This paper reports preliminary field and laboratory work on Dinoponera longipes from upper Amazonia Peru. Field study objectives were to examine: 1) nest site location relative to topography, vegetation, and Paraponera clavata nests, 2) nest architecture and population, and 3) foraging time and behavior. After captive husbandry techniques were developed, laboratory objectives were to observe: 4) foraging and feeding, and 5) nesting behavior.

The Neotropical ponerine genus Dinoponera contains some of the world's largest ants; individuals reach 3 cm or more in body length. The only ants slightly more massive are queens and major workers of the Borneon carpenter ant Camponotus gigas (Moffett, pers. comm.).

Dinoponera studies began relatively recently. Kempf (1971) preliminarily described nine species native to various savanna, dry forest, and wet forest habitats. Haskins and Zahl )

Paiva (1993) described D. australis nest architecture, population, and arthropod guests, including other ants. D. australis is native to savanna-like vegetation; nests are rather deep, with up to 15 large chambers, each big enough to contain the entire colony, which is queenless and at most contains 30 large workers. Paiva also noted laboratory colonies housed in Gypsy nests. Mendez (pers. comm.) maintained a captive D. gigantea colony for several years in both hydrostone and natural soil nests; mating was observed in the nest, and the ants' diet included "pinkie" or newborn mice.

METHODS AND MATERIALS

Field Work: Studies began June 1992 and occurred during three brief visits to the Amazon Center for Environmental Education and Research (ACEER). The ACEER is located about 3° South latitude, approximately 70 km northeast of Iquitos, Peru, and is positioned along the Quebrada Grande, a narrow tributary of the Sucusari River, just north of where it joins the Napo River, a major tributary of the Amazon River.

Much of the surrounding Amazon River basin floods annually, but ACEER habitat is upland forest (terra firma) on lateritic soils; vegetation is characterized by high taxonomic diversity (Gentry 1980, 1989). Except in a relatively few disturbed areas, this is primary forest habitat, consisting of mostly mature trees and relatively open understory. Temperatures near the forest floor are generally stable and warm (around 24°C). Rainfall is heavy and non-seasonal, occurring with about equal frequency throughout the year.

During a two-day visit spanning 24-26 June 1992, a nest of the giant ponerine ant, Paraponera clavata was excavated for captive study and possible Insectarium display. While digging, it was surprising to find agitated Dinoponera females periodically emerging from an exposed subterranean passage.

During a four-day visit spanning 12- 16 July 1993, a Dinoponera nest site field survey was conducted. Because of its accessibility, the ACEER Canopy Walkway Loop Trail was chosen as a survey baseline, and was measured along its length with a fiberglass measuring tape.

Throughout the study period, Dinoponera nest associated arthropods and foraging behaviors were noted.

Laboratory Work: Captive study was carried out at the Insectarium during the same time period, and is ongoing. The rearing room is maintained at 24-27°C, 40-60% relative humidity, and illuminated primarily with fluorescent lights on a 16-hour day cycle. Stable shelves support observation nest boxes contained within standard sized glass aquaria.

Various size glass aquaria contain nest boxes and serve as foraging areas. Aquaria are bedded with a layer of cypress bark mulch, used because it is relatively clean and rot-resistant. A flat brick provides a stable, elevated, mulch-free area for dish placement. A petri dish, regularly supplied with fresh drinking water, is routinely visited by the ants. A wide variety of prey items are provided; however, honey or sugar water is offered infrequently since the ants show little to no interest in these sweet liquids.

Before aquarium use, silicon sealant in the inner corners was removed with a razor blade to prevent the ants from finding footholds, although the ants proved to be extremely poor climbers. They often slip when crossing smooth surfaces, can not climb vertical glass, and therefore do not escape aquaria. Tanks typically are left uncovered to facilitate routine management.

Colony D-1 currently is housed in two large interconnected aquaria. One serves as the nesting area and is nearly filled by three

The nesting tank is connected to a smaller foraging area tank via two 3.75-cm diam x 15 cm 1ong clear plastic tubes positioned horizontally, and is treated by the ants like tunnels. Foraging area mulch is kept dry to retard decomposition of old, discarded prey items. Ants place colony trash at the tank's far end, facilitating periodic debris removal.

Nest D-7 population, recently collected and consisting of only seven individuals, was established in an observation nest providing room for colony growth.

Voucher specimens deposited at the Harvard Museum of Comparative Zoology, and later examined by W.L. Brown, Jr. at Cornell University, were preliminarily identified as D. longipes Emery. However, slight anatomical differences, especially denser body pubescence, suggest that the specimens may in fact represent a new undescribed Dinoponera species near longipes (Cover, pers. comm.).

RESULTS

Survey Baseline Description: The ACEER Canopy Walkway Loop Trail is approximately 810 m long and is topographically variable. The trail starts on one hilltop, near the lodge's westerly wing, descends to a narrow creek, ascends to the top of an adjacent hilltop, follows the ridgetop, then sharply descends back towards and crosses the same creek, and finally ascends to its end near the lodge's easterly wing.

The hilltop supporting the canopy walkway access tower was chosen for this purpose because it is the area's highest point, while the creek, which rarely floods (Donahue, pers. comm.), traverses the study site's lowest region. The elevation change between this hilltop and creek roughly was estimated to be about 25 m. most of the baseline traversed primary forest. However, about the last 50 m passed through an area of tree cut and natural tree fall exhibiting dense secondary growth.

Dinoponera Nest Sites: A total of 22 nests (D-2 to D-23) were found along the survey baseline. Nest position seemed independent of the area's variable topography. Overall, the Dinoponera nest site pattern suggested that an apparent minimal spacing was maintained between nests.

Nesting sites were found at all elevations, from the low lying stream bank (D-20), through intermediate levels, and at the highest hilltop (D-16). Several steeply sloping areas evidently lacked nests, possibly reflecting the ants' poor climbing ability. The distance between neighboring nests, estimated from the survey baseline, ranged 14-69.5 m and averaged about 35 m apart (Table 1).

Nest Site Vegetation: Associated plant species were diverse and seemed unrelated to any particular plant, since nests were close to at least 24 plant species in 12 families (Table 1). Nests were variably positioned around, between or under plant bases or roots. Most nests seemed heavily shaded by overhead vegetation, although at least two nest mounds received dappled sunlight at mid-day.

All nests were located in areas characterized by sparse understory vegetation and lightly littered ground, except for two. Nests D-7 and D-8 were found in dense secondary growth vegetation and, judging from the number of surface openings (Table 1) and the excavation of one nest, both contained small or young colonies.

Nest Site Relative to Paraponera: An obvious association between Dinoponera and Paraponera nest sites was not apparent, since only three or about 13% of the Dinoponera nests (n=23 including D- 1), were within 10 m of Paraponera nests. These Dinoponera-Paraponera nest pairs (n=3) ranged 1-10 m apart and averaged 4.4 m apart, curiously much closer than the distances found between adjacent Dinoponera colonies (Table 1).

Nest Architecture: Most, but not all nests exhibited slight to well-developed earthen mounds. These apparently were formed from an accumulation of excavated, rain-compacted soil, and seemed to be most noticeable when surrounding terrain was fairly level, or when vegetation provided structural support and protection from rain erosion. Nest D-1 atypically lacked an obvious surface mound and was positioned on a lightly vegetated, gently sloping hillside.

The number of nest surface openings (tunnel entrances) may be a relative measure of nest size. The number of openings per nest (n=23) ranged from 1-30, and averaged about 11 openings per nest (Table 1). Nest openings were 2-3 cm diameter holes penetrating the mound or ground surface, and frequently were surrounded by thinly spread piles of newly excavated clay soil particles. Often some openings were partially hidden under a light covering of forest litter.

Nest D-1 had 21 surface openings roughly scattered over the ground throughout a 2.1 x 0.6-m area. Most nest opening tunnels gently angled downward, and broadly interconnected with nine subterranean chambers. These were horizontally oriented, roughly linearly aligned, and all shallowly positioned, generally about 30 cm from the surface.

Nest chambers were round or oval; most were 10-15 cm wide, but two were about 25 cm wide at their longest axis. The chambers tended to be slightly domed and about 3-4 cm high from floor to ceiling. Some chamber floors were lightly covered with a layer of what appeared to be dry vegetation bits.

During nest D-1 excavation, several spot temperature measurements made at the nest surface and within nest chambers were similar. These generally ranged closely to 24°C both during the day and at night.

Nest D-7 had a single surface opening surrounded by freshly excavated soil. The opening accessed a 13-cm deep vertical tunnel leading to a single subterranean chamber. The chamber was oval, horizontally positioned, and measured about 10x13 cm wide, and about 3 cm high from floor to ceiling.

Colony D-8 was situated in the side of what appeared to be a dead, large Dinoponera nest mound. While two nest openings appeared actively maintained, the remainder of the nest was deteriorating, was heavily overgrown with a mix of herbaceous vegetation, and seemed abandoned.

Nest Population: Nest D-1 population (less 13 females collected 10 months earlier) was 120 females (all morphological workers, hence no distinct queen), 5 winged males, 57 cocooned pupae, six large larvae, over 30 small to medium-sized larvae, and 25 eggs. The relatively numerous nest surface openings (Table 1) and male presence suggests that nest D-1 contained a large, reproductively mature colony.

Nest Associated Arthropods: During Colony D-1 and D-7 excavations, unidentified tiny yellow polymorphic ants were abundant in the soil around nests, and complicated work since their stings were somewhat painful.

One D-1 nest chamber contained a dead centipede and lampyrid larva, both presumably prey items.

A digger wasp and a 4-cm. long dark cockroach inhabited old surface openings of the apparently dead Dinoponera nest associated with Colony D-8. Both insects were seen repeatedly over a three-day period. The same roach species was also observed twice elsewhere in other active Dinoponera nest openings.

Natural Foraging Behavior: Females appeared to forage alone, remain on the ground, be more active at night, and readily exhibited observer avoidance behavior. During the seven-day field study period, over 30 observations were made of ants away from their nests, presumably foraging. All were solitary, were crawling on the ground surface, and were never seen on plant bases or elevated vegetation.

Foragers seemed to be principally nocturnal, since most were observed at night, but a few were active at day, both mornings and afternoons. During a "friaje" or brief cold front, night air temperatures became atypically cool (approximately 19°C), and foragers were conspicuously absent.

Foragers also seemed highly visual and timid; in response to nearby observer movement, they usually darted under leaf litter and temporarily hid, or paused in position and crouched. Without further disturbance, their activity usually soon resumed. Coquinche and Donahue (pers. comm.) confirmed that this was typical behavior.

Most foragers carried nothing in their mandibles, except as follows. During nest P-1 excavation at night, two ants carried exposed, injured earthworms. On other nights, one ant carried a cricket leg, while another held what appeared to be a dry leaf fragment.

Captive Foraging Behavior: The ants behave as predator-scavengers, and appear to be almost exclusively carnivorous on a wide variety of live, injured, or recently dead animals. Small live prey is bitten with the mandibles and carried whole to the nest. Strongly struggling prey may be grappled with and stung. Dinoponera longipes' sting reportedly is much less severe than that of Paraponera (Coquinche, pers. comm.), although D. gigantea's sting was very painful and paralytic (Mendez, pers. comm.).

Food items too large to carry are first cut into pieces. The mandibles are serrated and seem adapted for cutting in two very different ways. Against large items, they are pushed open and used like a curved saw blade with back and forth head movements. Against narrow objects, when mandibles are mostly closed, a ratchet or leverage mechanism seems to be used, evidenced by audible clicking noises in conjunction with head lurches.

Food includes both invertebrates (mostly insects) and various vertebrates. Invertebrates eaten included Dictyoptera (roaches, mantids), Phasmatoptera (stick insects), Orthoptera (crickets, grasshoppers, katydids), Isoptera (giant termites), Neuroptera (dobsonflies), Hemiptera (back swimmers, giant water bugs), Homoptera (cicadas), Coleoptera (mealworms), Diptera (flesh flies), Lepidoptera (waxworms), Hymenoptera (male Paraponera, honey bees), Arachnida (orb spiders), Chilopoda (centipedes), and Annelida (earthworms).

Observation Nest Behavior: Colonies appear to be highly mobile, and when first released, readily explore newly found nest openings and within minutes carry their brood inside. Eggs most often are held and moved in small clumps, while larvae and pupae are transported individually.

Similarly, if nest chambers are opened or otherwise disrupted, the ants carry their brood to secure areas. Observation nest chamber temperatures generally remained close to 24°C, but when one chamber was slowly warmed, the colony began to move out as the temperature approached 27°C.

Once settled in a nest, workers often partially cover and hide, or sometimes reduce or temporarily block nest openings with mulch. This behavior perhaps helps camouflage nest openings, or makes them more defensible.

Fecal droplets are used to line nest chamber walls, and to a lesser degree, crevices and spaces outside observation nest boxes. The droplets are applied as a dark gray liquid but soon dry and blacken. This or a similar material slowly accumulates as an oily film on the underside of viewing glass (nest chamber ceilings), eventually interfering with clear observation and requiring cleaning.

The brood tends to be segregated according to stage and size. Eggs are narrowly elongate (about 1 x 3 mm), and are grouped on chamber floors or held with the mandibles, either singly or in small clusters. The eggs readily adhere to one another, apparently facilitating their movement. Several observations suggest that small numbers of eggs normally are present in nests, even when developing larvae and pupae are not.

Two repeatedly observed behavioral interactions between nestmates are noteworthy. The first, called antennal drumming, is exhibited in the nest between meeting or facing ants. The antennae of one are drummed rapidly for about 1 second, against its nestmate's head region; this frequently is followed by an immediate, reciprocating response.

Females seem long-lived and apparently can begin egg production after months of inactivity. In D-cf, eleven or about 85% of the originally collected females (n=13) were alive 14 months later. Presumably, these were relatively old when captured, since they initially behaved as nest defenders. After about six months in captivity, this group began egg production and routinely tended about a dozen eggs. Several hatchling larvae were observed at various times, but all failed to develop; one was observed shriveled and dead, and all disappeared after 1-2 days.

DISCUSSION

Dinoponera nesting sites at the ACEER seemed relatively unaffected by topography, associated vegetation. However, given the ants' large size and poor traction, natural barriers to individual or colony movement probably include steep terrain, dense ground surface debris or vegetation, and standing waterways.

Nesting sites also seemed unrelated to Paraponera nest position. Paraponera and Dinoponera, both giant ponerine ants, look somewhat alike, except that the latter is larger. In Peru, Paraponera is locally called "isula", and Dinoponera "isula vieja", meaning old lady isula.

Despite the species' outwardly similar appearance, the close (1 m) position of D-1 to P-1 may have been a chance result of niche partitioning. Paraponera mostly forages high in the canopy, and though predatory, is predominantly nectivorous (Young and Herman, 1980; Bennett and Breed, 1985; Tobin, 1991). Conversely, Dinoponera seems restricted to the ground surface and appears to be almost exclusively carnivorous.

Since Dinoponera nests appeared over seven times more abundant in the study area, and assuming relatively non-aggressive species' interactions, some nests likely would be positioned closely to Paraponera by coincidence. However, the possibility remains that one or both species benefit when closely associated. For example, the jumping ant Gigantiops destructor sometimes nests with Paraponera, apparently for defensive advantages (Tobin, 1989).

Finding two, seemingly young colonies (D-7 and D-8) in recent secondary growth suggests the possibility that nest initiation may be linked with natural forest canopy gaps, a reproductive strategy used the local leaf cutting ants Atta sp. (pers. obs.), and by many plants (Hartshorn, pers. comm.). However, the apparently dead, heavily vegetated nest associated with D-8 also suggests that large colonies may not flourish in secondary growth areas. The overall D. longipes nest site pattern suggests that colonies are over dispersed, that is, they maintain a minimum acceptable distance for colony location relative to other existing nests. It would not be surprising to find that colonies maintain and defend relatively small, distinct foraging territories around their nests and defend these from conspecific colonies. The ants are well adapted to exploit a wide variety of small live prey or larger animal carcasses. One likely important food source is dead or dying animals raining down from the prolific rainforest canopy.

Dinoponera longipes have numerically small, queenless, and potentially highly mobile nest populations. The ants evidently prefer a fairly narrow temperature range, centered at about 24°C, for both foraging and nesting. Colony social organization seems to be governed, at least in part, by dominance interactions within the nest. The nests are generally large, spacious, and shallowly positioned, possibly to avoid flooding in the wet forest habitat.

The ants seem to be highly adapted for primary forest habitat, especially since disturbed areas around and near the ACEER lodge conspicuously lacked colonies. Given their extreme specialization and timid nature, D. longipes may be among the first species to disappear with significant human disturbance and primary forest habitat alteration. Long-term monitoring of D. longipes population trends might serve to measure the impact of ACEER program development relative to site habitat preservation.

Captive Dinoponera colonies are ideal for laboratory study or public educational display, since the ants are relatively easy to manipulate, feed, and confine. Individuals are huge, simplifying behavior observations, while their colonies readily move into and utilize observation nests.

ACKNOWLEDGMENTS

The Cincinnati Zoo and Botanical Garden supported this work in part. The ACEER Board approved the research program; Steve Timme shared reference material and gave guidance. International Expeditions, Explorama Lodge, and ACEER staff generously facilitated logistics. International Rainforest Workshop participants assisted some of the fieldwork. ACEER Resident Manager Theresa Wood provided study site orienteering. Naturalist Guide Reninger Coquinche greatly helped Paraponera nest excavation. Ethnobotanist Jim Duke and Naturalist Guide Segundo Inuma graciously identified plants. U.S. and Ohio Departments of Agriculture approved importation permits. Curator Milan Busching provided laboratory photography. Insectarium Entomology Technicians Karen Creamer and Theresa Austing assisted captive husbandry. Wildlife artist Joyce Turner provided artwork. Wife Kathy Beil-Morgan provided field photography, and tremendously contributed to survey work and manuscript preparation.

MEMORIAL ACKNOWLEDGEMENT

Dr. Alwyn Gentry tragically died in an airplane crash on August 3, 1993 in Ecuador while performing an aerial reconnaissance survey for Conservation International. Dr. Gentry, Senior Curator of the Missouri Botanical Garden, was a leading authority on Neotropical plants and was revered as a biodiversity conservation pioneer and as a fantastic teacher. He generously served on the ACEER Scientific Advisory Board where is was instrumental in program development. As a tropical tree climber, biting and stinging ants were his nemesis. He once told me, only slightly facetiously, that there wasn't a plant in the tropics that did not exploit ants for self-protection, a comment that I will always remember with fondness. Al will be dearly missed by all who know him.

REFERENCES

BENNETT, B. and BREED, M.D. (1985): On the association between Pentacletna macroloba (Mimosaceae) and Paraponera clavata (Hymenoptera: Formicidae) colonies. Biotropica 17(3):253-255.

COQUINCHE, R. (pers. comm.): Naturalist Guide, Explorama Lodge, Peru.

COVER, S. (pers. comm.): Depart. of Entomol., MCZ Labs., Harvard Univ., Cambridge, MA.

DONAHUE, P. (pers. comm.): Resident Manager, ACEER, Peru.

GENTRY, A. (1980): The flora of Peru: A conspectus. Fieldiana. Bot., n.s. 5:1 -11.

GENTRY, A. (1989): Northwest South America (Colombia, Ecuador, and Peru). In: Floristic inventory of tropical countries. The status of plant systematics, collections and vegetation, plus recommendations for the future (Campbell, D.G and Hammond, H.D.; eds.). New York Botanical Garden, N.Y.

HARTSHORN, G. (pers. comm.): World Wildlife Fund, Washington D.C.

HASKINS, C.P. and ZAHL, P.A. (1971): The reproductive pattern of Dinoponera grandis Roger (Hymenoptera, Ponerinae) with notes on the ethology of the species. Psyche 78(1-2):1-11.

HOLLDOBLER, B. and WILSON, E.O. (1990): The Ants. Cambridge, MA: Belknap Press of Harvard Univ. Press, 732 pp.

KEMPH, W.W. (1971): A preliminary review of the ponerine ant genus Dinoponera Roger (Hymenoptera, Formicidae). Studia Entomologica, n.s., 14(1-4): 369-394.

KEMPH, W.W. (1972): Catalogo abreviado das formigas da Regiao Neotropical (Hymenoptera: Formicidae). Studia Entomologica,n.s.,15(1-4):3-344.

MENDEZ, R. (pers. comm.): Exhibit Consultant, Portal, AZ.

MOFFETT, M. (pers. comm.): Depart. of Entomol., MCZ Labs., Harvard Univ., Cambridge, MA.

OVERAL, W.L. (1980): Observations on colony founding and migration of Dinoponera gigantea. Journal Georgia Entomol. Soc. 15(4):466-469.

PAIVA, R.V.S. (1993): Another case of association between ants: Dinoponera and Pheidole. Notes From Underground, A Myrmecological Newsletter, No. 8. (Tobin, J.E, Chavarria, G., and Alpert, G.D., eds.), MCZ Labs., Harvard Univ., Cambridge, MA.

TOBIN, J.E. (1989): On the association between Gigantiops destructor and Paraponera clavata: An unlikely symbiosis. In: Notes From Underground, A Myrmecological Newsletter No. 3:3-6, MCZ Labs, Harvard Univ., Cambridge, MA.

TOBIN, J.E. (1991): A Neotropical rainforest canopy, ant community: Some ecological considerations. In: Ant-Plant Interactions (Huxley, C.R., and Cutler, D.F., eds.), pp. 536-538, Oxford Univ. Press, New York.

WHEELER, G.C. and WHEELER, J. (1976): Ant larvae: review and synthesis (Memoirs of the Ent. Soc. of Washington, No. 7). Entomology Society, Washington, D.C., 108 pp.

YOUNG, A.M. and HERMAN, H.R. (1980): Notes on foraging of the giant tropical ant Paraponera clavata. J. Kansas Entomol. Soc. 53:35-55.