Trail clearing and foraging behaviors in two species of Leaf Cutter ants Atta cephalotes and Acromyrmex coronatus Katy Zaksek Department of Environmental, Population and Organismic Biology University of Colorado Boul der ____________________________________________________________________ ABSTRACT Leaf cutting ants play an integral role in tropical forest communities. Foraging strategies impact the harvest efficiency of a colony and in turn rates of herbivory on a forest c ommunity. Trail construction and foraging efficiency were compared between two species of leaf cutting ants Atta cephalotes and Acromyrmex coronatus. Atta cephalotes constructed major trail systems and Acromyrmex coronatus, a species with smaller colony size, traveled between harvest sites on simple foraging lines across leaf litter. Significant differences existed across species with respect to traveling speed and burden ratios ANOVA, p < 0.0001 and 0.0071, respectively in dicating benefits in maintaining unobstructed trails. Atta cephalotes trail clearing behaviors were investigated further by simulating natural obstructions. Colonies exhibited rapid removal of all objects placed on trails and removal time was significantly correlated with worker activity and distance from nest entrance ANOVA, p = 0.0016 and 0.0163. This study attempts to demonstrate that A. cephalotes gains a competitive advantage over other leaf cutting species, and reaps benefits, from trail constructio n and maintenance despite large initial investment. RESUMEN Las zompopas juegan un papel muy importante en l os bosques tropicales. La s Estrategias de forrajeo influyen la eficiencia de cosecha de hojas de una colonia de z ompopas y en tomo a la proporcin d e herbvora en l as plantas del bosque. Construccin de senderos y eficiencias de forrajeo fueron comparados en dos especies de zompopas Atta cephalotes y Acromyrmex coronatus. Atta cephalotes construyen grandes sistemas de senderos y Ac. coronatus una especie de menor tamao colonial, viajan entre el sito de cosecha en simple lnea de forrajeo a travs de un desorden de hojas. Se encontraron diferencias significativas en tre las especies con respecto a la velocidad y razones de carga ANOVA, p < 0.0001 y 0.0071, respectivo ind icando beneficios en mantenimiento de senderos sin obstrucciones. El compartimiento de con respeto A, cephalotes de limpiar los senderos fue investigado simulando obstrucciones naturales. Se encontraron correlaciones significativas entre el tiempo que tomaron en limpiar el sendero y el numero de trabajadoras y la distancia del nido ANOVA, p = 0.0016 y 0.00163, respectivo Este estudio trato demostrar que A. cephalotes obtiene un a ventaja competitiva sobre otras especies de zompopa s porque construyen senderos y los mantienen a pesar de una gran inversin primaria.
INTRODUCTION Leaf cutting ants Formicidae: Myrmicinae: Atta cephalotes and Acromyrmex coronatus are found abundantly in lowland tropical forests and pastures Stevens 1983. Preferring relatively humid environments, they are distributed from the southern United States through northern South America Hanson and Gauld 1995, Hogue 1993. Leaf cutter ant s harvest leaves, flowers, and fruits and transport them back to underground nests Stevens 1983, Wilson 1971. Atta cephalotes colonies operate through complex social hierarchy consisting of different castes of workers; primarily the reproductive queen wh o produces all colony individuals, soldiers who defend the colony and clear trails, minima who clean parasites from leaves entering the colony, and workers, a polymorphic caste with up to seven sizes, which apparently serve various functions, including for ag ing and tending larvae Hoyt 1996 Colonies generally consist of thousands of individuals but can grow to several million Hanson and Gauld 1995, Hoyt 1996 Similarly, Ac. coronatus colonies consist of three castes. The reproductive queen is only slightly larger than the individuals in other castes, distinguishing Acromyrmex from other leaf cutter genera, which possess la rge queens Hogue 1993, Hoyt 1996 Acromyrmex coronatus minima tend the fungal garden, and brood but are also involved in colony defense since the soldier caste is absent in this species. Workers are polymorphic and forager size in Ac. coronatus is similar to the worker sizes found in small A. cephalotes colonies. Acromyrmex coronatus exhibits a narrower range size of work ers. Small colonies of Ac. coronatus are found in mounds of organic debris and larger colonies resemble the large underground nests of A. cephalotes Wetterer 1995. Because of their slightly smaller size, Ac. coronatus regularly cuts thin, soft leaves of herbs rather than tougher leaves from large trees as A. cephalotes does Wetterer 1991 1995 Colonies can consist of up to 150,000 workers but colonies will never get as large as mature A. cephalotes colonies Wetterer 1995. Evidence of the extensive, underground A. cephalotes colonies are large, bare mounds of dirt with trunk trail systems branching from multiple exit holes Stevens 1983. The trunk trail system resembles a tree with one main trail branching out into numerous limbs, leading to foraging sites. This wide branching allows thousands of colony members to forage rapidly, and bring large amounts of organic matter into the nest Wetterer 1995. When a designated worker scout discovers a new food source, the ant secret es a pheromone trail to guide fellow colony members to the plant Holldobler and Wilson 1990. The pheromone trail is sensed when an individual waves its antennae and picks up the chemical traces of the path originally traveled by the scout. As a plant's l eaves and flowers are harvested, successive ants lay their own pheromones, creating a strong chemical path Holldobler and Wilson 1990. The trail is subsequently cleared to soil level by designated castes of workers and soldiers to ensure unobstructed tra vel to the nest Howard 1991,
Hoyt 1996 Foraging trails can reach lengths of up to several hundred meters from the nests Hogue 1993. The plant material collected by A. cephalotes workers is cut into small pieces, and combined with feces to form a sticky mass. This mass is used to grow the colonies' fungus, possibly Leucocoprinus Basidimycota : Agaricaceae, and the swollen hyphae are used to feed brood Stevens 1983, Wilson 1971. Acromyrmex. coronatus colonies also cu ltivate what is assumed to be Leucocoprinus fungus for its larvae Wilson 1971. Workers in each species feed directly on plant sap, receiving only five percent of their food supply from the gongylidia of their cultivated fungus Holldobler and Wilson 1990 In larger colonies, A. coronatus exhibits trunk trail construction but lacks the resources to define trails clearly when the colony is smaller. In these smaller colonies only one or two foraging lines of ants spread from the nest. These foraging lines are presumably marked when a worker d rags its abdomen across the leaf litter with secretions from the Dufour's gland Evans 1984. Acromyrmex coronatus also constructs subsurface trails, which may be an alternate to visibly clear trails and may provide the colony with added protection. These two factors often lead to the assumption and field observation that Ac. coronatus doesn't construct trails Wetterer 1995. Extensive trail construction and maintenance are costly to both species but once cleared, trails are often used for months of foragi ng Howard 1991. The number of ants foraging and the speed they are able to travel along trails plays an important role in the survival of the colony. In this study traveling speeds were compared across the two species with the hypothesis that A. cephalot es, which regularly construct visible trails, would exhibit faster speeds. Furthermore, since A. cephalotes are larger ants and consistently travel on unobstructed trails, it was predicted that A. cephalotes would carry larger loads than Ac. coronatus for their body size. Specific trail clearing behaviors in A. cephalotes were also examined. It has been found in a previous study that no relationship exists between small distances under four meters from the nest and time to clear leaves von Waldburg 2000 However, it was hypothesized that when a larger spatial scale was looked at, a positive correlation would exist between distance from the nest and time to clear objects from the trail. Predictions for shorter distances from the nest included: activity of ants being greater, more ants being involved in trail clearing, and time involved in trail clearing being lower. Lastly, a comparison was made between two different size trail obstructions, expecting that removal of a larger object would require more ants and more tim e.
METHODS Study Sites This study was conducted in p remontane wet forest in Cerro Pl ano, Los Llanos, Monteverde, Puntarenas, Costa Rica, and Caitas Guanacaste, Costa Rica. Four A. cephalotes nests were located in pastures and regenerating forests in Los Llanos, two nests on Caitas farmland and secondary forest and the final two colonies were found in Cerro Pl ano secondary forest. Four Ac. coronatus colonies were found in Los Llanos secondary forest, one in a Caitas pasture, and two in secondary forest above the Estacin Biol gica de Monteverde. Acromyrmex coronatus foraging lines were the only trails evident and therefore colonies were assumed to be of small size. Colonies were found in warme r areas and, with the exception of one A. cephalotes trail destroyed by a landowner, were undisturbed throughout the duration of the study. Data Collection Data was taken over a period of 22 non consecutive days between April 12 and May 8. At each A. cepha lotes colony, the number of natural trail obstructions was recorded. At intervals of one, five, eight, 16 and 32 meters from the nest entrance, activity of workers, soldiers and minima was recorded for three minutes by counting the number of ants traveling out to forage. Two objects, a leaf and a twig, were placed on the trail at the same meter intervals. Leaves used were less than three inches long and twigs were no longer than four inches but possessed a diameter large enough to prevent cutting behavior. Number of soldiers and workers assisting in removal of the object were recorded at time intervals 1 5 10,20,30,45, 60, and 75 minutes until the object was remove d from the trail. An object was considered removed when it no longer impeded trail traffic, usually two to three inches off the trail. To compare across species, six ants from each of the fifteen colonies, A. cephalotes and Ac. coronatus, were timed over o ne meter as they returned to the nest with their harvest load. Times were taken as the ant traveled between four and three meters of the nest, to ensure equal activity levels across species since Ac. coronatus colonies did not exhibit extensive trails. Tim ed ants were then collected and weighed. The leaf material being carried was also weighed and data collected was used to determine traveling speed and the burden ratio leaf/ant weight for each individual. RESULTS A significant difference existed between traveling speed of A. cephalotes and Ac. coronatus Unpaired t test, p < 0.0001, A. cephalotes mean speed = 0.653 minutes/meter
Ac. coronatus mean speed = 0.349 min/m. A significant difference also existed between burden ratios leaf/ ant weight across species Unpaired t test, p = 0.0071, A. cephalotes mean ratio = 3.051, Ac. coronatus mean ratio = 2.037. When relative speed was combined with the burden ratio no significant difference existed between the two spec ies Unpaired t test, p = 0.0748. Mean ant and leaf weight, burden ratios, and maximum load carried across species can be found in Table 1. Mean data for the above interspecies relationships can be found in Figure 1. Natural trail obstructions averaged 16 sticks, 16 leaves, and 15 leaf fragments over a 32 meter trail section. Obstructions were found in higher concentrations farther away from the nest personal observation. Data for A. cephalotes colonies yielded no significant difference between type of o bject and the time required for its removal ANOVA, p = 0.1000, Figure 2. Significant differences existed overall between time and activity of workers and time and distance from the nest ANOVA, p = 0.0016, p = 0.0163, respectively, Fig. 2. All distance relationships, with respect to removal time, were significant with the exception of one and five post hoc Fisher's PLSD, p = 0.2577, five and eight p = 0.2018, and 16 and 32 meters p = 0.1335, Fig. 2. Maximum number of workers involved in trail clear ance was not significantly different for clearance of different trail obstructions ANOVA, p = 0.7309 or for distance from the nest p = 0.6312, Figure 3. Additionally, when the covariant of worker activity was considered, there was still no significance between maximum number of workers, and objects and distance ANOVA, p = 0.2643. Significant differences did exist between distances of one and 32 post hoc Fisher's PLSD, p = 0.0022, and five and 32 meters p = 0.0107, Fig. 3. A significant trend was s een between the maximum number of workers and the two objects used in trail obstruction post hoc Fisher's PLSD, p = 0.0523. Significant differences existed overall between worker, soldier, and minima activity and distance from nest ANOVA, p values < 0.0 001, Figures 4 5, 6 All specific worker/distance relationships are significant with the exception of one and five post hoc Fisher's PLSD, p = 0.0583, five and eight p = 0.0601, 8 and 16 p = 0.2510, and 16 and 32 meters p = 0.2243, Fig. 4. Soldier/distance relationships were only significant between one meter, over all distances post hoc Fisher's PLSD, p values < 0.007, Fig. 5. All minima/distance relationships were significant with the exception of one and five post hoc Fisher's PLSD, p = 0.3353, five and eight p = 0.1073, 8 and 16 p = 0.0563, and 16 and 32 meters p = 0.1867, Fig. 6. DISCUSSION The polymorphism in A. cephalotes colonies allows for multiple divisions of labor, which in turn allows a colony to exhibi t best fit foraging, where a worker harvests from the food resource to which her size is best suited Wetterer 1991. This behavioral pattern can be extended to trail clearance and used to explain why there is a significant difference
between removal time and distance from the nest Fig. 2. Closer to the nest more individuals are present, and subsequently a wider range of body sizes. This broad size range produces greater probability that the individual best suited for efficient removal of a certain size t rail obstruction will contact the object and remove it quickly. Results show that when more individuals are present to encounter obstructions, objects are removed faster than those objects which fewer ants contact those farther away from the nest. In a s tudy by von Waldburg 2000 no relationship was found between distance and removal time, but this can be attributed to the fact that her study only included distances of up to four meters from the nest, where individuals are still highly concentrated on th e trunk of the trail. But in support of her findings, no significance was found between specific distance relationships closer to the nest due to relatively equal worker activity. Time of object removal is therefore only significant across A. cephalotes tr ails on larger spatial scales. The non significance between the 16 and 32 meter removal times indicates the greater cost of sending workers to distant foraging sites. Natural trail obstructions were more abundant at farther distances from the nest suggesting that when A. cephalotes inve sts in harvesting food sources it is not as concerned with rapid trail clearance since the main cost of foraging is travel time Howard 1991. Keeping the trail constantly clear closer to the nest is a more economical investment because a greater number of ants travel this section of the trail. Higher frequency of contact with obstructions, and subsequent efficient removal, can also explain the significant relationship between time and activity of workers. When more workers are present, they will move an ob ject faster simply there is a higher probability that an ant will pass by. Since no significance exists between maximum number of workers and distance with respect to activity, no i nduced or recruiting response in used in trail clearance Fig. 3. This dem onstrates that workers are utilized uniformly in removal of all trial obstructions, regardless of their location on the trail. There was no significance in clearance time of different size, indicating that leaves and sticks are cleared at equal rates. Only one or two ants were involved in removal of a stick because an ant would simply grab an end with its mandibles and drag the stick across the trail. With leaf removal, many ants would stop and chew on the leaf, presumably to see if it was an acceptable foo d source. Many ants stopping were most likely feeding off the leaf nutrients rather than assisting in the removal of the obstruction Holldobler and Wilson 1990. This leaf sampling and feeding led to longer times of removal and added unanticipated numbers of workers to total counts. In addition, each leaf removal involved cutting and dragging by multiple ants. The cutting process also added time and helps to explain why removal of leaves was not significantly faster, and did not exhibit significantly lower worker counts than stick removal.
Significant relationships in activity of workers, soldiers and minima are explained by the trunk trail behavior of A. cephalotes Howard 1991. More ants overall are concentrated near the nest because of the utilization of one main trail, with individuals dispersing as trails branch off. This is supported by the significantly higher numbers workers, soldiers, and minima found near t he nest Fig. 4 5, 6 Dispersal of individuals can contribute to the explanation of the significant relationship between higher levels of activity and less time involved in obstruction removal. Concentrations of soldiers near the nest can be explained by their involvement in nest hole defense. Soldiers near the nest entrance can better protect a colony from invaders and can also regulate individuals that are entering the nest hole. The significant distance/minima indicate concentra tion near the nest because the minima focus on cleaning leaves before they enter the colony. Minima hitchhike on leaves and if they were concentrated farther from the nest their weight would add greater energy cost to workers as leaf matter is transported back to the colony Hoyt 1996 Foraging strategies across species of leaf cutting ants vary to partition available resources Wetterer 1995. Howard 1991 found the foraging strategy of trail construction demonstrated by A. cephalotes ants increases the ir running speed four to ten fold, allowing ants a traveling speed of one to two meters per minute, with laden workers traveling slightly slower. Atta cephalotes possesses faster running speeds that Ac. coronatus. Laden A. cephalotes ants observed possess ed increased running speeds predicted by Howard 1991, Fig. 1 but this increase on cleared trails may also be evident in larger Ac. coronatus colonies. Present results though demonstrate that A. cephalotes speeds are almost twice as great as Ac. coronatus supporting Howard's findings and the hypothesis that investing in a trail allows more efficient travel to harvest sites and therefore increased foraging benefits Wetterer 1995. A trail also allows more ants to reach the nest in a given time. In his comparison of Acromyrmex and Atta colonies, Howard 1991 found that foraging at greater distances involved little loss of net energy, with the main cost stemming from time taken to tr avel. In A. cephalotes colonies, ants are larger and possess a higher degree of polymorphism. Polymorphism allows division of labor, with workers specializing in tasks at which that are more efficient, which in turn increases overall efficiency of the colo ny Carroll & Janzen 1973, Wetterer 1995. Time cost of foraging at greater distances can be reduced through trail construction allowing ants to be recruited faster to foraging sites Wetterer 1995, Howard 1991. Also, polymorphism broadens the diversity o f resources available for harvesting because different size ants can cut different amounts of leaf toughness Wetterer 1991. Conversely, construction of trails is not economically worthwhile in the Ac. coronatus colonies studied because of small colony si ze, and subsequently fewer foragers Wetterer 1991, 1995. Efforts are more economical when focused on nearby, smaller vegetation rather than large trees, because of the narrow size range found in Ac. coronatus colonies. Individuals of this species are
jus t large enough to cut soft, herbaceous leaves and it would be inefficient for them to forage to large tress sources as A. cephalotes does Wetterer et al. 1998. Additionally, for Ac. coronatus colonies to send workers out to distant sites without the avai lability of rapid travel would incur large costs because less vegetation would arrive at the nest in a given time period. This study also found that A. coronatus workers were smaller but carried a much larger load weight than A. cephalotes. In contrast, Wetterer 1995 concluded that Ac. coronatus workers weigh less and exhibit a much smaller burden size Table 1. Possible explanations for the reduced A. cephalotes burden size found in this study include the finding that means burden size de creases with increasing mass Howard 1991, Wetterer 1991. Furthermore, forager size in A. cephalotes changes with colony size and sampling of various colony sizes with few individuals may have affected mean weight obtained Wetterer 1995. Explanations of why results indicated that Ac. coronatus foragers carry more weight across their trails can be explained with the idea that the species was attempting to compensate for slower travel by investing more energy in each harvesting trip. Since Ac. coronatus co lonies studied were small, and possibly expanding, a larger supply of leaf matter may have been required, forcing greater investments in foraging trips. These explanations point to the large influence colony size has on forager size and efficiency. Determi ning colony size is difficult, often requiring nest excavation, but may be necessary to further define these relationships. Finally, when relative speed was compared with burden ratio no significance was found because each species exhibits dominance in one variable speed or burden ratio creating a null relationship. A key factor in this study was colony size, especially in Ac. coronatus. Interspecific competition exists between the two species studied since foraging habits are rather similar Wetterer 199 5. Acromyrmex coronatus colonies are at a disadvantage against mature, larger A. cephalotes because of their sheer lack of numbers and a defensive soldier caste. Furthermore, aggressiveness of the colony increases with age and where mature A. cephalotes c olonies are present they often halt or severely limit the colony size of nearby leaf cutting species Carroll and Janzen 1973; Wetterer 1995. This limitation is a possible explanation for why only smaller Ac. coronatus colonies were found. In future studies, large and small colonies of Ac. coronatus should be looked at to determine if a true benefit, of increased and more efficient foraging rates, exist when a trail is constructed. Both size colonies of Ac. coronatus should also be lo oked at to support Wetterer's findings 1995 that Ac. coronatus does in fact carry a smaller burden than A. cephalotes. Foraging ecology of leaf cutting ants is understudied because of the difficulties in determining exact cost and benefit from trail con struction. Benefits of trail construction are spread over time since colonies use trails over long periods of time. Trail construction, and the foraging habits that result, influence niche partitioning of Costa Rican leaf cutter ants. Understanding foragin g, efficiency, and partitioning relationships
better will yield important insights into community ecology and leaf cutter ants' major role as herbivores in tropical forests. ACKNOWLEDGEMENTS I would like to thank Mauricio Garcia for his assistance in formulating this project, its many alterations and for supporting me through my "is this really happening to me the night before my paper is due and already written?" experience. Thanks also to An drew Rodstrom for bringing me smiles through the tough stuff and for his help in strengthening my meager Stat View skills; thanks to the families in Caitas and Los Llanos for use of their farms; and thanks to the Familia Santa Maria Brenes for their willi ngness to search far and wide for my "zompopas". And finally, let me not forget thanks to Club Party and the chicky titas that I roomed with because without them I would have had such an uneventful and serious month. = LITERATURE CITED Carroll, C.R. and D .H. Janzen 1973. Ecology of foraging by ants. Annual review of ecology and systematics. 4: 231 257. Evans, H. E. 1984. Insect Biology a textbook of entomology. Addison Wesley Publishing Company, Reading, Massachusetts, pp. 127 129. Hanson, P.E. and I. D Gauld 1995. The Biology of Hymenoptera and Economic Importance of Hymenoptera In: The Hymenoptera of Costa Rica. Oxford University Press, Oxford, England. Hogue, C. L. 1993. Latin American Insect and Entomology. University of California Press, Berkeley, California, pp. 446 449. Holldobler B. and E.O. Wilson. 1990. The ants. The Belknap Press of Harvard University Press, Cambridge, Massachusetts.' Howard, J .J. 1991. Resource quality and cost in the foraging of leaf cutter ants. In: Ant Plant Interactions. Huxley C.R. and D. F. Cutler eds.. Oxford Science Publications, Oxford, England, pp.42 50. Hoyt, E. 1996. The earth dwellers: adventures in the land of ants. Touchstone, New York, New York. Stevens, G.C. 1983. Atta cephalotes Zompopas, L eaf cutting ants In: Costa Rican Natural History. Janzen, D.H ed.. The University of Chicago Press Chicago Illinois. Von Waldburg, J. 2000. Removal time of obstacles and ant activity in Leaf cutter ant trails, CIEE student papers, summer 2000. Wetterer, J.K. 1991. Foraging Ecology of the leaf cutting ant Acromyrmex octospinosus in a Costa Rican rain forest. Psyche 98: 361 371. Wetterer, J.K. 1995. Forager size and ecology of Acromyrmex coronatus and other leaf cutting ants in Costa Rica. Oecologia 104: 409 415.
Wetterer, J.K., D.S. Gruner, and J.E. Lopez. 1998. Foraging and nesting ecology of Acromyrmex octospinosus Hymenoptera: Formicidae in a Costa Rican tropical dry forest. Florida Entomologist 81:61 67. Wilson, E.O. 197 1. The Insect Societies. The Belknap Press of Harvard University Press, Cambridge, Massachusetts.