INITIAL EFFORTS TO COLLECT AND MAINTAIN A LIVE COLONY OF GIANT FOREST ANT, CAMPONOTUS GIGAS (HYMENOPTERA: FORMICIDAE)
AT THE PENANG BUTTERFLY FARM

Randy C. Morgan1, David Goh2, B.T. Chin2, Kuennie Lee2 and Tan Poai Ean2

1 Insectarium, Cincinnati Zoo and Botanical Garden, 3400 Vine St., Cincinnati, OH 45220 USA
2 Penang Butterfly Farm, 830, Jalan Teluk Bahang, 11050 Penang, Malaysia

Abstract: Attempts to field collect and keep a live colony of C. gigas in captivity are described. Twelve satellite nests were found in 0.5 ha of disturbed forest and determined to be part of the same colony by observing forager movements or combining workers from different nests without negative interactions. Nests were fully or mostly excavated over a several day period. Major workers had sharp powerful mandibles, bit painfully and often cut collectors’ skin. About 800 workers were captured and funneled into clear plastic field containers containing 4 cm diam cardboard tubes for perching. Neither the queen nor brood of any stage were located and probably resided in undiscovered nests. The ant-laden cardboard tubes, along with a few sections of wood naturally containing resident ants, were placed within laboratory housing. This consisted of four large aquariums linked with ant runways and containing three types of artificial nest chambers. Captive workers consumed sugar and honey water but generally ignored small insect prey. Most of the ants remained clustered within the cardboard tubes and natural nest cavities, and largely ignored the artificial nest chambers. Initial worker mortality was extremely high and apparently related to large quantities of formic acid secreted by agitated ants when captured, then spread to other ants, at least partly by collectors’ fingers. Regular hand rinsing with water while collecting seemed to improve survivorship. Significant mortality continued for six weeks post collection. The group stabilized at approximately 50 workers then gradually declined during subsequent months. Considerations and plans for future work are noted.

INTRODUCTION

     We attempted to excavate and collect a live colony of the giant forest ant, Camponotus gigas Latreille 1802 (subgenus Dinomyrmex) and maintain it in the laboratory. We hoped to eventually create a public educational display showcasing an active colony of these giant ants at the Penang Butterfly Farm. Previous work with two other giant ant species, Dinoponera longipes and Paraponera clavata, both Neotropical ponerines, had been successful and publicly well-received at the Cincinnati Zoo and Botanical Garden’s Insectarium (Morgan 1993, 1997).
      Camponotus gigas (Fig. 1) is broadly endemic to Southeast Asia, ranges from Sumatra to Thailand and commonly occurs in Malaysia. It lives in tropical rain forests and is found in habitats varying from low lying peat-mangrove swamps to mountain forests around 1500 m above sea level (Pheiffer and Linsenmair 2000). This conspicuous ant has been the subject of various behavioral, ecological and life-history studies (Tho 1981, Gault 1987, Chung and Mohamed 1993, Levy 1996, Orr and Charles 1994, Orr et al. 1996, Yemane et al. 1996, Pfeiffer 1997, Pfeiffer and Linsenmair 1997, 1998, 2000, 2001).
      The giant forest ant is one of the world’s largest ants, and by some measures is the largest of all ants. The worker caste is bimorphic or consists of size-specialized minor and major sub-castes (Fig. 1), with the largest major workers reaching 3 cm long and weighing up to 400 mg (Pfeiffer and Linsenmair 2000). These major workers have about the same body length as the largest Dinoponera species but are more massive, and are noticeably bigger than the largest Paraponera workers or queens (Morgan pers. obs.). However, the queens of Camponotus gigas are even somewhat larger and more massive than the major workers, making these reproductive females the largest known ants (Moffet pers. comm.).
      Camponotus gigas belongs to the highly-evolved and species-rich ant subfamily Formicinae. Formicine ants are distinguished from those in other subfamilies by their ability to spray or otherwise secrete concentrated formic acid from an acidipore, a glandular opening at their abdominal tip. The formic acid is generally used offensively to help incapacitate prey and/or defensively to repel or kill enemies (Hölldobler and Wilson 1990).
      Established C. gigas colonies have relatively moderate worker populations, are polydomous and maintain huge three-dimensional territories (Pfeiffer 1997, Pfeiffer and Linsenmair 2000, 2001). The best-known of several colonies studied by these researchers consisted of about 7000 workers variably occupying 17 (on average 11) separate nests within a ground surface territory of 0.8 ha (8000 sq m or about 2 acres). Other colonies occupied 8-14 nests in 0.35 to 0.66 ha territories. Nests were located in the soil at tree bases, under fallen logs and in living tree cavities near the ground. Workers traveled between active nests on trails passing through the forest canopy.
      Territorial borders are maintained by long-lasting, highly-ritualized fights between specialized majors from adjacent colonies (Pfeiffer and Linsenmair 2001). The majors may meet every night at relatively permanent tournament sites used for up to several months. They face off, then box or strike at each other with their forelegs, and fight for hours without harming one another.
      Foraging by C. gigas is predominantly nocturnal. Shortly after dusk, large numbers of workers leave their nests and most ascend trees into the canopy, returning later with food (Pfeiffer and Linsenmair 2000). The vast majority of the food appears to be either honeydew or extrafloral nectar carried within the ants’ gasters. Other foragers use their mandibles to carry various small insects or bird droppings, the latter thought to be an important source of nitrogen. C. gigas is a central place forager using an efficient communication and recruitment system to optimize foraging. Some workers specialize as transporters to carry food amassed at peripheral nests to the main nest occupied by the queen (Pfeiffer and Linsenmair 1998).
      The nuptial flight time of C. gigas is highly unusual among ants, being aseasonal and phase-shifted, occurring on average every 188 days (Pfeiffer and Linsenmair 1997). The end result is that the time of the mating flight changes from year to year. New colonies evidently are started by the claustral mode of colony founding (Pfeiffer and Linsenmair 2000) typical of most higher ants. Newly mated queens drop their wings, seclude themselves in a cavity and rear their first brood on nutrients stored within their own bodies (Hölldobler and Wilson 1990). The first few workers to emerge are tiny but soon begin to forage and help the queen raise additional brood. Established C. gigas colonies appear to be monogynous or contain only a single mated queen which generally resides in one of the centrally located nests (Pfeiffer and Linsenmair 2000).
When we began this project with C. gigas, we expected that it would be challenging to locate and excavate all of the nest sites occupied by a large well-established colony, especially given limited field time. And based on prior work with another challenging ant, Paraponera clavata (Morgan 1996, 1997), we also expected that eventual success would require multiple efforts incorporating trial and error experimentation and technique refinement based on what we learned in both the field and laboratory.

MATERIALS AND METHODS

Study site: Our study took place 03-10 February 2006 on Penang Island, a 285 sq km island near the western coast of peninsular Malaysia. Penang is home to the modern port city of Georgetown, and much of the island has been developed and urbanized. However, several large tracts of forest preserve and public parkland remain and have endemic populations of C. gigas.
      The project was based at the Penang Butterfly Farm located in Teluk Bahang about 17 km from Georgetown. Our laboratory was a small, screen-sided building situated in a non-public space normally used for insect rearing. Our field site was about a 200 m walk from the Butterfly Farm and consisted of both disturbed forest and lightly managed parkland, the latter with a relatively open under story and little ground cover. Earlier observations by Butterfly Farm staff had found C. gigas actively foraging in this area.
      There were two key preliminary tasks. We designed, purchased materials and constructed prototype laboratory containment for C. gigas. And we conducted nocturnal field surveys in our study area to locate as many active nest sites as possible.
      Laboratory nest design and set-up: Four, large, all-glass aquariums, two each with 227 and 303 liter (60 and 80 gal) capacity, were linked together providing what we hoped would be sufficiently spacious nesting and foraging arenas for C. gigas (Fig. 2). To create ant runways between the aquariums, we first drilled a single 5 cm (2 in) diameter hole in the end of each tank, just above the floor line, using a special bit for cutting glass. Standard PVC couplings were then snugly fitted into these holes, and in turn supported horizontal sections of clear PVC tubing, all joined together with additional sections of tubing and couplings. The aquariums rested on stable wooden benches and were positioned in a U-shaped configuration, creating a central servicing aisle (Fig. 2). Bench legs sat in oil pans to prevent access by small crawling animals, particularly various tiny ants that commonly foraged in the laboratory area.
      Excess silicon sealant protruding from inner tank corners was removed with a straight-edged razor blade to prevent the ants from gaining footholds. The upper inner perimeter of each tank was smeared with a band of 3-in-1 Household Oil to create a slippery barrier, an effective means of temporarily preventing the escape of many ant species (Morgan 1991a, b, 2004). Brick stacks provided feeding platforms (Fig. 2) to facilitate servicing by keepers needing to reach into the tanks. Various crickets, flies and other small insects reared at the Butterfly Farm as feeder animals were offered as prey, and presented to the ants freshly killed to encourage feeding.
      Three types of artificial nest chambers were placed within the aquariums to determine if the ants would find them acceptable or show a preference. These were: 1) small, clear, polystyrene boxes with single 2.5 cm diam entrance holes, 2) 10 cm diam, clear, vinyl hose sections lined with screen for perching, and 3) horizontal, narrowly-tiered boards with dark inner cavities (Fig. 4).
      To prevent the intense tropical sunlight from shining directly into the laboratory and potentially overheating captive ants, three layers of shade cloth were hung on the exterior wall screens. This greatly limited insolation while allowing adequate lighting and airflow within the workspace.
      Field-surveys and colony determination: Starting around dusk each evening, and armed with powerful flash lights, our field team extensively combed the study site in ever-widening circles, searching for ant activity, foraging trails and nest sites. Nests were flagged as found with brightly-colored marking tape, making it easier to find them again later when we returned to test for colony membership, or to excavate and collect resident ants.
      It was not immediately obvious that all of the nests we found were part of the same colony since associated foraging trails often went up the sides of trees into the canopy. To determine whether or not these nests were part of the same colony, we performed a simple field test similar to that used by Pheiffer and Linsenmair (2001). We collected several workers from undetermined nest sites and placed them into a small container already holding workers from our subject colony. We then watched for fighting, other aggressive interactions or strong avoidance behavior.
      Colony field collection: Soil nests were carefully excavated with shovels and hand trowels, and those in stumps or fallen logs broken apart with axes and pry bars. Exposed ants either ran or tried to hide in partial nest cavities. The ants were captured one by one with fingers or forceps, or physically directed to run into small plastic cups, then quickly placed into field containers.
      Field containers consisted of clear plastic boxes with well-ventilated lids modified to hold an upright funnel (Fig. 3), similar to those used previously to collect Paraponera clavata in South America (Morgan 1996, 1997). The ants were relatively easily funneled into the container and returned back through the opening only with difficulty, thus were little prone to escape. When not in use, the funnel opening was plugged with a small rubber stopper. Field containers were partially-filled with stacked, 5 cm diameter cardboard tubes (from rolled paper towels) providing perching for captured ants (Fig. 3), and were kept shaded to prevent overheating from sunlight.

RESULTS

Field-survey and colony determination: After dusk, C. gigas workers were found actively moving about on tree trunks, vegetation and the ground surface. We located 12 separate nesting sites in an area roughly 50 x 100 m (0.5 ha) in size. The nests were generally about 10-30 m apart and variably positioned in the ground at the bases of trees and stumps, in and under fallen and sometimes partially-rotten logs, and in apparently pre-existing holes in living trees, both near the ground surface and in one case about 3 m above ground. All of the nests were determined to belong to the same colony since workers were either observed regularly moving back and forth between them, or did not exhibit aggressive interactions when groups were confined together.
     Colony field collection: When nest tunnels and chambers were exposed, resident workers often swiftly scattered thus needed to be collected quickly. Despite their size, even the major workers were soft-bodied and easily damaged if handled too roughly with fingers or pinched with a forceps, sometimes resulting in kinked legs or dented abdomens. Using fingers to grab the ants seemed to work the best since it could be done relatively rapidly and provided the necessary degree of control to keep from injuring the ants. The major workers were particularly challenging in this regard since they have powerful mandibles that easily and painfully cut through one’s skin, often drawing blood. Major workers were normally grabbed such that their heads were wedged between the collector’s thumb and forefinger, preventing them from turning and biting. None-the-less, cries of human pain and curses often echoed through the forest! Capturing running ants by trying to direct them one by one into a small cup worked adequately, but was much more time-consuming.
      More than 50% of the workers we collected the first few days were found dead on subsequent days. We soon came to suspect that our fingers were painting formic acid from one captured ant onto the body of the next, since it was obvious that agitated or captured ants typically released significant quantities of concentrated formic acid, in some cases producing visible droplets. We then began carrying water buckets with us into the field, and rinsing our fingers and hands after every capture. Following this procedural modification, subsequent worker mortality appeared to be significantly reduced, at least in the short term.
      At the end of each collecting session, the cardboard tubes containing ants clustering within were gently removed from field containers and stacked within one of the aquariums (Fig. 4). We also collected a few pieces of wood containing mostly-intact nest chambers housing tightly packed groups of workers, and added these to the laboratory nest system. Eventually we collected a total of over 800 C. gigas workers, including a wide array of minor and major sub-castes. However, we did not find the queen or any brood in the nests we excavated.
      Captive colony history: Workers continued to have a strong affinity for their natural nest chambers within wooden objects, and also for the cardboard tubes that they had began perching in when first placed in the field containers (Fig. 4). For the most part, they remained clustered within these objects and showed little to no interest in moving into any of the three types of artificial nest chambers provided for their use (Fig. 4). Some workers foraged for honey and sugar water, mostly after dark, but otherwise exhibited very little interest in prey items.
      For about six weeks post-collection, captive workers continued to experience significant mortality, and then the colony fragment stabilized at about 50 individuals. During subsequent months the remaining group of ants slowly declined.

DISCUSSION

     As expected, Camponotus gigas was challenging to collect and maintain in the laboratory. Despite intensive searching for several days in our study area, we did not find the queen nest or nests containing any immature stages. Capturing the queen and brood would likely improve group cohesion and chances for successfully establishing a laboratory colony. Future attempts to excavate and capture a colony need to invest more time and effort delineating colony boundaries, finding all or most of the nesting sites, and particularly the queen nest.
      Also challenging were C. gigas’ defensive behaviors, both biting and formic acid secretions. Major workers were especially painful biters, frequently drawing blood. Both field collectors and laboratory colony keepers needed to handle or otherwise manipulate these ants with great care. Gloves were generally not helpful, since the majors clamped onto the fabric or leather material with their mandibles, and then were difficult to remove without injuring them.
      It seems unlikely that C. gigas’ formic acid is more concentrated or toxic than that produced by smaller formicines. However, individual C. gigas workers clearly have the ability to secrete much more of it. Once we suspected that formic acid was being transferred between ants held trapped in our fingers, we modified our field technique to include frequent hand and tool rinsing. This seemed to reduce the impact of formic acid on the ants we collected, but was not a perfect solution, since ants continued to exhibit higher mortality than we would have normally expected.
      The caustic property of the C. gigas’ formic acid was dramatically demonstrated several days into our field work, when both R. Morgan and a field assistant, who had done most of the collecting by hand, experience serious after-effects. On both, thick layers of skin first blistered and then sloughed off the thumb, index and middle fingers that had been used to grasp ants. The end result was severe second-degree burns and, in a few spots, third-degree chemical burns that blackened the underlying tissue. These injuries were painful, greatly limited the use of the hands, and took several weeks to completely heal.
      The large quantities of concentrated formic acid produced by agitated C. gigas workers remain a significant challenge for future attempts to collect and maintain live colonies. We plan to continue this work in October 2006. We hope to collect, with minimal disturbance, sections of wood occupied by resident ants and gently transfer these into our laboratory nest system, and only handle individual ants when absolutely necessary.

Acknowledgments: The Penang Butterfly Farm generously supported this work, and many employees of the Butterfly Farm provided helpful assistance at various stages of the project. David Goh and B.T. Chin graciously hosted R. Morgan during his stay in Penang.

REFERENCES

Chung, A.Y.C. and M. Mohamed (1993): The organization and some ecological aspects of the giant forest ant, Camponotus gigas. Sabah Society Journal 10:41-55.
Gault, D. (1987): Feeding and Foraging Behavior in the Giant Forest Ant, Camponotus gigas, in a Malaysian Rain Forest. Tropical Biology Newsletter 52:2.
Hölldobler, B. and E.O. Wilson (1990): The Ants. Belknap Press of Harvard University Press, Cambridge, MA, 732 pp.
Levy, R. (1996): Interspecific colony dispersion and niche relations of three large tropical rain forest ant species. In: Edwards, D.S., W.E. Booth, and S.C. Choy (eds.), Tropical Rainforest Research 47: 331-340. Kluwer Academic Publishers, Dordrecht, Boston, London.
Moffet, M. (pers. comm.): Department of Entomology, Museum of Comparative Zoology, Harvard University, Cambridge, MA.
Morgan, R.C. (1991a): Natural history, field collection and captive management of the Honey ant, Myrmecocystus mexicanus. International Zoo Yearbook 30:108-117. Zoological Society of London, U.K.
Morgan, R.C. (1991b): Natural history and captive management of leaf-cutting ants in the genus Atta (Hymenoptera: Formicidae). 1991 AAZPA Annual Conference Proceedings: 321-328. American Zoo and Aquarium Association, Silver Springs, MD.
Morgan, R.C. (1993): Natural History Notes and Husbandry of the Peruvian Giant Ant, Dinoponera longipes (Hymenoptera: Formicidae). 1993 Invertebrates in Captivity Conference Proceedings: 140-151. Sonoran Arthropod Studies Institute, Tucson, AZ
Morgan, R.C. (1996): Quest for the giant tropical bullet ant, Paraponera clavata. 1996 Invertebrates in Captivity Conference Proceedings: 13-20. Sonoran Arthropod Studies Institute, Tucson, AZ.
Morgan, R.C. (1997): Giant Tropical Bullet Ant, Paraponera clavata, Natural History and Captive Management. 1997 AZA Regional Conference Proceedings, 8 pp. American Zoo and Aquarium Association, Silver Springs, MD.
Morgan, R.C. (2004): Biology, husbandry and display of the diurnal honey ant Myrmecocystus mendax Wheeler (Hymenoptera: Formicidae). 2004 Invertebrates in Captivity Conference Proceedings: 91-104. Sonoran Arthropod Studies Institute, Tucson, AZ.
Orr, A.G. and J.K. Charles (1994): Foraging in the giant forest ant, Camponotus gigas (Smith) (Hymenoptera: Formicidae): evidence for temporal and spatial specialization in foraging activity. Journal of Natural History 28:861-872.
Orr, A.G., J.K. Charles, H.R. Hj Yahya, N. Hj Sharebini (1996): Nesting and colony structure in the giant forest ant Camponotus gigas (Latreille) (Hymenoptera: Formicidae). The Raffles Bulletin of Zoology 44:247-251.
Pfeiffer, M. (1997): Sozialstruktur und Verhaltensokologie von Riesenameisen Camponotus gigas Latreille 1802 im Regenwald Malaysias uaf Bornea. Thesis, Bayerische Julius-Maximilians-Universitat Wurzburg 1996, Verlag Wissenschaft and Technik, Berlin, 191 pp.
Pfeiffer, M. and K.E. Linsenmair (1997): Reproductive synchronization in the tropics: the circa-semiannual rhythm in the nuptial flight of the giant ant Camponotus gigas Latreille (Hymenoptera: Formicidae). Ecotropica 3:21-32.
Pfeiffer, M. and K.E. Linsenmair (1998): Polydomy and the organization of foraging in a colony of the Malaysian giant ant Camponotus gigas (Hymenoptera: Formicidae). Oecologia 117:579-590.
Pfeiffer, M. and K.E. Linsenmair (2000): Contributions to the life history of the Malaysian giant ant Camponotus gigas (Hymenopter, Formicidae). Insectes Sociaux 47:123-132.
Pfeiffer, M. and K.E. Linsenmair (2001): Territoriality in the Malaysian giant ant Camponotus gigas (Hymenoptera/Formicidae). Journal of Ethology 19:75-85.
Tho, Y.P. (1981): The giant forest ant Camponotus gigas, Malaysia. Nature Malaysia 6:32-35.
Yamane, S., T. Itino and A.R. Nona (1996): Ground ant fauna in a Bornean dipterocarp forest. Raffles Bulletin of Zoology 44:253-262.