Trail complexity and foraged plants for the leaf-cutting ant Atta sexdens (Hymenoptera: Formicidae) in a caatinga area of Bahia

ABSTRACT

Leaf-cutting ants (LCAs; Atta, Acromyrmex, and Amoimyrmex) are key herbivores in Neotropical ecosystems, preferentially harvesting leaves from palatable plant species such as pioneers and colonizers. However, the extent to which the spatial and temporal availability of plant resources influences the structure and complexity of LCAs foraging trail networks remains poorly understood. We investigated the foraged plant community and the foraging network system of five Atta sexdens colonies in a human-modified landscape of the Caatinga domain, a Brazilian tropical dry forest, over a one-year period. Colonies harvested leaves from 104 plant individuals belonging to 18 species. Most plants (63 individuals, 60.6%) were collected during the dry season, compared with 41 plants (39.4%) in the wet season. Notably, more than half of the plants foraged during the dry season (62%) still exhibited green leaves. The total number of foraged plants positively influenced nest attributes-such as the number of foraging holes, trails, and branching points-and measures of foraging effort, including seasonal cumulative foraging area and trail length per colony. In contrast, seasonality itself had no detectable effect on nest attributes or foraging effort. Because LCAs are common in disturbed landscapes such as the study area, our results suggest that A. sexdens may intensify plant harvesting during the dry season by maintaining and expanding trail networks directed toward green-leaf plants. This behavior may play an important role in shaping plant dynamics, potentially influencing plant growth, survival, and recruitment in Caatinga ecosystems.

KEYWORDS:
Atta foraging behavior; dry forest; optimal foraging theory; foraging system.

INTRODUCTION

In natural environments, animals constantly make strategic decisions to access resources efficiently (MacArthur and Pianka 1966). These choices help them gather the energy needed for reproduction and offspring survival, which, in turn, influences population dynamics and reproductive success (Krebs and Davies 1997). Their ability to adapt and focus foraging efforts based on the spatial and temporal distribution of resources is a key element of optimal foraging theory (Pyke et al. 1977). Optimal foraging theory intends to explain how animals choose which foods to eat (optimal diet), which patches to feed in (optimal patch choice), how to divide their time among different patches, and the most efficient way and speed to move (Pyke et al. 1977).

The leaf-cutter ants (genera Atta, Acromyrmex and Amoimyrmex; hereafter LCAs) are a good example of how organisms adapt their foraging strategies to access plant material across different forest layers, including the forest floor, understory, and canopy (Vasconcelos and Cherrett 1997, Wirth et al. 2003, Silva et al. 2007, Corrêa et al. 2010, Meyer et al. 2011a, 2011b, 2013, Silva et al. 2012a). It is already known that LCAs’ foraging network is highly adaptable, allowing workers to adjust their foraging based on the availability of palatable food sources (Silva et al. 2013), such as pioneer plants in tropical moist forests (Farji-Brener 2001) and herbs and pioneer or colonizing plants in Caatinga dry forests (Siqueira et al. 2017), especially during the dry season when the number of plants bearing green leaves is strongly reduced (Machado et al. 1997).

In human-disturbed landscapes, changes in resource availability and quality are expected, including an overabundance of pioneer plants in forest fragments in tropical rain forests (Wirth et al. 2008; see Laurance et al. 2006 for the distribution of fast-growing trees in fragmented landscapes) and the proliferation of some families, such as Euphorbiaceae which increase in disturbed Brazilian dry tropical forests (Rito et al. 2017). These changes probably lead to a reorganization of LCAs’ foraging strategies. For example, Atta cephalotes (Linnaeus, 1758) colonies in secondary Atlantic Forest patches, which contain more pioneer plants, expanded their network complexity of foraging trails (Silva et al. 2013). This increased network complexity enhances foraging efficiency in a smaller area, enabling them to access more plants and possibly escalate herbivory pressure (Silva et al. 2013).

In the Caatinga dry forest, the availability of plants with green leaves, which fluctuates seasonally, becoming increasingly scarce during the dry period (Machado et al. 1997), may impact the LCAs’ network of foraging trails and the size of foraging areas. A study on how vegetation seasons affect foraging areas in the Caatinga dry forest found that the average foraging area of eight Atta opaciceps Borgmeier, 1939 colonies was nearly five times larger during the dry season compared to the wet season (Siqueira et al. 2018). This suggests that LCAs may need to expand their foraging efforts (foraging areas) by creating new trails to locate the reduced supply of food plants during the dry season. Another interesting result was that LCAs colonies consumed significantly more plant material during the dry season, both in terms of biomass and leaf area (Siqueira et al. 2018). The scarcity of trees and shrubs with leaves in most Caatinga during the dry season likely forces the LCAs to compensate by accessing additional areas and alternative sources to maintain the amount of vegetation needed to sustain their fungus garden (Siqueira et al. 2018). In this case, the reduced number of trees and shrubs with leaves in most Caatinga during the dry season may lead to expanded foraging area and greater complexity of trail networks, such as increased number of foraging holes, trails, and branching points.

Most studies that examine the factors influencing the LCAs’ foraging trail network (Wirth et al. 1997, Silva et al. 2012b, Silva et al. 2013), nest distribution (Vasconcelos 1990, Vasconcelos and Cherrett 1995, Farji-Brener and Illes 2000, Wirth et al. 2007, Dohm et al. 2011), and herbivory rate (Urbas et al. 2007, Silva et al 2012b) were conducted in tropical rainforests. In contrast, non-forested ecosystems such as the Brazilian Cerrado savanna (Vasconcelos et al. 2006, Vieira-Neto et al. 2016), the Caatinga dry forest (Siqueira et al. 2017, Siqueira et al. 2018, Cruz et al. 2020, Knoechelmann et al. 2020, Oliveira et al. 2024), the Venezuelan savanna (Farji-Brener and Silva 1995a, 1995b, 1996), and the semi-arid steppes in Argentina (Farji-Brener 2010) have been comparatively less studied. Some researches indicate that road construction can increase Atta nests occurrence nearby, as seen in the Cerrado savanna (Vasconcelos et al. 2006, Vieira-Neto et al. 2016) and in the Caatinga dry forest (Siqueira et al. 2017, Cruz et al. 2020). This probability is linked to the growing presence of palatable food sources, such as herbs and pioneer or colonizing plants, in degraded areas like pastures and roadside regions (Siqueira et al. 2017).

In this context, studies have shown that the herbivory and soil changes caused by nest building and maintenance by LCAs can negatively impact forest regeneration in disturbed areas of the Caatinga dry forest (Knoechelmann et al. 2020, Oliveira et al. 2023; but see Oliveira et al. 2024 for a potential species-specific effect of LCAs on beneficial plant reproduction). However, to understand their role as key herbivores, it is essential to determine how LCAs regulate their trail network when exploiting plant resources in the Caatinga.

To date, no study has examined how vegetation influences the trail network complexity of Atta sexdens (Linnaeus, 1758) in the Caatinga or affects foraging effort (but see Siqueira et al. 2018 for information on how vegetation seasonality influences the size of foraging areas of A. opaciceps). To expand the understanding of A. sexdens behavior and ecology in a human-altered landscape within the Caatinga, we examined how this species organizes its trail system and the plants it collects across seasons. For both dry and wet seasons, we identified the plant species and individuals that were foraged, including plants with green leaves.

Additionally, we investigated how seasonality and the number of foraged plants influence (1) nest attributes associated with foraging activity, which serve as indicators of trail complexity, such as (a) the number of foraging holes, (b) trails, and (c) branching points, and (2) foraging effort, measured by (d) seasonal cumulative foraging area and (e) trail length per colony. Considering the limited availability of plants with leaves during the dry season (Machado et al. 1997), we predicted that LCAs would create new foraging trails and increase the complexity of existing ones (e.g., by adding more foraging holes and branches) to support their high demand for plant material. In this context, we expected LCAs to forage from a larger number of individual plants during the dry season, which would be associated with an expansion of their foraging network, mainly focused on evergreen species.

MATERIAL AND METHODS

Study area

The Brazilian Caatinga is part of the largest seasonally dry tropical forest region in South America (sensu Queiroz et al. 2017). It is characterized by a xerophytic and deciduous thorn forest (Prado 2003, Silva et al. 2018). The climate is typically semiarid, with low annual precipitation ranging from 300 to 800 mm per year and irregularly distributed (IBAMA 2006, Silva et al. 2018). As a result, only about 30% of plant species retain their leaves during the dry season (i.e., evergreen plants; Queiroz et al. 2017), while deciduous species are the most common in the Caatinga (Prado 2003). Mean annual temperature is 23 °C, and mean altitude is 350 m a.s.l.

This study was conducted in a degraded area, a privately owned dry forest (21.83 ha), located in the municipality of Contendas do Sincorá, southwestern Bahia state, Brazil (13°54’08”S, 41°07’55”W, Fig. 1A). Cattle have overgrazed this site for many years (we estimate at least 20 years). The vegetation consists of numerous small patches of scattered trees and shrubs, mostly deciduous species, with few or any litter cover on the ground (Fig. 1B, C). In the region, plant families such as Fabaceae, Euphorbiaceae, Malvaceae and Cactaceae are very common (Vitório et al. 2019). Some plant species retain their leaves during the dry season and are relatively frequent in this degraded area, such as Senna acuruensis (Benth.) H.S. Irwin & Barneby (Fabaceae) (personal observation).

Figure 1
(A) Maps showing the location of Bahia State, Brazil, in South America, and the municipality of Contendas do Sincorá. (B) Aerial photograph of the study area showing the location of the five studied Atta sexdens colonies inside a privately-owned farm subjected to overgrazing caused by loose housed cattle, also showing, at the right side of the picture, a small portion of the preserve Floresta Nacional Contendas do Sincorá (FNCS). The pictures depict (C) the general aspect of the degraded Caatinga area where the study was carried out, and (D) a foraging trail of A. sexdens. Photos by L.C. Marinho.

Determining wet and dry seasons for the study area

According to IBAMA (2006), in the region the wet season starts in November and lasts until April. To verify if this pattern holds true in face of the actual climate changes, we used a 14-year historical record (from January 2010 to December 2023) provided by the Brazilian Water Agency (ANA). This dataset, available at https://www.snirh.gov.br/hidroweb/serieshistoricas, includes data from a meteorological station in Contendas do Sincorá, about 18 km from our study area.

Mapping and analyzing foraging trail system

Five A. sexdens colonies were monitored for one year, from October 2010 to September 2011, with the aim of mapping the foraging trail network and foraging activity on plants. Average colony size, based on the area of the nest mound (i.e., external area with more loose soil removed by ant workers during excavations), was 17.78 ± 3.89 m². Mean distance between the colonies was 196 ± 110.16 m, ranging from 62 m to a maximum of 362 m.

The peak of foraging activity was determined for each colony as described by Wirth et al. (1997). For this, at 1-hour intervals over a 2-minute period, all workers passing a fixed point near the foraging hole were counted on each active foraging trail (Wirth et al. 1997). The survey began at 06:00 a.m. (on October 23^rd, 2010) and ended at 6:00 a.m. of the following day. Ant foraging activity was predominantly centered overnight. Leaf harvesting started in the late afternoon with the decline of sunlight and temperature. The highest peak activity was observed from 11:00 p.m. to 1:00 a.m., reducing drastically after 4:00 a.m. In the following months, constant observations during filed work (day and night) confirmed whether the colonies continued to favor this foraging pattern at night.

Subsequently, once monthly, each colony was surveyed during the foraging activity peak (from 11:00 p.m. to 02:00 a.m.). The following data were gathered: (1) number of foraging holes (i.e., the nest entrance with foraging activity), (2) number of trails (total number of trails starting from each foraging holes per nest), (3) number of branching points (bifurcation points, where a trail is subdivided into two or more trails), (4) number of foraged plants (with and without green leaves), and (5) total foraging trail length, which represents the sum of the total length of all trails per season for each colony, including the primary and secondary trails. To evaluate the number of trails, we considered only the trails with foraging activity. Primary trails begin at each foraging hole and extend to their respective endpoints, often a leaf source, and are maintained by ant foragers (Silva et al. 2013). Secondary trails branch off from these main trails and may be more temporary (Silva et al. 2013). Two or more trails starting from the same foraging holes were considered as separate trails. Although partially correlated, these metrics serve as complementary indicators of the spatiotemporal complexity of trail networks (Silva et al. 2013).

Furthermore, to confirm that specific foraging holes located far from the nest mound belonged to the evaluated colony, they were examined and tracked using pieces of distinctly colored drinking straws. These straws had been previously soaked in a mixture of vegetable oil, orange juice, and cornflakes (see Silva et al. 2012b for more details on this methodology). Later, the colored straw fragments were searched for in the nest refuse piles by carefully scanning the soil surface 24 and 48 hours after the offering (Silva et al. 2012b). Each color represented a different foraging hole. Only foraging holes associated with the identified drinking straws were included in this study. Each active foraging trail was tracked from the foraging hole to the stem of the respective foraged plant, where the ants ascend into the canopy to gather vegetative (i.e., leaves) and non-vegetative plant materials (i.e., floral parts and fruits). It was possible to observe the LCAs cutting plant material, such as leaves or flowers, in plant crowns, since the trees were no more than 2-3 m tall. Harvested plants with green leaves (i.e., leaves neither wilted nor dried) were also mapped and identified.

Foraging trails were mapped by measuring the compass bearings and lengths of all quasi-linear trail segments following the trails until the food source or the respective endpoint (for more details, see the method used by Silva et al. (2009, 2013). All foraged plants were marked for later gathering of botanical material and identification of species. The plant specimens were deposited in the HUESBVC herbarium, with duplicates sent to HUEFS (acronyms according to Thiers, 2021). Valid species names were verified on the website of the project ‘Flora e Funga do Brasil’ (http://floradobrasil.jbrj.gov.br/).

At the end of our study period, the 12 maps generated monthly were combined into a single cumulative map per colony, following the approach of Kost et al. (2005) and Silva et al. (2009, 2013). Similarly, season maps were also obtained by grouping six monthly maps each. The dry season map included data from May to October, while the wet season map included data from November to April. For calculating the annual size of foraging areas, we first superimposed monthly foraging trail systems and then generated contour maps that excluded those sectors not visited by ant workers (Wirth et al. 2003). Finally, we estimated the percentage of the study area (%) under A. sexdens influence based on the multiplication of the average foraging area (this study) by colony density for this same area (3.6 colonies ha^-1, Cruz et al. 2020).

Data analysis

A one-way ANOVA, followed by Fisher post-hoc test, was used to indicate significant differences in monthly precipitation. General Linear Models (GLMs) were applied to assess seasonal effects (dry and wet seasons) and the number of foraged plants on the mean values of the following response variables: (1) nest attributes associated to foraging activity, such as (a) the number of foraging holes, (b) trails, and (c) branching points, and (2) foraging effort, measured by (d) seasonal cumulative foraging areas and (e) trail length per colony of five A. sexdens colonies in an area of Caatinga, municipality of Contendas do Sincorá.

A non-metric multidimensional scaling (NMDS) ordination of the five A. sexdens colonies was conducted, using Bray-Curtis similarity coefficients, to compare the foraged plant composition between the wet and dry seasons. Statistical analyses were performed as described by Sokal and Rohlf (1995) and Hair et al. (2006) using R (version 4.3.2; R Core Team 2023).

RESULTS

Determining the wet and dry seasons for the study area

From January 2010 to December 2023, the average monthly precipitation (mm) varied between months (F = 4.89, G.L = 11 and p < 0.0001, Fig. 2). The period with higher average precipitation, the months from November to April, was considered the wet season here. In contrast, the other six months (May to October) were considered as the dry season (Fig. 2, Fisher post-hoc test, p < 0.05).

Figure 2
Average precipitation (mm) per month based on a historical data series, from January 2010 to December 2023, for a meteorological station located 18 km away from the study area. The data were collected by the Brazilian Water Agency (ANA) and is available through the website: https://www.snirh.gov.br/hidroweb/serieshistoricas. Different letters above the confidence interval indicate significant differences according to Fisher’s post hoc test.

Foraged plants

The five A. sexdens colonies harvested plant material from 104 plant individuals belonging to 18 species (Table 1), of which 63 (60.6%) were harvested in the dry season and 41 (39.4%) in the wet season. During the dry season, LCAs collected plant material from 15 of the 18 identified species (83%), whereas in the wet season they collected material from 11 species (61%). Ants collected not only fragments of green or dry leaves (61% and 34% of the harvested material, respectively), but also floral parts (4%) and seeds (1%).

Table 1
List of plant species harvested by five colonies of Atta sexdens throughout one year (from October 2010 to September 2011) in a degraded area of Caatinga, municipality of Contendas do Sincorá, Bahia, Brazil. The species are listed in descending order according to the number of individuals foraged by LCAs.

Species	Family	No. individuals/season		Species with green leaves during the dry season
		Dry	Wet
Senna acuruensis (Benth.) H.S. Irwin & Barneby	Fabaceae	20	10	Yes
Combretum glaucocarpum Mart.	Combretaceae	5	10	No
Pereskia bahiensis Gürke	Cactaceae	4	5	No
Croton argyrophyllus Kunth	Euphorbiaceae	8	0	No
Senegalia ricoae (Bocage & Miotto) L.P. Queiroz	Fabaceae	5	3	Yes
Anemopaegma laeve DC.	Bignoniaceae	7	0	Yes
Passiflora setaceae DC.	Passifloraceae	4	1	Yes
Waltheria sp.	Malvaceae	1	4	No
Pterocarpus rohrii Vahl	Fabaceae	2	2	Yes
Fridericia sp.	Bignoniaceae	2	1	No
Canavalia brasiliensis Mart. ex Benth.	Fabaceae	0	2	Yes
Sideroxylon obtusifolium (Humb. ex Roem. & Schult.) T.D. Penn.	Sapotaceae	1	1	No
Annona vepretorum Mart.	Annonaceae	0	1	No
Dalbergia cearensis Ducke	Fabaceae	1	0	No
Hymenaea sp.	Fabaceae	1	0	No
Randia calycina Cham.	Rubiaceae	1	0	No
Schinopsis brasiliensis Engl.	Anacardiaceae	1	0	Yes
Ziziphus joazeiro Mart.	Rhamnaceae	0	1	Yes
Total species number: 18 spp.		63	41	Yes = 8 / No = 10

The three species with the most foraged individuals across both seasons were Senna acuruensis (Benth.) H.S. Irwin & Barneby (Fabaceae) with 30 individuals, followed by Combretum glaucocarpum Mart. (Combretaceae) with 15 and Pereskia bahiensis Gürke (Cactaceae) with nine individuals (Table 1). Also, the leguminous S. acuruensis was the only species used by all colonies studied. We also highlight the euphorb species Croton argyrophyllus Kunth, represented by eight individuals in LCAs diet, because, in this case, only dry leaves were harvested and during the dry season (Table 1).

Of the 18 foraged plant species, eight (44.4%) presented green leaves during at least a part of the dry season (Table 1). We observed LCAs collecting green leaves of seven of these eight species during the dry season. Except for Canavalia brasiliensis, which, although presenting green leaves, was only collected in the wet season. Six species of foraged plants were harvested only in the dry season (Croton argyrophyllus, Anemopaegma laeve, Dalbergia cearensis, Hymenaea sp., Randia calycina and Schinopsis brasiliensis). However, only two of them presented green leaves in this season (A. laeve and S. brasiliensis) (Table 1). In the case of the other four species (C. argyrophyllus, D. cearensis, Hymenaea sp. and R. calycina), fragments of dry leaves were harvested directly under the plants.

In the dry season, 39 individuals (62%) of the 63 foraged plants had green leaves, while the other 24 (38%) had no green leaves. When comparing the species composition of foraged plants in both seasons via NMDS (Bray-Curtis similarity), a clear tendency to segregation was observed, forming two groups (i.e., dry and wet seasons, Fig. 3).

Figure 3
Non-metric multidimensional scaling (NMDS) ordination of foraged plant species of the five Atta sexdens colonies in Contendas do Sincorá, Bahia, Brazil. NMDS is based on Bray-Curtis similarity coefficients and compares plant composition between the wet season (filled circles) and dry season (open circles).

Nest attributes and foraging effort

As expected, the trail attributes examined in this study as indicators of foraging trail complexity (such as the number of trails, foraging holes, and branching points) were positively influenced by the number of plants foraged (Table 2, Fig. 4A-C). However, no changes were observed between the dry and wet seasons (Table 2). The seasonal cumulative foraging areas were 1.5 times larger during the dry season (6154.5 m²) compared to the wet season (4133.6 m²). The mean seasonal cumulative size of the foraging areas was 1028.82 ± 241.27 m², ranging from 583.98 to 3059.89 m² (Fig. 5). Although no significant differences were observed in the size of the foraging areas between the dry and wet seasons (1230.9 ± 455.68 and 826.73 ± 184.1 m², respectively; Table 2), the size of the foraging areas was positively influenced by the number of foraged plants (Table 2 and Fig. 4D).

Figure 4
Effect of the number of foraged plants on the number of foraging holes (A), trails (B), branching points (C), size of foraging areas (D), and trail lengths (E) of five Atta sexdens colonies, during the dry and wet seasons in Contendas do Sincorá, Bahia, Brazil.

Figure 5
Annual foraging trail systems, size of colony foraging areas (m²), and cumulative length of trails (m) of five Atta sexdens colonies (A, B, C, D, and E) during the dry and wet seasons in Contendas do Sincorá, Bahia, Brazil. The lines represent foraging trails, the open dots correspond to the nest entrance, and the crossed dots to the foraged plants.

Table 2
General Linear Models (GLMs) considering the effects of two seasons (dry season and wet season) and the number of foraged plants on (1) nest attributes, such as (a) the number of foraging holes, (b) trails, and (c) branching points, and (2) foraging effort, measured by (d) seasonal cumulative foraging areas and (e) trail length per colony of five Atta sexdens colonies in an area of Caatinga, municipality of Contendas do Sincorá, Bahia, Brazil. Significant effects are bold.

Response variables	F	p	R2 (Adj.) (%)
No. foraging holes
No. foraged plants	9.68	0.02	49
Season	0.11	0.75
No. trails
No. foraged plants	67.02	<0.001	89
Season	0.04	0.85
No. branching points
No. foraged plants	92.58	<0.001	92
Season	2.26	0.18
Foraging areas
No. foraged plants	72.28	<0.001	90
Season	1.03	0.34
Trails length
No. foraged plants	45.17	<0.001	83
Season	2.12	0.19

Total trail length was 291 m higher in the dry season than in the wet season (1180 m vs. 889 m, respectively). The mean seasonal cumulative length of foraging trail per colony (linear foraging distance) was 206.9 ± 53.39 m², varying from 46 to 548 m. Furthermore, the seasonal cumulative length of foraging trail also followed the same pattern as described for foraging areas, with no significant changes observed between seasons (dry and wet seasons, 236 ± 102.33 and 177.8 ± 43.96 m, respectively, Table 2). However, it was also positively influenced by the number of foraged plants (Table 2 and Fig. 4E). Together, the workers of these five colonies accessed at least 37% of the entire degraded area via foraging trails. The statistical analysis showed that the number of foraged plants accounted for a relatively large portion of the variance in the number of foraging holes (49% explanation), trails (89%), branching points (92%), foraging areas (90%), and the trail length (83%).

DISCUSSION

Several studies have shown that leaf-cutting ants can direct their trail networks to sites with high-value resources (e.g., Wirth et al. 1997, Hölldobler and Wilson 2011, Silva et al. 2012b, Silva et al. 2013). In this study, we found that A. sexdens adjusted its foraging trails to optimize access to plants with green leaves, especially during the dry season when this resource is scarce. Siqueira et al. (2018) observed that the congener A. opaciceps can maintain high rates of biomass consumption even at the peak of the dry season in another Caatinga area.

It is worth noting that, in our study, six plant species were harvested by LCAs only during the dry season. Furthermore, during the dry season LCAs typically gather plant material from a greater variety of species compared to the wet season. Contrary to our initial predictions, nest attributes related to foraging trail complexity (e.g., foraging holes, trails, and branching points) and foraging effort (e.g., foraging areas and trail lengths) did not differ between seasons. However, they were positively correlated with the number of foraged plants.

Another study reported a similar pattern, showing that A. cephalotes can build a more extensive, branched, and complex trail system in response to an increase in pioneer abundance (Silva et al. 2013). Both results (ours and those of Silva et al. 2013) are consistent with optimal foraging theory. As resources become increasingly scarce and patchily distributed during the dry season, A. sexdens likely requires fine-tuned adjustments to its foraging network in order to sustain high biomass consumption.

In both seasons, we also observed a preference of A. sexdens workers for harvesting leaves of Senna acuruensis, a deciduous shrub species endemic to the Caatinga (Oliveira and Garcia 2021). This may be related to relatively low defense against herbivory in this species, assuming it follows the general pattern described for deciduous plants (Dirzo and Boege 2008). Additionally, S. acuruensis, together with Combretum glaucocarpum, which was the most harvested species in the wet season, presents a higher density of individuals in our study area compared to the native vegetation found in the preserve ‘Floresta Nacional Contendas do Sincorá’ (M.M. Corrêa, unpublished data, 2018).

Combretum glaucocarpum has an arboreal habit (Vasconcelos et al. 2017) and is known as a toxic species for cattle in the Caatinga (Itakura et al. 1987). Despite the known presence of tannins (Itakura et al. 1987), a possible chemical defense against herbivores (Dirzo and Boege 2008), C. glaucocarpum leaves do not appear to be harmful to the symbiotic fungus of the colonies studied here. However, a high removal rate of C. glaucocarpum has already been observed in other studies conducted in the same area (G.V. Oliveira, unpublished data; see also Cordeiro et al. 2021). Additionally, another study assessing the preference of A. sexdens laboratory colonies for young and mature leaves of this species observed that colonies apparently preferred mature leaves over young ones, and that their interest in both leaf types decreased with increasing drying time (Cordeiro et al. 2021). Thus, it is possible that C. glaucocarpum is a delayed greening species (sensu Coley and Kursor 1996), meaning that its young leaves contain more tannins than mature ones and are therefore less likely to be consumed by herbivores (Cordeiro et al. 2021).

We did not evaluate the proportion of young and mature leaves collected by ants from C. glaucocarpum or other species. Nevertheless, this is a promising aspect for better understanding the foraging behavior of A. sexdens in the Caatinga biome in future studies. It is known that LCAs workers continuously make decisions about the quality of a given resource (i.e., whether it is harmful to fungus growth or not), favoring those that promote maximal fungal growth as well as immediate energy availability to support foraging (Roces 2002). In this context, it would be interesting to investigate whether ants prefer to collect dry rather than green leaves for a given species. Evergreen species are probably less preferred by LCAs during the wet season because they exhibit higher leaf toughness compared to deciduous species (see Dourado et al. 2016 for a comparison in our study area of herbivory intensity between six deciduous and evergreen species).

It is worth noting that during the dry season, green leaves in the Caatinga are provided mainly by evergreen species or by deciduous species with longer leaf longevity (Silva et al. 2018). In our study, some plant species with green leaves during the dry season had fewer specimens harvested in the wet season, including S. acuruensis, C. glaucocarpum, S. ricoae, and A. laeve. Additionally, no individual of A. laeve had leaves harvested during the wet season. Considering the relatively low cost of trail construction and maintenance due to the large number of workers per colony (Howard 2001), building new trails, especially in the dry season, may be advantageous for accessing plants with green leaves or other available resources. Furthermore, the low cost of trail construction and maintenance is likely facilitated by the minimal plant biomass and sparse leaf litter covering the ground. We suggest that this diversification of foraging during the dry season, including plants with green leaves, may be crucial for colony survival during periods of relative scarcity of green plant material.

Considering that LCAs species benefit from anthropogenic disturbances in the Caatinga biome and preferentially occur in areas with low vegetation cover (Siqueira et al. 2017), high herbivory pressure and significant impacts are expected in degraded areas. In our study area, A. sexdens colonies accessed at least 37% of the entire degraded region through their foraging trails. Studies conducted in another Caatinga area on the effects of A. opaciceps on vegetation found that: (1) leaf consumption and herbivory rates by LCAs colonies were higher in more disturbed areas (Siqueira et al. 2018); (2) nest mounds and foraging areas of LCAs showed lower abundance of seedlings and saplings compared to areas not affected by these ants, probably related to nest building/maintenance and foraging activities (Knoechelmann et al. 2020); and (3) LCAs also reduce seed deposition in the areas they affect, especially on nest mounds (Oliveira et al. 2023). Because A. sexdens may cause similar effects, it may potentially influence vegetation dynamics in our study area. This may lead to homogenization and decline of plant communities associated with human disturbance (Cruz et al. 2020). Therefore, further research is necessary to understand why LCAs foraged fewer plants during the wet season, to evaluate the palatability of foraged plants, and to monitor LCAs distribution in Caatinga dry forests. Understanding the factors that influence foraging activity is crucial for managing their impact as ecosystem engineers and herbivore pests in the Caatinga dry forest region.

ACKNOWLEDGEMENTS

The authors thank the Universidade Estadual do Sudoeste da Bahia (UESB) for financial support and Instituto Chico Mendes de Biodiversidade (ICMBio) for providing infrastructure and logistic support during the fieldwork. Special acknowledgments to Avaldo Soares Filho for his help with plant identification, and Filipe Ribeiro Sá Martins, Jan Borges Santos, and Antônio Correia Freire for their important contribution during field work.

LITERATURE CITED

• Coley PD, Kursor TA (1996) Anti-herbivore defenses of young tropical leaves: physiological constraints and ecological trade-offs. In: Mulkey SS, Chazdon RL, Smith AP (Eds) Tropical forest plant ecophysiology. Springer, Boston, 305-336.
• Cordeiro JJ, Costa IS, Salgado VJ, Costa-Filho RG, Santos MP, et al. (2021) Foraging behavior of Atta sexdens (Hymenoptera, Formicidae) on leaves of Thiloa glaucocarpa (Mart.) Eichler (Combretaceae) in a Brazilian seasonally dry tropical forest. Arthropod-Plant Interactions 15: 737-745. https://doi.org/10.1007/s11829-021-09847-z
  » https://doi.org/10.1007/s11829-021-09847-z
• Corrêa MM, Silva PSD, Wirth R, Tabarelli M, Leal IR (2010) How leaf-cutting ants impact forests: drastic nest effects on light environment and plant assemblages. Oecologia 162: 103-115. https://doi.org/10.1007/s00442-009-1436-4
  » https://doi.org/10.1007/s00442-009-1436-4
• Cruz IA, Silva GS, Bottcher C, Bieber AGD, Corrêa MM, Silva PSD (2020) Occurrence of the leaf-cutting ant Atta sexdens L. (Hymenoptera: Formicidae) related to unpaved roads in two Brazilian semiarid areas with contrasting disturbance degrees. Revista de Ciências Ambientais 14: 35-41. https://doi.org/10.18316/rca.v14i1.5980
  » https://doi.org/10.18316/rca.v14i1.5980
• Dirzo R, Boege K (2008) Patterns of herbivory and defence in tropical dry and rain forest. In: Carson WP, Schnitzer SA (Eds) Tropical forest community ecology. Wiley-Blackwell, Chichester, 63-78.
• Dohm C, Leal I, Tabarelli M, Meyer S, Wirth R (2011) Leaf-cutting ants proliferate in the Amazon: expected response to forest edge? Journal of Tropical Ecology 27: 645-649. https://doi.org/10.1017/S0266467411000447
  » https://doi.org/10.1017/S0266467411000447
• Dourado ACP, Sá-Neto RJ, Gualberto AS, Corrêa MM (2016) Herbivoria e características foliares em seis espécies de plantas da Caatinga do nordeste brasileiro. Revista Brasileira de Biociências 14(3): 145-151. https://seer.ufrgs.br/rbrasbioci/article/view/114671
  » https://seer.ufrgs.br/rbrasbioci/article/view/114671
• Farji-Brener AG (2001) Why are leaf-cutting ants more common in early secondary forests than in old-growth tropical forests? An evaluation of the palatable forage hypothesis. Oikos 92: 169-177. https://doi.org/10.1034/j.1600-0706.2001.920120.x
  » https://doi.org/10.1034/j.1600-0706.2001.920120.x
• Farji-Brener A (2010) Leaf-Cutting Ant Nests and Soil Biota Abundance in a Semi-Arid Steppe of Northwestern Patagonia. Sociobiology 56: 549-557.
• Farji-Brener AG, Illes AE (2000) Do leaf-cutting ant nests make “bottom-up” gaps in neotropical rain forests? A critical review of the evidence. Ecoogy Letters 3: 219-227. https://doi.org/10.1046/j.1461-0248.2000.00134.x
  » https://doi.org/10.1046/j.1461-0248.2000.00134.x
• Farji-Brener AG, Silva JF (1995a) Leaf-cutting ants and forest groves in a tropical parkland savanna of Venezuela: facilitated succession? Journal of Tropical Ecology 11(4): 651-669. https://doi.org/10.1017/S0266467400009202
  » https://doi.org/10.1017/S0266467400009202
• Farji-Brener AG, Silva JF (1995b) Leaf-cutting ant nests and soil fertility in a well-drained savanna in Western Venezuela. Biotropica 27(2): 250. https://doi.org/10.2307/2389001
  » https://doi.org/10.2307/2389001
• Farji-Brener AG, Silva JF (1996) Leaf-cutter ants’ (Atta laevigata) aid to the establishment success of Tapirira velutinifolia (Anacardiaceae) seedlings in a parkland savanna. Journal of Tropical Ecology 12(1): 163-168. https://doi.org/10.1017/S0266467400009378
  » https://doi.org/10.1017/S0266467400009378
• Hair JF, Black B, Babin B, Anderson RE, Tatham RL (2006) Multivariate data analysis. Pearson Education, Upper Saddle River, 734 pp.
• Hölldobler B, Wilson EO (2011) The leafcutter ants: civilization by instinct. WW Norton & Company, New York, 160 pp.
• Howard JJ (2001) Costs of trail construction and maintenance in the leaf-cutting ant Atta columbica Behavioral Ecology and Sociobiology 49: 348-356. https://doi.org/10.1007/s002650000314
  » https://doi.org/10.1007/s002650000314
• IBAMA (2006) Plano de Manejo Floresta Nacional Contendas do Sincorá. Informações gerais sobre a Floresta Nacional. Ministério do Meio Ambiente, Brasília, vol. 1.
• Itakura Y, Habermeh G, Mebs D (1987) Tannins occurring in the toxic Brazilian plant Thiloa glaucocarpa Toxicon 25: 1291-1300. https://doi.org/10.1016/0041-0101(87)90007-9
  » https://doi.org/10.1016/0041-0101(87)90007-9
• Knoechelmann CM, Oliveira FMP, Siqueira FFS, Wirth R, Tabarelli M, Leal IR (2020) Leaf-cutting ants negatively impact the regeneration of the Caatinga dry forest across abandoned pastures. Biotropica 52: 686-696. https://onlinelibrary.wiley.com/doi/10.1111/btp.12782
  » https://onlinelibrary.wiley.com/doi/10.1111/btp.12782
• Kost C, Oliveira EG, Knoch TA, Wirth R (2005) Spatio-temporal permanence and plasticity of foraging trails in young and mature leaf-cutting ant colonies (Atta spp.). Journal of Tropical Ecology 21: 677-688. https://doi.org/10.1017/S0266467405002592
  » https://doi.org/10.1017/S0266467405002592
• Krebs JR, Davies NB (1997) Behavioural Ecology: An Evolutionary Approach. Wiley-Blackwell, Oxford, 3^rd ed., 480 pp.
• Laurance WF, Nascimento HEM, Laurance SG, Andrade AC, Fearnside PM, et al. (2006) Rain Forest fragmentation and the proliferation of successional trees. Ecology 87(2): 469-482. https://doi.org/10.1890/05-0064
  » https://doi.org/10.1890/05-0064
• MacArthur RH, Pianka ER (1966) On optimal use of a patchy environment. American Naturalist 100: 603-609.
• Machado IC, Barros LM, Sampaio EVSB (1997) Phenology of Caatinga species at Serra Talhada, PE, Northeastern Brazil. Biotropica 29: 57-68.
• Meyer ST, Leal IR, Tabarelli M, Wirth R (2011a) Ecosystem engineering by leaf-cutting ants: nests of Atta cephalotes drastically alter forest structure and microclimate. Ecological Entomology 36: 14-24. https://doi.org/10.1111/j.1365-2311.2010.01241.x
  » https://doi.org/10.1111/j.1365-2311.2010.01241.x
• Meyer ST, Leal IR, Tabarelli M, Wirth R (2011b) Performance and fate of tree seedlings on and around nests of the leaf-cutting ant Atta cephalotes: Ecological filters in a fragmented forest. Austral Ecology 36: 779-790. https://doi.org/10.1111/j.1442-9993.2010.02217.x
  » https://doi.org/10.1111/j.1442-9993.2010.02217.x
• Meyer ST, Neubauer M, Sayer EJ, Leal IR, Tabarelli M, Wirth R (2013) Leaf-cutting ants as ecosystem engineers: topsoil and litter perturbations around Atta cephalotes nests reduce nutrient availability. Ecological Entomology 38: 497-504. https://doi.org/10.1111/een.12043
  » https://doi.org/10.1111/een.12043
• Oliveira FM, Knoechelmann CM, Wirth R, Tabarelli M, Leal IR (2023) Leaf-cutting ant nests support less dense and impoverished seed assemblages in a human-modified Caatinga dry forest. Biotropica 55: 444-453. https://doi.org/10.1111/btp.13198
  » https://doi.org/10.1111/btp.13198
• Oliveira FR, Centeno-Alvarado D, Oliveira FM, Wirth R, Lopes AV, Leal IR (2024) Divergent effects of leaf-cutting ant herbivory and soil engineering on the reproductive success of plants in a Caatinga dry forest. Ecological Entomology 49(1): 128-137.
• Oliveira LDGRD, Garcia FCP (2021) Senna (Leguminosae-Caesalpinioideae) in Minas Gerais state, Brazil. Rodriguésia 72: e01482019. https://doi.org/10.1590/2175-7860202172104
  » https://doi.org/10.1590/2175-7860202172104
• Prado EP (2003) As caatingas da América do Sul. In: Leal IR, Tabarelli M, Silva JMC (Eds) Ecologia e conservação da Caatinga. Editora Universitária da UFPE, Recife, 3-74.
• Pyke G, Pulliam HR, Charnov GL (1977) Optimal foraging: a selective review of theory and tests. Quarterly Review of Biology 52: 137-154.
• Queiroz LP, Cardoso D, Fernandes MF, Moro MF (2017) Diversity and evolution of flowering plants of the caatinga domain. In: Silva JMC, Leal IR, Tabarelli M (Eds) Caatinga: The largest tropical dry forest region in South America. Springer, Switzerland, 23-63.
• R Core Team (2023) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna. https://www.R-project.org
  » https://www.R-project.org
• Rito KF, Tabarelli M, Leal IR (2017) Euphorbiaceae responses to chronic anthropogenic disturbances in Caatinga vegetation: from species proliferation to biotic homogenization. Plant Ecology 218: 749-759.
• Roces F (2002) Individual complexity and self-organization in foraging by leaf-cutting ants. The Biological Bulletin 202: 306-313. https://doi.org/10.2307/1543483
  » https://doi.org/10.2307/1543483
• Silva JMC, Leal IR, Tabarelli M (2018) Caatinga: The largest tropical dry forest region in South America. Springer, Switzerland, 482 pp.
• Silva PSD, Bieber AGD, Knoch TA, Tabarelli M, Leal IR, Wirth R (2013) Foraging in highly dynamic environments: leaf-cutting ants adjust foraging trail networks to pioneer plant availability. Entomologia Experimentalis et Applicata 147: 110-119. https://doi.org/10.1111/eea.12050
  » https://doi.org/10.1111/eea.12050
• Silva PSD, Bieber AGD, Leal IR, Wirth R, Tabarelli M (2009) Decreasing abundance of leaf-cutting ants across a chronosequence of advancing Atlantic forest regeneration. Journal of Tropical Ecology 25: 223-227. https://doi.org/10.1017/S0266467409005823
  » https://doi.org/10.1017/S0266467409005823
• Silva PSD, Leal IR, Wirth R, Melo FPL, Tabarelli M (2012a) Leaf-cutting ants alter seedling assemblages across second-growth stands of Brazilian Atlantic forest. Journal of Tropical Ecology 28: 361-368. https://doi.org/10.1017/S0266467412000259
  » https://doi.org/10.1017/S0266467412000259
• Silva PSD, Leal IR, Wirth R, Tabarelli M (2007) Harvesting of Protium heptaphyllum (Aubl.) March. seeds (Burseraceae) by the leaf-cutting ant Atta sexdens L. promotes seed aggregation and seedling mortality. Revista Brasileira de Botânica 30: 553-560. https://doi.org/10.1590/S0100-84042007000300019
  » https://doi.org/10.1590/S0100-84042007000300019
• Silva PSD, Leal IR, Wirth R, Tabarelli M (2012b) Spatial distribution and fruiting phenology of Protium heptaphyllum (Burseraceae) determine the design of the underground foraging system of Atta sexdens L. (Hymenoptera: Formicidae). Neotropical Entomology 41: 257-262. https://doi.org/10.1007/s13744-012-0052-x
  » https://doi.org/10.1007/s13744-012-0052-x
• Siqueira FFS, Ribeiro-Neto JD, Tabarelli M, Andersen AN, Wirth R, Leal IR (2017) Leaf-cutting ant populations profit from human disturbances in tropical dry forest in Brazil. Journal of Tropical Ecology 33: 337-344. https://doi.org/10.1017/S0266467417000311
  » https://doi.org/10.1017/S0266467417000311
• Siqueira FFS, Ribeiro-Neto JD, Tabarelli M, Andersen AN, Wirth R, Leal IR (2018) Human disturbance promotes herbivory by leaf-cutting ants in the Caatinga dry forest. Biotropica 50: 779-788. https://doi.org/10.1111/btp.12599
  » https://doi.org/10.1111/btp.12599
• Sokal RR, Rohlf SR (1995) Biometry. WH Freeman, New York, 887 pp.
• Thiers B (2021) Index herbariorum: a global directory of public herbaria and associated staff. New York Botanical Garden’s virtual herbarium. http://sweetgum.nybg.org/ih/ Accessed: 22/02/2021
  » http://sweetgum.nybg.org/ih/
• Urbas P, Araújo Jr MV, Leal IR, Wirth R (2007) Cutting more from cut forests: edge effects on foraging and herbivory of leaf-cutting ants in Brazil. Biotropica 39: 489-495. https://doi.org/10.1111/j.1744-7429.2007.00285.x
  » https://doi.org/10.1111/j.1744-7429.2007.00285.x
• Vasconcelos ADM, Henriques IGN, Souza MP, Santos WS, Santos WS, Ramos GG (2017) Caracterização florística e fitossociológica em área de Caatinga para fins de manejo florestal no município de São Francisco-PI. Agropecuária Científica no Semiárido 13: 367-376. https://doi.org/10.30969/acsa.v13i4.967
  » https://doi.org/10.30969/acsa.v13i4.967
• Vasconcelos HL (1990) Habitat selection by the queens of the leaf-cutting ant Atta sexdens L. in Brazil. Journal of Tropical Ecology 6: 249-252. https://doi.org/10.1017/S0266467400004405
  » https://doi.org/10.1017/S0266467400004405
• Vasconcelos HL, Cherrett JM (1995) Changes in leaf-cutting ant populations (Formicidae, Attini) after the clearing of mature forest in Brazilian Amazonia. Studies on Neotropical Fauna and Environment 30: 107-113. https://doi.org/10.1080/01650529509360947
  » https://doi.org/10.1080/01650529509360947
• Vasconcelos HL, Cherrett JM (1997) Leaf-cutting ants and early forest regeneration in central Amazonia: Effects of herbivory on tree seedling establishment. Journal of Tropical Ecology 13: 357-370. https://doi.org/10.1017/S0266467400010567
  » https://doi.org/10.1017/S0266467400010567
• Vasconcelos HL, Vieira-Neto EHM, Mundim FM, Bruna EM (2006) Roads alter the colonization dynamics of a keystone herbivore in Neotropical savannas. Biotropica 38: 661-665. https://doi.org/10.1111/j.1744-7429.2006.00180.x
  » https://doi.org/10.1111/j.1744-7429.2006.00180.x
• Vieira-Neto EHM, Vasconcelos HL, Bruna EM (2016) Roads increase population growth rates of a native leaf-cutter ant in Neotropical savannahs. Journal of Applied Ecology 53: 983-992. https://doi.org/10.1111/1365-2664.12651
  » https://doi.org/10.1111/1365-2664.12651
• Vitório C, Marinho LC, Costa G, Aona LYS (2019) Flowering plants of Contendas do Sincorá National Forest (Caatinga, northeastern Brazil). Brazilian Journal of Botany 42: 717-725. https://doi.org/10.1007/s40415-019-00564-9
  » https://doi.org/10.1007/s40415-019-00564-9
• Wirth R, Beyschlag W, Ryel RJ, Herz H, Hölldobler B (2003) The herbivory of leaf-cutting ants. A case study on Atta colombica in the tropical rainforest of Panama. Springer Verlag Berlin, Heidelberg, 230 pp.
• Wirth R, Beyschlag W, Ryel RJ, Hölldobler B (1997) Annual foraging of the leaf-cutting ant Atta colombica in a semidecious rain forest in Panama. Journal of Tropical Ecology 13: 741-757. https://doi.org/10.1017/S0266467400010907
  » https://doi.org/10.1017/S0266467400010907
• Wirth R, Meyer ST, Almeida WR, Araujo Jr MV, Barbosa VS, Leal IR (2007) Increasing densities of leaf-cutting ants (Atta spp.) with proximity to the edge in a Brazilian Atlantic forest. Journal of Tropical Ecology 23: 501-505. https://doi.org/10.1017/S0266467407004221
  » https://doi.org/10.1017/S0266467407004221
• Wirth R, Meyer ST, Leal IR, Tabarelli M (2008) Plant herbivore interactions at the forest edge. Progress in Botany 69: 423-448.

ADDITIONAL NOTES

• ZooBank register
  https://zoobank.org/1B678807-7859-471D-95A2-C8B4CF7279C3
• Data Availability Statement
  Datasets generated or analyzed in this study are available from the corresponding author on reasonable request.
• Funding
  This research received no external funding.
• Ethical Statement
  This study did not involve live vertebrate animals and therefore did not require approval by an ethics committee. Field activities were conducted under collection permits issued by SISBIO (permit 29146).
• AI Statement
  Artificial intelligence tools were used solely to assist with language editing and grammar.
• How to cite this article
  Carmo TNN, Oliveira GV, Marinho LC, Bieber AGD, Corrêa MM, Silva PSD (2026) Trail complexity and foraged plants for the leaf-cutting ant Atta sexdens (Hymenoptera: Formicidae) in a caatinga area of Bahia. Zoologia 43: e25051. https://doi.org/10.1590/S1984-4689.v43.e25051
• Published by
  Sociedade Brasileira de Zoologia at Scientific Electronic Library Online - https://www.scielo.br/zool

Edited by

• Editorial responsibility
  Marcel Gustavo Hermes

Data availability

Datasets generated or analyzed in this study are available from the corresponding author on reasonable request.

Data citations

R Core Team (2023) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna. https://www.R-project.org

Publication Dates

• Publication in this collection
  20 Apr 2026
• Date of issue
  2026

History

• Received
  21 July 2025
• Accepted
  31 Oct 2025