Systematics and ecology of North American bee-associated mites: potential threats to native and introduced pollinators

"...the last 5 years of losses of honey bee colonies in North America leave us with fewer managed pollinators than at any time in the last 50 years and ... the management and protection of wild pollinators is an issue of paramount importance to our food supply system" (Allen-Wardell et al., 1998).

    This project develops baseline data regarding the diversity, host associations, geographic distribution and ecological associations of mites associated with native and introduced North American bees. This information is necessary for agricultural interests to develop and manage a new generation of native and non-native pollinators following the disastrous losses to feral and managed honey bee populations due to attack by parasitic mites. In this introduction, I will briefly document the economic importance of pollination services in agricultural production, discuss the reason for the recent precipitous decline in honey bee populations, and detail efforts to alleviate this situation through the development of replacement pollinators. I will then show how these "new pollinators" may face the same kind of peril and how knowledge of the naturally occurring community of associated parasitic, cleptoparasitic, commensal and mutualistic mites can aid in the management of these bees.

I. Economic importance of pollination by bees.

    McGregor (1976) listed 130 commercially grown crops in the United States that depend upon bee pollination to both increase yield and sometimes increase fruit size. Delaplane and Mayer (2000) list the following crops grown in the United States as having the greatest dependence upon bees for pollination: alfalfa seed, almond, apple, asparagus seed, avocado, beet seed, blueberry, cabbage seed, canola, cantaloupe, carrot seed, cherry, clover (various types), cotton seed, cranberry, cucumber, kiwifruit, onion seed, peach, pear, plum/prune, raspberry, soybean, squash, strawberry, sunflower, tomato, and watermelon. The actual economic value of bee pollination has been estimated by different authors over the years based on various formulae that take into account the difference in yields observed in experimental situations where bees are excluded or allowed to pollinate. Robinson et al. (1989) calculated the economic impact of honeybee pollination alone at over US$9 billion annually, while Southwick and Southwick (1992) reduced this figure to between US$1.6 and $5.7 billion annually by taking into account pollination by non-managed bees. Putting it succinctly, Buchmann and Nabhan (1996) summarized these and other figures by stating, "Even in the fedspeak of billions and billions of dollars, this is not small change."

II. Pollinator declines and the aftermath.

    Prior to the late 1980's, most commercial pollination in the United States was accomplished by the honey bee, Apis mellifera, a species introduced from Europe in colonial times. Managed and feral honey bees had their problems, notably bacterial, viral and fungal diseases that caused some headaches to beekeepers (most recently reviewed in Morse and Flottum [1997]), but nothing like what was to come. At some point following the importation of European Apis mellifera into eastern Asia in the late 19th century, a mite now known as Varroa destructor colonized these bees from their original host, the native Asian honey bee, Apis cerana (see Anderson & Trueman, [2000] for use of this name over the name V. jacobsoni for these mite populations). The subsequent history of dispersal of this mite pest is reviewed in Wienands (1988) and Bailey & Ball (1991), but involved transportation of Varroa-infested A. mellifera back to Europe and ultimately into the United States prior to 1987 despite proposed draconian measures designed to keep the pest out of this country (Lawrence, 1989). This mite pest, and secondarily another introduced mite pest, the tracheal mite Acarapis woodi, caused the virtual extinction of non-managed honey bee colonies in the United States in the 1990's.

    Following this catastrophe, the pollination industry took a two-fold approach to the problem: attempting to control pest mites on honey bees by chemical, cultural and genetic methods, and examining other bee species to see if they could be managed for specific pollination tasks (Torchio, 1994). Fifteen years after the introduction of Varroa destructor into North America, honey bees have not recovered, and a new industry has developed to manage and use "new pollinators." Among the "new pollinators" are a number of old, native North American species which have been both encouraged to pollinate crops and in some cases, actually managed, i.e. reared in artificial nest aggregations and delivered to crop sites. Chief among these are the orchard mason bees, family Megachilidae, genus Osmia, and most notably Osmia lignaria. Another approach to "new pollinators" has been to purposefully introduce exotic species. The model for this approach was the alfalfa-leafcutter bee, Megachile rotundata, which was introduced in the early 20th century. This bee is easily managed in soda-straw nests and has largely taken over the pollination of alfalfa seed from native species such as the alkali bee, Acunomia melanderi, and can also pollinate clovers (Richards, 1987, 1995). The megachilid bee, Osmia cornuta, has been recently introduced from Europe as a pollinator of orchard crops, notably almonds. In addition to these non-social bees, species of social bumblebees (Bombus spp.) have been managed in both North America and Europe, and some European bumblebees have been introduced into Australia and Mexico.

    In the midst of these developments, a number of biologists noticed a general decline in the numbers and diversity of pollinators of both crop plants and wild species. These declines have been attributed to many factors, notably habitat destruction, the elimination of floral diversity through monoculture, and overuse of chemical pesticides with unintended consequences for non-target species (Buchmann & Nabhan, 1996). All of the above changes to the pollination landscape have resulted in a much greater awareness of the importance of pollination and the need for conservation of pollinators. Several symposia have been held in the past few years to bring together specialists in these issues (e. g., Stubbs & Drummond, 2001; Strickler & Cane, 2002). In their "white paper" commissioned by the Society for Conservation Biology, Allen-Wardell et al. (1998) stated, ".... Among the most critical priorities for future research and conservation of pollinator species are (1) increased attention to invertebrate systematics, monitoring and reintroduction as part of critical habitat management and restoration plans..." Several of the recommendations of this group of experts with respect to non-Apis bees are relevant to this project, notably their recommendations for "increased funding for studies of a diversity of economically important plant-pollinator relationships, including analyses of the life histories and the ecological and economic roles of flying insects and their pollen or nectar sources;...long term monitoring protocols for use in protected areas that allow the collection of baseline data to establish benchmarks in order to assess changes in the diversity and abundance of pollinators through time;...[and] a greater investment in training invertebrate systematists in order to overcome the lack of taxonomic expertise." (Allen-Wardell et al., 1998).

    An important component of this understanding involves the relationship between non-Apis bees and mites. Information on the native fauna of bee associated mites in North America is scattered, incomplete, and difficult to access by the non-specialist. These factors, coupled with a serious decline in the number of specialists on mites in North America (OConnor, 1990), have led to the present project. An understanding of the diversity and the role played by the various mite associates of native bees in their natural situations is necessary in order to monitor for host shifts onto economically important species. Secondly, as native or introduced bee species are brought under management, the role of mite associates may shift. For example, Bosch (1992) found that the mite, Chaetodactylus osmiae, existed at a less than 10% infestation rate in wild populations of the orchard mason bee, Osmia cornuta in its native habitat in Spain. After developing managed colonies of this bee for almond pollination in Spain, he found that incidence of this cleptoparasite increased sixfold in the managed colonies. This bee species has been introduced into North America, but knowledge of North American Chaetodactylus mites is very limited, and it is not possible to say whether the European mite also now exists in North America. Anecdotal evidence also indicates that native North American Chaetodactylus are on their way to pest status in managed Osmia colonies (see Collaborative Arrangements).

    The information on mite associates needed to properly assist in the management of wild and domesticated pollinators includes basic taxonomic information, that is, accurate species identification. Knowledge of the natural host associations is also needed to provide baseline data prior to any geographic moves of native bees or introduction of exotic species. Such baseline data will allow the recognition of host shifts or geographic introductions when such events occur. Finally, knowledge of the natural geographic range of each species is also necessary in order to avoid introduction of potential pest mites into new areas.

A second type of information required for proper pollinator management is ecological, that is, what is nature of the association between mite and bee species? There is a need to be able to identify mite species that can be detrimental to bees, and an equal need to be able to identify those which are not. Identification of detrimental mites can lead to control measures being implemented, while identification of commensal or beneficial mites obviates the need for control.

III. Current status of knowledge of the North American bee associated mites.

    Most systematic work published to date on mites associated with native North American bees took place between 1960 and 1990, accomplished by systematists associated with the USDA and academic institutions. Preston Hunter of the University of Georgia studied the laelapid genus Pneumolaelaps, associated with bumblebees, describing most of the known North American species (Hunter, 1966; Hunter & Husband, 1973). Earl Cross of the University of Alabama studied the systematics of heterostigmatid mites, including bee associated taxa now placed in the families Pyemotidae, Pygmephoridae and Trochometridiidae (Cross, 1965; Cross & Moser, 1975), and also studied the astigmatid mite associates of the alkali bee, now known as Acunomia melanderi (Cross, 1968). Cross also collaborated with bee researcher George Bohart on biological studies on these mites (Cross & Bohart, 1969, 1978). George Eickwort of Cornell University primarily studied the systematics and ecology of bees themselves, but also studied bee-mite associations involving several groups of mites including the Laelapidae (Eickwort, 1966), Heterostigmata (Rack & Eickwort, 1979), and Histiostomatidae (Eickwort, 1985). He also wrote or collaborated in reviews of mites associated with Halictidae (Eickwort, 1979) and honey bees (De Jong et al., 1982; Eickwort, 1988). Hunter, Cross and Eickwort are all now deceased, with Cross and Eickwort both victims of auto accidents while still active in research. A final academic researcher, Robert Husband of Adrian College in Michigan, studied mite associates of bumblebees as well as publishing on the biology of the bees themselves (Hunter & Husband, 1973; Husband & Sinha, 1970). Husband is now retired, but still actively publishing on podapolipid mites, primarily associated with beetles.

    Two of the most prolific North American workers on bee-associated mites were Edward Baker and Mercedes Delfinado-Baker, both associated with the USDA in Beltsville, Maryland; Baker with the Systematic Entomology Laboratory, and Delfinado-Baker ultimately with the Bioenvironmental Bee Laboratory (now the Bee Research Laboratory). Both acarologists published on a wide variety of taxa. In the 1960's, Baker collaborated with hymenopterist Karl Krombein on systematic and biological studies of astigmatid mites associated with Hymenoptera, including bees (Baker, 1962a, 1962b; Krombein, 1962a, 1962b), and in the 1970's he began collaborating with Delfinado on insect associated mites (Delfinado & Baker, 1976). After their marriage, Baker and Delfinado-Baker concentrated on mites associated with social bees, primarily Apis and stingless bees, publishing descriptions of numerous gamasine and astigmatid taxa. Delfinado-Baker, herself, published primarily on the systematics of Apis-associates including Varroa, Tropilaelaps and Acarapis. Following their retirement, they moved to the Philippines, where declining health prevented any further research activity. With the retirements of Baker and Robert Smiley from the Systematic Entomology Laboratory, and with budget constraints preventing the permanent replacement of the two mite specialists, no further work on bee-associated mite systematics is currently underway in that unit. Ronald Ochoa, who did postdoctoral work on bee- associated mites under the PD's direction (Ochoa & OConnor, 1996, 2000), is currently under contract at the SEL to handle identifications and conduct research on plant-feeding mites of agricultural importance. Systematic work at the Bee Research Laboratory also ended with Delfinado-Baker's retirement. Acarological work continues in that unit specifically on the biology and control of Apis-associated mites, particularly Varroa destructor, under the direction of Jeffrey Pettis.

    The PD conducted his dissertation research at Cornell University under the direction of George Eickwort and has continued to publish on the systematics and biology of bee-associated mites since moving to the University of Michigan. He is currently the only active systematist in North America working on mites associated with bees other than Apis. He has published mainly on astigmatid mites (OConnor, 1988, 1991, 1992, 1993a, 1993b, 1993c, 1997; OConnor & Daneshvar, 1999; OConnor & Eickwort, 1988; OConnor & Klompen, 2000; Lombert et al., 1987), but has also published on gamasine mites (OConnor et al., 1991, 1997), and heterostigmatid mites (Ochoa & OConnor, 1996), the three main groups containing bee associates.

Following are brief summaries of the known diversity of obligate bee-associated mite taxa in North America, including unpublished observations of the PD. Known diversity, host associations, feeding habits and known or potential pest status are discussed. All of the mentioned taxa exhibit phoretic associations with adult bees. In addition to these taxa, other species also occur as facultative or accidental associates in bee nests, but do not exhibit phoretic associations with adult bees (e. g. Macrochelidae [Richards & Richards, 1977], Ereynetidae [Hunter & Cross, 1968], and other taxa noted in Eickwort [1979, 1988] and OConnor [1988]).

    A. Order Parasitiformes, Suborder Gamasina

        1. Family Parasitidae. This large family of mainly soil dwelling predators includes one lineage of obligate associates of Bombus. These species were formerly placed in the large genus Parasitus, but are now currently recognized as a separate genus, Parasitellus (Hyatt, 1980). Knowledge of this genus in North America is limited to the work of Richards (1976) and Richards and Richards (1976) who studied these mites in Alberta, Canada. Four species were recognized, all of which were described as new (Richards, 1976). The species are not host specific, with species often co-occurring in individual Bombus nests. Like other Parasitidae that practice phoretic dispersal, Parasitellus species are phoretic as deutonymphs that occur primarily on overwintering queens. The exact nature of the association between these mites and their bumblebee hosts is uncertain, although predatory behavior toward acarid mites and other parasitid mites was suggested (Richards & Richards, 1976). Additional study is needed to determine if other species of this genus exist in North America, and to determine their geographic ranges and actual feeding preferences. If these mites feed preferentially on potentially damaging acarid mites, they may be beneficial to colony health. The lack of host specificity in this group may be the result of host switching in flowers, as has been demonstrated for the European species, Parasitellus fucorum (Schwarz & Huck, 1997).

        2. Family Laelapidae. This extremely diverse family includes free-living predators and various commensals and parasites of arthropods and vertebrates. Among the numerous genera of laelapid mites associated with bees, most are associated with social species (Apis, Bombus, Meliponini), but others are found with colonial nesting Halictidae and Xylocopini. Notable among this group are the parasitic genera Varroa (see Casanueva [1993] for the current systematic position of this genus), species of which are responsible for the worldwide decline in Apis mellifera populations, and Tropilaelaps, also parasitic on Apis species in Asia. In North America, a rather restricted fauna of native bee-associated Laelapidae is known. Nine species of Pneumolaelaps are all associated with Bombus (Hunter, 1966; Hunter & Husband, 1973). These mites are not host specific, and are phoretic as adult females on bumblebee queens. In the nest, they have been observed to feed on pollen and nectar stores. They also feed on hemolymph of dead or injured bees (Hunter & Husband, 1973). The known species are widely distributed in eastern North America; the western fauna has not been studied. As with Parasitellus, lack of host-specificity in Pneumolaelaps species may be the result of host-switching in flowers, as has been demonstrated for the European species, P. hyatti (Schwarz & Huck, 1997).

The other obligate bee-associated laelapid in North America is Laelaspoides , forming a monobasic genus associated specifically with halictid bees of the genus Augochlorella in the central United States. These mites feed on pollen and are phoretic as adult females (Eickwort, 1966, 1979).

    3. Family Ameroseiidae. This family includes fungivores and pollen feeders, with many Old World species associated with flowers and dispersing on pollinating insects and birds. Species of Neocypholaelaps are commonly encountered in Apis hives as well as in flowers across the native range of Apis in the Old World (reviewed in Delfinado-Baker et al., 1989). Described New World species all belong to fungivorous genera living in non-flower habitats, but the PD has recently obtained specimens representing one of the Old World genera from flowers of red mangrove trees in Florida and the West Indies. Species in this genus are known to be phoretic on honey bees in the Australian and tropical Pacific regions, and it is possible that the presence of this mite in the Caribbean region represents an introduction from the Old World.

    B. Order Acariformes

        1. Suborder Sarcoptiformes, Infraorder Astigmata. Astigmatid mites share an ancestral life-history pattern involving exploitation of patchy habitats and dispersing via a highly specialized deutonymphal instar (often termed "hypopus") (OConnor, 1982). Such dispersing deutonymphs are morphologically quite different from the preceding and following instars and do not feed, but attach phoretically to adult insects or vertebrates (Houck & OConnor, 1991). A number of families of Astigmata include obligate bee-associates. In every case where life cycles are known, the habitat for the feeding stages of the mites is the nest of the bee.

            a. Family Histiostomatidae (=Anoetidae). Feeding stages of species in this large, diverse family have mouthparts highly specialized for filter-feeding, and all require a water film in which to feed. Generic level taxonomy is confused, especially regarding the large genus Histiostoma (with over 200 named species) and Anoetus, a small group of obligate associates of halictid bees including two described North American species. Some authors have used the name Anoetus for the entire assemblage (e.g., Mahunka, 1968), others split the halictid associates between Anoetus and Histiostoma (Woodring, 1973), but Mahunka (1974) separated all the halictid associates as the genus Anoetus. McGinley (1986) noted the presence of Anoetus deutonymphs on a large number of species in the halictid genus Lasioglossum (s. lat.) in North America, but these mites have not been studied. Phoretic Anoetus deutonymphs are often associated with an "acarinarium", a specialized area on the bee host's propodeum or first metasomal tergum, which serves as an attachment site. Eickwort (1979) regarded the acarinarium as evidence for a mutualistic association, a conclusion also based on his observations that feeding stage Anoetus mites feeding on the surface of the bee's larval provisions may remove potentially harmful microorganisms.

    Other obligate associations between histiostomatid mites and bees include Glyphanoetus nomiensis with alkali bees (Cross, 1968; Cross & Bohart, 1969) and Histiostoma inquilinum with squash bees (Peponapis) (Cross, 1968). The life cycles and the nature of the associations of these species with their bee hosts are unknown.

            b. Family Winterschmidtiidae (=Saproglyphidae). Species in the genus Vidia are obligate associates of leaf-cutter bees of the genus Megachile (Megachilidae). Although the PD has recovered this genus on these hosts from all continents, only five species have been described (3 from North America) and only one life cycle is known (OConnor & Eickwort, 1988). Eickwort's observations suggested that these mites are probably scavengers/fungivores on the leaf pieces that make up the nest cell linings. Unpublished observations suggest a lack of host specificity which may be mediated by the fact that Vidia mites are often found phoretic on parasitic bees of the genus Coelioxys, obligate, but not species-specific parasites of leafcutting Megachile species.

            c. Family Chaetodactylidae. Species in this family are primarily obligate cleptoparasites. Two genera occur in temperate North America, Chaetodactylus primarily with Megachilidae, and Sennertia with Xylocopini, but some records suggest these mites may occasionally use other hosts (Baker & Delfinado-Baker, 1983). There is only one described species in each genus in North America, but the PD has obtained specimens of many others. This group is becoming particularly important with management of species of Osmia and Megachile. Chaetodactylus species may destroy entire nests of Osmia and Anthidium bees (Fain, 1966; Maeta, 1978). This seemingly "imprudent" parasitic behavior is countered by the ability of these mites to form two types of deutonymphs, one with the typical phoretic morphology, and the other highly reduced and cyst-like. These non-phoretic deutonymphs are highly resistant to environmental change and remain in the nest cavity to infest a new generation of cells when the cavity is re-used. With management of Osmia species entailing mass rearing in artificial soda-straw nests, these mites have the potential of overrunning such managed systems. Studies in Europe (Bosch, 1992) and anecdotal evidence from managers using these "new pollinators" in North American orchards suggest that these problems may already be occurring (see Collaborative Agreement with Dr. Evan Sugden).

    Species of Chaetodactylus exhibit some degree of host specificity, but serial use of nest cavities among different bee species may lead to host-shifting. The PD has observed almost identical mites associated with species of Osmia and Hoplitis that serially used the same nest cavity in New York.

The PD has revised the genera of this family, providing a phylogenetic framework for species-level taxonomic work (OConnor, 1993b). Aside from the numerous new species that need to be described, the taxonomic status of the one nominal North American species, C. krombeini, is uncertain. Fain (1981) suggested that this name is probably a synonym of C. claviger, described from Europe. The PD has attempted to re-collect specimens of C. claviger from museum specimens of its type-host without success, suggesting that the original collection of this species may have resulted from an accidental association.

    The second North American genus of Chaetodactylidae, Sennertia, is associated with Xylocopini. The only published biological observations on the one described North American species, S. americana from Xylocopa virginica, are equivocal regarding the nature of the association. No live host larvae were recovered from cells infested with this species, however, the cells also harbored the acarid mite, Horstia virginica, a known cleptoparasite (Lombert et al., 1987). The PD has obtained specimens of a number of undescribed Sennertia species from southern and western Xylocopa species, and from Ceratina species in Mexico. These collections suggest a high degree of host specificity among mites of this genus.

            d. Family Gaudiellidae. This family comprises five genera associated with social bees, four with tropical meliponines and one, Cerophagus, with Bombus. The PD described the first, albeit partial life cycle in describing the one North American species, C. nearcticus, but comparison with specimens from Europe are needed to verify the specific status of this species (OConnor, 1992). The two nominal species are both known from more than one species of Bombus. There have been no observations on the biology of these mites.

            e. Family Suidasiidae. One genus, Tortonia, in this relatively small family comprises obligate associates of Hymenoptera. In North America, one species of Tortonia is described, T. quadridens, which is a relatively non-specific cleptoparasite of both wasps and bees (Krombein, 1962a; Lombert et al, 1987) The PD has obtained several undescribed species of Tortonia from Megachilidae, the primary host group for this genus in other parts of the world, including a species from the alfalfa leafcutter bee, Megachile rotundata, an economically important pollinator. This undescribed species, for which we have the entire life cycle, is a cleptoparasite, capable of destroying entire nest aggregations of this bee.

            f. Family Acaridae. This large family includes several discrete lineages with obligate bee associations. The most diverse lineage is the subfamily Horstiinae, in which all species are obligate bee-associates. The PD has recently published a phylogenetic analysis of relationships among the genera of this group (OConnor, 2001). North American genera belonging to the Horstiinae include Horstia, species of which are cleptoparasites of Xylocopa. Only one species is described in North America, H. virginica, a cleptoparasite of Xylocopa virginica (Krombein, 1962a), but the PD has recovered several new species from other Xylocopa species. The genus Medeus includes one species described as a probable cleptoparasite of Anthophora (OConnor, 1997), and undescribed species have also been recovered from Diadasia species. The PD described the new genus Diadasiopus from deutonymphs associated with Diadasia species (OConnor, 1997). Newly collected material includes the undescribed adults and additional new species of this genus from both Diadasia and Anthophora species. The genus Sennertionyx, which is associated with the megachilid tribe Anthidiini, has not been previously recorded from North America, however, the PD has records of this genus from many North American anthidiines; nest-inhabiting stages are unknown as is the nature of the association with the bee host. Another genus with no published North American records is Cerophagopsis, species of which are primarily associated with resin collecting species of Megachile (formerly Chalicodoma), family Megachilidae. Surveys of collections to date suggest this genus is largely restricted to these bees, but there are literature records from social bee nests in other geographic regions (Fain & Heard, 1987), and the PD has a collection from an Apis mellifera colony from Thailand. These records suggest that megachilid bees raiding apid nests for wax or resin may leave mites behind. There have been no observations on the nature of the association between Cerophagopsis mites and their bee hosts.

    The second lineage of acarid mites including bee-associates is the subfamily Tyrophaginae, most genera of which are associates of social insects, primarily ants and termites. Within this lineage, the genus Kuzinia comprises obligate Bombus associates. Only two species are described from North America, both only from the phoretic deutonymph (Delfinado & Baker, 1976). The PD has many collections of this genus from many species of Bombus in North America. Based on observations of the European K. levis, these mites are generalized scavengers in the nest, feeding on pollen, honey, fungi and other detritus (Chmielewski, 1969). Another genus possibly belonging to this lineage is Schulzea. No species are described from North America, but the PD has obtained specimens of one species from malaise trap residues in Michigan. Associations of Old World species include Halictidae as well as other Hymenoptera (OConnor, 1988).

    The third lineage of Acaridae including bee associates is the subfamily Rhizoglyphinae, a very large group of fairly generalized mites inhabiting moist habitats and showing little specificity to particular insect carriers. Some species of Sancassania (=Caloglyphus) are known from bee associations, both in North America and elsewhere. Cross (1968) described S. boharti from all stages found in nest aggregations of Acunomia melanderi. As is common with other Sancassania species that are generalist scavengers, these mites were observed feeding on dead bee larvae. Eickwort (1979) reported unidentified Sancassania deutonymphs phoretic on other species of Halictidae.

        2. Suborder Trombidiformes, infraorder Heterostigmata. The Heterostigmata is a very diverse lineage of trombidiform mites that in many ways converges with the Astigmata in habitat preferences and dispersal mode. Many of these mites specialize in patchy habitats, including the nests of insects and vertebrates, and disperse by phoresy on the adult insects. Unlike the Astigmata, where the dispersing stage is the deutonymph, heterostigmatid mites disperse as adult females. In some groups, two distinct female morphs are formed, one that does not disperse, and another, termed "phoretomorph" which has modified forelegs for attaching to a host (Cross & Moser, 1975). In many heterostigmatid families, species have adopted more widespread habitat preferences and do not disperse via phoresy. Most heterostigmatid mites have a very abbreviated life cycle, consisting only of egg-larva-adult. In some groups, such as Pyemotidae, even the larval stage is suppressed. Obligate parasitism of insect hosts has evolved in a number of heterostigmatid lineages (Kaliszewski et al., 1995).

            a. Family Tarsonemidae. This very large, ecologically diverse family includes the economically important Apis-parasitizing genus Acarapis. Other tarsonemid genera with specific bee associations occur in other geographic regions such as some Tarsonemus with Xylocopa (Magowski, 1986), Pseudacarapis with Apis in India (Sumangala, 1999), and Crossacarapis with Euglossa in the neotropics (Ochoa & OConnor, 1996). Species of Acarapis are permanent parasites of adult honey bees, either in the tracheal system (A. woodi) or externally on the body (other species) (Delfinado & Baker, 1982). Pseudacarapis indoapis feeds on provisions in the nest of Apis cerana and was regarded as a cleptoparasite (Sumangala, 1999). The ecology of the other bee-associated Tarsonemidae is unknown, although the presence of only adult females on bees in genera other than Acarapis suggests a phoretic relationship, with the feeding taking place in the nest cells.

            b. Family Podapolipidae. This large family comprises exclusively parasitic mites of a variety of insect groups, primarily Coleoptera. One species, Locustacarus buchneri, is a tracheal endoparasite of several Bombus species (Husband & Sinha, 1970).

            c. Family Trochometridiidae. The single known species in this family, Trochometridium. tribulatum, is a polyxenous cleptoparasite of ground nesting bees. It has an unusual relationship with the bee hosts in that the female mites disperse via phoresy on adult bees from nesting aggregations. The dispersing mites carry fungal spores in structures termed "sporothecae," and upon entering a new host cell, they kill the bee egg and inoculate the provisions with the spores. The new generation of mites develops upon the fungal garden thus produced, with second generation females dispersing through the soil to seek new bee hosts (Cross & Bohart, 1978; Lindquist, 1985).

            d. Family Scutacaridae. This is another very large family with many ecological associations. The genera Imperipes and Scutacarus both contain species of fungivores in bee associations. In North America, I. apicola is known to be polyxenous, being found on both solitary and social bees. In Europe, the polyxenous bee-associate, Scutacarus acarorum, disperses in flowers by hyperphoresy on the parasitid mite, Parasitellus fucorum (Schwarz & Huck, 1977). All scutacarid mites whose feeding behavior has been observed feed on fungi, and mites may introduce fungi into bee provisions in the manner of Trochometridium.

            e. Family Pygmephoridae. In this family of primarily beetle phoretics, the genus Parapygmephorus comprises obligate bee associates. Based upon observations on the Costa Rican species, P. costaricanus, these mites probably develop on waste material in the cells of the host bees (Rack & Eickwort, 1979) and are not likely to be cleptoparasites. Few species have been described in North America (Cross, 1965), but the PD has obtained many more from a variety of bee families. The level of host specificity is unknown in this group, since all described species are known only from the original collections.

            f. Family Pyemotidae. Species in the single genus, Pyemotes, are obligate, polyxenous, larval/pupal parasites of many insect groups, including bees, with some species also able to feed on adult insects. Pyemotes females have a toxic venom that immobilizes insect prey. Once attached, female Pyemotes swell enormously ("physogastry"), and produce fully formed adult male and female offspring. In the nests of bees, young females move freely through cells and are capable of destroying entire colonies (Krombein, 1967).

    The rationale behind the research project proposed here is straightforward. The role of parasitic mites in the drastic decline of honey bee populations has been documented above along with the two processes that allowed the particular pests to colonize these bee hosts. Varroa destructor was originally a relatively non-pathogenic parasite of Asian Apis cerana, then colonized Apis mellifera when the latter was introduced into the range of A. cerana. On the new host, V. destructor is highly pathogenic. The pest was then moved into other geographic regions with its new host and entered the United States as a classic "invasive species." The two necessary processes in the making of this pest were the host shift, followed by inadvertent introduction into new areas. The pest status of the second honey bee mite, Acarapis woodi, in North America, is mainly the result of simple introduction. This mite appears to be a natural parasite of Apis mellifera, but the mite was not known in North America until recently, suggesting that early introductions of honey bees into North America did not also introduce the mite. The fact that North American honey bee populations show greater damage due to this parasite now that it is here may be due to the fact that these bee populations had been free of the parasite for a large number of generations and had lost some native resistance.

    With the loss of so many honey bee colonies due to these pest mites, the previous utility of this species in crop pollination was compromised. This has led to the development of the "new pollinators" as noted above. These new pollinators also have potentially serious mite pests, and their effects may be exacerbated by the double threat of host switching and geographic spread due to introductions. Knowledge of the associations between native bees and their mites remains woefully incomplete. Many of the mites remain undescribed, most life cycles are unknown, and even the nature of many of the associations, whether damaging, neutral or even beneficial, remains poorly understood. This project seeks to remedy this situation through a serious survey of the diversity of bee-mite associations among native and introduced North American bee species.

    The primary method to be used in documenting the bee-mite associations involves examining existing museum collections of both bees and mites. Fortunately, all of the taxa of obligate bee associated mites noted above have at least one life stage which is spent on the adult bee host. These mites typically remain dried on the host following collection and preservation in a museum. Thus, the simplest method for obtaining associational data is to examine museum collections. As part of our previous project revising the genera of the family Acaridae funded through NSF's PEET program, the PD and trainees in that program have visited many of the major bee collections in the United States, examined specimens for the presence of mites, borrowed the infested specimens and removed and prepared acarid mite specimens. Collections and relevant parts already surveyed include some of the major bee collections in the country: Cornell University, Ithaca, NY (all long-tongued bees except Bombus), University of Kansas, Lawrence, KS (all bees except Bombus and Andrenidae), USDA Bee Biology and Systematics Laboratory, Logan UT (all long tongued bees except Bombus), Michigan State University (all bees), University of California, Davis, CA (all long tongued bees except Bombus), University of Michigan, Ann Arbor, MI (all bees). We have also examined a few selected genera of bees at the US National Museum of Natural History, Washington, DC, the Illinois Natural History Survey, Urbana, IL, the California Academy of Science, San Francisco, CA, and the University of California, Berkeley, CA.

    A second source for bee associated mites which has yielded specimens of non-phoretic stages are collections made by individuals who worked specifically with bee nests. The PD has on hand collections of bee associated mites and copies of associated field notebooks obtained from the late George Eickwort (Cornell University), notably correlated series of life-stages of a number of species in the families Chaetodactylidae, Histiostomatidae and Acaridae. Eickwort's notebooks contain unpublished observations on a number of mite associations with bee nests. Additional material has been obtained from the collection of the late Earl Cross, now housed at the Florida State Collection of Arthropods in Gainesville, and from Dr. Robert Husband, who has deposited his collection of Bombus and associated mites in the University of Michigan Museum of Zoology. Bee workers in a number of states have also sent numerous other collections of bee-associated mites to the PD over the past few years.

    In order to complete the survey of North American taxa, we propose to re-visit several of the collections indicated above to examine taxa not previously surveyed. These collections include the University of California collections at Davis and Berkeley. The Davis collection includes large series of Bombus from the research of Dr. Robin Thorpe, while the Berkeley collection is particularly strong in Xylocopini from the research of Dr. Howell Daly (both researchers now retired). We also propose to re-visit the Illinois Natural History Survey, a collection strong in Andrenidae from the work of retired curator, Dr. Wallace LeBerge. We will also visit three large collections which we have not previously surveyed, those of the U.S. National Museum of Natural History in Washington, D.C. (the largest bee collection in the country, and notable for collections obtained by former curators, Drs. Paul Hurd and Ronald McGinley), the American Museum of Natural History in New York, housing collections obtained by current curator Dr. Jerome Rozen, and the University of California, Riverside, housing the collections of Dr. P. H. Timberlake. While in Washington, we will also spend time at the mite collection of the U.S. National Museum, housed at the Systematic Entomology Laboratory in Beltsville, MD. This collection houses material obtained by Drs. Baker and Delfinado-Baker which remained unstudied at the time of their retirements, as well as type specimens of species described by those and other researchers. The PD has already conducted a brief survey of this material, much of which remains scattered, unprepared, and generally in need of curation. Drs. Ronald Ochoa and Douglass Miller of the SEL have agreed to assist with this aspect of the project (see Collaborative Arrangements).

    In addition to surveying museum collections, we will also continue to cooperate with bee and mite workers across the country, a number of whom have sent specimens to the PI for identification in the past. These include Drs. Robin Thorpe, University of California, Davis, Evan Sugden, University of Washington, Pullman, Michael Engel, University of Kansas, Lawrence, Robert Husband, Adrian College, MI, and Ms. Virginia Scott, Colorado State University, Boulder. An especially important collaboration has been developed with the USDA Bee Biology and Systematics Laboratory in Logan, UT. The research staff of this lab, including collaborators Drs. James Cane and Terry Griswold, have agreed to pay special attention to mites encountered in their large scale survey of nests of Western North American bees. This collaboration is so important that we propose to hire an undergraduate student at Utah State University to specifically examine both freshly collected material and museum specimens not previously examined by us (see Collaborative Arrangements).

    Following discovery of mite-infested bee specimens, in the laboratory, the bees are individually labeled with voucher labels reading "Mites removed, B.M. OConnor" followed by a unique number. This number along with the identification of the bee, all collection data, and the source collection for the specimen are entered into a database maintained by the PD. This database already contains over 1500 records of mite-infested bees examined by the PD throughout his career. Bees with mites are photographed prior to mite removal to document the attachment site, mode of attachment, and density of mite populations. In the past, we have used a still camera mounted on a stereo microscope for this purpose, but the need for high magnification in order to see the mites clearly means very short working distances and poor resolution of the larger field of view. We propose to remedy this by purchasing the Auto-Montage system from Syncroscopy. This system uses a high resolution digital camera coupled with automated mechanical stage shifting to take hundreds of photos, each a different "slice" through the depth of the field. Software in this system then combines the images into a single, high resolution photo in which the entire three-dimensional view is in focus. This system is in use in a number of entomological laboratories, with beautiful and amazingly detailed photos now becoming fixtures on internet sites.

Following examination of the bee host, a series of mite specimens (not all) will be removed from each bee, with the location of each mite species on the host recorded in the database. Mite specimens are cleared in lactophenol, mounted in Hoyer's medium, dried and sealed with Red Insulating Paint (GC Electronics, Inc.). Slide labels with host identification, collection information, source collection, and voucher number are generated from the database. The unique voucher number permanently associates each mite specimen with its original host specimen. (Such vouchering was practiced by only two prior workers on this fauna, George Eickwort and Robert Husband. Unfortunately, other workers, notably Baker and Delfinado-Baker, did not voucher host specimens, and sometimes described mites from undetermined hosts. Host associations for these species must be verified with new material.) Mite specimens will then be sorted, identified, and new species grouped for taxonomic description. Illustrations of all available life stages will be prepared for each species. Photographs of the mites themselves will also be taken using the Auto-Montage system mounted on a compound microscope equipped with differential interference contrast optics. Keys for identification of all known life stages will be developed. A catalog of host bee species and geographic ranges for each species will also be developed, providing the necessary baseline data on the native fauna of bee associated mites prior to planned or accidental introductions of new bee species (and potential associated mites) into the region. At the end of the study, bee specimens and series of mite specimens will be returned to the source institutions following the loan agreements that we have worked out in advance with each institution. Specimens of all mite species will be retained in the Museum of Zoology, University of Michigan, and also deposited in the collection of the U.S. National Museum in Beltsville, MD, as permanent repositories.

    Information will be disseminated in three ways. Taxonomic works will be published in peer-reviewed journals. These papers will include full descriptions of each taxon, keys to North American species, and biological information when available. The format will be similar to that used in the study of Vidia mites (OConnor & Eickwort, 1988 - copy attachedas Appendix I). We will also utilize developing technology to make this, and other information available over the internet. We will develop pictorial keys, incorporating our Auto-Montage photographs into the LUCID system. These will allow bee biologists or commercial interests to identify mite species using a type of expert system built into the LUCID system. LUCID has the advantage over traditional dichotomous keys in that the user is presented with all available character information in both textual and visual formats, and can choose the state of the most easily visible characters in their specimen. Through reference to the library of photographs, the user can see each character state on a photograph of an actual specimen, allowing totally inexperienced people to accurately identify species. Several works on different mite groups (as well as many other groups of invertebrate animals) using the LUCID system have been published either on the internet or on CD-ROM (e.g. Hunt, Norton et al., 1998; Hunt, Colloff, et al., 1998; Walter & Proctor, 2001).

    Finally, the accumulated taxonomic information will be used at a later date to conduct phylogenetic analyses of the various mite taxa. These analyses can be coupled with complementary studies on the phylogenies of the bee hosts to develop and test hypotheses regarding the evolution and historical association of these systems. The PD has utilized these methods in the past to document ancient host shifts among bee associated mite lineages (OConnor, 1988) and to study the evolution of acarinaria on hymenopteran hosts (OConnor & Klompen, 1999). These future studies are noted here as potential value-added products from the data generated during the research proposed here. These studies do not form part of the research to be funded by this project, however, such higher order studies are totally dependent upon basic taxonomic and associational data that will be generated during the present project. This project has as its most basic goal, the documentation of the bee-mite associations to assist managers in dealing with potential and actual pest problems in agricultural pollination systems.

    The schedule proposed here should allow us to complete the survey, prepare the taxonomic works and the photographs necessary to implement the LUCID system within the three year time frame. We will conduct our museum surveys in the first two years, with the largest and so far unsurveyed collections of the U.S. National Museum of Natural History and the American Museum of Natural History being conducted in the first year. In the second year, survey of the smaller collections will be completed. In the third year, we will concentrate on dissemination of information using the database and photographic library we will have developed. In order to accomplish these goals in a timely manner, Dr. Pavel Klimov will be hired as a postdoctoral research associate. Dr. Klimov has worked with the PD on his NSF-PEET project revising the genera of the Acaridae (funding for which ends in June 2002), and he has excellent knowledge of all of the taxonomic groups to be covered in this project, particularly with bee associated mites in the Russian Far East. He has excellent qualifications and experience in preparing species descriptions, scientific illustration, database management, and has developed computer-based identification systems. Dr. Klimov's assistance is integral to this project, as the PD is also involved with teaching, curation, and administrative service at the University of Michigan, and in other research projects.