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 M39-00246
Snowden, Robert, O.
Forrajeo ptimo de las zompopas (Atta cephalotes)en la seleccin de hojas basado en la proximidad del nido
Optimal foraging in leafcutter ants (Atta cephalotes) selection of leaves based on proximity to the nest
The foraging tendencies of Atta cephalotes, a leaf-cutting ant, have been widely studied. Optimal foraging theory
dictates that ants should maximize rate of energy intake while minimizing costs by harvesting from suitable food sources closer to the nest. However, actual foraging patterns are more disparate and do not always reflect optimal strategies. Here I measure harvesting rates and recruitment speeds at different distances from the nest of A. cephalotes colonies by offering six leaf disk samples every 15 minutes at 5, 10, and 15 meters from the nest entrance. Harvesting rate was significantly higher closer to the nest, with 54% of leaf disks selected at 5 meters. Recruitment was also faster closer to the nest, as ants selected leaf disks within the first three minutes in 63% of the trials at 5 meters. Communication between workers may determine selection of closer leaf offerings and thus optimize foraging on a colonial level. Other factors, such as nutritional qualities from a plant, can influence foraging choices, but when leaf offering vary solely by distance, A. cephalotes exhibit time and distance-maximizing foraging patterns.
Las tendencias de forrajeo de Atta cephalotes, una hormiga cortadoras de hojas, han sido ampliamente estudiadas. La teora del forrajeo ptimo dicta que las hormigas pueden maximizar la energa que obtienen mientras minimizan los costos al cultivar fuentes de alimento adecuados cercanos al nido. Sin embargo, los patrones de forrajeo actuales son distintos y no siempre reflejan las estrategias ptimas. Aqu yo mido las tasas de recoleccin y velocidades de reclutamiento a diferentes distancias del nido de las colonias de A. cephalotes, al ofrecer seis muestras de hojas en forma de disco cada 15 minutos a 5, 10 y 15 metros de la entrada del nido. Las tasas de recoleccin fueron significativamente mayores cerca del nido, con un 54% de las hojas seleccionadas a 5 metros. Tambin, el reclutamiento fue ms rpido cerca del nido, a medida que las hormigas seleccionaron las hojas dentro de los primeros tres minutos en 63% de las pruebas hechas a 5 metros. La comunicacin entre las obreras puede determinar la seleccin de las hojas cercanas ofrecidas y por lo tanto optimizar el forrajeo a nivel de la colonia. Otros factores, como las cualidades nutricionales de la planta, pueden influenciar la escogencia del forrajeo; pero cuando la hoja ofrecida varia solamente por distancia, A. cephalotes exhibe patrones de forrajeo maximizados en tiempo y distancia.
Text in English.
Costa Rica--Puntarenas--Monteverde Zone--Cerro Plano
Alimentos de animales
Costa Rica--Puntarenas--Zona de Monteverde--Cerro Plano
Tropical Ecology Spring 2011
Foraging behavior in animals
Ecologa Tropical Primavera 2011
Conducta de alimentacin en animales
t Monteverde Institute : Tropical Ecology
Optimal foraging in l eafc utter a nts ( Atta cephalotes ) : s election of leaves based on proximity to the n est Robert O. Snowden Department of Biology, University of Puget Sound ABSTRACT The foraging tendencies of Atta cephalotes a leaf cutting ant, have been widely studied. Optimal foraging theory dictates that ants should maximize rate of energy intake while minimizing costs by harvesting from suitable food sources closer to the nest. However, actual foraging patterns are more disparate and do not always reflect optimal strategies Here I measure harvesting rates and recruitment speeds at different distances from the nest of A. cephalotes colonies by offer ing six leaf disk samples every 15 minutes at 5, 10, and 15 m eters from th e nest entrance. Harvesting rate was significantly higher closer to the nest, with 54% of leaf disks selected at 5 meters Recruitment was also faster closer to the nest, as ants selected leaf disks within the first three minutes in 63% of the trials at 5 meters Communication between workers may determine selection of closer leaf offerings and thus optimize foraging on a colonial level Other factors, such as nutritional qualities from a plant, can influence foraging choices, but when leaf offering vary so lely by distance, A. cephalotes exhibit time and distance maximizing foraging patterns. RESUMEN Las tendencias de forrajeo de Atta cephalotes una hormiga cortadoras de hojas, han sido ampliamente estudiadas. La teora de forrajeo ptimo dicta que las hormigas pueden maximizar la energa que obtienen mientras minimizan los costos al cultivar de fuentes de alimento adecuados cercanos al nido. Si n embargo, patrones de forrajeo actuales son distintos y no siempre reflejan estrategias ptimas. Aqu yo mido tasas de recoleccin y velocidades de reclutamiento a diferentes distancias del nido de colonias de A. cephalotes al ofrecer seis muestras de h ojas en forma de disco cada 15 minutos a 5, 10 y 15 metros de la entrada del nido. Las tasas de recoleccin fueron significativamente mayores cerca del nido, con 54% de hojas seleccionadas a 5 metros. Tambin, el reclutamiento fue ms rpido cerca del nid o, a medida que las hormigas seleccionaron hojas dentro de los primeros tres minutos en 63% de las pruebas a 5 metros. La comunicacin entre obreras puede determinar seleccin de hojas cercanas ofrecidas y por lo tanto optimizar el forrajeo a nivel de co lonia. Otros factores, como cualidades nutricionales de la planta, pueden influenciar la escogencia de forrajeo; pero cuando la hoja ofrecida varia solamente por distancia, A. cephalotes exhibe patrones de forrajeo maximizados en tiempo y distancia. INTRO DUCTION O PTIMAL FORAGING THEO RY energy intake when searching for a nd obtaining food (MacArthur & Pianka 1966). Organisms that optimize by gaining the most energy, while minimizing time spent foraging, should be favored by natural selection, as maximizing energetic efficiency can lead to greater fitness. Over evolutionary time, species should refine mechanisms to improve their energy intake to output rat io when foraging. (MacArthur & Pianka food choice, its choice of which area to feed in, time allocation in different areas, and optimal speed of movements (Pyke et al. 1977). Foraging becomes a more complex process in social insect colonies, wh ere both individual and social aspects interact to determine food intake. Not
only must individuals collect and transport food, but they must also transmit information about its source to others (Roces & Hlldobler 1994). The leaf cutting ant Atta cephalo tes (Formicidae: Myrmicinae ) is a social herbivorous insect. Rather than directly consuming vegetation, they harvest leaves from trees to cultivate a fungus upon which they feed, relying on systems of cleared trunk trails le a ding to resource plants that e nhance food intake rate back to the central nest ( Traniello 1989, Kost et al. 2005). They also exhibit division of labor among size based castes and remarkable task partitioning for foraging activi t i es ( Wilson 1980 ) Foraging behavior of eusocial leaf cutting ants can be influenced by many factors, and A. cephalotes foraging tends to be quite complex, with the distribution and quantity of selected plant species often varying among colonies ( Rockwood 1976 Traniello 198 9 ). Overall, A. cephalotes colonies are generalist herbivores, typically 80% of trees in a location (Rockwood 1977). However, colonies seem to focus most of their fo raging effort on a limited n umber of plant species, with one third of harvested plant species consisting up to of 87% of diet (Rockwood and Hubbell 198 7) In light of their complex foraging behavior, whether A. cephalotes optimally for ages is an important question. If optimal foraging theory does apply to an A. cephalotes colony, foraging workers should maximize their energy budgets by selecting preferred trees closer to the nest. According to Traniello (1989), leaf cutters are efficient by cutting trees closer to nests and optimize their diet by specializing on certain tree species when they are foraging in resource rich areas and by generalizing on tree species when the quality of reso urce trees are low However, other evide nce suggests that trees can be selected independent ly of distance from the nest, and that foraging is not evenly distributed thus ants are not minimizing retrieval costs (Rockwood 1976). Ants have even been observed to forego resource trees near the nest for an individual of the same species further away (Rockwood and Hubbell 1987) Workers could abandon trees farther from the nest for others closer to the nest if these closer trees, in fact, higher rates of energy return (Roces 2002), in line with optimal foraging theory. Colony foraging is further complicated when taking into account its individual components. If colonies are foraging optimally as a collective sum of its parts, then individuals should optimize performance. Workers face a trade off been t ime spent harvesting and time invested to recruiting nestmates to a new food source (Roces and Hlldobler 1994). Caste divided systems where specific workers have specific tasks in theory, allow workers to more efficiently partition leaf harvesting based on body size, as larger workers harvest larger loads ( Rudolph 1986, Wetter er 1990 ). However, it has been shown that workers do not maximize individual foraging efficiency, typically taking suboptimal leaf fragment size s ( Rudolph 1986, Burd 2000). Previous research has been inconclusive in determining whether energetic optimization applies t o leafcutter foraging strategies In this study I examine whether A. cephalotes optimally forage on equally sized leaf fragment samples based on distance from th e nest using tree species harvested by the colony If the ants maximize foraging efficiency, then palatable leaf fragments closer to the nest entrance should be selected more often than similar fragments further away. Additionally, if the theory applies, recruitment to leaf samples may be faster closer to the nest as well.
METHODS Study Sites and L eaf S election Three ant colonies near Cerro Plano, Costa Rica were observed between April 17 th and May 3 rd 2011, in secondary, premontane moist forest. Two colonies were located in Valle Escondido, at about 1350 meters above sea level (masl) and one at the property of Frank Joyce at about 1380 m asl Experiments at t he two Valle Escondido colonies occurred after 1700 hours wh en the colonies were active and foraging active both day and night was tested in the afternoon. Trail segments where experimental procedure occurred tended to be in human made clearings with relatively little surrounding undergrowth a nd leaf litter. Ant foraging trails were followed to determine which tree ind ividuals the ants were foraging from and one tree individual for each colony was selected for leaf samples Selected trees were identified to species by Willow Zuchowski and Bill Haber. Leaves were obtained from the approximately the same area of the tree Experimental T rials Leaf disks 1 cm in diameter were made from sample leaves using a hole puncher Disks were offered to the ants at approximately 5, 10, and 15 mete rs from the nest entrance on the same trail that led to the tree used for samples. For each trial, six disks were placed several centimeters apart in a column parallel to the trail Disks were placed far enough from the center of the trail to avoid interf erence with ants, but still within the flow of traffic on the trail. After 15 minutes, the number of leaves taken at each distance was rec orded and any remaining leaf disk s were removed. Six more leaf discs were subsequently placed along the trail at each site, and the 15 minute trial was repeated. To determine differences in recruitment, every trial in which at least one leaf disk was taken within the first three minutes was recorded for each distance. Occasionally, ants were obse rved to purposely move lea f disk s away from the trail to the leaf litter edge; because they were not being taken to the nest they were not counted as removed leaf disk s. This behavior was recorded every time it occurred at each distance RESULTS Species Identities of Leaf Samples Leaves from a different tree species were used for each colony. One colony at Valle Escondido was offered leaves from a Viburnum costaricanum ( Caprifoliaceae) individual 26m from the nest, and the other Oreopanax xalapensis ( Araliaceae) from an individual 35m away. The Symphonia globulifera ( Clusiaceae) from an individual 52m from the nest. Harvesting Rate Overall, 107 15 minute trials were run at each distance from the nest for a total of 320 trials The re were differences in leaf disk extraction rate among the different colonies; Valle Escondido
colony 1 had the highest pickup rates. However, all colonies picked up the greatest percentage of disks 5 m from the nest entrance Distance from the nest had a significant effect on leaf extraction (One Way ANOVA, F 2, 320 = 14.7, P < 0.0001). For each trial ants 5 meters from the nest e ntrance took 53.6% 3.3 of leaf disk s on average (Figure 1) At 10 meters from the nest, ants remov ed 43 .4% 2.8 of leaf disk s were per trial, while at a distance of 15 meters, only 30.7% 2.8 of leaf disk s were removed Post hoc tests revealed the pickup rates of disks were significantly different at each distance (Tukey Kramer HSD, P < 0.05). FIGURE 1 Harvesting rate of A. cephalotes at three distances from the nest entrance, represented by mean percent leaf disks picked up per 15 minute trial (n = 107 trials at each distance, 6 disks offered per trial). Data were recorded at three different c olonies near Cerro Plano two at Valle Escondi do and one on per 15 minute trial varied significantly among distances with highest pickup rates occurring closest to the nest. Error bars represent one sta ndard error of the mean Recruitment Speed Distance also had a significant effect on the recruitment speed of ants ( Chi squared Goodness of Fit, X 2 = 6.65, df = 2 P = 0.036 ), as ants tended to recruit faster to disks closer to the nest. At 5 meters f rom the nest entrance, leaf disk s were extracted within three minutes of placement 63.3% of the time (Figure 2) At a distance of 10 meters, leaf disk s were taken within three minutes 46.5% of the time, while 15 meters from the nest, leaf disk s were taken within three minutes only 38.4% of the time. 0 10 20 30 40 50 60 5 10 15 % Leaf Disks Harvested/15 minute trial Distance From Nest Entrance (m)
FIGURE 2. Recruitment speed of A. cephalotes at three distances from the nest entrances represented by percentage of trials in which leaf disks were picked up within the first three minutes (n = 99 trials at each distance). Data were recorded at three different colonies, two at Valle Escon dido and o ne on property near Cerro Plano Recruitment speed varied among distance s with leaf removal occurring within the first three minutes most frequently closest to the nest. Rejection Frequency Ants were also obse rved to purposely move leaf d isk s away from the trail to the edge of leaf litter in some cases, and the frequency of this rejection varied significantly by distance ( Chi squared Goodness of Fit, X 2 = 19.77 df = 2 P < 0.0001 ), as ants further from the nest were more likely to do this. Ants at a distance of 5 meters and 10 meters f rom the nest entrance moved disk s away from the trail for only 10.8% and 15.7% of the trials, respectively (Figure 3) Howeve r, at a distance of 15 m disk s were moved to trail margins 31.3% of the tim e. 0 10 20 30 40 50 60 70 5 10 15 Percent Trials With Pickups Within First 3 Minutes Distance From Nest Entrance (m)
FIGURE 3. Rejection frequency A. cephalotes at three distances from the nest entrance, represented by percentage of trials in which leaf disks were purposely moved to the sides of the trail by ants (n = 83 trials at each distance). Data were recorded at three different colonies, t wo at Valle Escondido and one on Frank at Cerro Plano Rejection frequency varied among distance s with leaf disks moved to trail margins more frequently farther from the nest. Additional Observatio ns Each colony tended to have several trunk trails, and harvested trees tended to be betwe en about 10 100 meters from the nest; the majority of resource trees were within 60 meters of the nest. The observed foraging area and distance of resource trees fro m he nest agreed with previous observations (Rockwood 1976, 1977). Leaf disks were always removed by t he smallest workers on the trail, the minims, consistent with studies of caste partitioned foraging (Rudolph 1986, Wetterer 1990), as the leaf disks were smaller than cut leaf fragments carried by larger workers. Additionally, the first ants to encounter leaf disks would not pick them up, but appeared to inspect them, before other ants arrived to remove the disks. DISCUSSION With higher leaf pickup rate as we ll as faster recruitment closer to the nest, A. cephalotes colonies favored leaf samples in close proximity to the nest These preferences re flect an optimal foraging strategy, as selecting nearby food source s minimizes search time and travel distanc e, and thus is more energy efficient for the colony Maximizing forag ing strategy in this way agrees with the optimal foraging hypothesis proposed by Traniello 1989) That the relatively small leaf disks were always harvested by smaller ants reflects evide nce of efficient partitioning of foraging tasks based on caste size (Rudolph 1986, Wetterer 1990) Previous research typically has explored optimal foraging patterns on a larger scale level examining host plant selection. However, hold ing constants of leaf size, leaf species and other factors equal, we see that ants exhibit a preference for harvesting leaves based on proximity from the nest. 0 5 10 15 20 25 30 35 5 10 15 Percent Trials With Disks Moved to Trail Edge Distance From Nest Entrance (m)
It should be noted that leaf samples were offered much closer to the nest than the act ual resource tree from which t hey were obtained I f workers searching for foraging material are distributed evenly throughout the trunk trail it would seem that leaf samples should be sel ected at the same rate regardless of distance. F oragers could sele ct leaves when they happened upo n them ; it would be energetically optim al on an individual level (Roces and Hlldobler 1994 ) Perhaps forager traffic was less dense as far as 10 15 meters from the nest, explaining foraging preference at 5 meters, but ants encountered leaf disks soon afte r placement regardless of distance The results indicate that there is a more complicated foraging dynamic within the colonies. A tta cephalotes like many social insects, rely on pheromones for communicati on, which can be used to transmit information about food sources and recruit workers to resource trees (Jaffe and How s e 1979, Traniello 1989 Farji Brener et al. 2010 ). The number of workers recruited to a food source depends on its quality, and by varying pheromone output ants can re gula te forager density on trunk trails and thus foraging efficiency (Jaffe and How s e 1979 Farji Brener et al. 2010 ). Because suitable resource trees can be widely dispersed, time spent searching for new food sources limits potential for optimizing foraging e fficiency (Rockwood and Hubbell 1987). Thus, w hen leaf samples are offered along the trunk trail near the nest, foragers have quick access to food information then more distant food sources. The closer the food is to the nest, the faster information exchan ge and recruitment to the new food source can occur faster. Research indicates that workers are more likely to harvest a leaf the sooner they arrive to it (Shepherd 1982), and that outbound ants are more likely to find and select food after colliding into a returning worker with the same type of food (Farji Brener et al 2010). These can explain higher selection rates and recruitment closest to the nest, as ants would have encountered the most leaf disk carrying workers in the first five meters of the trail Though ants may not optimize as individuals, pheromone signaled r ecruitment can enable more efficient foraging on a colony level by directing more ants to the nearest food source in order to minimize foraging time. Additionally, further from the nest, it may have been more energetically valuable for an ant to continue to and harvest from the actual food source (the tree), then to select a relatively small food item. Occasionally, workers went out of their way to cl ear introduced leaf disks from the trail off to the edge of the leaf litter T his form of rejection must be considered That workers would spend time and energy moving a suitable food item away from the trail seems counterintuitive from an efficiency stand point. An explanation pertinent to optimal foraging theory may reflect communication between workers. Ants further away may receive information about food available closer to the nest and forgo harvesting and thus remove unneeded leaf disks from the trail If unneeded, removing t hese disks may have been a behavior done to clear the trails of obstructions to maintain efficient traffic flow to and from food sources (Roces and Hlldobler 1994 Howard 2001 ). Trails also tended to be narrower at sites 15 m from the nest e ntrance than the sites closer, so keeping clear pathways may have been more important where traffic is constrained. However, the leaf disks were relatively small, and would not seem to be an impediment to the ant s. Evidence indicates leafcutters will remove trail obstacles as small as 0.02 g, and that doing so only incurs low energetic cost (Howard 2001 ). Unused leaf disks may not have been significant obstructions, but ants would not compromise efficiency in removing them either The f oraging st rategy for the colony as a whole often involves other factors that can not be entirely explained by experimental studies; actual patterns are more complex. Atta cephalotes
often forage in patchy environmen ts, where high quality resour ces can be sparsely distributed (Kost et al. 2005). P references for certain plants may be influe nced by more than just distance, as so me studies indicate leafcutters occasionally avoid apparently suitable trees close to the nest (Rockwood and Hubbell 1987 ). Rather than optimi ze efficiency by m inimizing foraging time at nearby trees, they may be selecting leaves to maximize nutritional benefits especially for fungal growth Leaf preference can be influenced by levels of secondary chemistry, water content, leaf toughness, nutri tional payoffs, and even salt content within the foliage (Hubbell et al. 1984 Berish 1986, Howard 1987, Nichols Orians & Schultz 1989, Farji et al 2010). Taking into account such qualities can help explain foraging strategy on the colonial level in a varied habitat Another h ypothesis explains that ants may avoid closer tre es to conserve resources (Cherrett 1983 ) reflecting longer term optimization over the lifetime of a colony (Shepherd 1982) Applying optimal foraging model s to A. cephalotes is further complicated by behavioral limit ations unpredictable resource distribution. Some models assume that forager knowledge of resource base is required for optim al strategy, but individual leafcutters typically have limited knowledg e of food sources (Traniello 1989) It is difficult to be optimal when many potential resource plants are unknown. Even mechanisms like trunk trails, which improve leaf transport efficiency, have trade offs in that they restrict ability to search for new f ood sources (Traniello 1989). Though this study did not consider l arger scale environmental foraging patterns, it reveals aspects of optimal foraging on a smaller scale, independent of the availability and distribution of re territory. Leafcutters may discriminate between plants due to nutritional qualities or other demands, but we see definitive foraging trends when variations in leaf type, size, and ease of extraction are constant. When offered equivalent selections of leav es, A. cephalotes exhibit a preference determined by proximity to the nest. This behavior is not necessarily the most efficient on an individual level, but reflects an optimization oriented foraging strategy for the colony as a whole. ACKNOWLEDGEMENTS I thank Anjali Kumar for her advising, statistics crunching, and guidance with my research. The rest of the CIEE Staff deserves thanks for their h elp as well, including Moncho Clderon and Gisella Fern ndez for supplying equipment. I thank Frank Joyce and th e staff at Valle Escondido for access to their respective properties, and Willow Zuchowski and Bill Haber for their plant identification knowledge. Flor and Ramn Solis Araya were exceptionally hospitable in providing comfortable accommodations and delicio us food during my homestay. Lastly, I thank Tim Curry for his accompaniment and photographic documentation of my research, and of course to the rest of my classmates for their perpetual morale boosting. LITERATURE CITED B ERISH C.W 1986. Leaf cutting ants select nitrogen rich forage. American Midland Naturalist. 115: 268 276. B URD M. 2000. Foraging behavior of Atta cephalotes (leaf cutting ants): an examination of two predictions for load selection. Animal Behaviour. 60: 781 788. C HERRE TT J.M. 1983. Resource conservation by the leaf cutting ant Atta cephalotes in tropical rainforest. In S.L.Sutton, T.C. Whitmore, and A.C. Chadwick (eds.). Tropical Rain Forest: Ecology and Management pp. 253 263. Blackwell Scientific Publications, Oxford UK.
C HERRETT J. M 1989. Leaf cutting ants, biogeographical and ecological studies. In H. Lieth and M. Werger (Eds.). Ecosystem of the World 14b, Tropical Rain F orest Ecosystems, pp. 473 488. Elsevier, Amsterdam, NL. F ARJI B RENER A 2001. Why are leaf cutting ants more common in secondary forests than in old growth tropical forests? An evaluation of the palatable forage hypothesis. Oikos. 92: 169 177. F ARJI B REN ER A., S. A MADOR V ARGAS F. C HINCHILLA S. E SCOBAR S. C ABRERA M.I. H ERRERA AND C. S ANDOVAL 2010. Information transfer in head on encounters between leaf cutting wo rkers: food, trail condition or orientation cues? Animal Behavior. 79: 343 349. H OWARD J.J 1987. Le afcutting ant diet selection: t he role of nutrients, w ater, and secondary chemistry. Ecology. 68:503 515. H OWARD J.J. 2001. Costs of trail construction and maintenance in the leaf cutting ant Atta columbica Behavioral Ecology and Sociobiology. 49: 348 356. H UBBELL S.P., J.T. H OWARD AND D.F. W IEMER 1984. Chemical leaf repellency to an attine ant: seasonal distribution among potential host plant species. Ecology, 65: 1067 1076. K OST C., E.G. DE O LIVEIRA T.A. K NOCH AND R. W IRTH 2005 Spatial temporal permanence and plasticity of foraging trails in young an mature leaf cutting ant colonies ( Atta spp.). Journal of Tropical Ecology. 21: 677 688 J AFFE K., AND P. E. H OWSE 1979. The mass recruitment system of the leaf cutting ant, Atta ceph alotes. Animal Behavior. 27: 930 939. M AC A RTHUR R. H. AND E. R. P IANKA 1966 On optimal use of a patchy e nvironment. The American Naturalist 100: 603 609. N ICHOLS O RIANS C.M. AND J.C. S CHULTZ 1989. Leaf toughness affects leaf harvesting by the leaf cutter ant, Atta cephalotes Biotropica. 21: 80 83. ONNELL S., J.M. G ARCA J. B EARD T. C HIWOCHA D. L EWIS C. L IU H. P HILLIPS AND T. W ILLIAMS 2010. Leaf cutter ants ( Atta cephalotes ) har vest baits offering sodium chloride rewards. Insectes Sociaux. 57: 205 208. P YKE G. H., H.R. P ULLIAM AND E. L. C HARNOV 1977 Optimal foraging: a selective review of theory and tests. The Quarterly Review of Biolog y 52: 137 154. R OCES F. 2002. Individual complexity and self organization in foraging by leaf cutting ants. The Biological Bulletin 202: 306 313. R OCES F., AND B. H LLDOBLER 1994 Leaf density and a trade off between load size s election and recruitment behavior in the ant Atta cepha lotes. Oecologia 97: 1 8. R OCKWOOD L 1976. Plant selection and foraging patterns in two species of leaf cutting ants ( Atta ). Ecology. 57: 48 61. R OCKWOOD L. 1977 Foraging patterns and plant selection in Costa Rican le af cutting ants. Journal of the New York Entomological Society. 85: 222 233. R OCKWOOD L., AN D S.P. H UBBELL 1 98 7. Host plant selection, diet diversity, and optimal foraging in a trop ical leafcutting ant. Oecologia. 74: 55 61. R UDOLPH S.G. AND C. L OUDON 1986. Load size selection by foraging leaf cutter ( A tta cephalotes ). Ecological Entomology. 11: 401 410. S HEPERD J. D. 1982. Trunk trails and the searching strategy of a leaf cutter ant, Atta colombica Behavioral Ecology and Sociobiology. 11: 77 84. T RANIELLO J 1989. Foraging stra tegies of a n ts. Annual Review of Entomology 34:191 210. W ETTERER J.K. 1 990 Load size determination in the leaf cutting ant, Atta cephalotes Behavioral Ecology. 1: 95 101. W ILSON E.O. 1980. Caste and division of labor in leaf cutter ants (Hymenoptera: Formicidae: Atta) Behavioral Ecology and Sociobiology. 7: 143 156.