Evolutionary and Ecological Pressures Shaping Social Wasps Collective Defenses

Abstract

Social insects are well known for their aggressive (stinging) responses to a nest disturbance. Still, colonies are attacked due to the high-protein brood cached in their nests. Social wasps have evolved a variety of defense mechanisms to exclude predators, including nest construction and coordinated stinging response. Which predatory pressures have shaped the defensive strategies displayed by social wasps to protect their colonies? We reviewed the literature and explored social media to compare direct and indirect (claims and inferences) evidence of predators attacking individuals and colonies of wasps. Individual foraging wasps are predominantly preyed upon by birds and other arthropods, whereas predators on wasp brood vary across subfamilies of Vespidae. Polistinae wasps are predominantly preyed upon by ants and Passeriformes birds, whereas Vespinae are predominantly preyed upon by badgers, bears, and hawks. Ants and hornets are the primary predators of Stenogastrinae colonies. The probability of predation by these five main Orders of predators varies across continents. However, biogeographical variation in prey–predator trends was best predicted by climate (temperate vs. tropical). In social wasps’ evolutionary history, when colonies were small, predation pressure likely came from small mammals, lizards, or birds. As colonies evolved larger size and larger rewards for predators, the increased predation pressure likely selected for more effective defensive responses. Today, primary predators of large wasp colonies seem to be highly adapted to resist or avoid aggressive nest defense, such as large birds and mammals (which were not yet present when eusociality evolved in wasps), and ants.

Social insects are perhaps best known for the fierce collective defense of their nests and the ability to deliver painful stings (e.g., Pseudomyrmex species [Hymenoptera: Formicidae] in ant acacias; Schmidt 2016, Sumner et al. 2018). Wasps in particular have a special reputation for stinging; indeed, for millennia, human cultures throughout the world have taught their children to recognize the local wasp phenotype (e.g., the aposematic yellow and black stripes of Vespinae in temperate regions) and to fear and avoid social wasps (Lester 2018, Sumner et al. 2018, Jones 2019). Despite widespread public recognition of the defensive talents of social wasps, we know remarkably little about the ecological and evolutionary forces driving the evolution of their defensive behaviors. The defensive behaviors of any organism evolve in response to selective pressures exerted by their predators; yet, little is known about the predators of social wasps (Fig. 1a). Although anecdotal reports are present in the literature, direct evidence of wasp predation is surprisingly scarce. To understand the (co)evolution of defensive behaviors in social wasps, we require a holistic understanding of who these predators are and to what extent these predator–prey relationships can be explained by ecological and/or evolutionary traits.

Fig. 1.

Predation on social wasps and defensive strategies to prevent colony-level predation. (A) A direct observation of an Otago skink (Oligosoma otagense McCann [Squamata: Scincidae]) capturing a Vespula germanica Fabricius (Hymenoptera: Vespidae: Vespinae) queen in New Zealand (Oceania). Photo by Ricardo Rocha Mello 2019. (B) A Vespula vulgaris Linnaeus nest (left) after being excavated from underground (by JMJ, not included as a predation event), and (right) after peeling away the paper envelope to reveal the combs full of brood. Photos by JMJ 2018. Nesting underground (out of reach) and/or building a nest envelope that blends in with the bark/stems of the tree in which it is built (crypsis) are examples of primary defensive strategies to avoid colony-level predation. (C) A Polistes major major Palisot de Beauvois (Polistinae: Polistini) foundress, on her nest, displaying a warning posture. Photo by KJL 2014. Visual behavioral displays are examples of secondary defensive strategies that deter nearby predators from attacking the colony.

The Defense Apparatus of Social Wasps

The sting apparatus of the Aculeata (ants, bees, and wasps) is an anatomical and physiological derivative of the ovipositor in female Hymenoptera, able to penetrate the skin of vertebrate predators and inject pain-inducing venom (Shing and Erickson 1982). In the Hymenoptera, the apparatus is theorized to have initially evolved in solitary wasps in the role of paralyzing prey, a behavior observable in extant solitary wasps (Schmidt 2004). However, in addition to using the sting apparatus during foraging, nonsocial hunting wasps (e.g., Pompilidae and Mutilidae) may sting vertebrate predators in self-defense, injecting highly painful venomous cocktails (Schmidt 2016). This defense mechanism helps to explain why nonsocial wasps seem to have very few natural predators (Schmidt and Blum 1977, Schmidt 2004).

In social Hymenoptera, stinging and biting are often used to deter potential enemies. When a colony perceives a threat, nonreproductive adult females engage and sting and/or bite potential predators, often cooperatively (Starr 1985, Nouvian et al. 2016). The venom injected via stinging often can translate into intense pain. Vertebrate predators associating social insect individuals or colonies with negative experiences may learn to avoid them (Schmidt 2016). This associative learning can be further reinforced by the aposematic coloration of many social bees and wasps (Vidal-Cordero et al. 2012).

Several lines of evidence support the effectiveness of the stinging response as an anti-predator strategy. In fact, the defensive stinging behavior is widespread among Hymenoptera Families (e.g., Formicidae, Kugler 1979; Apidae, Nouvian et al. 2016; and Vespidae, Starr 1985). Also, male hymenopterans (lacking the sting apparatus) may mimic the stinging abdomen movements of females when caught by a predator (Giannotti 2004, Schmidt 2016). Similar behavior is displayed by many nonstinging insects that are Batesian mimics of the aposematic coloration of bees and wasps (e.g., arctiid moths [Lepidoptera: Erebidae], Simmons and Weller 2002; hoverflies [Diptera: Syrphidae], Rashed and Sherrat 2007, Penney et al. 2014; robber flies [Diptera: Asilidae], Brower et al. 1960).

Social Wasps as Prey

Eusocial wasp (referred to from here on simply as ‘social wasp’) colonies are by definition characterized by overlapping generations, a reproductive division of labor, and cooperative brood care (Wilson 1971). At the peak of its cycle, a typical social wasp colony includes a nest (the architecture of which varies across taxa), one or a few reproductives (‘queens’), nonreproductive workers, and developing brood (Spradbery 1973). The brood is defenseless, and immature wasps are essentially parcels of lipid and protein, a bounty for a skulking predator (Spradbery 1973, Ying et al. 2010). Nonreproductive workers protect the brood through primary or secondary colony defenses. Primary defenses, usually related to the choice of nesting site or nest architecture, operate before a predator initiates any prey-catching behavior (Edmunds 1974), and decrease the chance that an encounter will take place between the colony and a potential predator. Secondary defenses, such as active behavioral responses, come into play after the potential predator has encountered a nest (see below).

Generally, predators that attack social wasp nests fall into two major groups: arthropods and vertebrates. Due to their size differences and different means of finding and attacking wasps, the two pose different kinds of threats. Because social wasps have evolved virtually nonoverlapping means of defense against these predator categories, we will often distinguish arthropod versus vertebrate predation when discussing social wasp defense strategies below.

Primary Defense: Before Predators Arrive

Nesting Out of Reach

Ants are cursorial hunters that reach their prey on foot. There is probably no natural substrate, especially in the tropics, that is entirely free of scouting-and-recruiting ants, but some sites experience lower levels of predation pressure than others. Wasps that form relatively small colonies and build small nests—many Polistes, Mischocyttarus, Leipomeles (Hymenoptera: Vespidae), and others—on the undersides of leaf blades likely experience lower exposure to patrolling ants than do those that nest on trunks and branches (Jeanne 1979). The mass-foraging army ants (Eciton spp. [Hymenoptera: Formicidae]) are extreme examples of ant predation in the tropics, as they cover the ground and climb vegetation in a moving mass, plundering whatever cannot escape (Chadab 1979a). However, because Eciton raid from the ground-up, it is possible that wasps that nest high enough in the canopy enjoy the benefit of a dilution effect. Additionally, trees and shrubs growing in water are likely completely immune from attack. The small trees, shrubs, and dead snags standing in a meter or more of water along the shores of Gatun Lake in Panama are favored nesting sites of Polistes canadensis Linnaeus (Rau 1933). Similarly, seasonally flooded habitats in the Amazon basin are probably largely Eciton-free (RLJ and SO, personal observation). Dolichovespula maculata Linnaeus (Hymenoptera: Vespidae) and Polistes spp. nests are particularly common in tree branches overhanging rivers in the southeastern United States (KJL, personal observation). Such sites could plausibly shield nests from mammalian predators as well, and though this has not been demonstrated, these hypotheses are eminently testable in environments where social wasp colonies face well-known predation pressures, as we later discuss.

Some paper wasp species (Hymenoptera: Vespidae) frequently nest in dense vegetation and on plants sporting spines or thorns, which can be inaccessible to birds and mammals (Richards 1978). Nesting on vertical and/or relatively flat surfaces, such as tree trunks, walls, and eaves of buildings, is typical for some Neotropical Epiponini genera (Metapolybia, Synoeca, Clypearia, and Nectarinella) and sometimes observed in other genera (some species of Polybia, Chartergellus, and Parachartergus, as well as many independent-founding Polistini; Jeanne 1975, Richards 1978, Edwards 1980). These sites likely reduce access to nests by climbing and flying vertebrate predators.

Among the vespine wasps there are aerial and cavity/ground-nesting species. Some species of Vespa and Vespula (Hymenoptera: Vespidae) use abandoned rodent holes or cavities in man-made structures or trees to initiate their nests in early spring (Akre et al. 1980, Edwards 1980, Matsuura and Yamane 1990). As the nests grow, workers gradually expand the cavities by excavating the surrounding dirt, insulation, or decayed wood (Fig. 1b). The only visual evidence of these nests is the traffic of foragers entering and exiting the nest, providing some camouflage against enemies, such as humans, that rely heavily on visual cues. Other vespine species build aerial nests. These can be as high as 40 m from the ground and thus inaccessible to most nonflying predators (Feás and Charles 2019).

Nesting in Association With Other Species

Dolichoderus and Azteca ants, especially Az. chartifex Forel (Hymenoptera: Formicidae) and allies, form huge colonies that can densely occupy entire trees and defend them and their surroundings against other ants, including Eciton (Delabie 1990, Somavilla et al. 2013). Some Neotropical paper wasps build nests regularly, and in some cases obligately, in close association with these ants. The most intimate associations involve some Epiponini wasps. Some species of Agelaia (Hymenoptera: Vespidae) build their nests in cavities inside the large, arboreal carton nest of Az. chartifex (Richards 1978). A somewhat less intimate but more common co-nester is Polybia rejecta Fabricius, which constructs its nest within a few centimeters of the ants’ nest, or even in contact with it (Servigne et al. 2020). Wasps respond behaviorally to ant foragers near their nests. A combination of physical ejection of ant workers, wing-buzzing behavior, and the elimination of their scent trails (see secondary defenses, below) maintains an ant-free zone around the wasps’ nest entrance (Servigne et al. 2020, Barbosa et al. 2021) and helps to offset opportunistic attacks by associated ant colonies. In drier habitats, where Az. chartifex is absent but Eciton occurs, wasps may nest in association with other ants. For example, in Quintana Roo, Mexico, and Guanacaste, Costa Rica, Epiponini wasps nest in myrmecophytic acacias occupied by Pseudomyrmex spp. (Espelie and Hermann 1988, Joyce 1993, Dejean et al. 1998). Numerous species of myrmecophyte have evolved intimate relationships with other species of ants, many of which can keep Eciton from their host plant. Several species of wasps in the genera Angiopolybia, Pseudopolybia, and Mischocyttarus nest on these plants, thereby gaining protection from army-ant predation (Herre et al. 1986).

Nesting in trees defended by Azteca ants, described above, may afford social wasps some protection against vertebrates as well as army ants. Some species of host ants sting vertebrates that contact their host trees (e.g., Pseudomyrmex [Hymenoptera: Formicidae] species in ant acacias; Young et al. 1990). The relationship is mutualistic: in a study in French Guiana, none of 42 Az. chartifex nests with one or more Polybia rejecta colonies in close association were attacked by birds, whereas 9 of the 88 wasp-free nests of the ant were attacked by woodcreepers (Dendrocolaptinae) and by a woodpecker (Celeus flavus Statius Muller [Piciformes, Picidae]; Le Guen et al. 2015). Polybia rejecta is especially aggressive and at least partially effective at defending against these and other vertebrate predators. By building its nest as close to the Azteca nest, and often in contact with it, the wasps provide protection for the ants against these birds and possibly anteaters. Several species of independent-founding polistines also nest not just on trees with Azteca nests, but on myrmecophyte plants (those in close association with ant colonies) such as Cordia spp., Tococa spp., and others. These are probably commensal relationships, with the wasps gaining protection from Eciton, but contributing nothing to benefit the host ants (Jeanne, 2020). Although nesting with ants is much rarer in Africa and Asia than in the Neotropics, the wasp Polybioides tabidus Fabricius [Hymenoptera: Vespidae] has been reported to nest with the aggressively stinging ant Tetramorium aculeatum Mayr [Hymenoptera: Formicidae] in Cameroon (Dejean and Fotso 1995).

Some arboreal wasp nests are joined by passerine birds that nest nearby; birds use existing wasp nests as cues for nest-site selection and will build near artificial Polybia rejecta nests (Joyce 1993). Avian nesting partners could offer protection to the wasps if the birds mob approaching vertebrate predators such as raptors and monkeys (Sandoval and Wilson 2012, Barbosa et al. 2021).

Tolerance to human presence in some species of social wasp may play a role on determining which predators target their colonies. For instance, some species of Polistini and Mischocyttarini wasps are synanthropic in the Neotropics, often nesting on man-made structures; on the other hand, most Epiponini wasps tend to rely more on natural structures such as vegetation, and their diversity tends to be relatively reduced in highly urbanized areas (Detoni et al. 2018). Thus, it is possible that the predators of synanthropic social wasps are somewhat limited to other animals that also tolerate human presence. On the other hand, nesting on buildings may result in nests that are more conspicuous to predators (e.g., young Dolichovespula colonies are much easier for humans to notice on buildings than in trees; KJL, personal observation). If this increases detection and later predation by predators such as birds, then the use of buildings as a nest site could be seen as an ecological trap.

Crypsis

Adaptive nest crypsis is a response to selective pressure imposed by visually hunting predators, almost always vertebrates. Paper wasp nests exhibit several putative adaptations that likely decrease their visual detection by predators by decreasing contrast from, or increasing their resemblance to, the background. In some cases, wasps may nest on substrates where their nests visually match the background. This effect can be enhanced by mosses and liverworts growing on the nest paper (Barbosa et al. 2016, Milani et al. 2020). Nest envelopes of the epiponine wasp Leipomeles dorsata Fabricius (Hymenoptera: Vespidae) constructed beneath leaf blades can closely resemble the leaf substrate, even mimicking the venation of the leaf (Richards 1978). Overall shapes of nests can evolve to lessen recognition by vertebrates, such as the stick-like nests of some Mischocyttarini and Polistini species (Vesey-Fitzgerald 1950, Richards 1978, Starr and Hook 2006, Silveira et al. 2015).

Nest crypsis may also be a strategy adopted by Vespinae. Vespula spp. may benefit from constructing their nests underground, but Dolichovespula produce large, aerial nests that often hang from trees (Akre et al. 1980). However, the nests can still be difficult to locate. The paper envelope can blend in with both the bark/stems of the tree it is built inside and may also be able to provide camouflage amidst vegetation or other background if hanging from trees (Feás and Charles 2019; KJL, personal observation).

Aposematism

An alternative strategy for both nests and groups of adult wasps is to be visually conspicuous. Well-defended nests are often placed relatively high in tree canopies, often along forest or river edges. Pale nest paper contrasts with surrounding vegetation. Polybia striata Fabricius, Polybia scutellaris White, Chartergus spp., and Epipona niger Brèthes (Hymenoptera: Vespidae) exemplify this pattern (Jeanne 1975, Richards 1978). For groups of individuals, the sunflower-like radial array of pale-colored adults of Apoica pallens Fabricius (Hymenoptera: Vespidae) resting on the underside of their open nest elicits a startle response, at least in humans, when first spotted from below (RJL, personal observation).

Chemical Repellents

Independent-founding Polistinae (Polistini, Mischocyttarini, and Ropalidiini) and foundress queens of young colonies of Vespinae apply an ant-repelling glandular product to the nest petiole. The ant-repellent secretions are produced by specialized exocrine glands on the terminal sternite of the gaster (van der Vecht’s gland; Jeanne 1970a, Post and Jeanne 1981, Keeping 1990, Kojima 1992, Martin 2017). This gland is often secondarily lost during the evolution of the Epiponini lineages that employ nonrepellent means of ant defense (see secondary defenses, below; London and Jeanne 2000, Smith et al. 2001). Similarly to the independent-founding Polistini, in the Vespini, at least one species (Vespa velutina Lepeletier) has been shown to utilize ant repellents on its nest during the foundation stage of the colony cycle (Martin 2017).

Sticky Traps

Members of the genera Nectarinella and Leipomeles (Epiponini) erect sticky-tipped stalks around the access to their nests (Jeanne 1975, Schremmer 1977, Mateus and Noll 1997). This appears to be an effective defense against scouting-and-recruiting ants for wasp species with small colonies and with body sizes in the range of these predators, for whom active defense is less reliable.

Hardened Nests

The exceptionally tough carton envelope of some arboreal-nesting swarm-founding species may be impenetrable to arthropod raiders, including chewing ants. In some species (e.g., Chartergus artifex Christ [Hymenoptera: Vespidae]) the inter-comb passageways within the nest are narrowed to the size of a brood cell (RLJ, personal observation), possibly an adaptation allowing a single wasp to effectively block access to brood-laden combs by ants that have entered the lower chambers.

The nature of the nest material can also have important effects on the capacity of the nest itself to withstand mechanical attacks by vertebrate enemies. The extremely dense and tough carton of C. artifex, for example, may be able to resist attacks by monkeys and birds (RLJ, personal observation). The transition from wood pulp (paper) to mud as a nesting material has occurred twice in the genus Polybia; in both mud-nesting lineages, nests are highly robust to mechanical damage, and may at least narrow the range of species that can penetrate them. Polybia emaciata Lucas workers often retreat into the nest upon mechanical disturbance, apparently relying on the nest as a fortress-like defense (O’Donnell and Jeanne 2002). Similarly, the hard mud nests of some stenogastrines (e.g., Liostenogaster flavolineata Cameron [Hymenoptera: Vespidae]) may afford them greater survival than the paper nests of their close relatives (e.g., Parischnogaster spp; SS, personal observation).

Secondary Defenses: Behavioral Responses to Predator Arrival

Physical Predator Removal

Due to the relevance of ants as threats to Neotropical colonies, wasps have developed anti-predator behaviors dedicated specifically to repel ant invaders from their nests. If a foraging ant makes its way onto a nest, a defending wasp may dart at it, grab it in the mandibles, and toss it from the nest or fly off with it and drop it (Chadab 1979b). If ant foraging persists, swarm-founding wasps (Epiponini) recruit nestmates to encircle the access point and repel the intruders with semisynchronous bursts of wing-buzzing. The mini-blasts of air are often effective in causing the ants to turn around and exit the nest (Chadab 1979b, Jeanne 1991). Following an ant invasion to their nest, wasps extensively mandibulate or lick the traversed surface, apparently expunging the ants’ trail pheromone or scent (West-Eberhard 1989). Whether the behavior removes the chemical or covers it up, possibly with the labial gland secretion used in nest construction, has not been determined. Although less well documented than the tropical examples, ground-nesting Vespula wasps defend their colonies from Argentine ants (Linepithema humile Mayr [Hymenoptera: Formicidae]) and Red Imported Fire Ants (Solenopsis invicta Buren [Hymenoptera: Formicidae]) by darting at and biting ant foragers exploring the nest entrance (KJL, personal observation). How wasps repel the abundant subterranean foragers of these species, particularly early in colony development, is unknown.

Vibro-acoustic Warnings

When facing bigger predators, namely vertebrates, adult epiponine wasps in at least three genera (Synoeca, Chartergus, Polybia rejecta, Polybia sericea Olivier) respond to vibrations of the nest or its substrate by rhythmically and synchronously striking or drumming against the nest carton, thereby generating characteristic sounds that are audible to humans over distances of five meters or more (Evans and West-Eberhard 1970, Taylor and Jandt 2020, SO, personal observation). Grazing livestock seem to learn to avoid patches of grass and bushes around Po. sericea colonies, possibly by associating their alarm sound to eventual stinging (F. Prezoto, personal communication). Protonectarina sylveirae de Saussure (Hymenoptera: Vespidae) makes a characteristic high-pitched sound when attacking (Richards 1978). When their nest is disturbed, Vespa mandarinia Smith workers closely approach the intruder in flight while loudly snapping their mandibles (Schmidt 2016; RLJ, personal observation).

Visual Displays

Several species of Polistinae in the genera Agelaia, Apoica, Brachygastra, Epipona, Polistes, Polybia, Synoeca, as well as Ropalidia revolutionalis de Saussure (Hymenoptera: Vespidae) (Hook and Evans 1982), engage in visual warning displays, such as gaster-flagging, when the nest is disturbed (O’Donnell et al. 1997; Fig. 1c). In response to a disturbance, workers on the nest (or on the vertebrate intruder that caused the disturbance) raise and wave the gaster, extrude the sting, and fan the wings. In some species, the gaster is conspicuously colored, suggesting its use as a visual signal, either as a threat to predators or to communicate with nestmates. In disturbed Polistes spp. nests (in Malaysia), both workers and males display raised forelegs, providing a visual signal to the potential predator (Turillazzi 2003).

Chemical Responses

The existence of alarm pheromones has been demonstrated in many large-colony species across different subfamilies (Maschwitz 1964, Jeanne 1981, Veith et al. 1984, Kojima 1994, Sledge et al. 1999, Cheng et al. 2017), as well as some small-colony species (Post et al. 1984, Bruschini et al. 2006). These compounds are released either at the nest or when workers sting their aggressor, and serve to recruit in-nest workers to engage in defensive behavior, as well as to attract defending workers to a specific target. Notably, experiments have failed to demonstrate venom-associated alarm pheromones in some small-colony wasps (Keeping 1995, London and Jeanne 1996), further suggesting that the selective pressures behind colony defense have been varied, and have produced a number of different strategies.

In what may be a specialized defense against small vertebrate predators, Parachartergus colobopterus Licht and Parachartergus fraternus Gribodo use chemicals in defense of the nest by spraying its venom in a fine mist that travels several centimeters (Jeanne and Keeping 1995, Mateus 2011). The nests are built on tree trunks and are visually cryptic, which may narrow the range of potential predators to small gleaning birds, for which an eyeful of the sticky venom may be a more effective deterrent than stinging (Jeanne and Keeping 1995).

Hiding

Contrary to popular perception of wasps being aggressive every time they are disturbed (Sumner et al. 2018), some species exhibit remarkable timidity, either fleeing or hiding in response to disturbance, despite possessing a functional stinging apparatus (Hermann and Chao 1984, Strassmann et al. 1990, O’Donnell and Jeanne 2002). Some species of Mischocyttarus, for example, will often hide behind their nests or, if sufficiently disturbed, fly away—avoiding a direct confrontation of any kind (Hermann and Chao 1984; RLJ and MD, personal observation). Even more drastic is the response of some Stenogastrinae wasps, which may simply ‘drop’ from the nest when threatened (SS, personal observation). While this may confuse a potential predator, or redirect their attention away from the nest, it also leaves the brood completely undefended.

Absconding

When all defensive tactics fail to repel a predator, it can be in the colony’s best interest for the adults to abandon the nest and their brood, thus saving themselves for the opportunity to re-nest elsewhere. In an intriguing overlap with responses to vertebrate attacks, Neotropical swarm-founder adults readily abandon their nests when threatened by army ants (mostly Eciton that raid above ground; in some cases, rapid absconding can be triggered by encounters with just a few Eciton workers, or by their odor alone; Chadab 1979b). Observations suggest that rapid absconding may be coordinated via the wasps’ alarm pheromone (Chadab 1979b).

Repeated strikes on nests by avian predators such as Red-throated Caracaras Ibycter americanus Boddaert (Falconiformes: Falconidae) initially induce coordinated stinging attacks, but eventually the defenders shift to rapid departure of all adults (McCann et al. 2013). Following absconding events, surviving adults in swarm-founder colonies first cluster, then move to re-nest in a new location by following a pheromone trail (Sonnentag and Jeanne 2009).

In contrast to most polistine wasps, temperate zone Vespula and Dolichovespula colonies almost never abscond in the face of persistent or catastrophic predator attacks. This is presumably because the short seasonality of their colonies limits them from obtaining fitness gains by rebuilding a new nest following the loss of the original. Similarly, in Neotropical genera such as Agelaia, the scarcity and quality nesting sites may select against absconding. While adults may leave the nest during ant raids, they return after the predators are gone, and try to rebuild the colony in the same site (O’Donnell and Jeanne 1990). Exceptionally, Vespa velutina adults may abandon underground nests and brood when under attack by ant predators and try to reestablish their colonies elsewhere (XF, personal observation).

Stinging

The most notorious response to disturbance in a social wasp colony’s arsenal is the painful and dangerous sting of large-bodied species such as Vespa and Synoeca (Xuan et al. 2010, de Castro e Silva et al. 2016), and the vigorous stinging attacks by tens to hundreds of workers in large-colony species like Vespula and Agelaia, in which potential predators can receive hundreds of stings (Vetter et al. 1999). Workers are initially alarmed by movement, vibration of the nest, or by volatiles in mammalian breath (Landolt et al. 1998, Jandt et al. 2020).

Despite being the most well-acknowledged active collective response to disturbance in wasps, the sting is the weapon of last resort against vertebrates. Aggressiveness in response to threats varies widely among species. The effectiveness of the sting as a defense against vertebrates is partly a function of the size of the colony and the size of the wasp. The sting of many wasp species is ineffective against a wide variety of vertebrate predators. Large numbers of epiponine colonies are taken with apparent impunity by monkeys and birds, and vespine nests are commonly attacked by large mammals and birds (see ‘Results’ section). Yet if the defenders are aggressive enough and numerous enough, they can be effective. A study in French Guiana found that nests of Polybia rejecta, remarkable for its aggressiveness when defending its nests, suffered no vertebrate attacks, whereas other less aggressive epiponines did (Le Guen et al. 2015).

If a colony is sufficiently provoked, collective or group stinging defense is coordinated, often by alarm pheromones comprised of volatiles released with venom. Similar to alarm pheromones, vibratory signals may be used to recruit wasps inside the nest to the outside in preparation for further action (Strassmann et al. 1990, Jeanne and Keeping 1995, Taylor and Jandt 2020). The intensity and duration of collective defense by some species suggests a strong selection pressure exerted by nest predators, though we have yet to understand the links between predation frequency and type and the interspecific variation in defensive behaviors exhibited by social wasps. Colonies within a species can also vary greatly in the magnitude of the collective defensive response, which can be, but is not always, associated with colony size and developmental stage (London and Jeanne 2003, Brito et al. 2018, Jandt et al. 2020).

Who Are the Primary Predators of Social Wasps?

Today, a complete spectrum of solitary to eusocial species is represented within the monophyletic lineage of Vespidae (Jandt and Toth 2015; Fig. 2), wherein eusociality evolved independently at least twice (Hines et al. 2007, Huang et al. 2019), making it one of the best systems for studying the co-evolution of predator–prey relationships and of coordinated colony defense. To understand why social wasps have evolved these primary and secondary methods of collective defense, we first need to ask: Who are the predators of social wasps? Various predators have been sparsely identified in the literature, including other social wasps (Turillazzi 1984, Gadagkar 1991, Jeanne and Hunt 1992), ants (Bruch 1923, Jeanne 1972, Tindo et al. 2002), birds (Birkhead 1974, Huang et al. 2004, McCann et al. 2010, van Bergen 2019), and mammals (Bigelow 1922, Perry and Manson 2008, Ying et al. 2010). Do social wasps that use different primary and secondary defensive strategies receive the same predation pressure from each of these taxonomic groups; and/or is that predation pressure consistent across different regions where these wasps are biogeographically located? Moreover, by including novel data sources in our search, such as social media (Nyffeler and Vetter 2018), we may find evidence of predator–prey relationships overlooked in the literature. Here, we review the published literature and compile records from social media on the predation on vespid wasp individuals and colonies to 1) identify the predators of social wasps; 2) explain any consistent patterns of variation in who the predators are with regards to biogeography, evolutionary and ecological traits; and 3) relate the findings above to the evolutionary history of social wasps and their predators.

Fig. 2.

Colony characteristics (colony cycle, size, nesting site, architecture, and climate) of the three social subfamilies of Vespid wasps (broken into the four tribes for Polistinae). Data are not provided for the three solitary subfamilies (Euparagiinae, Masarinae, or Eumeninae), as they were not included in the data presented in this study. The primary colony predator taxon is based on data shown in Table 2.

Methods

Literature Search

We searched ISI Web of Science, Scopus, Google Books, and Google Scholar databases for publications on the predation of specific wasp taxa by systematically combining the taxonomic group (e.g., ‘Wasp’, ‘Vespidae’, or ‘Polistinae’) with the terms ‘predator of’, ‘predation on’, and ‘attack on’ (e.g., [‘Vespula’] AND [‘predator of’ OR ‘predation on’ OR ‘attack on’]). The terms ‘predator’ and ‘predation’ alone were avoided due to social wasps being predators themselves, with most of the search results turning up literature focused on their ecological role as predators. Books, chapters, and reviews on social wasp biology were scanned for additional references of predation of social wasps. For each reference (from literature to other media searched), we noted: 1) the predator taxa, 2) wasp prey, 3) whether the predation event targeted an adult wasp or the colony, 4) the geographic location of the observation, 5) the medium in which the reference was found, and 6) which type of evidence did the reference constitute (defined in ‘Data Analysis’, section). All reference searches (including other media below) were carried out between April and June of 2020.

Social Media Search

We used the social media platforms YouTube (youtube.com), Facebook (facebook.com), and Twitter (twitter.com) to collect additional evidence (accounts and videos) of predation events on wasp individuals and colonies. Nonnaturally occurring predation events (e.g., artificial setups, such as arenas) were excluded from our data.

YouTube

Using the YouTube video database, we initially combined the search terms ‘predator of’, ‘predation on’, or ‘attack on’ with the common names for social wasps (e.g., ‘hornets’, ‘hover wasps’, ‘paper wasps’, ‘social wasps’, ‘wasps’, and ‘yellowjackets’). We then conducted a focused search using the common names for specific predator taxa (based on data from the literature search) + predation terms + common name of social wasp (e.g., ‘badger attack on social wasps’). This facilitated the database’s search mechanism, while normalizing the amount of data obtained for predators which are reported less often or lack empirical evidence in the literature (Supp Table 1 [online only]). For each search, we scanned the first hundred results, though the most relevant were usually shown within the first 10–20 results. That means, even though we found abundant videos of other arthropods being preyed upon, we sometimes found videos of wasps as prey being mislabeled as bees or other insects, despite never including ‘bee’ or the other descriptor as a search term.

Facebook and Twitter

We posted our request for evidence of wasps being preyed upon to two Facebook ‘groups’: ‘Enthusiasts of Social Wasps’ ([URL 1]) and ‘Ecology of Vespinae’ ([URL 2]) on 17 April 2020 (Supp Fig. 2a [online only]), and posted a general call on Twitter (tagging those authors with a Twitter account) on 16 May 2020 (Supp Fig. 2b [online only]). For both posts, we included an image of a Vespula queen on a pink Grevillea flower. This image allowed us to avoid biasing responses (e.g., an image of a bird eating a wasp might have drawn the attention from those with other examples of bird predation), while also increasing the exposure of the post, given that the image was clear and brightly colored.

Direct e-mail to Wasp Researchers

We contacted wasp researchers directly or via wasp-focused listservs (Supp Fig. 3 [online only]). All responses that included unpublished observations of individual or colony-level predation were categorized as ‘unpublished accounts’.

Organization of Evidence

Taxonomy of Predators

Predation records were classified as direct or indirect evidence, as well as media type (literature, social media, video, or unpublished accounts). Direct evidence is represented by both ‘Empirical’ studies (experiments or systematic observations where quantitative data were collected) and ‘Observations’ (a description, image, or video of a direct observation of a predator attacking a wasp or colony) records. ‘Indirect Evidence’ refers to predation claims, common lore, or instances where predation was inferred through evidence left behind following a predation event. Predators were categorized by Phylum, Class and Order, and compared across the four tribes of Polistinae, one tribe of Vespinae, and Stenogastrinae.

Social wasp subfamilies (and Polistinae tribes) were further categorized by colony size, nest architecture, colony cycle, and coordination of colony response. We discuss patterns in predation based on these colony characteristics.

Biogeography of Predators

When location of direct evidence of colony predation was provided, records were grouped by continent. We included Central America as a separate ‘continent’ since the social wasp fauna in the area differs greatly from the rest of North America (Hunt 2007); and used ‘Oceania’ instead of ‘Australia’ to allow the grouping of records made in western Pacific areas (combining Australia, New Zealand, and Guam). Each record was counted once for each area cited by an observer; for example, we counted D. Santoro’s personal observation on the predation of humans on Vespula in Japan and China as two distinct observations in Asia.

Evolution of Predators

Using the results from the literature and social media searches, we mapped the evolutionary origins of vertebrate predators with the most records of predation attempts of social wasp colonies: three Classes (Aves, Mammalia ,and Reptilia), and the top two Orders in the Classes Aves (Accipitriformes and Passeriformes) and Mammalia (Carnivora and Primates). Among invertebrate predators, we mapped the evolutionary origins of hymenopteran Family Formicidae (ants) and Genus Vespa (hornets, Family: Vespidae). Along this evolutionary map, we also plotted the evolutionary origins of Aculeata (stinging wasp, bee, and ant ancestor), Vespidae, the common ancestor of Polistinae and Vespinae (the hypothesized evolution of eusociality in those groups), and the origins of the three social wasp subfamilies in Vespidae (Vespinae, Polistinae, and Stenogastrinae).

Results and Discussion

Search Results

Our sampling efforts yielded a total of 720 records of direct evidence (‘Observation’ and ‘Empirical Studies’ data; n = 376, 52.2%) and ‘Indirect Evidence’ (n = 344, 47.8%) of predation events on social wasps (summarized in Table 1; Supp Table 4 [online only]). Most events were obtained from the literature search (n = 509, 70.7%), followed by YouTube video data (n = 119, 16.5%), e-mail correspondence (n = 72, 10.0%), and Facebook + Twitter (n = 20, 2.8%). Although most of the data were published in scientific literature, our diverse methods of data collection increased our sample size and allowed the inclusion of previously unpublished predator observations (e.g., predation of Ropalidiini wasps by macaques; A. Brahma, personal communication), and surprising new observations (e.g., attack on a stenogastrine wasp by a predatory nematode; S. Turillazzi, personal communication).

Table 1.

Summary of records collected on individual- and colony-level wasp predation (combined)

			Social media
	Evidence type	Literature	Facebook + Twitter	YouTube	E-mail correspondence
Direct evidence	Empirical study  Experiments or systematic observations where quantitative data were collected	61	0	0	0
	Observation  A description of a one-off or multiple predation events that were observed first-hand.	229	18	114	67
Indirect evidence	Common lore regarding predation, or a suggestion of a predator event based on postpredation damage	219	2	4	6

Values represent total number of references for each data collection type.

Although empirical studies were the least common reference type (n = 61, 8.5%), all were published in the scientific literature (Table 1). ‘Observations’ (n = 428, 59.4%) and ‘Indirect evidence’ (n = 231, 32.1%) dominated the reference types reported. This shows a clear opportunity for future empirical research on predation on social wasps.

Predation on wasp colonies was recorded most often in Polistinae (52.7%), whereas predation on individual wasps was recorded most often in Vespinae (42.0%). This is likely due to the fact that Polistinae nests are often exposed (Fig. 1c) when compared with Vespinae, which often build nests in cavities or underground (Spradbery 1973, see ‘Patterns of Predator Variation With Regards to Social Wasp Traits’ section).

Taxonomy of Predators

Predator type varied across the subfamilies within Vespidae, and by whether they preyed upon individuals or colonies (Table 2). Direct evidence on individual polistine and vespine wasps shows they were primarily preyed upon by arthropods (Classes: Araneae and Insecta) and birds (Class: Aves). Amphibians, carnivores, and reptiles were also observed eating individual wasps. Only one observation of predation on a Stenogastrinae individual was observed, and it was particularly unusual. Stefano Turillazzi reported an observation of a predatory nematode consuming (not parasitizing) a stenogastrine wasp.

Table 2.

Summary of direct evidence of (A) individual-level and (B) colony-level predation on the three social subfamilies of Vespidae (Polistinae, Stenogastrinae, and Vespinae)

	Polistinae					Stenogastrinae	Vespinae	Unknown
A. Individual predation	Epiponini	Mischocyttarini	Polistini	Ropalidiini	Unknown		Vespini		Total
Arthropoda	5	4	30	0	2	0	59	2	102
Arachnida	1	4	14	0	1	0	16	1	37
Insecta—Hymenoptera	0	0	5	0	0	0	18	0	26
Insecta—Other	4	0	8	0	1	0	25	1	39
Chordata	21	1	9	0	1	0	37	45	114
Amphibia	0	0	1	0	0	0	3	0	4
Aves—Accipitriformes	1	0	1	0	0	0	3	1	6
Aves—Passeriformes	10	0	1	0	0	0	10	27	48
Aves—Other	10	1	4	0	1	0	6	16	38
Mammalia—Carnivora	0	0	1	0	0	0	11	1	13
Mammalia—Primates	0	0	0	0	0	0	0	0	0
Mammalia—Other	0	0	0	0	0	0	0	0	0
Reptilia	0	0	1	0	0	0	4	0	5
Nematoda	0	0	0	0	0	1	0	0	1
Total individual predation	26	5	39	0	3	1	96	47	217
B. Colony predation
Arthropoda	23	7	28	14	5	8	10	6	101
Arachnida	0	0	1	0	0	0	0	0	1
Insecta—Hymenoptera	23	7	22	13	5	8	10	6	94
Insecta—Other	0	0	5	1	0	0	0	0	6
Chordata	30	9	31	4	3	1	71	22	171
Amphibia	0	0	0	0	0	0	0	0	0
Aves—Accipitriformes	4	0	5	1	0	0	19	8	37
Aves—Passeriformes	10	2	13	1	1	0	4	6	37
Aves—Other	11	6	3	1	0	0	3	4	28
Mammalia—Carnivora	0	0	4	0	0	0	29	3	36
Mammalia—Primates	3	1	4	1	1	0	11	1	22
Mammalia—Other	2	0	0	0	0	0	5	0	7
Reptilia	0	0	2	0	1	1	0	0	4
Nematoda	0	0	0	0	0	0	0	0	0
Total colony predation	53	16	59	18	8	9	81	28	272

Direct evidence includes instances when predation was directly observed and reported in the literature, social media, or other sources. Columns listed as ‘Unknown’ represent wasp prey that could not be confirmed from video or was described vaguely in text as ‘social wasp’ or ‘Vespid wasp’. ‘Paper wasps’ were categorized as ‘Polistini’, as that is the accepted common name for Polistes wasps, and ‘Yellowjackets’ and ‘Hornets’ were categorized as ‘Vespini’, as those are common names for Vespa, Vespula, or Dolichovespula. Data are presented for predator Phylum (gray rows), which are broken into Class and most commonly observed Orders (white rows).

At the colony level, direct evidence of predation on Polistinae consists mostly of records of hymenopteran predators (Fig. 3); ants were most often observed preying upon polistine wasps (Jeanne 1972, Strassmann 1981, O’Donnell and Jeanne 1990, Barbosa et al. 2021), followed by Vespa hornets (Matsuura and Sakagami 1973, Matsuura 1991). Vespa will attack and collect adults as well as brood as prey items to bring back to the colony; Vespa mandarinia can recruit nestmates to join them in the attack on the colony (Ono et al. 1995). Passeriformes (e.g., crows) had the second highest record of colony predation (Fig. 3; Raw 1997). In spite of collecting the most direct evidence of Hymenoptera as predators, indirect evidence of Hymenoptera as predators of polistine wasps in the literature were overpresented, whereas Passeriformes predators were never identified (indirectly) as predators of polistine colonies (Fig. 3).

Fig. 3.

Direct versus indirect evidence of predation on social wasp colonies by the top five predators (see Table 2), and all other predators combined (‘Other’). Percentage of direct evidence of predation is shown with filled bars (gray = predation on polistine colonies, dark gray = predation on vespine colonies, light gray = predation on stenogastrine colonies); and percentage of indirect evidence is shown with dashed bars. There was no indirect evidence of Passeriformes predation on vespine colonies. Percentages are calculated based on the total evidence (direct or indirect percentages calculated separately) for each subfamily.

Carnivores were the most common predators of colonies of Vespinae, followed by Accipitriformes (eagles, hawks, and kites; Fig. 3). The highly specialized Pernis honey buzzard can individually raid vespine colonies for brood combs, seemingly unaffected by the wasps’ aggressive response (Gamauf 1999, Huang et al. 2004). Among the Carnivore predators, we found the most direct evidence of badgers and weasels (Family: Mustelidae; Blackith 1958, Lanszki and Heltai 2007) preying upon vespine wasps, followed by bears (Family: Ursidae; Mealey 1980). These carnivores possess strong paws with long claws, effective in digging up and tearing apart the wasp nest envelope (in the case of Vespula), undoubtedly playing a role in the success of carnivores as colony predators. Indirect evidence of Carnivora predators in the literature were slightly overrepresented; whereas indirect evidence of Accipitriformes predation were slightly underrepresented relative to the direct evidence we found (Fig. 3).

There were only nine records of direct evidence of predation on stenogastrine colonies (Table 2). Among these, Vespa hornets were identified as the most common predator (n = 6), followed by ants (n = 2) and one record of a reptile consuming a stenogastrine colony. Additionally, Hymenoptera were also the only taxonomic group reported in indirect evidence as predators on stenogastrine colonies.

We compared traits characteristic of the three subfamilies of Vespidae, and four tribes of Polistinae (colony cycle, colony size, nest building patterns, and climate) with the primary colony predator taxon (Fig. 2). Social wasps varied in terms of colony cycle (seasonal and perennial), colony size (2–10⁷ adults produced in a colony), and the types of nest that were constructed. Most subfamilies and tribes were primarily preyed upon by Hymenoptera, and this seemed most closely linked to the fact that those groups were predominantly located in tropical climates. Vespinae tend to be found in temperate climates, and they were preyed upon mostly by Carnivora. Polistini are found in both temperate and tropical climates, yet epiponines (exclusively tropical) had more direct evidence of predation by mammals among the Polistinae tribes.

How Does Predation Pressure Across These Taxonomic Groups Vary Geographically?

We found direct evidence of predation on social wasp colonies in every inhabited continent (all except Antarctica; Fig. 4). Predation on polistine colonies was recorded in every continent, whereas predation on vespine colonies was only recorded in North America, Europe, and Asia. Stenogastrine wasps are only found in Asia, and that is where its predation events were recorded (Turillazzi 1991).

Fig. 4.

Relative frequency of direct evidence of predation events on the three social subfamilies of Vespid wasps across the six continents where Vespidae are found. A separate pie chart was made for each subfamily in each of the continents where colony predation was recorded and the predator was defined. Pie charts are linked to the continent where its data were recorded by black lines. The most common Orders of predator are shown with solid colors, remaining Orders are combined and represented by ‘Class: Other’ as striped colors. A gray gradient was used to differentiate continents with shared borders. Continent map was modified from original template generated in R (version 4.0.3, R Core Team 2020) using the package ‘maps’ (version 3.3.0, Becker and Wilks 1993).

Evidence of colony predation in Africa and Oceania was limited to 6 and 5 records respectively, compared with >30 records for each of the remaining continents. This small amount of evidence suggests our findings probably do not represent the full picture of prey–predator relationships in those regions. For instance, despite the presence of Vespinae wasp populations in Africa and Oceania (Lester 2018), along with predator clades that are widely reported to prey on wasps elsewhere in our data (e.g., Pernis apivorus Linnaeus [Accipitriformes: Accipitridae] honey buzzards in Africa, Bijlsma 2002); weasels in Oceania, King 2017), we could find no evidence of these predation events occurring in such areas. The same bias can be observed by separating records by countries, which also varied in terms of predation records. For example, Brazil, Costa Rica, French Guiana, and the United States each had at least 15 records, whereas the remaining countries had few or none.

Africa

In the African continent, ants were the primary predators of polistine colonies. Of note, we also uncovered a record of predation of Belonogaster petiolata DeGeer (Hymenoptera: Vespidae: Polistinae, Ropalidini) by Hoplostomus fulgineus Oliv (Coleoptera: Scarabicidae). Although there are Vespinae in Africa (Tribe and Richardson 1994), we found no direct evidence of colony predation events.

Central America

In Central America, ants and passeriform birds were the main predators of polistine colonies. Eciton army ants comprised two-thirds (67%) of the hymenopteran examples. Army-ant colonies can raid entire wasp colonies without facing significant opposition. Birds are also reported predators of Central American Polistinae, though with fewer direct observations. Records of predation by non-Accipitriformes, non-Passeriformes birds (Ibycter americanus), and Passeriformes (especially in the Family Corvidae) on Epiponini colonies were common. Central America also had the most predation on wasp colonies by nonhuman Primates, all in the Family Cebidae (capuchin monkeys).

North America

In North America, Passeriformes were the most common predator of polistine colonies. Still, the taxonomic diversity of polistine predators in North America was the highest for all continents (which may be linked to the higher number of predation events recorded in this region). Along with other insects (including ants) and various chordates, we also found direct evidence of a black widow spider (Latrodectus mactans Fabricius [Araneae: Theridiidae]) preying upon a colony of Polistes apachus Saussure [Hymenoptera: Vespidae] (Gibo and Metcalf 1978). Carnivora, namely bears (Family: Ursidae), skunks (Family: Mephitidae), and badgers (Family: Mustelidae), were the primary predators of vespine colonies in North America.

South America

In South America, ants were the primary predators of polistine colonies, although nonpasseriform/nonaccipitriform birds were also commonly recorded. Bird predation in South America was mostly recorded for the red-throated caracara (Ibycter americanus) upon Epiponini, Polistini, and Mischocittarini (McCann et al. 2010, McCann et al. 2013, McCann et al. 2015). In South America, we found the only record of predation by bats, Phylloderma stenops Peters (Chiroptera: Phyllostomidae), upon Polybia sericea (Jeanne 1970b), and by bees, Trigona hypogea Silvestri (Hymenoptera: Apidae), upon Agelaia flavissima van der Vecht (Hymenoptera: Vespidae) and Polybia emaciata (Mateus and Noll 2004). Among the ant predators, Eciton, Camponotus, and Crematogaster were common predators of wasp colonies in South America. Like Africa, although vespine populations have established in the temperate regions of Chile and Argentina (Masciocchi and Corley 2013), we did not find evidence of predation on vespine colonies in South America.

Asia

Asia is the only continent where all three social wasp subfamilies are present, and where predation on each was recorded. Ants are common predators of polistine colonies in Asia, but here, two of the most seemingly specialized groups of social wasp predators emerged: Vespa hornets and Pernis honey buzzards. Predation by Vespa tropica Linnaeus is one of the foremost causes of colony failure in Polistes chinensis Fabricius in Japan, and hornets may play a significant role in controlling Asian polistine populations (Miyano 1980). Although we only found seven records of direct evidence for stenogastrine colony predation, V. tropica was the primary predator of these colonies. Still, it is worth noting that an unidentified gecko (Reptilia: Gekkota) was recorded preying upon a colony of Parischnogaster mellyi. It is likely that predators of Stenogastrinae are largely unreported in the literature; however, predation on stenogastrine colonies has been described as remarkably rare in nature (S. Turillazzi, personal communication). Finally, vespines were almost only recorded being preyed upon by mammals (Primates, all Hominidae; and Carnivora), with a smaller proportion of hymenopteran predators recorded (including Vespa). Notably, human consumption of wasps was the primary cause of vespine predation in Asia, showing a unique prey–predator dynamic as well as an important cultural relationship between human societies and social wasps compared with other continents (Nonaka 2010, see ‘Humans as Predators Across Continents’ section).

Europe

In Europe, similarly to Asia, ants, Vespa hornets, and the European honey buzzard (i.e., ‘specialized predators’) were the primary predators of polistine wasp colonies. Honey buzzards were also a major predator of Vespinae. The European badger Meles meles Linnaeus (Carnivora: Mustelidae) was the most common carnivore predator on European vespines, consistent with the carnivore predation data in North American on vespine colonies.

Oceania

In Oceania, ants were the primary predators of polistine nests, and Ropalidini colonies were the only tribe to be repeatedly recorded with direct evidence of predation. As in Asia and North America, Oceania had one record of geckos attacking Polistinae colonies (DS, personal observation). Similarly to South America and Africa, despite vespine wasp colonies having established in this region, we found no evidence of colony-level predation.

Humans as Predators Across Continents

Human predation on social wasps was recorded in North America and Asia (not including nest removal for the purpose of population control or other nonconsumption-related purposes). Insects are recognized as an important food source for early hominids (Arnold 2017). In fact, contemporary consumption of edible insects by humans is still common among one-third of the world’s population (Raheem et al. 2019). In the specific case of social wasps, researchers have documented long traditions of harvesting wild nests to eat larvae and pupae. Moreover, collectors have also developed practices that can be understood to some extent as domestication, such as the rearing and keeping of wasps in human-made enclosures for their entire life cycle (Payne and Evans 2017, Saga 2019). Wasp brood is highly nutritious also for humans, being rich in proteins and containing essential amino acids (Ying et al. 2010). Wasps are notably appreciated as food in parts of Asia, being commercially available at high prices depending upon their species (Nonaka 2010). Another aspect of human exploitation of social wasp products can be seen in the use of nests in medicine recipes since ancient times (Chinese Pharmacopoeia Commission 2010). Referred to as Nidus Vespae, contemporary science has investigated the therapeutic use of nests for human health, suggesting medical significance for the treatment of rheumatoid and psoriatic arthritis, dental disease, respiratory disorders, cervical erosion, and other disorders (Wang et al. 2013).

The Co-evolution of Predators and Their Social Wasp Prey

Stinging aculeate hymenopterans evolved ~200 million years ago (mya; Peters et al. 2017, Huang et al. 2019, Tang et al. 2019). These solitary stinging ancestors to social insects likely used their stings to immobilize prey before carrying it back to the nest to feed their larvae. Vespidae evolved ~166 mya, which is also when stenogastrines split from the rest of the Vespid wasps (Huang et al. 2019). There were two separate origins of eusociality in Vespidae (Hines et al. 2007, Huang et al. 2019), once in the common ancestor of Vespinae and Polistinae (approx. 75–80 mya; Huang et al. 2019, Tang et al. 2019) and once in Stenogastrinae (between 166 and 29 mya; Huang et al. 2019). Vespinae and Polistinae originated ~62 and 55 mya, respectively (Huang et al. 2019).

Among the five main predators of social wasps, the ants (Family: Formicidae) originated before either origin of eusocial evolution in Vespidae (135–115 mya; Brady et al. 2006, Huang et al. 2019). Primates (74 mya) and Carnivora (63 mya; Class: Mammalia) originated before the three social subfamilies radiated (Fig. 5; Springer et al. 2003) and Passeriformes (47 mya) and Accipitriformes (44 mya; Class: Aves) originated after Vespinae and Polistinae (Nagy and Tökölyi 2014, Oliveros et al. 2019). In other words, the coordinated response to low levels of carbon dioxide or vibrations—both indicators of a vertebrate predator—likely evolved before these extant predator lineages did.

Fig. 5.

Evolutionary origins of predators and prey. (Top) The evolutionary origins of the top three Classes of Chordates (Reptilia, Mammalia, and Aves), and the top four Orders (from Mammalia and Aves) of colony predators. Ants and Vespa hornets are shown separately to represent evolutionary origins of the top hymenopteran predators. (Bottom) Evolutionary origins of Aculeata, Vespidae, and the three main subfamilies of Vespidae (Polistinae, Stenogastrinae, and Vespinae). It is hypothesized that eusociality evolved twice in the Vespidae Family, once in Stenogastrinae (sometime between 166 and 29 mya) and once in the common ancestor of Vespinae + Polistinae (approximately 80–75 mya). References: [1] (Laurin and Reisz 1995); [2] (Kemp 2005); [3] (Huang et al. 2019); [4] (Peters et al. 2017); [5] (Tang et al. 2019); [6] (Brady et al. 2006); [7] (Lee et al. 2014); [8] (Springer et al. 2003); [9] (Oliveros et al. 2019); [10] (Nagy and Tökölyi 2014).

Coordinated defensive behavior toward vertebrates may have evolved in the early social vespid ancestor in response to predation pressure from small mammals (225 mya; Kemp 2005), birds (116 mya; Lee et al. 2014), and possibly lizards (>300 mya; Laurin and Reisz 1995) and other reptiles, many of which may be extinct today. This predation pressure may have driven the evolution of the response of 10s or 100s of individuals to leave the nest and attack an intruding predator, and the prevalence of defensive strategies with varied degrees of aggressiveness employed throughout Vespidae, and not limited to one subfamily or tribe. Those wasps that responded fast enough and with strong enough venom would have succeeded in deterring the potential predator.

Individual stinging insects can be consumed by a variety of predators (e.g., Table 2). Some of these predators have evolved strategies to avoid the harmful effects of toxic venoms. Although they were not found to be a common colony predator, reptiles were observed to prey upon individuals and/or colonies of all three subfamilies of social vespids. Horned lizards have evolved an innate resistance to harvester ant venom, one of the most toxic mammalian venoms, through a combination of a factor in their plasma (Schmidt et al. 1989), and a slippery and viscous mucus that lines their mouth and digestive system (Schmidt 2016). In spite of being able to avoid the painful or lethal effects of venom, this resistance does not protect the lizard from a colony response (Schmidt 2016). In other words, in the co-evolutionary arms’ race of predator and prey, small predators may have evolved tolerance to an attack from a single stinging insect, but the evolutionary response of coordinated defensive attack (e.g., harvester ant attack en masse by biting) means that small vertebrate predators could be quickly overwhelmed by their potential prey.

Among the predators identified here, two orders of birds (Passeriformes and Accipitriformes) and two orders of mammals (Primates and Carnivora) originated after eusociality evolved in polistine and vespine wasps (Fig. 5). These large, robust creatures would have stumbled across a social wasp nest, filled to the brim with proteinaceous larvae, and tried their luck at having a treat. For instance, the observations of hawks attacking vespid colonies suggests that the birds are somewhat aware of the wasps’ retaliation, but overall unfazed and undeterred (van Bergen 2019, video 39, Supp Table 5 [online only]). Similar observations of carnivore predators also show that these mammals can withstand multiple stings to obtain their prize (video 33, Supp Table 5 [online only]). Social wasps’ defensive response may have evolved alongside predation: as colonies became larger, they attracted more, larger and more sting-tolerant predators, and experienced greater predation pressure. This, in turn, may have selected for stronger defensive responses by wasp colonies, which had greater numbers of workers to allocate toward defense. This positive feedback may explain the wide variation in defensive response across species: while most social wasps can sting in response to a threat, many species, particularly with small colonies, often flee rather than defend (Mischocyttarus spp., Metapolybia spp.; Hermann and Chao 1984; KJL, personal observation), or mount a relatively weak defense (Vespula consobrina Saussure; Akre et al. 1982, Gaul 1952). However, their close relatives produce impressive defensive responses (e.g., Vespula germanica; Jandt et al. 2020, Synoeca;de Castro e Silva et al. 2016). The Vespa colonies, with their large workers, strong venom, and terrifying aggressiveness may be the ultimate evolutionary response to attacks on vespine colonies.

Conclusion

Wasp colonies are preyed upon by a diverse variety of natural enemies. Social wasps have evolved various primary (nest construction) and secondary (coordinated behavior) responses to defend the colony and avoid predation. The most common predators of social wasps tend to be other Hymenoptera (specifically ants and hornets), carnivores (Order: Carnivora), primates (Order: Primates), perching birds (Order: Passeriformes), and hawks (Order: Accipitriformes) are commonly observed preying upon wasp colonies. The prevalence of these different predator groups varied across different geographical regions, and we highlight gaps in some regions where colony predation has not been recorded, despite the presence of social wasps. The specialized behavior of some predators (e.g., honey buzzards) may be a good starting point to deepen our understanding of the co-evolutionary relationships between predators and social wasp prey.

Acknowledgments

We would like to thank Aglaia Bouma, Amy Toth, Anindita Brahma, B. G. Freeman, Christopher K. Starr, Christopher Walker, D. T. Rankin, D. Windsor, Fábio Prezoto, Fernando Noll, Julie Mills Harrison, Justin Schmidt, Kléber Del-Claro, Marcos Magalhães de Souza, N. Derstine, Penny Jacks, Raghavendra Gadagkar, Ricardo Rocha Mello, Robert Jacobson, Terry Prouty, and Stefano Turillazzi for contributing to this study with their personal observations and/or providing additional references, and Ricardo Rocha Mello for providing a photograph of a social wasp predation event used here. We thank Clare I. M. Adams for her assistance in designing the biogeography map. We thank Amy Toth, Phil Lester, and Sheri Johnson for insights during the writing, and the Behavioural Ecology and Evolution reading group and members of the Jandt Lab at the University of Otago for feedback on the manuscript. We want to thank James Crofts-Bennet for sharing the publication that inspired the idea for using social media to collect observations in this study, and Karen Sun, Hongmei Li-Byarlay, and Thomas O’Shea-Wheller for inviting us to the Entomological Society of America symposium that made the idea a reality. This work was partially funded by doctoral research departmental funds provided to MD by the Department of Zoology of the University of Otago.