1 Trail Fidelity in Atta cephalotes Amy Romer Department of Biology, Occidental College ABSTRACT Trail fidelity in Atta cephalotes was tested through marking experiments by comparing the movement of 500 marked workers that were removed but replaced on the same trail with the movement of a separate group of 500 marked workers that were displaced to a new trail. Trail fidelity was very high for replaced workers: 94% of workers observed at a census 24 hours after marking remained on the original trail. Trail fidelity was such a strong force in displaced ant behavior that by 24 hours after marking 74% had returned to their original trail. This indicates that worker ants are orientated to a specific trail within the colony, most likely due to orientatio n cues such as trail specific pheromones. Rate of disappearance of workers from their original trail was compared between replaced and displaced moved workers to illustrate that once displaced, ants return to their original trail, exhibiting similar trail movement behavior to the replaced ants. Lastly, the distance from the original trail to the new trail was significantly correlated with the proportion of switched ants returning to their original trail by 24 hours. (Spearman rank correlation, Rho = 0.8982, P = 0.0024, n = 8). This may suggest that orientation cues for specific trails are weaker with increased distance from the trail opening, reducing the number of ants able to return to their original trail. RESUMEN La fidelidad de sendero en At ta cephalotes fue determinada con experimentos de marca comparando el movimiento de los trabajadores marcados que fueron removidos y despuÃ©s substituidos en el mismo sendero con el movimiento de un grupo separado que fue substituido en un sendero diferente . La fidelidad de sendero de las hormigas que no fueron trasladadas fue muy alta: 94% de las trabajadoras observadas despuÃ©s de 24 horas de haber sido marcadas permanecieron en el sendero original. La fidelidad de sendero de fue tan fuerte en el comporta miento de las hormigas trasladadas que, 24 horas despuÃ©s de haber sido marcadas, 74% habÃan regresado a su sendero original. Esto indica que las hormigas trabajadoras estÃ¡n orientadas a un sendero especÃfico dentro de la colonia, muy probablemente debido a seÃ±ales de orientaciÃ³n tales como feromonas especÃficas del sendero o a seÃ±ales visuales. El Ãndice fue comparado entre las trabajadoras que no fueron trasladadas y las trasladadas para ilustrar que, despuÃ©s de haberse movido, las hormigas vuelven a su sendero original, exhibiendo comportamiento similar al movimiento de las hormigas que no fueron trasladadas. La distancia del sendero original al sendero nuevo estuvo correlacionada perceptiblemente con la proporciÃ³n de hormigas cambiadas que volvÃan a su sendero original (CorrelaciÃ³n de rangos de Spearman, rho = 0.8982, P = 0.0024, n = 8). Este resultado indica que las seÃ±ales de orientaciÃ³n para los senderos especÃficos son mÃ¡s dÃ©biles conforme aumenta la distancia a la abertura del sendero, reduciendo el nÃºmero de hormigas capaces de volver a su sendero original. INTRODUCTION Leaf cutter ants of the neotropical species Atta cephalotes are eusocial organisms that form expansive subterranean colonies. Workers collect plant material which is used to fa rm and fertilize a basidiomyocyte fungus Leucocoprinus gongylophorus that is fed to the developing ant larvae (HÃ¶lldobler and Wilson 1990). Atta cephalotes are highly
2 polymorphic with mature colonies containing three worker castes differentiated by size: soldiers, foragers, and minima (Janzen 1980). Workers forage on well established branching trails cleared of vegetation that can persist for months (HÃ¶lldobler and Wilson 1990). Trails are established by the laying of a short lived recruitment pher omone, Methyl 4 methylpyrrole 2 carboxylate, deposited by scout workers. Once the trail is established, an orientation pheromone is produced by foraging workers to maintain the trail (HÃ¶lldobler and Wilson 1990) Success of the colony is based on maximizin g the efficiency of fungus production through leaf collection and processing. Efficiency of leaf collection is based on maximizing the size of leaves collected, the quality of the resource, the speed at which workers return to the nest, and the time worke rs spend outside the colony foraging. The assertion that workers forage to maximize resource delivery is known as the optimal foraging hypothesis (Rockwell and Hubbell 1987, Roces 2002). Support of this theory n Atta sexdens the foraging caste is made up of ants with head widths of 1.8 2.4 mm, which was calculated to be the most efficient size in terms of net energetic yield in foraging. In a study by Rockwood and Hubbell (1987), colonies of Atta cephalotes wer e found to exhibit higher specificity and constancy in leaf choice when the average quality of resources was higher. This study also found that ants were more selective about foraging sites at greater distances from the nest. In this case, foraging effic iency is optimized by concentrating worker energy on high quality resources. In contrast, a variety of recent studies have found results not supporting the optimal foraging theory. Burd (1996) examined foraging efficiency in Atta colombica by measuri ng the size of leaves cut and collected by workers compared to the energy expended in this process. This study found that workers were collecting leaves smaller than those that would yield maximum efficiency. Workers have also been observed in numerous s tudies to bypass resources for those of equal quality farther from the nest, and often abandon resources before completely de foliating them (Cherrett 1983). These results have led to the development of the resource conservation hypothesis, which suggests that ants limit resource exploitation to conserve and manage resources over the course of the colony lifetime (Cherrett 1983, Roces 2002). One element of foraging efficiency deserving further examination in leaf cutting ants is worker trail fidelity. S hepherd (1982) conducted one of the only studies addressing trail fidelity in workers the genus Atta. He marked 3200 workers and observed their movement over the course of six days and found that workers showed short term trail fidelity: 74% of marked ind ividuals observed after one day were found on the same trail. Yet, by the end of six days, marked workers were equally likely to be found on a different trail as on their original trail. Trail fidelity can increase foraging efficiency in two main ways, by establishing worker familiarity with a trail, and by allocating specific individuals to a site based on the amount and quality or the resource. Visual, spatial, and chemical familiarity with a trail may allow workers to more efficiently navigate the trai l and select a cutting site, as was found to be the case with the foraging wood ant Formica aquilonia (Cosens and Toussaint 1985). Trail fidelity may also occur due to specialization of workers on a resource according to its quality, quantity, and physica l attributes. Increased quality of a resource has been found to elicit more active recruiting behavior by workers thus
3 increasing the number of new workers coming to the site (Roces 2002). Maintaining this high density of workers through trail fidelity w ould increase overall foraging efficiency of the colony by concentrating worker effort on a high quality resource. In addition, it is possible that workers are allocated to specific harvest specific resources based on their morphology. If this is the cas e, trail fidelity would increase efficiency by concentrating workers on resources that they would best suited to exploit. Lastly, trail fidelity may exist simply because colony architecture makes it quicker and easier for workers to enter the colony, depo sit leaves, and leave through the same trail entrance. Mature Atta cephalotes nests are highly compartmentalized and can contain hundreds of small fungus gardens, suggesting that leaf drop off most likely occurs close to trail entrances to the colony. Th erefore, time spent foraging would be maximized by simply entering and exiting through the same entrance. This study more thoroughly investigated the topic of trail fidelity by not only recording average trail fidelity of foraging workers as did Shepherd ( 1982), but also examining whether trail fidelity was due to active orientation to a trail or simply a passive response to the architecture of the colony. This was accomplished using marking experiments for two test groups at each colony: the first group of 500 foraging individual was marked and replaced on their original trail, while the second group of 500 was marked and displaced to a new trail. Censuses of the replaced ants provided information about the rate at which average worker switches trails. Movement of the displaced ants between trails provided information about the strength, nature, and cause of trail fidelity based on whether ants showed fidelity to their original trail, their new trail, or neither. Fidelity to their original trail would i ndicate that ants are specialized, or oriented to a specific trail. Fidelity to their new trail would indicate that compartmentalization of the colony causes workers to enter and exit from the same trail entrance due to the proximity of the leaf drop off site. Lack of a trend in trail fidelity would suggest that the disturbance by displacing workers causes them not to exhibit the same behavior that keeps their unmoved counterparts faithful to a given trail. ed that in addition to a trend of decreasing trail fidelity over time, the total number of marked ants observed on all trails decreased steadily over the six day study period. Rate of disappearance of marked ants, or rate of decay, may be due to a variety of factors including worker death, or cycling within the colony. This study used rate of decay to quality whether trail fidelity was a short or long term phenomenon. Lastly, this study examines the effect of distance moved from original to new trail on r ate of return of displaced ants to their original trail. This will be used to determine if the distance required to move within the colony has a significant effect on worker trail choice. MATERIALS AND METHODS Study Sites Trail fidelity experime nts were carried out on eight Atta cephalotes colonies in Monteverde and Santa Elena, Costa Rica in the Lower Montane Moist Forest Holdridge life zone between 1300 and 1470 m elevation (Hayes and Laval 1989) during October -
4 November, 2004. Four colonies we re found on the forest edge, including three colonies at the Ecological Farm in Monteverde and one in a forest fragment across the street from Hotel Don Taco in Santa Elena. Four colonies were found in pasture across from the Municipalidad de Monteverde i n Santa Elena. Marking Studies At each colony two marking experiments were conducted. Two trails extending from different colony entrances were selected and will be referred to as trails A and B, and the distance between the two trail openings was recor ded. Replaced Ants 500 foraging workers from trail A were marked on the abdomen with xylene based Ideal 563 Speedry paint marker one meter from the entrance to the colony, and replaced. Once marking was completed, ten minute censuses for marked ants w ere conducted at the opening of this trail, and a nearby trail B. Displaced Ants On trail B one meter from the entrance to the colony, 500 foraging workers were marked on the abdomen with a different color of paint, but in this case, they were displaced to a different trail, trail A. Once marking was completed, ten minute censuses for marked ants were conducted at the openings of both trails A and B. Censuses of both groups of ants were repeated the following morning, the following afternoon, the third , and the fourth morning. RESULTS Trail Fidelity of Replaced vs. Displaced Ants Ants replaced on their original trail exhibited high trail fidelity. An average of 94% of ants observed at the 24 hour census had remained on their original trail. Displac ed ants showed fidelity to their original trail with an average of 74% of ants observed at the 24 hour census had switched back to their original trail. Both groups show a trend of increased switching over time (Fig. 1). There was a statistically signifi cant difference in the amount of trail switching at the 24 hours census between the replaced and displaced test groups at the eight sites (Sign test, P < 0.01, n = 8). The average rate of trail switching at the 24 hour census was 6% for the replaced ants and 74% for the displaced ants, as stated above. By the time marking of displaced ants was completed, an average of 20% of displaced ants observed at the zero hour census had already returned to their original trail. Over the course of the next day di splaced ants continued to switch trails at a high rate because an average of 74% of ants observed at 24 hour census had returned to the original trail. There was a statistically significant difference in the percent of ants switching trails between the ce nsuses zero and 24 hours (Sign test, P < 0.05, n = 6).
5 Replaced ants showed only a slight increase in the average percentage of ants switching trails between the census directly after marking and the census at 24 hours. Directly after marking was complet ed, an average of 2% of ants observed in the ten minute censuses had switched trails, while at 24 hours, 6% had switched trails. An increase in trail switching between zero and 24 hours was observed in only five of the eight colonies studies. Therefore t he difference in percent of trail switching between zero and 24 hours for replaced ants was not significant (Sign test, P > 0.05, n = 8). Figure 1. Atta cephalotes workers that switched trails by zero, 24, and 48 hours after marking. Displaced ants, those that were transferred to a different trail, showed significantly higher percent of trail switching than replaced ants at all three censuses. In switching trails, displaced ants were returning to their original trail. Results were obtained from eigh t colonies in Monteverde and Santa Elena, Costa Rica. Decay In experiment one, the number of ants found on their original trail at each census was used to construct a decay curve of the average rate of disappearance of workers. Over the first 24 hours of data collection, there was a significant variation in rate of disappearance between the study sites between 3.1 ants/hour at Hotel Don Taco and 0.083 ants/hour at Ecofarm 1a. Yet after 24 hours, the rate of disappearance is similar for all sites varyi ng only from 0.714 ants/hour at Hotel Don Taco to 0.208 ants/hour at Santa Elena 2a (Fig. 2).
6 Figure 2. Presence of marked Atta cephalotes from the replaced study group on their original trail at six colony sites in Monteverde and Santa Elena, Costa Ric a. After 24 hours, the rate of disappearance for all six sites converges to an average of 0.55 ants lost/hour and varies only by Â± 0.2 ants/hour. Two sites were omitted due to limited data. The number of ants present on the original trail over all ce nsuses was plotted for both replaced and displaced ants at each site to determine whether the displaced ant group exhibited a normal rate of decay once they had returned to their original trail (Fig. 3). For the six colonies used in this comparison, displa ced ants showed an increasing presence on their original trail over the first 30 hours as they switched back from the new trail, but after that point, both displaced and replaced ants disappear at similar rates. The average rate of disappearance of displa ced ants from their original trail was 0.51 Â± 0.2 ants lost/hour after the 30 hour census, compared to replaced ants that showed an average rate of 0.55 Â± 0.3 ants lost/hour.
7 A) B) C) D ) E) F) Figure 3. Decay curves of the number of Atta cephalotes observed on their original trails during each ten minute censuses at six of the eight study sites: A) Ecofarm 1a, B) Ecofarm 2, C) Hotel Don Taco, D) Santa Elena 1, E)Santa Elena 2a, F) Santa Elena 3. For all sites, the replaced and displaced ants show a similar rate of decay after 30 hours. Two sites were omitted due to limited data. Trail Fidelity vs. Distance The proportion of displaced ants that switched back to their orig inal trail by the 24 hours census was negatively correlated with distance displaced (Fig. 4; Spearman rank correlation, Rho = 0.8982, P =.0024, n = 8). When ants were displaced to a trail 0.8 meters from their original trail, nearly all, 36 of 38, of ant s observed at the 24 hour census
8 were found on their original trail. Conversely, when ants were displaced 13 m apart, only half, four of eight, ants were found on their original trail at the 24 hour census. Figure 4. Correlation of proportions of Atta c ephalotes replaced on a different trail after marking that had returned to their original trail after 24 hours, and distance between the openings of the original and new trails at eight colonies in Monteverde and Santa Elena, Costa Rica. Note: there are t wo data points at distance two meters. The proportion of replaced ants switched to a new trail by the 24 hour census was generally smaller when the two trails were farther apart, but the correlation was not statistically significant (Spearman rank corre lation Rho = 0.5181, P = 0.1884, n = 8). For seven of the eight sites, the proportion of switched ants varied from only 0 and 0.08, while the distance between the old and new trails varied from two to 13 meters. Santa Elena 2a, in which the original and new trails were only 0.8 m apart, had a significantly different rate of switching at the 24 hours: almost one quarter, 0.23, of replaced ants had switched to a new trail. DISCUSSION Trail Fidelity of Replaced vs. Displaced Ants The results indicate t hat Atta cephalotes workers show strong trail fidelity in both the replaced and displaced test groups. Trail fidelity is so strong that displaced ants will attempt to reconnect with their original trail within the colony. A majority of displaced workers returned to their original trail by 24 hours, suggesting that some sort of cue, either visual, olfactory, or other, orients workers to a specific trail. Trail fidelity appears to a drive displaced worker behavior for at least 48 hours because there was ne t movement of displaced workers back to their original trail through the 48 hour census. Decay The rate of decay of replaced and displaced ants from their original trail were very similar after 24 hours. This suggests that once displaced ants ret urn to their original trail, they behave like the undisturbed replaced group. Trail fidelity appears to be strong only in the short term because displaced ants start to disappear from their original trail after 24
9 to 48 hours. The cause of disappearance from trails is unclear. There was a slight increase in the number of replaced ants viewed on other trails, but this cannot account for the total decay over the four days of data collection. Other factors may include worker death or loss of paint marking. Workers were never observed to have a partially lost mark, therefore it is unlikely that a significant number of workers lost their marks within four days. Trail Fidelity vs. Distance The distance ants were displaced had a negative impact on the numb er of ants finding their way back to their original trail by 24 hours. This may suggest that the cues that orient ants to their original trail are less frequently recognized and followed at larger distances from the original trail. For instance, workers displaced to a far trail may be visually unfamiliar with the new portion of the colony, reducing their ability to find their original trail. Atta cephalotes workers are known to deposit nest exit pheromones around trail openings to direct workers (Holldob ler and Wilson 1986). If these pheromone deposits are involved in worker orientation within the colony, workers will be less able locate their original trail when displaced large distances because the pheromone deposits will be less concentrated and there fore harder to follow. It is also possible that increased displacement distance affects only the time required to return to the original trail, but the total number of workers that will return to their original trail is the same. Replaced ants switched to trails varying distances from their original trail at a low, but relatively constant rate. This suggests that trail and resource specialization makes trail switching inefficient regardless of the distance of nearby trails. If the disadvantage of trai l switching was simply the reduction of foraging efficiency due to increased time spent within the colony, ants would switch to near trails with a much higher frequency than far trails, which was not the case in this study. Results from this study refut e the hypothesis that trail fidelity is due to minimizing the time spent within the colony, because displaced workers showed strong fidelity to their original trail, not their new trail. Short term trail fidelity even by workers displaced up to 13 m sugge sts that foraging efficiency is maximized when workers specialize on a trail. Specialization, supported by the optimal foraging hypothesis, would indicate that workers best suited to harvest a specific resource will continue to forage on only that resourc e until it is depleted or a better resource is found. Specialization could be based on worker size, i.e. larger workers harvesting tougher leaves. Yet, during my study both large and small workers were observed foraging on tough leaves. This may indicat e that either only a portion of workers specialize, or that other factors besides size determine how workers specialize. Lastly, specialization may be independent of worker characteristics and due only to workers establishing fidelity to whichever trail t hey are first recruited. While trail fidelity was found to be very strong in the short term in this study, it was unclear how long trail fidelity lasts. Marked ants show a high rate of disappearance from their original trail after 24 hours, but further s tudy is required to discover if the disappearance of marked individuals is due to worker death, cycling within the colony, the loss of paint marks, or movement to different foraging trails.
10 The mechanism by which displaced ants are able to locate and retu rn to their original trail is unknown but deserves further study to obtain a more thorough understanding of trail fidelity. It is likely that the same mechanisms that direct displaced ants to their original trail maintain trail fidelity in unmoved ants. Possible mechanisms include the presence or visual, spatial, or chemical cues that are trail specific. Since chemical cues, including recruitment and orientation pheromones, are known to be important in directing ants along foraging trails, they may also be involved in directing ants between trails. HÃ¶lldobler and Wilson (1990) assert that visual cues as well as pheromone trails guide foraging ants. Therefore visual and spatial cues may be important for maintaining trail fidelity. This study provided un ique information about trail fidelity in Atta cephalotes , but continued investigation is required to tease out how trail fidelity and allocation of workers within the colony affects foraging efficiency. ACKNOWLEDGEMENTS Thank you to Alan Masters and the CIEE students involved in preliminary studies in Atta trail fidelity. I am grateful to Javier MÃ©ndez, Alan Masters, Oliver Hyman, and Matt Gasner for their help in reviewing this paper with special thanks to Javier for his help with Spanish translation, a nd Alan for his help with organization of my study and statistical analysis. Thank you to the Ecological Farm in Monteverde and the Municipalidad de Monteverde for the use of their land in this study. LITERATURE CITED Burd, M. 1996. Foraging performan ce by Atta colombica, a leaf cutting ant. The American Naturalist . 148(4): 597 612. Cherrett, J.M. 1968. The foraging behavior of the Atta cephalotes (Hymenoptera, Formicidae) foraging pattern and plant species attacked in tropical rain forest. The Journa l of animal ecology. 37(2): 387 403. 1983. Resource conservation by the leaf cutting ant Atta cephalotes in tropical rain forest. In: Tropical Rain Forest: Ecology and Management . Eds. S.L. Sutton, T.C. Whitmore, and A.C. Cahdwick. Blackwell Scientific P ublications, Oxford, pp. 253 263. Cosens D. and N. Toussaint. 1985. An experimental study of the foraging strategy of the wood ant Formica aquilonia. Animal Behaviour. 33(2): 541 552. HÃ¶lldobler, B. and E.O. Wilson. 1986. Nest area exploration and recog nition in leaf cutter ants Atta cephalotes. Journal of Insect Physiology. 32(2) 143 150. 1990. The Ants . The Belknap Press of Harvard University Press, Cambridge, MA. Hayes, M and R Laval. 1989. The mammals of Monteverde . Tropical Science Center, Montev erde, Costa Rica. Roces, F. 2002. Individual complexity and self organization in foraging by leaf cutting ants. Biology Bulletin. 202: 306 313 Rockwood, L.L. and S.P. Hubbell. 1987 Host plant selection, diet diversity, and optimal foraging in a tropical le af cutting ant. Oecologia . 74: 55 61. Shepherd, J.D. 1982. Trunk trails and the searching strategy of a leaf cutter ant, Atta colombica. Behavioral Ecology and Sociobiology . 11:77 84. Stevens, G.C. 1983. Atta cephalotes (Zompopas, Leaf Cutting Ants). In: C osta Rican Natural History. Ed. D.H. Janzen. The University of Chicago Press, Chicago, IL, pp. 688 690. Wilson, E.O. 1980. Caste and division of labor in leaf cutting ants (Hymenoptera:Formicidae: Atta ), I: The overall pattern in A. sexdens . Behavioral Eco logy and Sociobiology . 7(2): 143 156 .
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xlink http:www.w3.org1999xlink xmlns:xsi http:www.w3.org2001XMLSchema-instance
leader 00000nas 2200000Ka 4500
controlfield tag 008 000000c19749999pautr p s 0 0eng d
datafield ind1 8 ind2 024
subfield code a M39-00388
Fidelidad del sendero en Atta cephalotes
Trail fidelity in Atta cephalotes
Trail fidelity in Atta cephalotes was tested through marking experiments by comparing the movement of 500 marked workers that were removed but replaced on the same trail with the movement of a separate group of 500 marked workers that were displaced to a new trail. Trail fidelity was very high for replaced workers: 94% of workers observed at a census 24 hours after marking remained on the original trail. Trail fidelity was such a strong force in displaced ant behavior that by 24 hours after marking 74% had returned
to their original trail. This indicates that worker ants are orientated to a specific trail within the colony, most likely due to orientation cues such as trail specific pheromones. Rate of disappearance of workers
from their original trail was compared between replaced and displaced moved workers to illustrate that once displaced, ants return to their original trail, exhibiting similar trail movement behavior to the replaced ants. Lastly, the distance from the original trail to the new trail was significantly correlated with the proportion of switched ants returning to their original trail by 24 hours. (Spearman rank correlation,
Rho = -0.8982, P = 0.0024, n = 8). This may suggest that orientation cues for specific trails are weaker with increased distance from the trail opening, reducing the number of ants able to return to their original
La fidelidad del sendero en Atta cephalotes fue determinada con experimentos de marca comparando el movimiento de los trabajadores marcados que fueron removidos y despus substituidos en el mismo sendero con el movimiento de un grupo separado que fue substituido en un sendero diferente. La fidelidad del sendero de las hormigas que no fueron trasladadas fue muy alta: 94% de las trabajadoras observadas despus de 24 horas de haber sido marcadas permanecieron en el sendero original. La fidelidad de sendero fue tan fuerte en el comportamiento de las hormigas trasladadas que, 24 horas despus de haber sido marcadas, 74% haban regresado a su sendero original. Esto indica que las hormigas trabajadoras estn orientadas a un sendero especfico dentro de la colonia, muy probablemente debido a seales de orientacin tales como feromonas especficas del sendero o a seales visuales. El ndice fue comparado entre las trabajadoras que no fueron trasladadas y las trasladadas para ilustrar que, despus de haberse movido, las hormigas vuelven a su sendero original, exhibiendo comportamiento similar al movimiento de las hormigas que no fueron trasladadas. La distancia del sendero original al sendero nuevo estuvo correlacionada perceptiblemente con la proporcin de hormigas cambiadas que volvan a su sendero original (Correlacin de rangos de Spearman, rho = -0.8982, P = 0.0024, n = 8). Este resultado indica que las seales de orientacin para los senderos especficos son ms dbiles conforme aumenta la distancia a la abertura del sendero, reduciendo el nmero de hormigas capaces de volver a su sendero original.
Text in English.
Costa Rica--Puntarenas--Monteverde Zone
Alimentos de animales
Costa Rica--Puntarenas--Zona de Monteverde
Tropical Ecology Fall 2004
Ecologa Tropical Otoo 2004
t Monteverde Institute : Tropical Ecology