Ecological characteristics of insects that affect symbiotic relationships with mites


Correspondence: Kimiko Okabe, Forestry and Forest Products Research Institute, 1 Matsunosato, Tsukuba, Ibaraki 305-8687, Japan.



Parasites and pathogens that begin as symbionts, i.e., organisms living together in the same habitat, are some of the most promising drivers of species evolution. Because insects are highly diverse and important as ecosystem service agents and because mites can exert large effects on insect populations (capable of killing at least juveniles), insect–mite interactions have been analyzed from various perspectives, including evolutionary, ecological and pest-management perspectives. Here, I review and examine insect–mite symbiotic associations to develop hypotheses concerning the factors that maintain and develop their relationships. Previous studies have hypothesized that insect sociality and mite richness and specificity affect insect–mite interactions. I found that both solitary and social insects, including parasocial and subsocial insects, harbor numbers of symbionts including species-specific ones but few dangerous mite symbionts in their nests or habitats under natural conditions. Nest size or the amount of food resources in a nest may affect mite richness. On the basis of this review, I hypothesize that the insect characteristics relevant for mite symbiotic hosting are sharing the same habitat with mites and living in a nutrient-rich habitat. I also suggest that many cases of species-specific symbiosis began with phoresy. To test these hypotheses, phylogenetic information on mites living with insect groups and quantitative analysis to characterize each insect–mite relationship are necessary.


Insects are highly diverse and important as ecosystem service agents in roles including as pollinators and pest control agents. Because mites can have large impacts on insect populations (capable of killing at least juveniles), numerous reports have documented insect–mite associations from ecological, evolutionary and pest-management perspectives. For example, mutualisms between mites and insects have been quantitatively described (Wilson & Knollenberg 1987; Okabe & Makino 2008b), coevolution between carpenter bees and chaetodactylid mites has been suggested (Klimov et al. 2007a), and biological control agents against insect pests by parasitism and/or predation are used in agriculture (Gerson & Smiley 1990). However, under natural conditions, the effects of mites on insect population control appear to be minimal in most associations at both the individual and population levels. From the mite perspective, living larval to adult insects might not be readily attainable prey, because the small body size of mites makes it difficult for an individual to attack insects. On the other hand, insects are important hosts for parasitism and dispersal (phoresy) (Athias-Binche 1991; Houck & OConnor 1991).

Parasites and pathogens probably began as neutral symbionts, i.e., organisms living together in the same habitat. These organisms are some of the most promising drivers of species evolution, as outlined by the Red Queen Hypothesis (Paracer & Ahmadjian 2000; Salathé et al. 2008). Because parasites and parasitoids are paramount natural enemies that must be overcome or avoided by host organisms, hosts have evolved many strategies against them, likely including sociality (Schmid-Hempel 1998; Wilson et al. 2003; Boomsma et al. 2005). Generally, social insects are thought to be able to effectively exclude invaders or foreigners from their nests through brood care and recognition of nest mates, although nests that provide stable habitats with large food resources for inquilines are attractive to many other organisms (Schmid-Hempel 1998). Although fewer parasites are expected in ant nests based on this hypothesis (e.g., Berghoff et al. 2009), social bees do not always eliminate symbionts or inquilines from their nests, perhaps because they actually function as commensals or mutualists (e.g., Eickwort 1994). Thus, it would be valuable to compare social organizations of a wider range of insect hosts using different habitats. Given that sharing the same habitat is the origin of various symbioses, it is also important for further evolutionary study to determine the characteristics of the habitats most likely to be shared by many insects and mites.

Here, I review and examine insect–mite symbiotic associations to develop hypotheses concerning the factors that maintain or develop their relationships, such as phylogenetic restrictions and ecological characteristics, including social organization and habitat. I first overview the relationships between mites and social insects. To the extent possible, I subdivide sociality into categories such as eusociality (all three of the following ecological traits are observed: overlap of adult generations, reproductive division of labor and cooperative care of young), parasociality (at least one of the three traits is present or sharing of common nest sites is observed) and subsociality (parents interact with young). Second, I review associations between mites and solitary insects that are closely related to their social counterparts. In addition, I focus on habitat characteristics that may be correlated with symbiotic mite richness and/or certain relationships between mites and insects (Table 2). I focus mainly on direct interactions because few reliable assessments have dealt with indirect interactions (e.g., an interaction through a food-web). Each insect group at the family or order level varies greatly because of individual entomologist or acarologist interests or because fewer interactions occur between particular insects and mites. For the taxonomies of insects and mites, I primarily consulted reviewed publications for insects (e.g., Hirashima 1989 for Japanese insects) and the work of Krantz and Walter (2009) for mites, except in cases for which strong evidence was available for revision.

Taxonomic and ecological overview of mites and types of interactions between mites and insects

When interested in species interactions, we usually observe whether two or more organisms live together. When they do, this is termed symbiosis in the original sense, although symbiosis is sometimes used restrictively for mutualism or only when discussing microorganisms (Cheng 1991). In this review, I follow the original meaning of symbiosis (living together) and symbionts (different species living together), but I also try to characterize the relationship further, i.e., into phoresy, parasitism, commensalism and/or mutualism (Roberts & Janovy 1996; also see Table 1).

Table 1. Symbiotic relationships between insects and mitesa
  1. aDefinitions based on Lincoln et al. (1998).
ParasitismSymbiosis between two different species, in which one (a parasite) metabolically depends on the other (a host) and the parasite adversely affects but rarely kills the host. A parasite that steals stored food in its ordinary state is called a cleptoparasite (kleptoparasite). Many cleptoparasitic mites are considered commensals (Walter & Proctor 1999). When the parasite kills the host, it is called a parasitoid.
CommensalismSymbiosis in which one organism benefits from, but does not adversely affect, the other.
MutualismSymbiosis in which both organisms benefit.
PhoresySymbiosis in which an organism is transported on the body of a different species without nutritional or developmental benefit for itself.
InquilinismSymbiosis in which a species lives in domicile, e.g., a nest, with another species. Adverse or favorable effects on the host are not necessarily discussed.

Mites (Acari) as a subclass are divided into two superorders. The superorder Parasitiformes includes four orders: Opilioacarida ancestrally ingested particulate foods such as fungi and dead arthropod bodies, most Holothyrida are scavengers, Ixodida are vertebrate parasites and Mesostigmata are primarily fluid feeders. The superorder Acariformes includes two orders: Trombidiformes (the former Prostigmata except for the new suborder Endeostigmata) including insect parasites and Sarcoptiformes (Endeostigmata, the former Oribatida and Astigmata, which is now considered a cohort of Oribatida) (Krantz & Walter 2009). The richest habitats for mites include the soil and litter layer, with 50–250 000 mites per square meter (Walter & Proctor 1999). Today, mites display a wider range of feeding habits than any other Arachnida, although the dominant habits are fungivores and scavengers of dead plant materials or algae and lichens (Evans 1992; Walter & Proctor 1999). Interactions with insects commonly occur in Mesostigmata typically as free-living predators and phoronts (organisms conveyed by others) and occasionally as parasites; in Prostigmata (Trombidiformes) as predators, parasites, and detritivores, all sharing the same habitat with insects; and in Heterostigmata (Trombidiformes) and Astigmata as parasites, fungivores and scavengers, perhaps living together with insects to later become phoronts. A common association with insects for mites of Mesostigmata, Heterostigmata and Astigmata is phoresy: an interspecific relationship between organisms in which one (phoront) attaches to another (host) for only the purpose of dispersal; hence, during the period of phoresy, neither nutritional exchange nor development of mites occurs (Farish & Axtell 1971; Houck & OConnor 1991). Although phoresy is a type of commensalism, I highlight this particular interaction for two reasons: (i) a different interaction may arise after phoretic dispersal during a certain developmental stage; and (ii) other relationships may have evolved through mediation by phoresy (Houck 1994; Houck & Cohen 1995).

Ancestral mite development involves egg, prelarva, larva with three pairs of legs, protonymph, deutonymph, tritonymph and adult. The prelarval stage has not been reported in Parasitiformes, and Mesostigmata suppress the tritonymph. An interaction between a mite and an insect may change in accordance with mite developmental stage: generally either deutonymphs or adults (typically females) are phoretic in Mesostigmata, deutonymphal astigmatids are usually phoretic, and only larval parasitigones are parasitic to insects (Walter & Proctor 1999). Astigmatids have highly modified phoretic deutonymphs lacking mouthparts and anal openings, and heterostigmatids tend to have phoretomorphs (specialized morphology for phoresy) in adult females (Moser & Cross 1975; Houck & OConnor 1991). Although most mesostigmatids do not exhibit distinctive morphological adaptations for phoresy, some deutonymphal uropdines have anal pedicels to attach to host insects (Walter & Proctor 1999).

Mite associations with social insects


Among ants (Formicidae), army ants, particularly Eciton burchellii (Westwood), are well known for associations with other organisms. Eciton burchellii temporarily constructs nests or bivouacs above ground, with 100 000 to 700 000 workers (one of the largest known ant colonies), and forages on the ground surface during the nomadic phase after a 20-day statuary phase (Hölldobler & Wilson 1990; Rettenmeyer et al. 2011). Mites, including scavengers and parasites, are the most common and diverse guests of these temporary nests and bivouacs (Rettenmeyer 1962; Gotwald 1996; Rettenmeyer et al. 2011). Some mite species reportedly show distinctive morphological adaptations to act as ant symbionts. For example, some Macrochelidae are parasitic specifically to E. burchellii and related species, as they take blood from hosts by attaching to membranous areas of host bodies (Schmid-Hempel 1998). Although mutualism has been suggested based on the morphological adaptation of the mite Macrocheles rettenmeyeri Krantz to attach specifically to the leg membrane of Eciton dulcius crassinode Borgmeier and on the unusual tolerance of the ant for the mite, possibly due to chemical mimicry by the mite (Eickwort 1990), there is no clear benefit to the ant (Krantz 1962; Rettenmeyer 1962; Krantz & Walter 2009).

Other army ants that use even more ephemeral shelters (i.e., shorter or no statuary phase) with generally fewer workers appear to have fewer symbiotic mites than E. burchellii (Table 2), although Gotwald (1996) reported that Neotropical army ants associate with more mite species than with myrmecophilous insects. Some mite species are reported to be specific to army ants. For example, adult Trichocylliba (=Circocylliba and Planodiscus) mites of Uropodidae are phoretic only on doryline and ecitonine ant hosts (Elzinga 1978). Female Perperipes ornithocephala Cross (Microdispidae: Heterostigmata) associated with army ants have an extremely elongate body, which probably represents mimicry of the ant larva, and these females are thought to feed specifically on ant eggs and larvae (Cross 1965). To date, no reports have documented lethal parasites among army ant mites.

Table 2. Ecological characteristics of insects associated with mites related to symbiotic or specific interactions
Insect groupSocialityNest characteristicsHabitat or nesting siteFoodLongevityGeographic distributionFood habits of associated mitesSpecific interactions with mite(s)Reference
ProvisionNutrition available for symbionts except for host hemolymphIndividualColony/nest
  1. –, no information or does not apply. References: 1 Hölldobler & Wilson 1990; 2 Krantz & Walter 2009; 3 Rettenmeyer et al. 2011; 4 Cross 1965; 5 Elzinga 1978; 6 Schmid-Hempel 1998; 7 Yamasaki et al. 2010; 8 Krombein 1967; 9 Cowan 1984; 10 Klompen et al. 1987; 11 Okabe & Makino 2003; 12 Okabe & Makino 2008a; 13 Eickwort 1990; 14 Sammataro et al. 2000; 15 Oldroyd & Wongsiri 2006; 16 Eickwort 1994; 17 Roubik 2006; 18 Michener 2000; 19 Okabe et al. 2008; 20 Phillipsen & Coppel 1977a; 21 Phillipsen & Coppel 1977b; 22 Schuster & Schuster 1997; 23 Kirkendall et al. 1997; 24 Okabe 2009; 25 Wilson & Knollenberg 1987; 26 Takaku et al. 1994; 27 Eggert & Müller 1997; 28 Hosoya & Araya 2005; 29 Okabe & Goka 2008; 30 Okabe et al. 2012; 31 Treat 1975; 32 Walter & Proctor 1999; 33 Gerson & Smiley 1990.
Ants (Formicidae) except for army ants and parasitic antsSocialExposedNest in forest to grassland, bare fieldsNoNest debrisMonths to several yearsSeveral yearsCosmopolitanParasite, scavenger, insect predatorParasitism, inquilinism (scavenging), phoresy, possibly mutualism1, 2
Eciton burchellii (an army ant)SocialTemporary nest or bivouacNest in trees above groundTemporaryNest debrisSeveral monthsSeveral yearsNeotropicsMaybe parasite, scavenger, insect predatorParasite, inquilinism (scavenging), phoresy, possibly mutualism2, 3
Army ants except E. burchelliiSocialTemporary bivouac to no nestNest in forests, swamps area, desertNoNest debrisSeveral monthsSeveral yearsOld and New tropicsMaybe parasite, scavenger, insect predatorParasite, inquilinism (scavenging), maybe mutualism1, 4, 5
Vespine wasps (Vespidae) except for parasitesSocialExposed or nesting in cavities, soilNest in forest to grasslandsNoMicroorganismFor a few weeks to monthsAnnualFrom tropics to temperate, Old and New WorldScavengerNot known2, 6
Polistine wasps except for the Polistes spp. aboveSocialExposedNest in forest edge to grassland, bare fieldNoMicroorganismAbout 10 days to 12 monthsAnnualCosmopolitanMaybe wasp parasite (K Okabe pers. obs., 2000)Not known2, 6
Polistes gallicus, Polistes snelleniSocialExposedNest in forest edges to grasslands, bare fieldsNoMicroorganismAbout 10 days to 12 monthsAnnualSouthern Europe and Russia, Northern Africa, middle-east to east Asia for gallicus and China, Korea, Japan for snelleniPerhaps scavengerInquilinism (perhaps scavenger), phoresy2, 7
Eumenine wasps (Vespidae)SolitaryExposed or nesting in closed (e.g., dead plant materials)Nest in forest edges to grasslandsAlways for juvenilesAnimal meat for larvae, microorganism, nest debris including dead larva and larval fecesOne to several monthsOne to several monthsCosmopolitanScavenger, fungivore, parasite, insect predatorParasitism, cleptoparasitism, inquilinism (scavenger), phoresy, mutualism8, 9, 10, 11, 12
Honey bees (Apis, Apidae)SocialExposed or nesting in cavitiesNest in forest edges to grasslands, bare field (e.g., on rock)Always for juvenilesFlower nectar, pollenAbout 1 month to several yearsSeveral yearsPalearctic region, Pacific maritime of Russia, Africa, Oriental regionsScavenger, fungivore, parasite, insect predatorParasitism, phoresy13, 14, 15
Stingless bees (Meliponini, Apidae)SocialCavities including trees and below groundNest in forestAlways for juvenilesFlower nectar, pollen, microorganism, nest debris, microorganismAbout 24–50 daysAbout 1 to several decadesCentral and South America, AfricaScavengerCleptoparasitism, Inquilinism (probably scavenger), phoresy13, 16, 17, 18
Bumble bees (Bombinae, Apidae)SocialBelow ground in soilNest in forest to grasslandsAlways for juvenilesFlower nectar, pollen, microorganism, nest debrisMonths to a yearAnnualEurope, Southeastern to eastern Asia, North and South AmericaScavenger, fungivore, parasite, insect predatorParasitism, inquilinism (probably scavenger), phoresy13, 16, 18
Large carpenter bees (Xylocopinae, Apidae)SocialDead plant materialsNest in forest, forest edgeAlways for juvenilesFlower nectar, pollen, microorganism, nest debrisMonths to a yearAnnualPantropic and some temperate regionsScavenger, fungivore, arthropod predatorCleptoparasitism, phoresy, maybe mutualism (by microphagy)16, 18, 19
Sweat bees (Halictidae)Social/ solitaryBurrow in soil, rarely in dead woodNest in forest, forest edges, grasslandsAlways for juvenilesFlower nectar, pollen, microorganism in solitary bee nest, nest debrisSeveral weeks to monthsOne to several months for solitary; about a year for socialTemperateScavenger, microphagousCleptoparasitism, phoresy, maybe mutualism (microphagy)13, 16, 18
MegachilidaeSolitaryMostly in cavitiesNest in forest, forest edge, grasslandAlways for juvenilesFlower nectar, pollen, microorganism, nest debrisSeveral weeks to months (reuse of a nest is possible for Osmia spp.)One to several monthsCosmopolitanScavenger, egg predatorsCleptoparasitism, inquilinism (scavenging), phoresy16, 18
Termites (Isoptera)SocialIn soil or dead woodNest in forest to grassland, bare fieldFungal cultivationMicroorganismSeveral years to decadesSeveral years to decadesTropical to temperate regionsScavengerInquilinism (scavenging), phoresy2, 13, 20, 21
PassalidaeSocial or colonialLiving in dead woodForestNoDead wood, microorganism, nest debrisSeveral month to yearsProbably several yearsPantropicalScavenger, microphagy, nematophagyInquilinism (scavenging, nematophagy), phoresy2, 22
Bark and ambrosia beetles (Scolitidae)SocialDead woodForestFungal cultivation for juveniles in ambrosia beetlesDead wood, microorganism,Several months to a yearSeveral months to a yearCosmopolitanFungivore, nematophagy, insect predatorPhoresy2, 23, 24
Asocial carrion beetles (Silphidae)SocialMaintaining larval food in soilForest to grasslandDead animal meatDead animalsSeveral months to a yearSeveral weeks to monthsCosmopolitan (mostly temperate zone)Scavenger, micropahgy, insect, parasite, predatorParasitism, phoresy, mutualism2, 25, 26, 27
Other SilphidaeSolitaryNo nestNest in forest to grasslandNoSeveral months to a yearCosmopolitan (mostly temperate zone)Microphagy, predatorProbably phoresy2, 26, 27
Dorcus beetles (Lucanidae)SolitaryNo nestForestNo(Host exudates)Several months to 2–3 yearsEurope, Northern Africa, Asia, Northern Australia, North and Central AmericaScavenger, parasiteParasitism, commensalism (scavenging)26, 29, 30
Noctuids associated with mitesSolitaryNo nestForest to grasslandNoHost tissuesSeveral weeksCosmopolitanParasiteParasitism2, 31
Non-social aquatic insectsSolitaryNo nestFresh waterNoSeveral weeks to monthsCosmopolitanParasite (larva)Parasitism, phoresy2, 32
Grasshoppers (Orthoptera)Solitary or colonialNo nestGrasslandNoSeveral weeks to monthsCosmopolitanParasite (larva)Parasitism2, 33

Non-legionary ants also associate with myrmecophilous mites. Among the mesostigmata, myrmecophily is relatively well described. For example, the cohort Uropodina includes the myrmecophilous Trachyuropodidae and Oplitidae, and several other families of the cohort Gamasina include moderate to specific associations with ants (Krantz & Walter 2009). However, many others are described as effective predators of ants. Immature mites of Macrodinychus sellnicki Hirschmann (Dinychidae: Mesostigmata) are ectoparasitic on the pupal crazy ant, Paratrechina fulva (Mayr), and can serve as a biological control agent based on their preference for this species (González et al. 2004; Krantz et al. 2007). González et al. (2004) reported that parasitism rates by M. sellnicki reached over 90% in the field in Colombia and that no mite populations were found in areas of crazy ant outbreaks, suggesting a strong effect of this mite species on ants. Parasitism on ants appears to have evolved independently in different mite lineages. The myrmecophilous Sphaerolaelaps holothyroides (Leonardi) (Pachylaelapidae: Mesostigmata) is specific to British Lasius ants nesting in soil, and the mite may obtain food by intercepting trophallaxis by its host (Eickwort 1990); thus, they could be defined as cleptoparasitic (Table 1). This mite is morphologically well adapted, with modified chelicerae for piercing and anchoring, so that ovoviviparous larvae can avoid ant predation without harming the host (Krantz et al. 2007). Recently, an unidentified uropodid that is parasitic on native ants has also been reported to parasitize invasive Pheidole megacephala (Fabricius) in Okinawa, Japan (Le Breton et al. 2006). Although high parasitism rates and mortality inflicted by this mite are expected to reduce ant populations, a low rate of parasitism on queen ants was observed in the field, probably because mites were removed by the extra grooming of queens by worker ants (Le Breton et al. 2006).

Phoresy to workers is most frequently described throughout all groups of mites. Mites that are frequently collected with ants, e.g., pygmaphorids and scutacarids of Heterostigmata and histiostomatids and acarids of Astigmata, are suspected consumers of refuse deposits in ant nests and are probably myrmecophilous (Eickwort 1990; Berghoff et al. 2009). In Europe, Imparipes hystricinus Berlese (Scutacaridae) is the most common mite species found in ant nests and in the surrounding soil, feeding on fungi that grows on the bodies of dead ants (Eickwort 1990). Most heterostigmatid and astigmatid mites have been suggested to be commensals; however, the effects of such “garbage collecting” mites on ant nest hygiene have never been assessed.

Overall, oribatid mites that are free-living within soil, litter and trees, including canopy trees where ants often encounter them, are a good food resource for small ants even though the mites have sclerotized outer protection (Masuko 1994; Wilson 2005). Indonesian ant species appear to maintain oribatids in their nests, perhaps as an alternative food source (Ito & Takaku 1994; Ito & Aoki 2003). Although the behavior of one inquiline oribatid, Protoribates myrmecophilus Aoki & Ito (Protoribatidae) associated with Myrmecina spp., is similar to that of other free-living oribatids, another ant inquiline, Aribates javensis Aoki et al. (Aribatidae), cannot survive without its specific host, an unidentified Myrmecina (Ito & Aoki 2003). Mutualism has been suggested as a mechanism to explain why these mites, particularly the latter, exhibit higher survivorship in ant nests than away from nests, as the ants could feed on the mites as emergency food (Ito & Takaku 1994; Ito & Aoki 2003; Ito 2013).

Some life history traits (e.g., nesting habits and diet) that differ across ant species may affect symbiotic mite richness, such as myrmecophilous mites belonging to Mesostigmata of Parasitiformes and Acariformes (Krantz & Walter 2009). For example, the nests of small colony species tend to harbor fewer symbionts (i.e., species of ponerines, dacetines and leptothoracines; Hölldobler & Wilson 1990). Ants have a large number of mite associates that are dominant in soil, litter, or dead wood, likely because most ant species nest in these same habitats.

Social wasps

Symbiotic relationships between social wasps and mites have rarely been described; specific associations are known only between Polistes gallica (L.) (Vespidae) and Sphexicozela connivence Mahunka (Winterschmidtiidae: Astigmata) and between Polistes snellei de Saussure and an unidentified Sphexicozela (Mahunka 1970; Yamasaki et al. 2010). Although Medeus vesparius Volgin (Acaridae: Astigmata) was originally described from Vespula germanica (Fabricius) (Vespidae), later collections indicated that Medeus mites are not wasp associates but are instead solitary bee associates (OConnor 1996, 2001). Parasitus of Mesostigmata, Pyemotes of Heterostigmata, and Glycyphagus and Acarus of Astigmata were found in Vespula and Polistes nests as scavengers or parasitoids, but they were only temporal migrants, as multiple individuals have rarely been reported to occur in wasp nests (Spradbery 1973; Schmid-Hempel 1998).

Social bees

Several types of relationships between social bees and mites have been reported: inquilinism for scavenging, parasitism including cleptoparasitism, commensalism and possibly mutualism. Among eusocial Apis honeybees, most reported mite associates are associated with Apis cerana Fabricius and A. mellifera L. (Apidae). These relationships have formed because of apiculture and not because of the ecological characteristics of the bees. However, recent increases in parasitic mites on non-native hosts could help our understanding of the nature of host specificity and the attributes of symbiotic relationships. Apicultural pests such as Varroa, Euvarroa (Varroidae, Mesostigmata), Tropilaelaps (Laelapidae, Mesostigmata) and Acarapis (Tarsonemidae, Heterostigmata) are European honeybee (A. mellifera) parasites, but they appear to have much less impact on native bee hosts (Asian honeybees), perhaps because of grooming behavior in native bees (Eickwort 1990; Boecking & Spivak 1999; Sammataro et al. 2000; Oldroyd & Wongsiri 2006 Warrit & Lekprayoon 2009). Four described species of Varroa are hosted by cavity-nesting Asian honeybees (A. cerana, A. koschevnikovi Enderlein and A. nigrocincta Smith) with less species specificity, whereas Euvarroa mites were originally parasitic to dwarf bees (A. andreniformis Smith and A. florea Fabricius) in Asia (Oldroyd & Wongsiri 2006; Warrit & Lekprayoon 2009). Varroa jacobsoni Oudemans and Varroa destructor Anderson & Trueman are two recently separated species based on distinctive differences in mtDNA and a slight difference in morphology (the latter is larger in body size, although overlap exists between them) (Anderson & Trueman 2000). These species may also potentially be distinguished by geographic distribution (the former was originally distributed in India and Southeast Asia including Indonesia and Malaysia, whereas the latter occurs in eastern to southeastern Asia including Japan, Korea and Thailand). However, the latter species has been collected beyond its original distribution in colonies of A. mellifera (Anderson & Sukarsih 1996; Anderson & Trueman 2000). Varroa mites (probably V. destructor based on its high pathogenicity and original distribution) transferred to the USA via Europe from far-eastern Russia, where they had been spread from native A. cerana colonies to A. mellifera by Ukrainian beekeepers in the 1950s (Guzman et al. 1997). The adaptation of V. destructor to A. mellifera was curiously rapid compared to that of V. jacobsoni, which was unable to recognize A. mellifera for oviposition until recently when a genetically distinct strain of the species began to reproduce on the western honeybee in Java (Oldroyd 1999). These invasion histories suggest that a reduction of pathogenicity is only effective between a parasite and the native host.

Tropilaelaps mites are also parasites of immature bees and potentially harmful to their native hosts, A. dorsata and A. laboriosa; the mites were originally distributed within the host distribution range (Oldroyd & Wongsiri 2006). Females of both varroids and laelapids are phoretic on host bees (Oldroyd & Wongsiri 2006). Acarapis woodi (Rennei) (Tarsonemidae: Heterostigmata) is a parasite that sucks blood on the tracheal surface of adult honeybees (A. mellifera and A. cerana), although three species belonging to Acarapis appear to be less harmful than mesostigmatid parasites in apiculture (Eickwort 1988). Acarapis woodi spread throughout North America in the 1980s after initially being found in England; however, the species had not been found in Scandinavia, most of Southeast Asia, China, Japan or the Caribbean countries until the late 1980s and is thought to exhibit low survival in the tropics (Bradbear 1988; Morse & Nowogrodzki 1990; Warrit & Lekprayoon 2009).

Although well-detailed mite–honeybee associations are limited to parasitism, most mite associations on honeybees are commensal and/or cleptoparasitic, whereas predators that probably feed on other arthropods are common in stingless bee nests; thus, predacious mites may be mutualistic (Schmid-Hempel 1998). Scavenger mites that naturally inhabit flowers (e.g., Neocypholaelaps and Afrocypholaelaps of Ameroseiidae) are sometimes introduced into bee colonies by phoresy, although maintained A. cerana colonies do not appear to be disturbed by them (Okabe et al. 2000). In Polish apiaries, 99% of honeybee colonies (A. mellifera) were infested with stored product mites that were scavenging hive products, dead hosts, fungi and other organic substances; the mites usually did not negatively affect the bees (Chmielewski 1991). Synchronized life cycles of phoretic mites and their Apis hosts have never been reported, but parasitic mites appear to prefer young adult hosts (Sammataro et al. 2000).

Stingless bees (Meliponini: Apidae), which are eusocial to parasocial (semisocial: among the three traits of eusociality, generation overlap is missing) bees, are a close relative of honeybees (Apis) that pre-date them by 65 million years (Michener 2000). Stingless bees nest in soil or cavities of existing structures, including hollow trees and active termite nests, and the nests are generally large and used for long periods to accumulate new cells (Michener 2000; Roubik 2006). Despite high brood density in nests, diseases are rare in stingless bees (Roubik 2006). In total, 43 mite species (including one morphological species) associated with honeybees and 42 species associated with stingless bees (e.g., several genera of Laelapidae, Melissotydeus of Tydeidae (Heterostigmata) and several genera of Gaudiellidae and Meliponocoptidae of Astigmata) have been documented as stingless bee specialists (Eickwort 1988, 1990; OConnor 2009). While 26% of these mites were obligate honeybee associates, 95% were obligately associated with stingless bees (Eickwort 1988). Less biological information exists for meliponine associates compared to Apis associates, but two types of mutualism have been suggested for meliponine associations: (i) pest control, for example, a predacious mesostigmatid (Neohypoaspis ampliseta Delfinado-Baker, Baker and Roubik) possibly feeding on cleptoparasitic astigmatids in Trigona fulviventris Guérin-Méneville; and (ii) pathogen removal, e.g., Neotydeolus therapeutikos Flechtmann & Camargo (Tydeidae, Heterostigmata) possibly feeding on fungi in nests of T. postica (Latreille) (Flechtmann & Camargo 1974).

Mites are also common and potentially dominant in and around bumblebee nests (Bombus spp., Apidae) in soil or wood cavities. Prasitellus of Parasitidae and Pneumolaelaps of Laelapidae are specifically associated with bumblebees as predators of other arthropods and/or cleptoparasites without harming the hosts (Royce & Krantz 1989; Eickwort 1994). The mites are usually found in large numbers in bumblebee nests (Costa 1966; Hunter & Husband 1973; Krantz & Walter 2009). In particular, the cleptoparasitic Parasitellus fucorum (De Geer) and Hypoaspis bombicolens (Canestrini) (Laelapidae) exhibit a range of host specificity during phoresy (B. terrestris (L.), B. lacorum (L.) and B. lapidarius (L.) are preferable hosts for the former mite and B. lapidarius is for the latter) in Switzerland, although ecological characteristics associated with the specificity remain unknown (Schwarz et al. 1996; Koulianos & Schwarz 1999). Macrochelids that have sometimes been collected from bumblebee and honeybee nests to serve as predators of other arthropods may be temporal migrants of soil-nesting insects, as they are not restricted to bee nests (Eickwort 1990).

Eusocial Bombus spp. are parasitized by Locustacarus buchneri (Stummar) or L. trachealis Ewing (Podapolipidae: Heterostigmata) but generally at very low infestation rates (Husband & Husband 1996). In North America, bumblebee nests are co-inhabited by L. buchneri (a truly parasitic tracheal mite) at low infestation rates and by Kuzinia americana Delfinado-Baker (Acaridae) and Parasitellus (=Parasitus) spp. (Parasitidae) (both are scavengers) at higher rates, although the bees are not harmed by these mites (Husband & Sinha 1970; Goldblatt & Fell 1984; Goka et al. 2001; Otterstatter et al. 2005). Although deutonymphal parasitoid mites overwinter with queen bees and are transported to their new nests, the life cycles of most associated mites may not be well synchronized with their host life cycles because of the constant emergence of workers, except in winter (Huck et al. 1998).

Interactions between xylocopine bees and associated mites are well described in large carpenter bees (Xylocopa) and small carpenter bees (Ceratina). Unlike honeybees, stingless bees or bumblebees, the sociality of large and small carpenter bees lacks divisions of labor, although the bees maintain a short period of maternal care of the brood; thus, they are considered to be subsocial. Their nests are similar to those of solitary bees in the use of dead plant materials: plant materials are excavated to form a tunnel for provisioned cells using a mixture of pollen and nectar (Michener 2000).

Old World Xylocopa, the subgenera Mesotrichia, Koptortosoma and Afroxylocopa females have pouch-like mite pockets called acarinaria on the first metasomal tergite, and these usually contain mites (Madel 1974; Eickwort 1994; for details of acarinaria, also see the “solitary wasps” subsection below). The metasomal acarinarium may have developed for Dinogamasus mites of Laelapidae because the structure appears to be highly specialized for mites of that size. The acarinarium of Xylocopa latipes (Drury) is also exploited for phoresy by smaller mite species such as Tarsonemus platynopodae Magowski (Tarsonemidae), Sennertia hipposiderus (Oudemans), S. koptorthosomae (Oudemans) (Chaetodactylidae: Astigmata) and Horstia helenae (Oudemans) (Acaridae) (OConnor 1993). On the other hand, some species, such as those in the genera Stigmatolaelaps and Xylocolaelaps of the Laelapidae family, are also restricted to large carpenter bees but do not use an acarinarium during phoresy, even when it is not occupied by Dinogamasus mites (Royce & Krantz 2003). Cheletophyes apicola Fain, Lukoschus and Nadchatram (Cheyletidae: Prostigmata) on X. latipes and Dinogamasus mites stay only in the metasomal acarinarium, whereas smaller mites have also been collected from thoracic acarinaria, the metasomal acarinarium, and a less-developed indentation on the first abdominal tergite (OConnor 1993; Okabe & Makino 2002, 2005). Based on the observation of Madel (1974) that Dinogamasus mites were present around the host juvenile but did not harm it, mutualism to remove microorganisms has been suggested, but clear evidence of pathogenic microorganisms and their removal by the mite is lacking (OConnor 1993; Eickwort 1994).

Unlike mites associated with honeybees or stingless bees, synchronized life cycles of a scavenging mite, Sennertia alfkeni (Oudemans) and its host carpenter bee, X. appendiculata circumvolans (Smith), has been reported, although the mite can also grow to the dispersal stage (deutonymph) in the absence of the host (Okabe et al. 2008). Sennertia alfkeni is morphologically well adapted to be phoretic on the host: whereas larger deutonymphs use claws to grasp host hair during phoresy, smaller mites attach to the smooth surface of the host exoskeleton using their attachment organ (Okabe & Makino 2002, 2005; Okabe et al. 2008). Although another common scavenger in carpenter bee nests, Horstia verginica Baker, is suspected to kill the carpenter bee brood and feed on nectar (Krombein 1962), the biology of Horstia mites is not well known.

Halictidae (sweat bees) are broadly distributed in temperate areas, where they nest in soil; they include social and solitary species (Sakagami & Munakata 1972; Michener 2000). Perhaps because of the soil habitat, mites associated with halictid bees include several species that are closely related to ant-associated mites; for example, cleptoparasitic Laelaspoides ordwayae Eickwort (Melittiphinae, Laelapidae), which is related to the ant-associated genus Laelaspis (also Melittiphinae), is phoretic on eusocial Augochlorella bees of Halictidae (Eickwort 1966). Also, Parapygmephorus species are restricted to halictids, but its close relative Petalomium is myrmecophilous (both pygmephorids are fungivorous). However, as seen in all Melittiphinae, scutacarids and Sancassania mites of Acaridae, many insect-associated mites in the soil at the genus or family level are symbionts of a broad range of soil-nesting insects including bees (Eickwort 1988; Krantz & Walter 2009).

A previous study suggested that fungivorous mites might serve as sanitary mutualists of the facultatively social sweat bees, Megalopta genalis Meade-Waldo and M. ecuadoria Friese (Halictidae), although neither the mite nor the fungal species was documented (Biani et al. 2009). On the other hand, fungivorous Trochometridium mites (Trochometridiidae: Heterostigmata), which are symbionts of soil-nesting bees including halictids, kill host eggs or larvae and then feed on mold that may have been introduced by the mite via sporothecae (a lobe-like structure to enclose fungal spores) into the cell (Cross & Bohart 1969, 1978; Lindquist 1985).

Phoresy of mites is also common for the social halictids. The life cycle of Parapygmephorus costaricanus Rack & Eickwort, for example, closely follows that of the parasocial host, Agapostemon nasutus Smith (Halictidae) (Rack & Eickwort 1980). Imparipes apicola (Banks) (Scutacaridae), which is also associated with halictids, exhibits a similar life cycle to that of P. costaricanus, but this mite more actively migrates in the soil to find a phoretic host (Eickwort 1979, 1994). Although Cosmolaelaps vacuus (Michael) and Hypoaspis queenslandicus (Womersley) (Laelapidae) are frequently found to be predators of halictid bee nests, they are not phoretic on the bees (Eickwort 1979).

Subsocial passalid beetles

Passalidae are pantropical beetles, numbering approximately 600 species. They primarily gather in dead wood with members of their own family as subsocial insects (Schuster & Schuster 1997). Most mites reported from passalids are considered passalid specialists, and whether incidental associates exist has not yet been determined. Members of both superfamilies of Fedrizzioidea and Diarthrophalloidea (Mesostigmata) primarily occur on passalid beetles (Krantz & Walter 2009). While Fedrizzioidea live in passalid pupation chambers and galleries and feed on nematodes and dead arthropods (Krantz & Walter 2009), interactions other than suspected commensalism on or under the elytra of beetles have not been quantified for any instar in Diarthrophallidae (Lombardini 1951; Hunter & Glover 1968; Schuster & Summers 1978). Similar associations have been observed in Apalotacarus and Passalophagus of Canestriniidae (Astigmata) that are associated only with adult beetles (Nesbitt 1976; Summers & Schuster 1981, 1982; OConnor 2009). All species of Heterocheylus (Heterocheylidae: Heterostigmata) have short styliform chelicerae, suggesting parasitism in the subelytral space, although this has never been documented (Lindquist & Kethley 1975; Hunter 1993). The tunnels of passalids house specialized astigmatids, such as Kanoetus and Scolianoetus of Histiostomatidae (filter feeders with modified chelicerae to feed on, for example, bacteria), Passaloglyphus and some Schwiebea of Acaridae (originally fungivores) (OConnor 1994); however, their effects on hosts remain unknown. Although parental care behaviors such as feeding and help with the construction of pupal cases have been observed in passalids, antagonistic behavior against mites has never been reported.

Mite associations with solitary insects compared to those with closely related social hosts

Solitary wasps

Compared to social wasps, solitary wasps make relatively fewer juvenile chambers, called cells (1 to approximately 10 per nest), which are partitioned with plant materials and/or soil; hence, no parental care is observed (O'Neil 2001). Unlike social wasps, solitary wasps exhibit mite associations, although less so than bees. Mason wasps (Vespidae), which make nests using soil, are well-known associates with specific winterschmitiid mites (Astigmata), but few other guests or migrants are found in nests (Krombein 1967; Mostafa 1970; Cross & Moser 1975; Eickwort 1988; OConnor & Klompen 1999). Obligate inquilines that live specifically with solitary hosts demonstrate synchronized life cycles with those of the hosts: the phoretic stage of a mite disembarks from its host during host provisioning of cells and then reverts to the phoretic stage by the time of host pupation or eclosion at the latest (Krombein 1967; Eickwort 1979; Cowan 1984; Klompen et al. 1987; Okabe & Makino 2003, 2008a; Okabe et al. 2008). When some solitary wasps spin a cocoon for pupation, scavenging mites that are usually spread across the cell wall come together on the prepupal host to get inside the cocoon (Krombein 1967; Okabe & Makino 2003). Truly parasitic mites such as Ensliniella parasitica Vitzthum (Winterschmidtiidae) exhibit a life cycle with low plasticity, and mite development halts when the hosts die (Okabe & Makino 2008a). The deutonymph of the empirically host-specific mite, Kurosaia jiju Okabe & OConnor (Winterschmidtiidae), which has both parasitic (tritonymphs and adult females) and scavenging stages (larvae and protonymphs) in its life cycle, rides on newly emerged Macrosiagon nasutum (Thunberg) (Rhipiphoridae: Coleoptera) that are predators of the mite host, Anterhynchium flavomarginatum micado (Kirsch) (Vespidae) in the same cell (K Okabe pers. obs., 2002). This observation suggests that inquiline mites might lose or not develop the ability to search for a specific host. Another example also suggests that phoresy on the host while living together is not necessary for the mites: tritonymphs and possibly the adults of Lackerbaueria krombeini Baker (Acaridae) kill eggs of the host, Psenulus atratus parenosus Pate (Diodontus atrantus parenosus in Krombein 1967) (Crabronidae), and deutonymphs of the mite become phoretic on surviving hosts, emerging from inner cells sequentially made in the same nest (Krombein 1967). On the other hand, the location of a host during phoresy is probably crucial for host-specific mites; in the case of Ensliniella parasitica, the mite seems to be able to reach an acarinarium without random migration (Okabe & Makino 2011).

Acarinaria range from shallow indentations to well-developed pocket-like structures that are externally located on insects (particularly hymenopterans) to facilitate mite transportation (Eickwort 1994; OConnor & Klompen 1999). In solitary wasps, acarinaria are usually developed on both sides of the scutellum, on the propodeum (posterior ventral mesosoma), and/or on the second metasomal tergites covered by the first or perhaps interiorly on a posterior metasomal tergite (OConnor & Klompen 1999; Makino & Okabe 2003). The structures are suspected to have evolved independently several times in Vespidae (OConnor & Klompen 1999). With the development of acarinaria, the enslinielline mite lineage may exhibit evolution of parasitism from commensalism as scavenging inquilines (Krombein 1967; Klompen & OConnor 1995). Interestingly, both Kennethiella trisetosa (Cooreman) (Winterschmidtiidae) associated with Ancistrocerus antilope (Panzer) (Vespidae) without an acarinarium and E. parasitica associated with Allodynerus delphinalis (Giraud) (Vespidae) with acarinaria are cleptoparasitic to paralyzed moth larvae (larval wasp food) as tritonymphs and truly parasitic to the juvenile host during other stages, except for the deutonymph (Cowan 1984; Okabe & Makino 2008a). However, the guarding mutualism of E. parasitica against the perilous juvenile parasitoid Melittobia acasta (Walker) (Eulophidae) is mediated only by acarinaria (Okabe & Makino 2008b, 2010).

Solitary bees

Compared to eusocial bees and solitary wasps, solitary bees are specifically associated with a wider range of mites including mesostigmatids, heterostigmatids and astigmatids. Megachilid nests are probably best known to host inquiline mites. Examples include Neocypholaelaps and Afrocypholaelaps, which are also associated with honeybees, as well as astigmatids including Vidia of Winterschmidtiidae; Chaetodactylus of Chaetodactylidae; Tortonia of Suidasiidae; and Neohorstia, Horstia, Cerophagopsis, Megachilops, Sennertrionyx, Schulzea and some Sancassania of Acaridae (Eickwort 1994). Chaetodactylus mites are moderately host specific to solitary bees, such as Ceratina (Apidae), Osmia and Megachile (Megachilidae), and they reportedly kill the eggs and juveniles of their hosts (Krombein 1962; Fain 1966; Maeta 1978) but perhaps only when initial mite populations are very high (Qu et al. 2002). Tortonia species also kill host eggs before becoming cleptoparasites (Eickwort 1994; Qu et al. 2002). Chaetodactylus nipponicus Kurosa has been reported as a pest of the commercialized pollinator Osmia cornifrons (Radszkowski) (Qu et al. 2002), even though the phoretic rates of chaetodactylids on bees are generally low in the field (Eickwort 1994). One potential reason for the mite infestation of commercial bees may be that farmers re-use the same nest materials (Eickwort 1994). Whereas Chaetodactylus deutonymphs have both phoretic- and inert-morphs in a single population and the latter remains in a nest after host eclosion, Sennertia mites of the sister genus have only phoretic deutonymphs, likely because of the host nesting behavior, that is, no re-use of old nests (Eickwort 1994; Qu et al. 2002; Okabe et al. 2008). Several instances of life cycle synchronization of Chaetodactylus with that of hosts have been reported (Qu et al. 2002).

Ctenocolletes (Stenotritidae) is one of the rare solitary bee groups known to develop acarinaria to carry specific Sancassania (=Ctenocolletacarus) mites (Fain & Houston 1986; Houston 1987; OConnor & Klompen 1999). The behavior of entering and detaching their acarinarium is very similar to that of other acarinarium users of both bees and wasps (Houston 1987; Okabe & Makino 2008a). Although Houston (1987) hypothesized that sanitary mutualism was at play, as seen in facultatively social halictids and their mites, acarinaria on some bee species may function to trap parasitic mites (Klimov et al. 2007b). Lasioglossum females of solitary halictid have acarinaria on the anterior surface of the first abdominal tergite, and they are frequently filled with Anoetus deutonymphs (Histiostomatidae) (McGinley 1986). Anoetus species feed on microbes on the surface of stored pollen and nectar and on juveniles of halictid bees in host nests; these mites are almost certainly mutualistic, as they remove pathogens and molds (Eickwort 1979, 1990).

Non-social lucanid beetles

Lucanidae is the sister family of Passalidae (Schuster & Schuster 1997). It contains more than 1000 species and is especially diverse in Asia (Hosoya & Araya 2005). The larvae are xylophagous, and adults of semivoltine species overwinter in dead wood materials such as logs and stumps (Michaels & Bornemissza 1999). The beetles are associated with parasitic and saproxylic micro- and macrosymbionts including fungi, nematodes and mites (Okabe & Goka 2008; Kanzaki et al. 2011). Although some genera include subsocial species, others such as Dorcus beetles are solitary (Hosoya & Araya 2005). Dorcus adults are associated with species-specific canestriniids, although some associations are based on false specificity that is more restricted by geographic isolation (Okabe & Goka 2008; Okabe et al. 2012). Most canestriniids associated with lucanids inhabit the sub-elytral space directly on abdominal surfaces or on the surface of the host exoskeleton, probably as commensals that feed on host exudates, but some parasites do feed on host hemolymph in the sub-elytral spaces (OConnor 2009). Compared to subsocial passalids, solitary lucanids likely do not host many phoretic mites, although their overwintering habitats have never been investigated (Okabe & Goka 2008; Okabe et al. 2012).

Conclusions and future perspectives

Social insects, including those that are parasocial and subsocial, have few dangerous mite symbionts (i.e., parasites that strongly reduce host fecundity, parasitoids and/or predators) in their nests or habitats under natural conditions, except for macrochelids associated with crazy ants. On the other hand, many social insects exhibit a wide variety of associations with fungivores and scavengers. Although few ecological studies have examined mite associations with termites, subsocial bark and ambrosia beetles (Scolitydae and Platypodidae), or subsocial burying beetles (Nicrophorus of Silphidae), no studies have reported negative effects of mites on host fecundity: 20 genera of mites are considered termatophilous (Eickwort 1990), mite species belonging to at least 31 families are associated with bark or ambrosia beetles (Okabe 2009) and more than 212 mite species have been reported on carcasses with animals including carrion beetles (Perotti & Braig 2009). An indirect antagonism mediated by a fungus has been suggested for only one bark beetle–mite association: Tarsonemus mites (Tarsonemidae) (T. ips Lindquist, T. krantzi Smiley & Moser and T. fusarii Cooreman), which are phoretic on the pine beetle Dendroctonus frontalis Zinmerman (Scolytidae). They carry fungal spores of Ophiostoma minus (Hedgcock) (Ophiostomaceae), which is a mite food resource (hence, a mutualism exists between the mites and the fungus). Hence, they decrease the reproductive success of the beetle through competition for the fungus, which is symbiotic to the beetle (Lombardero et al. 2003).

Although adequate information is lacking for the quantitative analysis of social and solitary insect interactions with mites, brood care by adults is likely crucial for reducing disadvantageous symbionts. For example, previous studies have found a positive correlation between Nicrophorus beetles and the presence of phoretic Poecilochirus sp. (Parasitidae) within their larval chambers, and the mite helps to eliminate flies that compete against the beetles (Wilson 1983; Wilson & Knollenberg 1987). Sanitary mutualism has also been suggested between Dinogamasus mites and subsocial carpenter bees, which can not directly care for their broods in cells; thus, non-eusocial insects that engage in less brood care might exploit symbiotic mites. To compare the effects of social level and social and solitary life histories on the frequency of symbionts, I suggest further quantitative analysis of the effect of brood care on both specific and non-specific but lethal mites.

Consequently, I identify three characteristics of insect–mite relationships:

  1. Host sociality does not consistently affect the diversity of mite associations (Table 2; Eickwort 1990, 1994). Social insects have large colonies or aggregates, but colony size is not always correlated with mite richness (appears to be positive in the army ant E. burchellii but no evidence exists for a similar pattern in other army ants; Hölldobler & Wilson 1990; Rettenmeyer et al. 2011). On both social and solitary insects, symbiotic and particularly host-specific mites reduce virulent or negative effects (reduction of virulence of parasites to social hosts; Hughes et al. 2008; Berghoff et al. 2009), and many other podapolipids on solitary insects also do not cause high mortality on hosts (Hochmuth et al. 1987; Gerson & Smiley 1990; Otterstatter et al. 2005). I suspect that fewer pathogenic parasites and commensals are not excluded from host nests, and eventually such nests exhibit mite richness depending on the amount of food resources.
  2. Sharing the same habitat promotes specific symbiosis; while soil-dwelling insects have the highest probability of encountering mites because of mite dominance in soil, few mites are associated with insects that have diversified on living plants as herbivores (Walter & Proctor 1999; Krantz & Walter 2009). Based on studies of insect–mite associations in patchy habitats such as on mushrooms, aggregation of host individuals may also function to mediate host-switching of mites as well as to maintain symbiotic mites with the same host because of the chance of infection or phoresy (OConnor 1984; O'Connell & Bolger 1997; Sueyoshi et al. 2007; Okabe 2013). Phylogenetic studies of symbionts (mites) and/or hosts are necessary to determine how the symbiotic relationship originated and when the host switch occurred (e.g., Klimov et al. 2007a). Because astigmatids derived from free-living oribatids have exhibited a wide range of associations with both vertebrates and invertebrates over the evolutionary period of associations and dependence on hosts, life-history modifications in astigmatid mites may represent good examples (e.g., OConnor 1994).
  3. Nests that maintain nutrition are crucial for symbiotic mite richness; for symbiotic mite richness, the accumulation of nutrition that mites can exploit appears to be more important than the time period of nest existence. For example, nests of E. burchellii (which exhibit low stability in bivouacs but are nutrition-rich based on population size) appear to have more mite associations than nests of most non-legionary ants (which show high stability but fewer food resources for mites). Nutrition-rich nests of social bees host more mites than those of social wasps without provisions. This hypothesis should be explored using comparisons of field nests of insects under different conditions including mite richness.

On the basis of this review, I suggest that many of the examples of species-specific symbiosis known today originated from phoresy, as mites tend to depend on these insects for dispersal. One future avenue of research would be to test whether phoretic mites have developed post-phoretic symbiosis or if inquiline mites have developed phoresy among specific insect–mite relationships, as hypothesized for ant–fungus mutualisms (the consumption-first model and the transmission-first model; Mueller et al. 2001). It is also necessary to collect phylogenetic information on mites living with insect groups as well as quantitative ecological data for both groups and to clarify and quantify symbiotic interactions. Species interactions are complex, and insect–symbiont systems generally include several organisms such as mites, nematodes and fungi (Cardoza et al. 2008). Thus, multiple species, including both macro- and micro-symbionts, must be included in future studies.