Cooperation is defined as any behavior that has evolved, at least in part, to enhance the fitness of other individuals (e.g., West et al., 2007; Gardner et al., 2009). The fitness-enhancing (helping) individual is called the actor, and the fitness-enhanced (helped) individual is called the recipient. Actors incur some fitness cost, either directly because of temporary reduction of their individual fitness or indirectly because of increasing the fitness of others (fitness is a relative indicator; thus, if the fitness of recipients is enhanced, individual fitness of the actor is reduced, in relative comparison). As predicted by pertinent theories, such as kin selection (Hamilton, 1964) and reciprocity (Trivers, 1971), cooperation is only evolutionary stable if the actor is more than compensated for the costs of helping and obtains a direct fitness benefit via the enhancement of its individual fitness, and/or an indirect fitness benefit via the enhanced fitness of recipients sharing genes with the actor (Sachs et al., 2004; West et al., 2007; Gardner et al., 2009). These two mutually non-exclusive ultimate drivers of cooperative behavior may be proximately subdivided according to the mode of cooperation and route of fitness gain, with authors differing in the terminology of subtypes yet often having similar, strongly overlapping, or identical meanings. Direct fitness benefits, which are based on shared interests in cooperation, may arise from byproducts of otherwise selfish behaviors of the actor, and/or enforced cooperation, that is, rewarding cooperation and punishing non-cooperation (Gardner et al., 2009; originally dubbed reciprocal altruism by Trivers, 1971; similar to directed reciprocation sensu Sachs et al., 2004). Indirect fitness benefits may arise from population viscosity (i.e., limited dispersal passively leading to actors locally interacting more likely with kin than non-kin recipients; Hamilton, 1964) and/or kin discrimination (i.e., helping actors recognizing and preferentially interacting with kin recipients) and/or green beard effects (i.e., helping actors recognizing genetically pre-determined cooperation intents of helped recipients, no matter of their relatedness at other genetic loci) (e.g., Sachs et al., 2004; Gardner et al., 2009). Sachs et al. (2004) used the terms kin fidelity for site-specific helping kin, that is, in a given site, kin are more likely to encounter each other than non-kin, without the need of kin recognition (this also includes population viscosity), and kin choice for active kin discrimination. Further, these authors subdivided byproducts into one-way, which are behaviors that are not necessarily selected for cooperation, two-way, i.e., fitness-enhancing behaviors when performed in a group and include synergism (sensu Queller, 2011), and byproduct reciprocity. Queller (2011) extended inclusive fitness theory (Hamilton, 1964) to describe how cooperation may evolve between kin, kith (selection of neighbors who are neither kin nor kind via manipulation, actor-recipient choice, or actor-recipient fidelity feedback), and kind (based on green beard alleles).

Here, we give an overview of cooperation and behaviors akin to cooperation by true spider mites (Tetranychidae), which clearly represent cooperative behaviors or which could qualify as cooperative behaviors yet we do not have enough information to judge whether these behaviors indeed qualify as cooperation according to the definitions described above. Spider mites (Tetranychidae) are globally distributed, mostly group-living herbivores (Helle and Sabelis, 1985 for review; Figure 1). Spider mites are highly rewarding model animals to view specific kin, non-kin, and heterospecific interactions from an evolutionary-grounded cooperation perspective, for a number of biological and ecological features. (i) As their name suggests, spider mites possess spinning glands in their mouthparts, with basic spinning types, such as Tetranychus spp., always and consistently spinning threads while walking (by every mobile life stage), and advanced types, such as some bamboo spider mites, being able to switch thread production on and off (Hazan et al., 1974; Saito, 1983; Clotuche et al., 2012). Over time, spinning threads result in three-dimensional webs on leaves and other plant parts, on, and under, which spider mites cohabit. The sophistication and complexity of jointly produced webs differ among spider mite genera and species, with the most advanced types being roof-like nests with protected entrances observed in grass spider mites Stigmaeopsis spp. (Saito, 1983). Accordingly, many cooperative behaviors of true spider mites and behaviors akin to cooperation are characterized and mediated by the joint use of webs produced by individual mites. Webs and single spinning threads are also beneficial as railroads and used for communication (Saito, 1983 for an overview). The ability to produce spinning threads is a decisive feature for the evolution of cooperative behaviors by spider mites. (ii) Most spider mite species are patchily distributed on their host plants and live in groups, with webbing being an important aspect of group formation and cohesion. (iii) The vast majority of spider mite species are arrhenotokous, that is, females produce haploid sons from unfertilized eggs and diploid daughters from fertilized eggs (Saito, 2010). Due to arrhenotoky, son-mother and, following mother-son mating, son/brother-sister coefficients of relatedness are 1, which has important implications to founder effects. (iv) Arrhenotoky allows colonizing host plants and the founding of patches/groups also by single immature or adult unfertilized females. While young mated females are the predominant dispersing life stage (Margolies and Kennedy, 1985; Li and Margolies, 1993; Azandémè-Hounmalon et al., 2014), also immatures, males and unfertilized females disperse (Brandenburg and Kennedy, 1982; Krainacker and Carey, 1990). For indirect fitness benefits selecting for cooperative behaviors of spider mites, the host plant colonization patterns seem as important as population viscosity in determining the kin structure within local patches and, in consequence, at regional levels. Due to arrhenotoky, any local patch founded by single females will, at least initially, result in high intra-group relatedness because of the possibility of mother-son and brother-sister mating. Patches founded by several females, or when other individuals later arrive on the host plant, may be composed of only kin if later arrivers come from the same source. Alternatively, they may represent mixed kin/non-kin patches if the founders and later arrivers come from the same genetically heterogeneous source or if the founders and later arrivers come from genetically different sources. Nonetheless, even if simultaneous colonizers are genotypically heterogenous, kin individuals are more likely to interact with each other than with non-kin, at least until their density gets too high. The reason is that females deposit and aggregate their eggs at their feeding sites, which inevitably results in local kin subgroups within larger groups/colonies/patches. Whether it is single, several, or many individuals colonizing the same plant or site on a plant is tightly linked to the mode of dispersal and spinning thread-following behaviors, as described below.

We restrict our perspective of spider mite cooperation to species of two widely studied genera, that is, Tetranychus and the nest-building grass mite Stigmaeopsis (syn. Schizotetranychus), and consider interactions among kin, non-kin, conspecific, and heterospecific spider mites (Figure 1). The contexts looked at from a cooperation perspective comprise host plant colonization and exploitation, web-sharing for anti-predator benefits, and dispersal by Tetranychus spp., and nesting-associated behaviors by Stigmaeopsis spp. infesting bamboo and other grasses. Each behavioral-ecological context is illustrated by examples from the literature. We describe the current state of knowledge of behavioral characteristics and proximate aspects, and we contemplate and discuss whether the described behaviors have evolved for direct and/or indirect fitness benefits, which subtype of cooperation they seem to represent, and whether they require a given degree of genetic relatedness to enhance the fitness of both actors and recipients.

Tetranychus spp. females disperse on and between leaves of their host plant primarily by walking (Brandenburg and Kennedy, 1982; Margolies and Kennedy, 1985). Ambulatory dispersing Tetranychus females often follow spinning threads, functioning as trails, left by preceding females. Follower females reinforce the trails with their own spinning threads, providing an opportunity for collective choice of dispersal direction (Yano, 2008). Although ambulatorily dispersing Tetranychus females do not consistently display collective choices of feeding and oviposition sites (Astudillo Fernandez et al., 2012b), collectively dispersing Tetranychus females may gain byproduct benefits from sharing webs at the new feeding site, while Tetranychus females not following trails become solitary founders of new colonies with initially high intra-colony relatedness (Yano, 2008). Local colonies founded by solitary females may later merge into extended high-density patches representing an ensemble of local kin patches. The reasons why such collective choices do not always occur are debated (Astudillo Fernandez et al., 2012b). Collective site choices in environments with specialist predatory mites that use spider mite threads for prey-searching are costly (Roda et al., 2001; Furuichi et al., 2005b; Shinmen et al., 2010). Therefore, whether collective dispersal is more advantageous than solitary dispersal is thought to depend on the strength of “positive group effects” in new habitats (Astudillo Fernandez et al., 2012a). Whether spinning thread-following behavior is influenced by genetic relatedness between pioneers and followers is unclear, but it is often kin individuals that disperse from the same patch. Bitume et al. (2013) showed that both increased local density and closer genetic relatedness increased the ambulatory dispersal distance of T. urticae. Since direct fitness benefits accrue anyway, indirect benefits arising from local kin neighborhoods may be considered jointly acting or secondary selective forces of thread-following behavior.

Besides ambulatory dispersal, Tetranychus females also disperse aerially, either alone (Smitley and Kennedy, 1985; Margolies, 1987) or as part of a woven ball (dubbed ballooning; Bell et al., 2005), which may contain both adults and immatures. Ballooning mites can also be phoretic if the balls are carried away by other animals (Brandenburg and Kennedy, 1982; Clotuche et al., 2011, 2013b). For collective ballooning, mites start to move to the apex of leaves and plants, and others follow the spinning threads to jointly produce webbing and form balls on the apex. Depending on the delay between the initiation of ball formation and take-off and the size of the balls, all ballooning mites survive and are carried away, or early arrivers are trapped inside and die and only those joining the ball at a later time survive until being carried away by the wind. Collective dispersal via ballooning could represent cooperation based on the expression and recognition of green beard alleles that may indicate kinship or not. Clotuche et al. (2013b) observed that Tetranychus individuals did not discriminate and segregate with kin during ball formation; however, this may have been due to mixed rearing before the experiment, allowing familiarization among kin and non-kin. Also, these experiments do not rule out a possible role of kin selection, because on a local scale Tetranychus individuals live more likely with kin than non-kin and, thus, may not need to discriminate who initiated or joins in ball formation. If usually formed by kin, individuals initiating ball formation could be considered kin-selected true altruists (indirect fitness gain outweighing direct fitness loss; Kay et al., 2019) because those individuals (actors) may be enclosed and die inside the balls, but may gain indirect benefits by helping kin recipients to disperse (Clotuche et al., 2011). Whether mites dying inside balls sacrifice themselves to aid in ball formation or are trapped accidentally by other mites requires close scrutiny. In any case, dying inside the balls just occurs if there is a long delay between initiating ball formation and being carried away by the wind; if the take-off occurs soon after initiation of ball formation, there are no dead individuals inside the balls (Clotuche et al., 2013b). One likely selective force of collective ballooning may be immediately acting Allee effects on the new host plant (byproduct cooperation), i.e., collective colonization of a new host plant increasing individual fitness because of positive group effects (synergism sensu Queller, 1985) as compared to solitary colonization (Clotuche et al., 2013a,b). Cooperation in forming high density aggregations on tips of overexploited host plants may also counter dehydration (byproducts), as has been shown for the house dust mite Dermatophagoides farinae (Glass et al., 1997).

Predation pressure is a strong selective force for the evolution of sociality (e.g., Wilson, 1971). Woven nests of Stigmaeopsis spp. provide some physical protection from predators. However, several predators, such as the predatory mite Typhlodromus bambusae, are able to intrude into the nests, especially through nest entrances. Stigmaeopsis spp. (Stigmaeopsis miscanthi, Stigmaeopsis sabelisi, Stigmaeopsis longus, Stigmaeopsis celarius, and Stigmaeopsis nanjingensis) that construct extended large nests show cooperative brood defense (counterattack) by males against intruders. Nests of S. miscanthi are sometimes only occupied by a single male and then the male and females jointly defend the nest (Saito, 1990). Nests of S. longus (Figure 1) are commonly inhabited by several males, which jointly defend the nests. Adult mites (biparental, i.e., both males and females, or just males) drive potential intruders away from nests by pursuit, jabbing, and beating, and, sometimes, even kill immature predators (Saito, 1986a,b; Yano et al., 2011; Saito and Zhang, 2017). The success rate of counterattack varies among species (Mori and Saito, 2005), and increases as the number of adult mites in a nest increases, i.e., the success of counter-attacks positively correlates with the size of nests (Saito, 1986a,b; Yano et al., 2011; Saito and Zhang, 2017). Such cooperative defense behaviors seem absent or less effective in species that construct separate small nests, such as S. takahashii and S. saharai (Mori and Saito, 2005). However, their nests are so small and the nest entrances are so narrow that predators are rarely able to intrude. As a consequence, the physical protection provided by the small nests is much higher than that of extended large nests (Mori and Saito, 2004). Moreover, separate scattered nests decrease the success of predators in detecting nests with live prey inside, because nests with sucked-out prey corpses can function as a trap for predators (Saito et al., 2008). Altogether, these studies suggest that differences in nest and group sizes in the genus Stigmaeopsis are associated with divergence in anti-predator strategies: cooperative defense by counterattacking predators in large groups and constructing smaller more protective nests in small groups. Counterattacks against potential intruders protect their own and the offspring of nestmates but incur the costs of being killed by predators. Therefore, nest size and cooperative defense are regarded as key traits in the evolution of grass spider mite sociality.

The group of S. miscanthi species (S. miscanthi HG and ML forms, S. sabelisi, S. continentalis, and S. formosa) infests Miscanthus spp. grass, enlarges and extends their nests over time, and counterattacks predatory intruders (Saito et al., 2018, 2019; Sato et al., 2019). Adult males are not only aggressive against predators but also against conspecific males and may even kill each other to establish a harem (Saito, 1990). Stigmaeopsis longus (Figure 1) engage in precopulatory mate guarding without lethal fighting, whereas S. miscanthi males may fight to death inside nests when competing for females. However, the intensity of male-male aggression, quantified by the frequency of lethal male fights, varies among species and among populations in the S. miscanthi species group and seems to correlate with winter harshness (Saito, 1995; Saito and Sahara, 1999; Sato et al., 2013). Winter harshness can mediate average genetic relatedness among nestmates in the S. miscanthi species group because mother-son mating during spring colony-establishment occurs more likely in colder than warmer regions (Saito, 1995; Sato et al., 2013). Therefore, kin selection is a plausible explanation of the geographic variation in male-male aggression. Non-lethal fighting may represent cooperation by non-killing actors helping kin recipients to survive and reproduce at the expense of a decrease in the direct fitness of the actor but an increase in indirect fitness (Saito, 1995; Saito and Mori, 2005). Alternatively, lethal fighting could represent spite (Hamilton, 1970; Foster et al., 2001; Gardner and West, 2006; Sato et al., 2013).

Social immunity is defined as “any collective and personal mechanism that has emerged and/or is maintained at least partly due to the anti-parasite defense it provides to other group members” (Meunier, 2015). In nest-building organisms, social immunity is achieved by nest sanitation behaviors to prevent or reduce disease transmission and keep the living space inside nests clean. Waste management is widespread from communal to eusocial species (Jackson and Hart, 2009) and is closely associated with the evolution of sociality in the Acari (Saito, 1997). Some species of the genus Stigmaeopsis show obvious waste management (Figure 1). For example, the S. miscanthi HG form, which lives on Miscanthus spp. grass, constructs large woven nests by continuously extending its nests. In exceptional cases, large nests may be inhabited by more than a hundred individuals with three overlapping generations (Saito et al., 2000). Inside the nests, one or several fecal piles, spaced at similar distances, may be found. Fecal piles emerge by two simple behavioral rules: mites deposit their feces at sites with previous feces; in absence of previous feces, they deposit their feces inside the nest close to one of the two entrances (Sato et al., 2003). The mites recognize fecal sites by volatile chemical cues and the nest entrance by tactile cues (Sato et al., 2003). Similar waste management has been observed in S. longus, which also constructs continuously enlarged nests but infests bamboo plants. However, in this species, the first fecal pile is deposited outside nests (Sato and Saito, 2006; Figure 1). Stigmaeopsis takahashii and S. saharai, which also infest bamboo plants but rather construct separate new nests than expand existing nests, deposit their feces outside the nest entrances, and do not respond to volatile chemical cues (Sato and Saito, 2006, 2008). Therefore, the use of volatile chemical cues in waste management is thought to have co-evolved with extending and enlarging existing nests. In S. longus, additional highly sophisticated nest cleaning behaviors have been reported. Females keep spinning threads after nest construction, which not only function to reinforce the nests but also to remove exuviae and other dust, possibly containing pathogens, scattered on the leaf surface inside nests (Kanazawa et al., 2011). To this end, females walk in a zigzag pattern and spin threads that are soft and sticky when fresh. These threads trap the exuviae and dust from the floor (the leaf surface) of the nests. Females push the trapped exuviae and dust up and glue them to the woven roof of the nest, resulting in a clean leaf surface inside the nests beneath the roof. Cooperation in nest building and social immunity activities have clear direct benefits, so arise from byproducts, but it is more than plausible to also assume a role of kin selection in these behaviors and indirect fitness benefits since it is usually and predominantly kin that live together and enlarge nests (kin fidelity sensu Sachs et al., 2004).