This paper: 1) reviews the biology of the honey ant M. mendax relevant to husbandry and educational interpretation, 2) summarizes live field-collection and captive management techniques for this species, and 3) describes a naturalistic display nest for M. mendax that is part of a larger exhibit comparing honey ants and honey bees at the CZBG Insectarium.

Introduction:
      The North American genus Myrmecocystus is one of five genera worldwide containing honey ants (Table 1). Honey ants develop specialized workers, called repletes or honey pots, that store nectar within balloon-like abdomens swollen to a centimeter or more in diameter (Fig’s. 1 & 4). When a colony needs food, the repletes regurgitate their reserves to their normally-proportioned sisters. All live in dry habitats and have polymorphic or variably-sized workers. In most polymorphic ants, the larger workers serve as colony defenders, but in honey ants they become repletes. These genera are not directly related to one another (Table 1) and thus are evolutionarily convergent in terms of their food-storing behavior (Conway 1981b).
      Myrmecocystus stands out because of its species’ richness and variability
(Table 2). It is a fascinating example of adaptive radiation and diversification, embracing nocturnal, crepuscular and diurnal species living in a wide variety of hot to somewhat cooler deserts and semi-arid habitats (Snelling 1976). These ants are as much an expression of western North America’s biologically-diverse deserts and ‘Sky Islands’ as Darwin’s finches are to the Galapagos Islands.
     This variation in Myrmecocystus is relevant to husbandry. For example: What is the best species to rear and display? Our first choice was M. mexicanus (Morgan 1991), while most of the 29 valid USDA permits for interstate transport of honey ants are also for M. mexicanus, followed by M. mimicus, depilis and melliger (Wehling pers. comm.). M. mexicanus is the ‘classic’ honey ant. It is a very large, pale-yellow, nocturnal species with spectacular golden repletes studied and popularized well over a century ago by McCook (1882). Even so, very few institutions have successfully kept or displayed M. mexicanus. Of note is a captive colony now temporarily on show at the California Academy of Sciences. This vigorous colony was started by Steve Prchal and was recently moved from Tucson by car. The M. mexicanus colony is doing well on display and is very popular with the viewing public (Evans, pers. comm.).
      Experience has shown that diurnal species such as M. mendax (Table 2) generally thrive better in captivity than M. mexicanus (Morgan pers. obs.; Prchal pers. comm.; Mendez pers. comm.). M. mendax in particular, given its hardiness, adaptability and relatively large size, may be the ideal honey ant for beginning keepers. It has been successfully exhibited for over five years at the Smithsonian’s Orkin Insect Zoo (Erwin pers. comm.) and is now on public display in a new naturalistic nest at the CZBG Insectarium. M. mendax will also be on view later this year at the Houston Zoo and Oakland Children’s Zoo Bug Building (Mendez pers. comm.).
General Methodology:
      In 1985-1986, we carefully excavated three large M. mexicanus nests (Morgan 1991), initially guided by John Conway who had worked with this species in Colorado (Conway 1983a). For comparative study, we also partially unearthed a nest of M. mimicus, which was sympatric with M. mexicanus at our field site. Field-work took place in southeastern Arizona near the Chiricahua Mountains in low elevation, semi-desert grassland (Brown 1982). Our original goal was to transplant established M. mexicanus colonies into laboratory observation nests at the CZBG Insectarium for study, husbandry trials and eventual public display. However, transplanted colonies declined slowly, perhaps partially due to continued problems with rupturing repletes. Even so, we also collected newly mated queens from our field site and induced these to initiate nests in captivity. During subsequent years these grew into populous M. mexicanus colonies (Morgan 1991, 1995).
      We also began working concomitantly with M. mendax, a large, reddish-orange and black, diurnal species found at moderate elevations within the nearby Chiricahua Mountains. Captive colonies were started with wild-caught queens (Morgan 1991), though one established nest was quickly excavated and studied. M. mendax proved to be hardier and more tractable in captivity than M. mexicanus and we eventually focused on it exclusively. Some of our field and laboratory observations are included in the following biological synopsis. Experimentation with captive queens and colonies at the CZBG Insectarium led to the husbandry procedures and public display colony for M. mendax described below.

BIOLOGY OF THE HONEY ANT, M. MENDAX

Taxonomic Considerations:
      The genus Myrmecocystus belongs to the evolutionarily advanced subfamily Formicinae. Ants in this taxon are characterized by their ability to spray concentrated formic acid via a terminal acidopore (Hölldobler and Wilson 1990). The caustic spray is used both defensively to repel enemies and offensively to help incapacitate prey. Myrmecocystus is distinguished from other formicines by its unusually long labial palps (finger-like mouthparts) and psammophore. The psammophore is a basket-like fringe of long setae or ‘hair’ on the underside of the head used by a few desert-dwelling ants for carrying sand grains.
      Myrmecocystus contains three subgenera, 28 described species (Table 2, Snelling 1976) and at least one new species (Cover & Johnson 2002). M. mendax belongs to the subgenus Endiodoctes and, like its sister species, is relatively large, diurnally active, darkly pigmented, and has small compound eyes and well-developed ocelli or simple eyes (Table 2).
     M. mendax is geographically variable in both color and form. It is bi-colored, with either a yellowish or reddish-orange head, transitional thorax, and dark abdomen. It also has a northern long-haired and a southern short-haired form that intergrade across its range (Snelling 1976).

Range and Habitat:
      Myrmecocystus occurs only in the western United States and Mexico (Snelling 1976). The genus extensively blankets arid regions extending from southern California to Texas, and from southern Washington State to south central Mexico. Its species live in a variety of habitats ranging from extremely hot, dry, lowland deserts and semi-deserts to somewhat cooler, moister mountainous regions. Many species range widely and associate with a variety of plant and animal communities at characteristic elevations and latitudes. M. mendax ranges from central Colorado to northern Mexico, and southern California to Texas (Creighton 1950; Snelling 1976). It lives in various semi-arid woodlands, savannahs and short-grass prairies (Snelling 1976) at elevations between 3600-6000 ft. in Colorado (Gregg 1963) and 3900-6500 ft. in Nevada (Wheeler & Wheeler 1986). In the Chiricahua Mountains, M. mendax occurs around 5500 ft. in Oak-Juniper Woodland situated on rocky soil (Morgan pers. obs.). Here, summer days are normally hot to warm, nights sometimes cool, and winters chilly to cold with occasional freezing temperatures. Some creeks and waterways are relatively permanent, and others flow for weeks or months once the rainy season begins.

Nest Site and Architecture:
      Myrmecocystus nests are constructed in substrates ranging from compacted sand to rocky soil and usually have a single opening to the surface (Wheeler & Wheeler 1973; Snelling 1976). The nest opening is frequently surrounded by a small tumulus or crater of either coarse sand or gravel, and its appearance is often species specific (Snelling 1976). Nests may extend several meters or more below the surface with the majority of repletes typically in the lower chambers, deep enough to be within the level of permanent soil moisture (Snelling 1976). M. mexicanus and mendax nest architecture is similar (Conway 2003, 1983a).
      M. mendax may nest beneath stones and nest entrances may or may not be surrounded by crateriform tumuli (Gregg 1963). One nest consisted of a shallow, subsurface labyrinth of tunnels and small chambers, and a single main vertical passage extending to about a meter below ground. Side tunnels radiated outward from the main passage to 36 larger domed chambers (Conway 2003). Another nest was also about a meter deep (Morgan pers. obs.). In our naturalistic display nest (Fig. 2), the subsurface labyrinth is used by the ants mainly to incubate pupae, being warmed from behind with a heating pad (Fig. 3) as if by the sun.

Mating swarms and nest initiation:
      The nuptial flights of most Myrmecocystus species follow rain and are crepuscular or nocturnal (Wheeler 1908, 1917; Snelling 1976; Conway 1980a). The males die shortly after mating and the newly fertilized females shed their wings and quickly tunnel into the ground. Some mornings after, mated queens can still be found running across the surface in search of nest sites or actively digging in (Morgan pers. obs). But only those burrowed deep enough in moist soil to escape the blazing mid-day sun and baked surface survive. Swarming and nest-founding queens usually experience high mortality, much of it due to predation by a wide variety of vertebrates and invertebrates. In one study,
M. mexicanus queens suffered a 98% mortality rate (Chew 1987). Depending on the species, queens either initiate nests alone (haplometrosis) or co-operate in small groups (pleometrosis); in the later case, all but one are usually eliminated by the colony (Snelling 1976). M. mendax appears haplometrotic based on limited field studies (Morgan pers. obs.).

Colony development:
      The queens of most highly-evolved ants rear their first brood in isolation, relying on their abdominal fat reserves and metabolized flight muscles for larval food (Wilson 1971). Once Myrmecocystus queens sequester themselves in small earthen chambers, initial colony development proceeds quickly (Morgan 1991). Most begin laying eggs within a day and have a small group of mature larvae within several weeks. As larvae prepare to pupate, queens surrounded them with a supporting framework of sand or fine debris, a behavior characteristic of ants with cocoon-spinning larvae (Wheeler 1910). The first few workers, all tiny, emerge about a month after nest initiation. For many ants, the first workers are unusually small, favoring rapid population growth (Wilson 1971), which increases the probability of colony success. In young captive colonies, these workers began foraging for nectar and prey within a few days (Morgan pers. obs.).
      Most ant colonies grow for at least several years before reaching maturity, that is, becoming large enough to produce numerous winged males and virgin queens (Wilson 1971). Captive M. mendax colonies mature in about five years (Morgan pers. obs., Mendez pers. comm.).
     Once colonies are well-established they are exceptionally long lived, with estimated average life spans for natural M. mexicanus colonies ranging from seven to 39 years (Chew 1987). A captive M. mexicanus colony started at the CZBG Insectarium and now thriving in Germany (Hölldobler pers. comm.) is about 15 years old. An established wild colony of M. mendax is known to be at least 26 years old (Mendez pers. comm.).

Social Organization:
      Mature Myrmecocystus nests typically contain a single, mated, wingless queen (Fig. 1) and several thousand or more workers (McCook 1882; Creighton & Crandall 1954; Snelling 1976). However, a large M. mimicus nest held three wingless queens in the same lower chamber (Weissmann pers. comm.). M. mexicanus nests contained about 5000 workers (Conway 1983a) and a nest of M. mendax nearly 2000 workers (Conway 2003). The workers are polymorphic or size-specialized (Snelling 1976) and together are responsible for nest construction and colony defense, care of the queen and brood (eggs, larvae and pupae), foraging and long-term food storage (Fig. 1). Depending on the season and colony cycle, nests may also contain hundreds of winged reproductive forms or males and virgin queens.

Food and Foraging Behavior:
      Myrmecocystus species are generalized predator-scavengers feeding on freshly killed and recently dead arthropods, nectar from desert plants, and homopteran honeydew (McCook 1882; Wheeler 1908; Snelling 1976; Conway 1980b; Hölldobler 1981). The ants carry liquid food in their crops and small prey with their mandibles, and groups of workers co-operate in retrieving large items. Diurnal species tend to be somewhat more insectivorous and less nectivorous than nocturnal species (Snelling 1976). Foraging is limited by both high and low temperatures (Kay & Whitford 1978). Foraging by the diurnal M. mimicus involves elaborate ritualized tournaments and raids between competing colonies (Hölldobler & Lumsden 1980; Hölldobler 1981; Lumsden & Hölldobler 1983).
      M. mendax forages during the day (Snelling 1976). Natural prey items tend to be tiny and no larger than small termites (Mendez pers. comm.). However, M. mendax has been seen retrieving caterpillars up to 25 mm. or one-inch long (Snelling 1976). Prey items carried or dragged by foragers to the nest opening in our naturalistic display are quickly transported or dropped down the main vertical passage to the lowest nest chambers (Morgan pers. obs.).

Repletes:
      The development of repletes is a distinctive feature of Myrmecocystus biology (McCook 1882; Wheeler 1908; Snelling 1976). Repletes can form from any soft, newly emerged worker (Snelling 1976), though they typically develop from the largest workers (Rissing 1984). Besides nectar, repletes also store lipids (Burgett & Young 1974), water, and body fluid from insect prey (Snelling 1976; Conway 1977, 1990). Depending on their contents, repletes may be clear or vary in color from light to very dark amber (Conway 1977, 2003).
     Repletes are fragile, easily ruptured, have difficulty moving, and spend their lives hanging within their nest (Fig’s. 1 & 4). They comprise a significant reserve of food in some colonies (Snelling 1976). At least 1030 repletes were found in a M. mexicanus nest (Conway 1983a), more than 1500 in a M. melliger colony (Creighton & Crandall 1954), and nearly 600 repletes in a M. mendax nest (Conway 2003). The relatively small group of repletes in the M. mendax colony together weighed 130.4 g. or contained about 4.6 oz. of stored liquid food.
      Captive study of M. mendax suggests that developing repletes must somehow be pressurized to maintain and perhaps create their balloon-like shape (Morgan pers. obs.). Some young repletes have bulbous abdomens half-filled with nectar and a large bubble uppermost (Fig. 4). These gas pockets are eventually replaced with liquid. A positive internal pressure seems necessary for the flexible membrane to remain inflated. How this pressure is created or regulated is not clear. Perhaps gas is generated from sugars fermented by micro-organisms inhabiting the alimentary canal. Bubbles also sometimes occur in M. mexicanus repletes (Conway 1984, Morgan pers. obs.). Repletes may hang in order to gravity-feed liquid into their gas-inflated abdomens while allowing excess or displaced gas to escape upwards through the mouthparts.
     Large M. mendax repletes are not completely immobile (Morgan pers. obs.). Occasionally they reposition themselves or move short distances. They sometimes slip and fall, but if they can get a foothold, can crawl and climb back to a hanging position. Fallen repletes that become trapped in awkward positions are soon eliminated. Their stored nectar, and that of any other dead repletes, is not allowed to waste. After the head and thorax are cut away, the nectar-filled abdomens sit on nest chamber floors for a week or more and are slowly emptied by the workers. The persistence of these sack lunches suggests that the nectar in the isolated abdomens is protected in some way from fermentation or microbial decomposition (Morgan pers. obs.).
     While repletes seem bizarre, they are merely an exaggerated expression of basic ant behavior. Many ants collect and temporarily carry liquid food within their abdominal crop, a flexible sack within the alimentary canal. The food is later shared with nest mates by oral trophallaxis or regurgitation (Wilson 1971). As their crops fill or empty, flexible inter-segmental membranes allow their abdomens to correspondingly swell or contract. In repletes, these inter-segmental membranes continue to stretch until they become highly distended and balloon-like. The original cuticular segments float atop the membranous orbs like a chain of small islands (Fig.1).

Myrmecophiles:
      Myrmecocystus nests are sometimes co-habited by other insects and mites (Snelling 1976). A M. mendax nest held an ant-loving cricket Myrmecophilia sp., a staphylinid beetle larva, collembolans, ants Pheidole longula, and mites Gymnolaelaps sp. (Conway 2003).

Cultural entomology:
      Early Indians of the American Southwest and Mexicans dug up honey ant nests, collected and ate the repletes, which they considered delicacies (Snelling 1976; Conway 1984; Vander Wall 1986). The repletes’ fore-bodies were held pinched with the fingers and the engorged abdomens nipped off with the teeth. The Indians sometimes pressed nectar from the repletes to use as a condiment with other food. Replete nectar was thought to have medicinal properties when applied to bruises or swollen tissue, and was also used to prepare an alcoholic drink. Today, repletes are sometimes eaten as novelties at insect fairs, and live colonies are increasingly being used as public educational displays.

ACQUISITION OF M. MENDAX LIVESTOCK (MATED QUEENS)
      Acquiring Myrmecocystus livestock for captive study or public exhibit ultimately involves some form of field-collection (Morgan 1991). Most ants produce mated queens only under a precise set of environmental conditions that are very difficult or impossible to duplicate in the laboratory (Hölldobler & Wilson 1990).
     Two very different approaches have been used to establish captive Myrmecocystus colonies:
1) field-collecting newly mated queens for ‘natural’ colony initiation in the laboratory, and
2) field-excavation of established nests and transplanting large colonies into laboratory nests (Morgan 1991). These approaches are directed at different stages of the colony cycle.
     Field-collecting queens is vastly simpler than nest excavation. The principal difficulty is the need to be in the field shortly following the nuptial flight. This typically occurs only once a year at any given location and being weather dependent is difficult to predict. Fortunately, several of our early expeditions to excavate M. mexicanus colonies coincided with the first monsoon rains (Morgan 1991). We were able to observe natural nest initiation and field-collect many newly mated queens, including those of M. mendax.
     Wingless thus probably mated queens found running over the ground surface were simply caught in small plastic vials. As other queens burrowed into the ground, they dug small holes that were soon encircled by a 4-5 cm. diameter ring of excavated soil bits, literally targeting their positions so they were easier to spot and collect. A spade was carefully used to unearth queens that were below ground or had recently sealed their burrows. Within a day this became ineffective as queens moved deeper. Subsequent rainfall removed all signs of nest initiation. Newly collected queens were kept shaded and cool. Those held in vials for more than a few hours were given a damp paper wad or cotton ball for moisture.
     In later years, local field collectors sometimes supplied us with mated queens. Today, queens of several species, including M. mendax, are often seasonally available from arthropod livestock collector-dealers.

LABORATORY COLONY INITIATION

     M. mexicanus colonies initiated in plaster-of-Paris nests declined after emergence of the first workers (Conway 1981a). M. mimicus queens successfully initiated colonies in horizontally positioned test tube nests (Hölldobler 1981; Bartz & Hölldobler 1982). Water trapped in the ends of the tubes with cotton provided moisture, while wooden applicator sticks fastened lengthways allowed repletes to suspend themselves. M. depilis, mimicus, navajo and placodops colonies were started and kept long-term in similar nests (Hölldobler & Wilson 1990).

Nest initiation:
      Our early trials with test-tube nests for M. mexicanus colony initiation were generally unsuccessful. However, placing queens in small plastic vials packed with moist sand allowed them to burrow and form ‘natural’ queen cells. This greatly facilitated laboratory colony initiation (Morgan 1991) and we extended this technique to M. mendax. Myrmecocystus queens seemed more likely to successfully initiate colonies when allowed to sequester themselves within moist sand firmly packed in just about any small jar or container (Prchal; Mendez pers. comm.). These are left undisturbed until the emergence of the first workers.
      Brood and worker augmentation: Providing brood can be used for starting new or bolstering weak colonies, since most ants readily adopt any brood of their species (Morgan pers. obs.). Offering pupae is especially efficient since workers will be produced quickly and with minimal effort by the queen or small colony. For example, a M. mendax colony was created by combining a newly mated queen with emerging pupae collected from a wild nest (Morgan 1991). Within a few weeks this step-family was performing like a one year old colony. Similarly, pupae from a large captive M. mexicanus colony were transferred to aid a smaller struggling colony (Morgan 1991). Pupae can easily be collected from a populous colony by environmental manipulation. A warm spot temporarily formed in an easily accessible location using a small incandescent bulb will often be utilized by the ants for incubating their pupae. The pupae are then simply collected with an aspirator and transferred to the recipient colony.
     Newly emerged or callow workers lacking full pigmentation can also be safely moved between colonies (Morgan pers. obs.). Older, fully-pigmented workers should not be transferred since they have developed a colony odor and will fight with foreign workers or attack the queen.

Queen manipulation:
Starting captive colonies with multiple queens is a potentially useful technique. Some Myrmecocystus species are naturally pleometrotic (Snelling 1976; Hölldobler and Wilson 1990), while others that are not often accept this situation in captivity (Morgan pers. obs.). The advantage of multiple founding queens over solitary queens is that they produce larger first broods and more workers in less time (Hölldobler and Wilson 1990).

OBSERVATION NESTS AND FORAGING ARENAS
      Laboratory and observation nests for ants have been reviewed by Hölldobler & Wilson (1990). Test tube nests or glass-covered nest chambers dug out of plaster-of-Paris were effective for many species. Most colonies behaved normally when exposed to standard indoor lighting as long as they were provided with moisture. Foraging arenas consisted of plastic tubs lined with Fluon (Table 3), petroleum jelly, heavy mineral oil or talcum powder to prevent escape.
     Observation nests for most ants ideally should: 1) be modular or simply disassembled into several, easily-moved, ant-contained sections, 2) permit keepers to manipulate moisture levels and heat sources, if present, 3) allow ready viewing of the entire colony, and 4) be soil-free to prevent ants from covering viewing surfaces (Morgan pers. obs.).
     Several types of observation nests have been used to house Myrmecocystus colonies, showing that the ants are highly adaptable to various nest chamber sizes and shapes as long as basic moisture needs are met. Vertically-aligned plaster chambers were employed by Cazier & Mortenson (1965) and Conway (1980a). Horizontally-oriented test tubes kept moist with damp cotton plugs worked well for four species (Hölldobler 1981; Bartz & Hölldobler 1982; Hölldobler & Wilson 1990). Extra tubes were simply added to accommodate colony growth.
     Several Myrmecocystus species have been kept long-term in small plastic boxes containing molded Hydrostone (Table 3) nest chambers and tunnels (Morgan 1991; Prchal pers. comm.). Cured Hydrostone is porous, wicks water, slowly releases moisture by evaporation, and must occasionally be re-hydrated. Roughened upper nest chamber surfaces or even fine mesh screen serve as perching for repletes. Moisture gradients for colonies are created by keeping some nest chambers damper than others (Morgan 1991).
     Foraging arenas used at the CZBG Insectarium are simply standard glass aquaria containing observation nest chambers. Any silicon aquarium sealant protruding from inner aquarium corners is first removed with a straight-edged razor blade to prevent ants from finding footholds. A 10 cm. wide band of 3-in-One Household Oil (Table 3) is smeared around the upper inner perimeter of the glass to create a slippery barrier and prevent escape. Compared to other ants, Myrmecocystus is not a strong climber and a single application of household oil confines the ants for several weeks or more (Morgan 1991). An aesthetic advantage is that this barrier is nearly invisible. Foraging arenas may be left uncovered to facilitate routine management.

FEEDING AND CLEANING CAPTIVE COLONIES
      Natural and synthetic diets for ants were reviewed by Carney (1970) and Hölldobler & Wilson (1990). Captive Myrmecocystus colonies grew better on a combined diet of insect fragments and honey-water than on either item alone (Snelling 1976). A restricted diet of honey-water seemed sufficient for adult ants, while protein from insect tissue appeared essential for larval development. A captive diet of insect parts and honey-water was also used by Hölldobler (1981), Bartz & Hölldobler (1982), and Lloyd et al. (1989). Conway (1981a, b) fed sugar-water dyed with blue food coloring to captive M. mexicanus foragers; the solution was later seen in workers, larvae, the queen and repletes. At the CZBG Insectarium, honey ant colonies are provided artificial nectar, fresh insect prey, and drinking water (Morgan 1991) as follows.
     Nectar: Artificial nectar for M. mendax and other ants is prepared by mixing 300 ml. purified water, four rounded tbsp. (50-60 g.) granulated table sugar, and a pinch (about 0.05 g.) of both Vanderzant Vitamin Mix and Wesson Salt Mix
(Table 3). This solution is refrigerated and fed out for 1-2 weeks. Only a fraction of this amount is needed to feed several large Myrmecocytus colonies. Film canister lids serve as small dishes. Uneaten nectar usually spoils within a day or so, either becoming too viscous from evaporation or fermenting under humid conditions, and thus is replaced daily (Morgan 1991).

Prey:
      Our staple prey items are wingless fruit flies Drosophila melanogaster, and week-old or second instar European house crickets, Acheta domestica. Various other prey items are offered as available. Workers easily capture small live prey such as flightless fruit flies, tiny crickets and termites. Larger, stronger insects such as partially-grown crickets, grasshoppers and katydids are first incapacitated by freezing or pinching before given to the ants. Freezing is also used to incapacitate small prey offered to young colonies, especially when they first began to forage. Workers have difficulty retrieving prey if mired in body fluid. Prey is provided daily or twice daily for populous colonies, in an amount roughly proportional to the number of larvae since these ultimately are the primary consumers.

Water:
      Drinking water is continuously available in shallow Petri dishes half-filled with course gravel, with the water level maintained near the surface of the gravel. This provides workers with easy access to drinking sites but limits the possibility of them slipping and drowning.

Colony waste and removal:
      Debris piles or garbage dumps, consisting of dry discarded prey items, empty pupal cocoons and dead ants, are usually placed by the workers in one or more back corners of their foraging arena. We remove debris every few weeks or as-needed using a small flat paint brush and note card for a dust pan. If a more thorough cleaning is required, colonies in modular nests are simply transferred to new foraging tanks (Morgan 1991).

NATURALISTIC DISPLAY OF M. MENDAX AT THE CZBG INSECTARIUM
      Our public display of M. mendax (Fig. 2) is part of a larger exhibit that compares and contrasts honey ants and honey bees, Apis mellifera. For example, both are highly social insects that collect and store liquid sugars long-term in very different ways within very different nests.
     The glass exhibit tank (Fig. 2) measures 36 x 27 x 24 in. (91.4 x 68.6 x 61.0 cm.) wide by high and front to back. It protrudes 4 in. (10.2 cm.) from the supporting wall, providing a feeling of dimensionality. The display tank is horizontally divided into an upper foraging arena and lower nesting site. The foraging arena ground surface is sculpted concrete tinted with tan mortar color (Table 3) supported on a reinforced Extren (Table 3) platform. The foraging arena is decorated with rocks, sticks and artificial plants depicting an arid habitat, and illuminated from above with ‘warm’ fluorescent tubes. These more closely mimic natural daylight than standard cool bulbs. The lights are run by timers set on a long (16 hour) day cycle.
     Our display nest mimics natural nest architecture. It has a small surface mound, subsurface labyrinth, and a single main vertical passage leading to larger domed chambers (Fig. 2). The nest chambers are backlit to attractively illuminate the repletes (Fig. 4). The lower back of the tank is open for access to display mechanics (Fig. 3). The nest is modular, consisting of seven removable Hydrostone blocks containing cast tunnels and chambers. The blocks were molded using water-based pottery clay rather than modeling clay, since the latter leaves a slippery oily residue that is difficult to remove (Mendez pers. comm.). Numerous bits and pieces of pumice were molded into the main vertical passage and nest chamber ceilings to provide workers and repletes with secure footing. Several water wells were also cast into each block so that the Hydrostone can easily be re-hydrated as needed with a squirt bottle. The blocks are heavy and rest on a small tier of Extren shelves supported by stainless steel threaded rods and nuts, vertically adjusted to align interconnecting tunnels. A heating pad (Fig. 3; Table 3) affixed behind the subsurface labyrinth creates a warm zone for pupal incubation. The lower front display glass is painted with a rocky dirt-like façade (Fig. 2) hiding internal display mechanics.
     In May 2004, we induced a prolific, five year old M. mendax colony to expand into the display. Their laboratory nest was joined to the lower main passage via a narrow hose (Fig. 3) and the ants began to move in. Now, less than five months later, the colony has more than doubled in size to roughly 900 workers, 230 repletes and numerous brood. The new display at the CZBG Insectarium is a wonderful opportunity for our guests to intimately view the fascinating details of honey ant colony life without having to dig for hours in blistering heat. What a sweet deal!

ACKNOWLEDGEMENTS:
      The Cincinnati Zoo and Botanical Garden supported this work. Zoo volunteers helped excavate colonies and collect queens. Captive management was greatly assisted by CZBG Insectarium keepers Karen Schmidt, Theresa Austing and Winton Ray. John Conway shared his field experience with M. mexicanus and co-led our first expedition to excavate nests. Steve Prchal and Ray Mendez were always ready to discuss ants and husbandry. Barney Tomberlin and Tony Snell collected some of the queens. The United States and Ohio Departments of Agriculture issued interstate transport permits. Milan Busching helped design and build the display observation nest for M. mendax, and Sure Thing Pest Control sponsored the exhibit. Joyce Turner illustrated colony life, Dave Jenike provided display colony photographs, and Bernadette Plair offered helpful manuscript comments.

REFERENCES

Bartz, S.H. & Hölldobler, B. (1982): Colony founding in M. mimicus Wheeler (Hymenoptera: Formicidae) and the evolution of foundress associations. Behav. Ecol. Sociobiol. 10:37-147.
Brown, D.E. (1982): Semi-desert grassland. In Desert plants 4 ( 1-4) Biotic communities of the American southwest-U.S. and Mexico: 123-131. Brown. D. E. (Ed.). Tucson: Univ. of Arizona.
Burgett, D.M. & Young, R. (1974): Lipid storage by honey ant repletes. Ann. ent. Soc. Am. 67: 743-744.
Carney, W.P. (1970): Laboratory maintenance of carpenter ants. Ann. Ent. Soc. Am. 63: 332-334.
Cazier, M.A. & Mortenson, M.A. (1965): Bionomical observations on myrmecophilous beetles of the genus Cremastocheilus (Coleoptera: Scarabaeidae). J. Kans. ent. Soc. 38: 19-44.
Chew, R.M. (1979): Mammalian predation on honey ants, Myrmecocystus (Formicidae). Southwest Nat. 24: 677-682.
—— (1987): Population dynamics of colonies of three species of ants in desertified grassland, southeastern Arizona, 1958-1981. Am. Midl. Nat. 118: 177-188.
Conway, J.R. (1977): Analysis of clear and dark amber repletes of the honey ant, Myrmecocystus mexicanus hortideorum. Ann. ent. Soc. Am. 70: 367-369.
—— (1980a): The seasonal occurrence of sexual brood and the pre-and post-nuptial behavior of the honey ant, Myrmecocystus mexicanus Wesmael in Colorado. Jl. N. Y. ent. Soc. 88: 7-14.
—— (1980b): Foraging behavior of the honey ant, Myrmecocystus mexicanus in Colorado. Trans. Ill. St. Acad. Sci. 72: 81-93.
—— (198la): A study of dealated queens of the honey ant, Myrmecocystus mexicanus Wesmael, and their progeny in captivity (Hymenoptera: Formicidae). J. Kans. ent. Soc. 54: 41-55.
—— (1981b): Honey ants, sweet swell of success. Sci. Digest 1981(8): 58-59, 117.
—— (l983a): Nest architecture and population of the honey ant, Myrmecocystus mexicanus
— (l983b): A study of winged queens of the Colorado honey ant, Myrmecocystus mexicanus in captivity. JI N.Y. ent. Soc. 93: 252-263.
Conway, J.R. (1984): Swollen sisters provide honey pots. New Scientist, May, 3 pp.
—— (1990): Notes on repletes, myrmecophiles, and predators of honey ant nests (Myrmecocystus mexicanus) (Hymenoptera; Formicidae) in Ariz.. JI. N. Y. ent. Soc. 98: 103-107.
—— (1991): The Biology and Aboriginal Use of the Honeypot Ant, Camponotus inflatus Lubbock, in Northern Territory, Australia. Aust. Ent. Mag. 18(2) Jun 1991.
—— (2003): Architecture, Population Size, Myrmecophiles, and Mites in an Excavated Nest of the Honeypot Ant, M. mendax Wheeler, in Arizona. Southwest Nat. 48(3):449-450.
Cover, S.P. & Johnson, R.A. (2002): Checklist of Arizona Ants, Ant Course 2002, Portal, AZ.
Creighton, W.S. (1950): The ants of North America. Bull. Mus. Comp. Zool., Harvard, 104:1-585.
—— & Crandall, R.H. (1954): New data on the habits of Myrmecocystus melliger Forel. Biol. Rev. City College N. Y. 16(I): 2-6.
Gregg, R.E. (1963): The ants of Colorado. Univ. Colo. Press, Boulder xvi + 792.
Erwin, N. (pers. comm.): Director, Orkin Insect Zoo, Museum of Natural History, Smithsonian Institute, Washington, DC.
Evans, J. (pers. comm.): Aquatic Biologist, Californian Academy of Sciences, San Francisco, CA.
Hölldobler, B. (1976): Tournaments and slavery in a desert ant. Science 192:912-914.
—— (1981): Foraging and spatiotemporal territories in the honey ant Myrmecocystus mimicus Wheeler (Hymenoptera: Formicidae). Behav. Ecol. Sociobiol. 9: 301-314.
—— & Lumsden, C.J. (1980): Territorial strategies in ants. Science 210: 732-739.
—— & Wilson, E.O. (1990): The ants. Cambridge, MA: Belknapp Press, Harvard Univ. Press, 732 pp.
Kay, C.A.R. & Whitford, W.G. (1978): Critical thermal limits of desert honey ants: possible ecological implications. Physiol. Zool. 51: 206-213.
Lloyd, H.A.; Blum, M.S.; Snelling, R.R. & Evans, S.L. (1989): Chemistry of mandibular and Dufour's gland secretions of ants in genus Myrmecocystus. J. Chem. Ecol. 15: 2589-2599.
Lumsden, C.J. & Hölldobler, B. (1983): Ritualized combat and intercolony communication in ants. J. theor. Biol. 100: 81-98.
McCook, H.C. (1882): The honey ants of the garden of the gods and the occident ants of the American plains. Philadelphia: J. B. Lippincott & Co. Philidelphia, 188 pp.
Mendez, R.A. (pers. comm.): Principal, Work-as-Play, Portal, AZ.
Morgan, R.C. (1991): Natural history, field collection and captive management of the Honey ant Myrmecocystus mexicanus. Int. Zoo Yb. 30:108-117.
—— (1995): Honey, I Swelled the Kids!- Ambitious Ants Store Food for Dry Times. Backyard Bugwatching 16:3-6.
Prchal, S. (pers. comm.): Director, Sonoran Arthropod Studies Institute, Tucson, AZ.
Rissing, S.W. (1984): Replete caste production and allometry of workers in the honey ant M. mexicanus Wesmael (Hymenoptera: Formicidae). J. Kans. ent. Soc. 57: 347-350.
Snelling, R.R. (1976): A revision of the honey ants, genus Myrmecocystus (Hymenoptera: Formicidae). Sci. Bull. Nat. Hist. Mus. Los Ang. Cty. No. 24: 1-163.
—— (1982): A revision of the honey ants, genus Myrmecocystus first supplement (Hymenoptera; Formicidae). Bull. Sth Calif. Acad. Sci. 81: 69-86.
Vander Wall, S.B (1986): Gift of the Mice People. Natural History 3/86: 22-28.
Weissmann, M. (pers. comm.): Director, Kalima Consultants, Inc., Northglenn, CO.
Wehling, W. (pers. comm.): Entomologist, USDA APHIS PPQ, Riverdale, MD.
Wheeler, G.C. & Wheeler, J.W. (1973): Ants of deep canyon. Riverside, CA: Univ. California.
Wheeler, G.C. & Wheeler, J.W. (1986): The Ants of Nevada, Allen Press, Lawrence, KS.
Wheeler, W.M. (1908): Honey ants, with a revision of the American Myrmecocysti. Bull. Am. Mus. Nat. Hist. 24: 345-397.
—— (1910): Ants: their structure, development and behavior. New York: Columbia University Press.
—— (1917): The pleometrosis of Myrmecocystus. Psyche. Boston 24: 180-182.
Wilson, E.O. (1971): The insect societies. Cambridge, MA: Belnap Press, Harvard Univ. Press, 548 pp.

Randy C. Morgan
Invertebrate Conservation Program Manager
Insectarium, Cincinnati Zoo & Botanical Garden (CZBG)
3400 Vine Street, Cincinnati, Ohio 45220 USA