nutrition for humans Insects are a viable healthy alternative to eating meat, being high in protein and essential amino acids. Insect farming offers many advantages over traditional meat production, using less space and water, emitting fewer greenhouse gases and ammonia, and generating a much higher foodconversion efficiency than livestock (van Huis, 2013). At least two billion people across the globe consume insect protein (entomophagy) as part of their diet: this includes over 2000 insect species, with the orders Coleoptera (31%), Lepidoptera (18%) and Hymenoptera (15%) being the most prominent (Jongema, 2017). Wasp larvae have an exceptional dry protein mass (ranging from 46% in Polistes sagittarius up to 81% in Polybia spp.), typically contain around 70% of required amino acids and have an extremely low fat content Ramos-Elorduy, 1997;Chen, Feng & Chen, 2009 Wasp ecosystem services et al., 2010). We searched the literature to gather information on the range of aculeate wasp species eaten around the world using the search terms: "Entomophagy" OR "edible" OR "human food" as topics, and "Insect" OR "wasp" OR "Aculeata" as titles, across all years available in Web of Knowledge. Wasps account for 4.8% of the insect species consumed: we detected 109 wasp species reported as being eaten across 19 countries (Supplementary data set S5). This is likely to be an underestimate since many wasps sold for human consumption can only be identified to genus level and data are only known for a small number of countries.
Typically, wasps are eaten in their larval or pupal stages and thus mainly social species are utilised, since the nests of solitary species are rarely large enough to merit the efforts of collection (Ruddle, 1973; Supplementary data set S5). During the autumn harvest in Japan, wasp nests typically sell for US$100 per kg and demand is often so high that markets must be supplemented by the importation of further nests from China, New Zealand, and the Republic of Korea (Nonaka, 2010). Wasps are a popular street food in East Asia, Africa, and South America (Ramos-Elorduy, 2009). In rural China, wasps are the most common edible insects for sale; 18 species have been recorded at market (although many more remain unidentified) (Chen et al., 2009). Furthermore, in tropical climates, the perennial colony cycles of wasps means they are reliable food source which can be collected year-round (Acuña et al., 2011).
Several wasp genera produce honey (e.g. Brachygastra, Parachartergus, Polybia, Mischocyttarus and Polistes) (Sugden & McAllen, 1994;Hunt et al., 1998;Guimarães, De Castro & Prezoto, 2008), which is a valued food source across Latin America (Onore, 1997;Bentley & Rodriguez, 2001): Brachygastra mellifica hoards excess nectar collected from floral visitation, and steals high-energy honey of various honeybee species (Sugden & McAllen, 1994). Honey stored by Brachygastra, Polistes and Polybia contains high levels of fructose and a diverse array of amino acids, suggesting that wasp honey is a nutritious source of energy (Sugden & McAllen, 1994;Hunt et al., 1998). However, wasp honey production is unlikely to be commercially viable because the honey output of a typical wasp hive tends to be far smaller than the 0.45 kg of honey produced per colony per day by the European honey bee (Rinderer, Collins & Tucker, 1985;Sugden & McAllen, 1994).
Finally, social wasps provide safe overwintering havens for wine-producer's yeast Saccharomyces cerevisiae. Yeast cells survive in the intestines of overwintering queens of Vespa crabro and Polistes dominula; the yeast is transferred to workers and future foundress queens through trophallaxis, and finally distributed to vineyards (Stefanini et al., 2012). The intestines of social wasps provide the appropriate conditions for yeast reproduction, with multiple strains mating with one another whilst inside their vector, maintaining local phenotypic variability of yeast strains  and suggesting an important role of wasps in the evolution and genetic diversity of S. cerevisiae (Stefanini et al., 2012. (2) The medicinal properties of aculeate wasps and their products The venom of aculeate wasps is a complex mixture of toxins, allergens, enzymes, and amines (de Graaf et al., 2009) which are utilised in a wide range of biological functions, such as self-defence (or colony defence in the case of social vespids), prey capture and host manipulation. Given their compositional complexity, it is unsurprising that wasp venoms contain many biologically active molecules which are of significant pharmacological interest. Of recent interest is the potential of wasp venom in the treatment of cancer (Heinen & Gorini da Veiga, 2011). The most extensively studied of these molecules is mastoparan, a family of amphipathic peptides found in the venom of social wasps (Hirai et al., 1979;Argiolas & Pisano, 1984;Wang et al., 2008b) which show considerable promise for the treatment of cancer (Yamada et al., 2005;Moreno, Zurita & Giralt, 2014;Leite et al., 2015).
The antibiotic properties of wasp venom, larval secretions (e.g. Turillazzi et al., 2004;Herzner et al., 2013;Moreau, 2013) and nest materials have long been recognised, with use in historic and contemporary folk medicine (Wang et al., 2008a). The venom of solitary and social wasps has antimicrobial properties. Solitary wasps use antibiotics to keep their paralysed prey healthy for the developing offspring to feed off. Anoplin, a peptide isolated from the venom of Anoplius samariensis (Pompilidae), and Eumenitin, isolated from the venom of Eumenes rubronotatus (Vespidae), show strong inhibitory activity against both Gram-positive and Gramnegative bacteria (Konno et al., 2001(Konno et al., , 2006. Furthermore, larvae of the emerald cockroach wasp, Ampulex compressa secrete large amounts of oral antimicrobials whilst inside their cockroach hosts, providing defence against the bacterium Serratia marcescens (Herzner et al., 2013), an opportunistic human pathogen.
The life-history challenges facing social wasps have driven the evolution of anti-microbial defence systems. Like all social insects, social wasps are vulnerable to pathogens due to the reduced genetic diversity within colonies and high chances of pathogen transmission among colony members (Cremer, Armitage & Schmid-Hempel, 2007). Social wasps combat this using anti-microbials in the venom and in the larval secretions with which workers coat themselves, the brood, and the nest (Turillazzi et al., 2004. Many of these antimicrobials may be of direct benefit to human health. For instance, Polydim-I, a peptide isolated from the venom of Polybia dimorpha, exhibited promising anti-microbial activity against Mycobacterium abscessus both in vitro and in vivo (Das Neves et al., 2016), a mycobacterium which commonly causes soft tissue and skin infections after invasive surgery (Lee et al., 2015). Further examples include Dominulin A and B, peptides isolated from the venom of Polistes dominula, which were effective against Gram-negative and Gram-positive bacteria  and the larval secretions of P. dominula, which showed anti-microbial activity against Bacillus subtilis and Escherichia coli (Turillazzi et al., 2004). Finally, wasp products have a history of use in traditional medicine. For instance, quinone 7,8-seco-para-ferruginone isolated from the nests of Vespa simillima exhibits potent inhibition of rat hepatoma cells (Fujiwara et al., 2008); the peptide Nidus Vespa Protein 1 found in nest comb collected from a variety of social wasps including Polistes olivaceous, P. japonicus and Parapolybia varia promotes apoptosis and inhibits the proliferation of hepatoma cells, implicating its potential as a future treatment for liver cancer (Wang et al., 2008a). Nest comb components also show potential as a novel oral anti-microbial, reducing the biofilm capabilities of Streptococcus mutans (a bacterium heavily associated with dental plaque and decay) (Xiao et al., 2007;Guan et al., 2012). Furthermore, the flavonoids quercetin and kaempferol isolated from nest comb were shown to considerably inhibit the growth of various Streptococcus, Actinomyces and Lactobacillus species (Guan et al., 2012). Amongst the solitary wasps, nests of the mud-daubers (e.g. Sceliphron and Synagris) are a popular source of essential minerals, such as magnesium, calcium, manganese, iron and zinc for pregnant women and children in parts of rural Africa as a form of insect earths (geophagy) (Fairhead, 2017). The effectiveness and active ingredients of other wasp products (e.g. 'hornet juice', which contains amino acids found in the trophollaxis fluid produced by V. mandariania hornets and is marketed for 'boosting endurance') remain undetermined.
Determining the extent of provisioning services offered by the medicinal properties of wasp products requires a more comprehensive research base, along with clinical trials. The evidence so far, however, suggests that wasp nests and wasp venom exhibit high medicinal potential, especially as antimicrobials and possibly for cancer treatment.

IV. SUPPORTING SERVICES
(1) Vespicochorywasp-mediated seed dispersal Insects are important for seed dispersal; for example, myrmecochory (ant-mediated seed dispersal) is a common and widespread method, occurring in at least 11532 angiosperm species across 334 genera and 77 families (Lengyel et al., 2009). The presence of a lipid-rich appendage (known as the elaiosome) on the seed causes ants to transport it to the nest, feed the elaiosome to their brood, and discard the intact seed, aiding in the seed's dispersal (Beattie & Hughes, 2002). A similar behaviour (known as 'vespicochory') has been reported in Vespula species (Pellmyr, 1985), suggesting that, along with ants, social wasps may also play a role in seed dispersal. We searched the literature for reports of vespicochory using the search terms "seed dispersal" OR "vespicochory" AND "wasp" OR "Vespidae" as topics, across all years available in Web of Knowledge. The search returned vespicochorous interactions occurring between 10 plant species and 12 social wasp species (either in the subfamily Vespinae or Polistinae) in Asia, Europe, and North America (Supplementary data set S6).
These vespicochorous interactions involve the direct dispersal of seeds by worker wasps (e.g. Zettler, Spira & Allen, 2001;Chen et al., 2017;Burge & Beck, 2019) or the theft and subsequent dispersal of seeds from foraging myrmecochorous ants (Bale et al., 2003). The degree of reliance on wasps for seed dispersal appears to differ among plant species. For instance, Vespa velutina represents the primary disperser of Stemona tuberosa seeds, with wasps dispersing 100% of available seeds from parent plants (Chen et al., 2017), whilst Vespa affinis dispersed 33% of Aquilaria malaccensis seeds, with parent plants relying on the action of wind and gravity for dispersal of the remaining seeds (Manohara, 2013). Interestingly, Vespula species were found to remove more Trillium discolor seeds than the ant species Aphaenogaster texana, Formica schaufussi and Solenopsis invicta, suggesting their importance for seed dispersal (Bale et al., 2003). Furthermore, the presence of Vespula in this myrmecochorous system seemingly increased the foraging efficiency of the ants, suggesting that Vespula may also indirectly impact seed dispersal (Bale et al., 2003).
The large foraging range of wasps suggests that the process of vespicochory may disperse seeds further afield than ants (average dispersal: 1.99 m; Gómez & Espadaler, 2013), thus influencing population genetic structure, densitydependent mortality, and range expansion of plant species (Jules, 1996;Zettler et al., 2001). Unfortunately, data on seed dispersal distance and seed fate from vespicochory are lacking and, thus, we are unable to verify the importance of this supporting service. Three studies have investigated dispersal distances from vespicochory to date, with distances ranging from 1.4 m (range = 0.1-2.6 m, N = 41) in vespicochorous interactions between Vespula and Trillium species (Zettler et al., 2001) to 80 m (N = 40) and 110 m (range = 5-150 m, N = 2138) in interactions between Vespa species and Aquilaria sinensis and Stemona tuberosa respectively (Chen, Liu & Sun, 2016;Chen et al., 2017).
We suggest that vespicochorous interactions between plants and social wasps may be more common than the literature currently suggests, with vespicochory representing an underappreciated service in ecosystems for social wasps, especially given the large dispersal distances of which they are capable (Chen et al., 2016(Chen et al., , 2017. The true value of this wasp ES is currently obscured due to the lack of data on seed dispersal efficiency, seed dispersal distance, and seed fate. (2) Decomposition and nutrient cycling The role of wasps in decomposition and nutrient cycling is almost entirely unstudied. Wasps primarily predate live prey. Social wasps, however, have been reported engaging in necrophagy (the eating of dead or decaying animals); carcasses are a source of amino acids, lipids, protein, carbohydrates, vitamins and minerals, although the quality of these resources depend on the taxon and stage of decomposition (Roubik, 1989). Reports of wasp necrophagy appear in the early literature (Ducke, 1910;Richards & Richards, 1951;Cornaby, 1974); e.g. Bertoni (1912)  carcasses can be stripped clean of flesh within hours by the actions of social wasps; anecdotal accounts of meat-eating by Agelaia wasps in South American countries abound, where these wasps have been given the local (Spanish) name of 'carniceras' (meaning 'butchers') (O'Donnell, 1995). Mammalian protein has also been detected in molecular analyses of wasp gut contents (Kasper et al., 2004). This raises the possibility that wasps may contribute to the decomposition of waste products and to the recycling of nutrients.
Previous evidence suggests necrophagy by polistine wasps in Central and South America, with the swarm-founding genera Polybia, Agelaia, Angiopolybia, Brachygastra, Parachartergus, Protonectarina and Synoeca feeding on a wide range of wild 'carrion' from large insects to birds, reptiles, and mammals, to agricultural waste such as poultry, pigs and rats, and tuna fish baits laid out for ants (O'Donnell, 1995;Gomes et al., 2007;Somavilla, Linard & Rafael, 2019). More recently, Apoica and Polybia have been reported scavenging on fresh and decaying pig carcasses under natural conditions in Brazil (Simões et al., 2013); six species of Epiponines in Brazilian rainforest were detected using carrion traps (Silveira et al., 2005) and 10 species were collected from carcasses of rats in similar environments , with different species compositions on fresh and decaying cadavers; interestingly, wasps visiting the heavily decomposed cadavers were also observed preying on the fly eggs and larvae of Sarcophagidae and Calliphoridae (typically the first flies to colonise such carcasses) that were present. Finally, Vespula species are reported to scavenge on honey bee corpses (Pusceddu et al., 2018).
It has been suggested that wasps are most likely to visit carcasses that are in an advanced stage of decay and thus may contribute little to the decomposition process itself. Reports of wasps attempting to extract tissue from live vertebrates, and also freshly killed carcasses suggest that this is not always the case and that the cues used to find the prey are not directly linked to the decay process itself. Wasps may additionally prey on the larvae laid by flies on carcasses, thus contributing a regulatory service by reducing fly populations (Gomes et al., 2007), but also a cultural service through forensic analyses (see Section V.1). The possible effects of social wasps as consumers of fresh and decaying flesh may be substantial, but further research into the extent to which their presence speeds up decomposition is required.

V. CULTURAL SERVICES
(1) Aculeate wasps as biological indicators Biological indicators are taxa that reflect the state of an environment, the effect of environmental change on a habitat or ecosystem, or the diversity of a subset of taxa within an area (Duelli & Obrist, 2003). Due to their important ESs, ubiquity and abundance, insects make excellent biological indicators (McGeoch, 1998). Theoretically, the predatory and parasitic aculeate wasps should be particularly useful in indicating the health of an ecosystem: by ranking highly in food webs, their presence in a habitat indicates the presence of an abundant and diverse community of prey arthropods . Moreover, unlike vertebrate keystone predators, sampling wasp populations and measuring their abundance is relatively cheap and easy .
Both solitary and social aculeate wasps appear to be useful bioindicators. For example, solitary apoid wasp diversity correlated well with habitat quality in a natural park in Portugal (Vieira, Oliveira & Gayubo, 2011): wasp diversity was highest in mixed juniper woodland, a wellprotected native habitat, and lowest in eucalyptus plantation and fallow land, habitats that have been ecologically damaged by human intervention. Similarly, social wasp diversity correlated significantly with vegetative diversity in three contrasting Brazilian ecosystemsmangrove swamp, Atlantic rain forest and lowland sandy regions (Santos et al., 2007). A further study in Brazil also found the social species Pseudopolybia vespiceps and Polybia fastidiosuscula to be excellent indicators of healthy forest ecosystems with high conservation value, whilst Mischocyttarus drewsenii was an indicator of the most disturbed habitats (de Souza et al., 2010). Social wasps are also useful indicators of environmental contamination: analysis of larval faecal matter of both Dolichovespula saxonica and Polistes dominula reveals levels of lead and other heavy metals in the environment (Kowalczyk & Watała, 1989;Urbini, Sparvoli & Turillazzi, 2006); there is some evidence that the degree of melanisation in the facial marking of Vespula wasps is influenced by the level of heavy metals (Skaldina et al., 2020). Such studies imply potential uses of solitary and social wasps for assessing ecosystem health, including the success of rewilding projects (Henson, Craze & Memmott, 2009).
Wasp larvae guts are promising units for arthropod biodiversity monitoring. Their generalist diets, and the sessile nature of wasp nests means sampling is likely to be holistic and geographically consistent. Metagenomic sequencing of the gut contents of parasitoid wasp larvae has revealed the identity of the larval host species of parasitoid wasps (Rougerie et al., 2011). Sampling larval guts of aculeate wasps throughout the season could potentially provide data on seasonal variation in insect abundance and diversity. Furthermore, wasps may be used as a method of 'biosurveillance' across ecosystems. For instance, the prey range of Cerceris fumipennis (Crabronidae) includes the emerald ash borer (Agrilus planipennis), a highly invasive and damaging pest species of ash trees (Fraxinus spp.) across northern Europe and North America, and this wasp has been used to detect the presence of emerald ash borer, a method which far surpasses comparable human technology (Careless et al., 2009).
The association of wasps with animal carcasses (see Section IV.2) also raises the possibility that wasps could be indicators of decay and thus of use in entomological forensics. Different species of wasps have been reported at the same carcasses in the early and late stages of decay, raising the possibility that the type of wasp species could indicate, albeit qualitatively, the time since death. Unfortunately, wasp necrophagy may also add confusion to forensic investigations: the insects tend to remove lumps of tissue from cavities such as noses, mouth, ears and anus, altering the skin texture, size of cavities, and apparent injuries that may confuse postmortem results, leading to errors in forensic investigations regarding the cause or mode of death (Moretti et al., 2008;Simões et al., 2013;Somavilla et al., 2019).
(2) Wasps in culture and recreation Wasps have a historically rocky relationship with humans, and this is apparent in cultures through the ages (Jones, 2019). The first published entomologist, Aristole, wrote: "Hornets and wasps…are devoid of the extraordinary features which characterize bees; this we should expect, for they have nothing divine about them as the bees have." God sends hornets as punishment to sinners in three books of the Bible (Deuteronomy 7:20; Exodus 23:28; Joshua 24:12). The novelist E.M. Forster (1924) in A Passage to India describes how his lead character on a discussion about which animals might be invited into heaven was able to accept that perhaps monkeys and jackals were welcome, but was uneasy with the idea that there may be a place in heaven for wasps. Science fiction writer Eric Frank Russell (1957) gave his 'terrorists' handbook' novel the title Wasp to depict the secretive, persistent and pernicious nature of the protagonist. The dislikeable nature of wasps perpetuates through film: the mistreated female star of the thriller, The Wasp Woman, becomes an evil, man-devouring 'wasp' by night (Corman, 1959); while over-sized fire-breathing wasps star as the antagonists in Dragon Wasps, an action movie set in the jungles of Belize. More positive media interpretations include that of the Marvel superhero 'The Wasp' (Reed, 2018). Wasps have also inspired art; e.g. wasps feature on ornaments and artwork from Asia, especially Japan, where evidence of appreciation of the beauty of wasps is apparent in pottery depicting the nest structures of social wasps, and as predators in scrolls dating back to the Qing dynasty (17th-18th century) (Jones, 2019).
Sadly, these few favourable depictions of wasps in art and culture are far out-numbered by the anger, hatred and general negativity expressed towards them from ancient to modern times. There are two reasons why people feel negatively towards wasps. The first is that they sting: in a survey of 750 members of the public, the word 'sting' was used overwhelmingly to describe wasps, but not bees (which also sting); the same study revealed that people believe that there is nothing useful about wasps, whereas the same group of people recognised the importance of bees as pollinators (Sumner et al., 2018). Facets of nature that are uncharismatic and of no perceived value to humans receive the least attention by conservation initiatives (Daily et al., 2000). Determining the ESs offered by a particular facet of nature has become the currency by which nature is valued and liked by humans (Redford & Adams, 2009).
The many and important ESs provided by stinging wasps (Table 1; Fig. 2) may provide the currency for humans to shift their perceptions of wasps to a more positive and tolerant one. Effective marketing of the benefits of wasps as a valuable, natural commodity that contributes to our health, food security and health (Hart & Sumner, 2020) may help elevate wasps towards a more valued status, alongside other beneficial insects.

VI. A RESEARCH ROADMAP FOR VALUING THE ECOSYSTEM SERVICES OF ACULEATE WASPS
We have collated from a vast, disconnected literature, a holistic evidence-base that aculeate wasps provide a wide array of important ESs (Table 1; Fig. 2). To date, these services have largely gone unappreciated. One of the reasons for this is that aculeate wasps are poorly studied relative to other insects whose ESs are well recognised; e.g. bees, hoverflies, beetles and even their close relatives the parasitoid wasps. We lack the necessary comprehensive information on aculeate wasp life histories and the experimental data to make quantitative and qualitative assessments of the magnitude of their impact.
Here we outline some of the key questions that would help address these knowledge gaps and put the ESs of wasps on the map, alongside their relatives.
(1) Quantifying the impact of wasps as regulators of arthropod populations The most important service provided by aculeate wasps is undoubtedly their role as nature's pest controllers in both natural and agricultural landscapes. In almost all cases, we lack direct quantification and experimentation. Outstanding questions include: (a) In which natural ecological networks are social wasps key predators, and how will these networks be affected by changing populations of wasps?
We have shown qualitatively the diversity of arthropod orders that are preyed upon by social and solitary wasps (Fig. 3); but we have few quantitative data on of the quantity of prey items removed by wasps in a life cycle. Exclusion experiments (e.g. see Gould & Jeanne, 1984) are required to quantify the extent to which arthropod populations are regulated by different wasp species; such data will also provide insights into the potential ecological consequences of declining wasp populations (Archer, 2015;Bueno, Souza & Clemente, 2019;Outhwaite et al., 2020) and the impacts of introduced wasp species across their invasive ranges (Beggs et al., 2011;Lester & Beggs, 2019 (b) What is the economic value of aculeate wasps as agents of biocontrol in agricultural landscapes?
Our review of wasp predation suggests that wasps hold great potential in controlling crop pests and we highlight the ways in which social wasps in particular may be especially effective biocontrol agents. Data are still sorely lacking on the extent to which wasps are effective in regulating pests in fieldrealistic situations, in the company of other predators and in complex ecological interactions. Moreover, the methods by which the predatory services of aculeate wasps can be harnessed and applied safely remain almost entirely unexplored. Vespiaries (wasp 'hives', akin to apiaries for bees) have been trialled only rarely (Gould & Jeanne, 1984;Donovan, 2003), and may only suit certain wasp species whose nests can be easily relocated; furthermore, managing wasps for biocontrol also demands sociological research in order to successfully implement strategies to surmount the negative cultural attitudes that people have towards wasps (Sumner et al., 2018). Addressing the logistical, sociological and husbandry issues may be a greater challenge than answering the biological questions.
(c) Can life history and taxonomy be used to predict the quality and scope of biocontrol services provided by aculeate wasps?
Our databases on prey types of solitary and social aculeate wasps (Supplementary data sets S1 and S2; Fig. 3) identify life history (solitary versus social) as a primary indicator of the type of biocontrol services that wasps can provide, but also highlight how taxonomy (at the family or subfamily level) can also indicate specialist prey choice. It is clear that social wasps are, without exception, generalists in their prey choice (Fig. 3B). Although there are a great deal more data available on the prey of solitary aculeate wasps than social wasps, we lack a comprehensive understanding of the life histories of the vast majority of subfamilies, and as a consequence an understanding of their ecological impacts are taxonomically biased and poor in breadth and depth, especially when compared to the parasitoid wasps (Hawkins, Cornell & Hochberg, 1997). This lack of knowledge may account for the low utilisation of aculeate wasps in biological control programmes when compared to the parasitoid wasps (Greathead & Greathead, 1992;O'Neill, 2001) and their lack of commercial availability in augmentative biological control programmes ( van Lenteren, 2012).
(2) Determining the contributions of aculeate wasps as pollinators in natural and agricultural landscapes We identified 962 plant species, of diverse taxonomy, which depend either solely or partially on wasps for their pollination (Fig. 4). These include a number of orchids and shrubs which are solely reliant on wasps for pollination, without which they would go extinct; there is also evidence to suggest that wasps sometimes are at least as effective crop pollinators as bees. Extensive plant-pollinator networks for aculeate wasps only exist for southern Africa and partial pollination networks exist for areas in South America, Asia, and the Crimea (e.g. Kato, 2000;Corlett, 2001;Fateryga, 2010;Clemente et al., 2012); such data are lacking for other locations where wasps are abundant. The plant-pollinator reports for aculeate wasps are largely limited to flower visitation records and rarely quantify the effectiveness of pollen transfer between flowers by wasps. Thus, we lack knowledge on the extent to which wasps contribute to the reproductive success of the plants they visit, and how wasps compare as pollinators to other well-studied species. It remains to be seen whether wasps may be effective back-up pollinators in poor-quality habitats where specialist pollinators such as bees and hoverflies may struggle to maintain populations.
(3) Developing social wasps as biological indicators Aculeate wasps offer potential as bioindicators of habitat quality, biodiversity, and pollution from the species abundance, larval gut contents and potentially their body markings, respectively. However, the evidence for all these utilities is nascent and each requires substantial experimental research before wasps can be exploited for these services in any meaningful way.
(4) Harnessing the provisioning services of aculeate wasps Societal benefits to humans offered by aculeate wasps include their role as a food source (Supplementary data set S5), providing a source of highly nutritious protein with low fat content in the diets of over 2 billion people worldwide. Their potential contribution to sustainable food production is underappreciated in most parts of the world; however, harvesting wasp colonies for nutrition is practiced by diverse communities in remote locations from rural India to Japan and Latin America, providing evidence that the logistical challenges of working with stinging insects can be surmounted if the prize is valued. At a time when sustainable food security for a growing human population is of global concern, alternative sources of protein are likely to become of critical economic and ecological importance. Along with other insect protein sources (e.g. soldier flies), social wasp larvae hold great potential due to the large quantity of brood to be harvested from a single colony. Finally, from a biomedical perspective, there is a growing research field exploiting the biochemical components of social wasp venom for antimicrobial properties and use in the treatment of cancer and other diseases.

(5) Documenting drivers of change for wasp populations
The impact of 'insectageddon' (Thomas, Jones & Hartley, 2019) is as relevant to the lesser loved insects, such as wasps, as it is to more beloved insects like bees and butterflies (Sumner et al., 2018). We herein provide an evidence-base on 'the point of wasps', encouraging scientists, policymakers, conservationists and members of the public to care about aculeate wasp populations and how they may be affected by changing environments. Occurrence records for aculeate wasps suffer the same paucity of attention as other aspects of wasp biology . Recent analyses of 100 years of records on UK species highlight how some species appear to be relatively resilient to anthropogenic change (e.g. Vespula vulgaris), whilst others are declining rapidly (e.g. Vespula rufa) (Jonsson et al., 2021;Outhwaite et al., 2019); this pattern is typical of most insect groups (Outhwaite et al., 2020) and calls for more tolerant approaches to dealing with wasps in human living spaces and in their conservation management. Given the potential contributions of aculeate wasps through the ESs they provide, tackling declines in wasp populations (e.g. Outhwaite et al., 2019) could prove as important as ameliorating the decline of other 'high value' insects, such as bees.

VII. CONCLUSIONS
(1) We have collated data from a broad and disparate literature on how the varied life-histories of aculeate wasps impact the ecosystems they inhabit through pest control and pollination, and their broad-reaching potential for economic impacts on agriculture and human health. Our synthesis of these data provides the evidence base to show that aculeate wasps offer significant value to natural and agricultural ecosystems, the global economy, human health and society.
(2) Aculeate wasps offer important ESs as predators of arthropods. The hunting behaviour of both parasitic and predatory aculeate wasps is likely to play important roles in regulating arthropod communities across habitats worldwide. This suggests they are a fundamental part of the wider functioning of ecosystems, with a key role in the maintenance of biodiversity. Our synthesis of the data affords the identification of key differences in the ecological roles of aculeate wasps that may be due to their life histories. Solitary wasps are renowned for being prey specialists (Grissell, 2010); our data analysis indicates that some families are in fact more holistic in their hunting habits than previously thought. Further analyses of these groups are required to determine the extent to which prey choice is specialised at the genus, subfamily or species level. Taxonomy, therefore, may be a key predictor of the type and extent of regulatory services that solitary wasps provide. Social wasps are, without exception, generalist predators; their large colony sizes, and opportunistic predatory nature may mean that their ecological impact is significant in any habitat where they are present, natural or farmed. (3) Anecdotal evidence has previously pointed to the roles of wasps as pollinators. Our data analysis reveals that wasps are involved in almost 1000 plant-pollinator relationships across a range of plants and habitats. Through obligate plant-pollinator relationships (where plants are solely reliant on wasps for pollination), wasps are vital to the survival and pollination of at least 164 plant species, perhaps most important of which are interactions between orchids and thyniid wasps, without which the orchids would almost certainly go extinct. Wasps also play roles in facultative plant-pollinator relationships (where plants are reliant on a range of insects, including wasps for pollination), potentially pollinating at least 798 plant species. Importantly, wasps appear to form highly generalist pollination networks and can function effectively as 'back-up' pollinators, ensuring that pollination can continue even when plants lose their primary pollinators (Hallett et al., 2017). (4) Wasp and wasp-derived products directly support human welfare. The larvae and pupae of social wasp species provide a nutritious dietary component for humans in at least 19 countries (although this is likely to be an underestimate), and the rearing of wasp nests as a marketable food source represents an important source of income for farmers in countries such as China and Japan, with nests selling for up to US$100 per kg in wet markets (Nonaka, 2010). Besides food, wasp-derived products such as venom, saliva and nest materials appear to present medicinal potential in the prevention of bacterial infections and in the treatment of cancer and other diseases. (5) Wasps indirectly support the functioning of ecosystems through roles in seed dispersal and carrion decomposition. We identified at least 10 plant species in Asia, Europe and North America that have formed vespicochorous interactions with social wasps, and as a result are reliant upon them for efficient seed dispersal. Further, as scavengers, wasps also play roles in the decomposition and recycling of nutrients from vertebrate carrion. (6) The cultural services provided by wasps include their use as biological indicators, along with their (largely negative) inspiration of the arts and appearance in the media. Given their predatory nature, both social and solitary wasp species diversity within an ecosystem can provide accurate measures of overall arthropod diversity, while some social wasp species have been shown to be accurate indicators of heavy metal contamination. Culturally, wasps have historically suffered from a negative relationship with humans, ranging from their delivery of punishment by God in the Bible, through to yearly tabloid coverage branding them with terms such as 'murder hornets'. Such coverage is unfortunate given the ESs provided by wasps, and we hope that public awareness of such services may lead to wasps being seen in a different light. (7) The aculeate wasps are a highly speciose group of insects ( Fig. 1)  evolutionary diversity. Unfortunately, it is one group of only 53 species that tarnishes our impression of wasps: the social wasps (which includes the genera Vespa, Vespula and Dolichovespula) account for just over 3% of aculeate wasps (Aguiar et al., 2013). Their sometimes aggressive, annoying nature and the dominance of research into their success as invasive species (Lester & Beggs, 2019), has led to a focus on the economic and ecological costs associated with wasps, largely overlooking any benefits (e.g. Clapperton et al., 1989;MacIntyre & Hellstrom, 2015). As a result, there is a general paucity of research on aculeate wasp life history, evolution and ecology, preventing an accurate quantitative and economic assessment of these roles. This is of pressing concern given recent evidence that wasp populations are declining at similar rates to other key insect ES providers (Outhwaite et al., 2019). Quantitative and qualitative data on wasp life-history evolution and biotic interactions are required to assess their overall importance at ecological and economic scales. We have highlighted key areas of promise for understanding and harnessing better the ESs of wasps. (8) A definition of ESs that includes 'both the benefits people perceive, and those they do not' (Costanza, 2016) is apt for the aculeate wasps. We look forward to a future where wasps are not viewed as 'pointless', but where their critical roles in multiple facets of human health and wellbeing are recognised, quantified and conserved.
Wasps are most certainly a flagship group for illustrating the broader impacts of ESs, and in expanding our awareness of the range of values conferred by nature.

VIII. ACKNOWLEDGEMENTS
We thank Jacob Podesta and Daisy Dunn for contributions to earlier drafts of this review, and members of the Sumner Lab for comments, especially Adam Devenish, Emeline Favreau and Chris Wyatt; we also thank Tristram Wyatt for his comments. This work was supported by the "Natural Environment Research Council (grant NE/M012913/2 awarded to S.S.) and a Marie Sklodowska-Curie Action Individual fellowship (grant 706208-SocParPhenoEvol awarded to A.C.) Supplementary data set S1.