Dynamics of the restoration of physical trails in the grass-cutting ant Atta capiguara (Hymenoptera, Formicidae)

ABSTRACT

Dynamics of the restoration of physical trails in the grass-cutting ant Atta capiguara. Leaf-cutting ants of the genus Atta build long physical trails by cutting the vegetation growing on the soil surface and removing the small objects they find across their path. Little is known on the dynamics of trail construction in these ants. How much time do they need to build a trail? To answer this question we selected six trails belonging to two different nests of A. capiguara and removed on each trail a block of soil of 20 cm × 15 cm that included a portion of the physical trail. This block was then replaced by a new block of the same size that was removed in the pasture near the trail and that was uniformly covered by the same type of vegetation as that found on the block of soil that was removed. The time required to restore the trail was then evaluated by the length of the grass blades found along the former location of the trail. The results show that ants rapidly restore the portion of the physical trail that was interrupted, which suggests that they could also do the same after their trails have been recolonized by the vegetation.

Keywords:
Leaf-cutting ants; Foraging; Saúva; Social insects

Introduction

Foraging in leaf-cutting ants involves the selection, cutting and transport of plant material to the nest (Della Lucia, 1993). It begins when a scout ant finds a plant to exploit and recruits its nestmates by laying trail pheromone on its way back to the nest (Jaffé and Howse, 1979; Shepherd, 1982; Roces, 1990). The main part of foraging however consists in cutting and transporting the plant material to the nest (Forti et al., 1987). The retrieval of the vegetation occurs along well-defined physical trails (Howard, 2001), which are an essential part of the foraging system of leaf-cutting ants (Shepherd, 1982; Hölldobler and Wilson, 2010).

In most species of leaf-cutting ants the physical trails form a dendritic network which consists of stable physical trails departing directly from the nest that branch successively in numerous relatively short and ephemeral trails which provide access to the resources (Wirth et al., 2003; Kost et al., 2005). The network of trails can be considered as a territory which delimits the foraging area of neighboring colonies and thus minimizes the frequency of agonistic encounters between alien foragers (Shepherd, 1982; Rockwood and Hubbell, 1987). Trail usage varies over time and ants use some trails more than others (Cherrett, 1968; Vasconcelos, 1990; Silva et al., 2013). In leaf-cutting ants of the genus Atta some foraging trails can be used for months or even years (Weber, 1972; Kost et al., 2005; Wirth et al., 2003). The combination of physical trails and of the trail pheromone signal laid down on the trail surface allows recruited ants to access easily the resources dispersed in their environment (Weber, 1972; Fowler and Robinson, 1979). The physical trails form corridors in the vegetation and they are used as physical guidelines by recruited workers to travel from the nest to the resources discovered by the scouts. They thus reduce the risks for naive recruited workers of losing their way while following the chemical trail. The construction and maintenance of physical trails involve energy and require time (Shepherd, 1982; Howard, 2001). Traveling on cleared trails however increases worker speed by 4–10 times compared to traveling on uncleared soil (Rockwood and Hubbell, 1987). In addition, it facilitates the transport of leaf fragments, thereby increasing the rate of return to the nest of the plant material.

Given their importance for the collection of plant material, one can hypothesize that physical trails should be rapidly restored by leaf-cutting ants whenever they are recolonized by the vegetation after they have been abandoned for some length of time. In this paper we present the results of a study investigating the dynamics of restoration of physical foraging trails in the leaf-cutting ant Atta capiguara.

Material and methods

The study was conducted on two nests of A. capiguara near the city of Botucatu, São Paulo state, Brazil (22° 50.833′ S and 48° 26.476′ W). The size of the nests, estimated by the area of loose soil over the nests, was 113.71 m² for nest 1 and 47.88 m² for nest 2. The nests were around 3–5 years old.

Foraging in A. capiguara is completely diurnal in the wet season and completely nocturnal during the dry season (Forti, 1985). Before the foraging activity began in the early morning, a 20 cm × 15 cm block of soil (thickness around 5 cm, preserving most of the root system of the plants) was removed from an active physical trail, at a distance of 3 m from the beginning of the trail.

This block was replaced by another block of exactly the same size (henceforth called experimental) which was removed in the same pasture in which the nests were located and which was uniformly covered by vegetation. We took care that the vegetation of the experimental block was the same as that found on the edges of the manipulated trail. In other words, if the manipulated trail was lined e.g. by Paspallum, the experimental block was covered by the same plant species.

We then removed a block of soil of the same size as that of the experimental block, at a distance of 2 m perpendicular to the manipulated trail, and replaced it by a block (henceforth called control) of the same dimension, which was removed at the same location as that of the block inserted on the manipulated trail. This allowed us to assess the vegetation growth in absence of ant traffic and to measure the background grazing and cutting activity of other animals in the pasture. To avoid the intrusion of big herbivores the experimental site was fenced during the evaluation period.

The two blocks of soil (experimental and control) were divided into twelve 5 cm × 5 cm quadrats by white nylon threads. Five centimeters corresponds to the average width of the physical trails of leaf-cutting ants in pastures (Lopes, unpublished results). They were checked every 12 h during 5 days and the length of the five longest grass blades in each of the three quadrats located along the median of the block, corresponding to the former location of the trail, was measured. Three trails from each of the two nests studied were manipulated during our study, yielding a total of 6 trails.

For each block of soil the mean length of the 15 grass blades measured was calculated. Since the initial state of the vegetation on the blocks varied among trails, in order to standardize the change in the vegetation over time for all blocks, we then calculated the percentage variation in the mean length of the grass blades between the first survey (at 12 h) and each successive survey for each block. We then tested with two-way within-subjects ANOVA whether the length of the grass blades varied differently over time in the control and experimental blocks for the 6 trails studied. Block type (with two levels: control vs. experimental) was entered as the first within-subjects factor, time of survey (with 8 levels) was entered as the second within-subjects factor, and trail as the random factor. If ants are actively cutting the grass to build their trail one should find a significant interaction between block type and time of survey: whereas the mean length of the grass blades in the experimental blocks should decrease over time, it should remain stable or slightly increase because of vegetation growth in the control blocks. The variation over time in the length of the grass blades was then tested for each block with a one-way within-subjects ANOVA with trail entered as a random factor. The software R 3.1.3 was used to analyze the data (R Core Team, 2015).

Results

The percentage variation in the mean length of grass blades on the control and experimental blocks of each trail throughout the duration of the survey is shown in Fig. 1. The length of the grass blades varied differently between the control block and the experimental block for each trail (ANOVA, interaction block type × time of survey: F_7,35 = 9.931, p < 0.001). Whereas the length of the grass blades did not vary over time in the control blocks (ANOVA: F_7,35 = 0.354, p = 0.923), it decreased significantly in the experimental blocks (ANOVA: F_7,35 = 22.570, p < 0.001). The vegetation disappeared completely for trail 1 of nest 1 (Figs. 2A and 3B) and was only partially cut for trail 3 of nest 1, probably because of the low traffic intensity recorded on this trail. The restoration of the physical trail can clearly be seen, even in absence of ant traffic. Only the layer of dry vegetation directly over the soil remained on the trail.

Fig. 1
Percentage variation in the mean length of grass blades for the experimental block (continuous lines) and control block (dashed lines) of each manipulated trail of Nest 1 (A) and 2 (B). The lines of same thickness correspond to the same trail.

Fig. 2
Restoration of the physical trails of nest 1. First and last day of survey of trail 1 (A and B, respectively), 2 (C and D, respectively) and 3 (E and F, respectively). The lines on B, D, F show the location where the trail was restored.

Fig. 3
Restoration of the physical trails of nest 2. First and last day of survey of trail 1 (A and B, respectively), 2 (C and D, respectively) and 3 (E and F, respectively). The lines B, D, F show the location where the trail was restored.

Discussion

Our results show that A. capiguara workers rapidly restore the portion of their physical trails that was interrupted and thus suggest that they could also do the same when their trails have been temporary abandoned and recolonized by the vegetation. The measure of the height of the grass blades on the control blocks shows that there was no cutting or grazing activity by other herbivores in the pasture during the time required for ants to restore their trail. For some control blocks (Nest 2: trail 1 and 3) a low growth of the vegetation was even registered, showing that the block transplantation did not affect the state of the vegetation.

Upon encountering the new block of soil placed across their trail, foraging ants stopped and formed a cluster around and on top of it. Such a clustering behavior has also been observed by Ribeiro (2009) in his study of the behavior of workers of A. sexdens rubropilosa facing an obstacle placed on their trail. This is probably due to the interruption of the chemical trail. Ants rapidly began to cut the grass blades they found across their path and the physical trail was fully restored for most of the trails after 5 days.

The vegetation removed during the reconstruction of the physical trails is probably retrieved to the nest and incorporated in the fungus garden. In other Atta species collecting dicotyledons on the other hand, particularly those found in the understory of tropical forests, the construction and maintenance of physical trails involve the removal of leaf litter (Wirth et al., 2003), which is not the most suitable material for the growth of their symbiotic fungus since it is mostly composed of dry leaves. Hence, in the species of Atta cutting monocotyledons, contrary to those cutting dicotyledons (Howard, 2001), the cost of construction and maintenance of the physical trails may be partially offset by the use of the vegetation that is cut along the trails and that can be used to feed the symbiotic fungus.

The fact that ants restored their physical trails exactly at the same location as that of the previous trails and that the reconstructed trails joined precisely with the existing trails on the other side of the block implies that they had a memory of the direction in which they had to travel to get back to the physical trail on the other side of the block. Since the portions of physical trails that were removed in our experiment were relatively straight, ants could simply maintain the direction in which they were traveling by using a celestial or geomagnetic compass (Banks and Syrgley, 2003). Alternatively they could use a visual memory of the terrestrial landmarks surrounding the trail (Collett et al., 2003) to guide their path across the block of soil or, possibly, a topochemical sense based on the odor of the surrounding landmarks (Jaffé et al., 1990; Helmy and Jander, 2003).

The capacity of leaf-cutting ants to rapidly restore their physical trails could explain the stability of the foraging networks of these ants (Wirth et al., 2003; Kost et al., 2005). Indeed, as noted by Vasconcelos (1990), leaf-cutting ants generally prefer to maintain their trails rather than construct new ones. The foraging networks thus consist of a core of physical trails that are regularly used by ants and the networks are then expanded through the prolongation of these trails or the construction of short ephemeral trails (Forti, 1985). The maintenance of the physical trails or their restoration after they have been covered by leaf litter or recolonized by the vegetation may allow ants to adjust to the seasonal changes of their host plants and to resume their exploitation as soon as the quantity and quality of the vegetation on these plants meet their selection criteria. In particular, it should allow ants to use perennial resources such as large trees, for which they express a preference (Cherrett, 1968).

In conclusion, our study shows that A. capiguara workers are able to rapidly restore their physical trails after these latter have been interrupted and thus suggests that they could do the same after their trails have been recolonized by the vegetation.

Acknowledgements

We are grateful to Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) for the financial support and stipends to the authors (07/07091-0, 07/04010-0). LCF is grateful to Conselho Nacional de Desenvolvimento Cientifico e Tecnológico for research assistance (301917/2009-4). During his stay in Botucatu VF was financed by a CAPES-COFECUB grant N° 633/08.

References

• Banks, A.N., Syrgley, R.B., 2003. Orientation by magnetic field in leaf-cutter ants, Atta colombica (Hymenoptera, Formicidae). Ethology 109, 835-846.
• Cherrett, J.M., 1968. The foraging behaviour of Atta cephalotes L. (Hymenoptera Formicidae) I. Foraging pattern and plant species attacked in tropical rain forest. J. Anim. Ecol. 37, 387-403.
• Collett, T.S., Graham, P., Durier, V., 2003. Route learning by insects. Curr. Opin. Neurobiol. 13, 718-725.
• Della Lucia, T.M.C., 1993. As Formigas Cortadeiras. Folha de Viçosa, Viçosa.
• Forti, L.C., Tese (Doutorado em Entomologia) 1985. Ecologia da saúva Atta capiguara Gonçalves, 1944 (Hymenoptera, Formicidae) em pastagem. Escola Superior de Agricultura Luiz de Queiroz, Universidade de São Paulo, Piracicaba.
• Forti, L.C., Crocomo, W.B., Guassu, C.M.O., 1987. Biologia e controle das formigas cortadeiras de folhas em florestas importadas. FEPAF, São Paulo.
• Fowler, H.G., Robinson, S.W., 1979. Foraging by Atta sexdens: seasonal patterns caste and efficiency. Ecol. Entomol. 4, 239-247.
• Helmy, O., Jander, R., 2003. Topochemical learning in black carpenter ants (Camponotus pennsylvanicus). Insectes Soc. 50, 32-37.
• Hölldobler, B., Wilson, E.O., 2010. The Leaf-Cutting Ants: Civilization by Instinct. W.W. Norton & Company.
• Howard, J.J., 2001. Cost of trail construction and maintenance in leaf-cutting ant Atta colombica Behav. Ecol. Sociobiol. 49, 348-356.
• Jaffé, K., Howse, P.E., 1979. The mass recruitment system of the leaf cutting ant Atta cephalotes (L.). Anim. Behav. 27, 930-939.
• Jaffé, K., Ramos, C., Lagalla, C., Parra, L., 1990. Orientation cues used by ants. Insectes Soc. 37, 101-115.
• Kost, C., de Oliveira, G., Knoch, T.A., Wirth, R., 2005. Spatio-temporal permanence of plasticity of foraging trails in young and mature leaf-cutting ant colonies (Atta spp.). J. Trop. Ecol. 21, 677-688.
• R Core Team, 2015. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria, http://www.R-project.org/
  » http://www.R-project.org/
• Ribeiro, P.L., Tese (Doutorado em Ciências/Area de Fisiologia Geral) 2009. Trilhas de saúvas (Atta sexdens rubropilosa): um método que impede a formação de fluxo bidirecional e mostra que as forrageadoras resolvem o problema. Instituto de Biociências da Universidade de São Paulo.
• Roces, F., 1990. Olfactory conditioning during the recruitment process in a leaf-cutting ant. Oecologia 83, 261-262.
• Rockwood, L.L., Hubbell, S.P., 1987. Host-plant selection, diet diversity, and optimal foraging in a tropical leafcutting ant. Oecologia 74, 55-61.
• Shepherd, J.D., 1982. Trunk trails and the strategy of a leaf-cutting ant Atta colombica Behav. Ecol. Sociobiol. 11, 77-84.
• Silva, P.S.D., Bieber, A.G.D., Knoch, T.A., Tabarelli, M., Leal, I.R., Wirth, R., 2013. Foraging in highly dynamic environments: leaf-cutting ants adjust foraging trail networks to pioneer plant availability. Entomol. Exp. Appl. 147, 110-119.
• Vasconcelos, H.L., 1990. Foraging activity of two species of leaf-cutting ants (Atta) in a primary forest of the Central Amazon. Insectes Soc. 37, 131-145.
• Weber, N.A., 1972. Gardening Ants: The Attines. The American Philosophical Society, Philadelphia.
• Wirth, R., Herz, H., Ryel, R.J., Beyschlag, W., Hölldobler, B., 2003. Herbivory of Leaf-Cutting Ants – A Case Study on Atta colombica in the Tropical Rainforest of Panama. Springer.

Publication Dates

• Publication in this collection
  Jan-Mar 2016

History

• Received
  23 Aug 2015
• Accepted
  15 Oct 2015