Leaf-cutting ants belonging to the genera Atta and Acromyrmex are primary pests and have been considered as the insects that cause the most damage to agriculture throughout the Neotropics (1). During foraging, leaf-cutting ant workers cut and transport plant fragments, which are then used to cultivate a symbiotic fungus inside the so-called fungus chambers (2). This fungus serves as the sole food source for the colony (3). Atta vollenweideri (Hymenoptera: Formicidae) is well suited for studies examining the negative effect leaf-cutter ants may have on extensive livestock practices in South American grasslands. They belong to the group classified as grass-cutters together with Atta capiguara and Atta bisphaerica (4, 5). Grass-cutting ants have traditionally been seen as the main cattle competitor because they forage on the functional group on which cattle mostly feed (6). A. vollenweideri does not show homogenous distribution on a local scale because it tends to nest on heavy clay soils used for livestock production due to its low productivity for cropping (7–9). Table 1 summarizes a comprehensive list of studies reporting plant consumption over the last decades. To date, three publications have analyzed consumption by A. vollenweideri and its possible impacts on livestock. While the study carried out by Robinson and Fowler (13) reported a consumption of 868 and 924 kg of dry weight/ha/year, Jonkman (12) estimated a consumption of only 38 kg of dry weight/ha/year. Based on those results, Robinson and Fowler (13) concluded that A. vollenweideri consumed twice the yearly requirement of cows. However, Jonkman (12) estimated it was below 5%. Both studies, nevertheless, suffer from having estimated the consumption based on only one A. vollenweideri nest. Additionally, they did not cite the aerial net primary productivity (ANPP) of the habitat, which determines the amount of net available forage source for which ants and cows should compete. The most comprehensive study was conducted by Guillade and Folgarait, who also measured ANPP and the portion consumed by cattle (14). They reported a minimum of 224 and a maximum of 1,867 kg of dry weight/ha/year, averaging 1,216 kg of dry weight/ha/year consumed by Atta vollenweideri. The authors concluded that competition between ants and cattle occurred only during the low-productivity periods of the year. The economic injury level for the study location was estimated at 0.29 nest/ha.

In addition to estimating consumption by ants and the pasture demands of cattle, a comprehensive estimation of competition between leaf-cutting ants and cattle should also include an estimation of nest density at a landscape scale, and not only the density of the study site where measurements have been performed. Leaf-cutting ant distribution is not homogeneous and tends to be aggregated at both local and landscape scales (7, 23, 24). When actual density at landscape scales is not determined, and the impact of leaf-cutting ants on cattle is extrapolated from the highly infested areas where consumption was measured, the negative effect of ants on livestock can be overestimated. An extrapolation to landscape level should also be carried out to properly discuss the impact of leaf-cutting ants.

Thus, this study aimed to study the potential impact of A. vollenweideri as a pest at both local and landscape scales. Locally, we assessed functional groups foraged by A. vollenweideri (grasses, shrubs, and trees) to estimate the potential impact of ants by considering the extent to which ant preferences might overlap with the known grass preference of cattle. We also calculated the impact of such overlapped consumption by considering the standard cattle management regime for the region. Second, working at a regional scale in midwestern Argentina, we estimated nest density through satellite images. This was carried out to assess the percentage of the area that surpass the critical density at which A. vollenweideri may negatively affect production.

Although grass-cutting ants have been traditionally seen as a pest of livestock production, our results show that the economic impacts of Atta grass-cutting ants are less important than what is commonly assumed. Although Atta is repeatedly cited as a grass-cutter (36, 44), our work shows that A. vollenweideri also cuts dicotyledonous leaves from shrubs and trees. In addition, results show that ant preferences for a specific functional group change throughout the year, with grass consumption greater in autumn but less representative in spring. These changing preferences are probably ruled by a high selectivity in response to fluctuations in palatable resources and distance to the nest, as known for other Atta species (45, 46). These plant preferences raise the question about competition between ants and cattle because cattle mostly consume grasses instead of shrubs and trees (28, 29). All studies to date ( Table 1 ) have not differentiated the functional groups consumed by ants when comparing them with the actual dietary preferences of cattle (10–21, 47). Those studies generally applied the foraging method (9 of 13), and loads carried by ants belonging to all functional groups were pooled. In our view, this overestimated the negative impact of ants on cattle consumption. Therefore, any calculation related to economic injury levels based on published data obtained by the foraging intake method would be biased. In our case, for instance, not classifying into functional groups as we did would result in an overestimation of the percentage of non-grass loads cut by A. vollenweideri by approximately 33%.

During this study, we decided to classify the loads into functional groups to focus our analysis on the fraction in which the competition was more likely to occur. By doing this, we intended to overcome the extended misleading practices of previous studies that applied the foraging method. Nevertheless, and although the exclusion method for estimating ant consumption is less used than the foraging method, we believe it also has some methodological issues worth mentioning. The main difficulty of the exclusion method is that the area around nests is not homogeneously cut. Colonies periodically change activity among trails (4, 48, 49). In fact, the extent of the current foraging area was less than 25% of the potential foraging area based on an assessment of the extension of the trails (50). Additionally, cages also excluded cow trampling (28, 51, 52) and forage by other small herbivores that consume grasses (53–55). Altogether, our study and those listed in Table 1 suffer from methodological issues that explain the great variability of values reported for Atta consumption and the impact they would have on livestock. However, from our point of view, our work improves on previous studies by considering functional groups using foraging methods over 24 h, increasing the number of surveyed trails, colony number, and considering the impact at a landscape level.

In addition to the method for estimating ant consumption, other factors to be taken into account are the consumption of ants and cattle in relationship with the primary productivity of the habitat and season. Farmers tend to maximize the management regime, i.e., the number of cows per hectare, for the consumption of all the ANPP under a high-pressure grazing regime (41). In this case, the probability that the consumption by any other herbivore negatively affects cattle intake is greater than when practicing a lower pressure grazing regime. However, competition occurs only if both herbivores forage on the same portion of the plants. It is known that cattle mostly cut portions of plants located above 5–10 cm (28, 56). On the contrary, there is no evidence that ants also focus on the biomass above 5–10 cm or that they cannot cut below this height. There is also no evidence to date indicating that forage consumed by ants could be taken by larger herbivores when ants are not present in that area, which seems to be an open question not only for ants but also for other small-large herbivore interactions (57). In fact, leaf-cutting ant colonies of both genera (Atta and Acromyrmex) are known to sustain stable populations in modified habitats (58–62), such as those aimed at livestock production, as commonly seen in other ant species (63). By lowering the vegetation, large herbivores promote an increase in soil temperature, which would be a limiting factor for the establishment of leaf-cutting ant colonies (62, 64), at least in southern South America (65). Intensive rangeland practices do not seem to imply the displacement of leaf-cutting ants (62, 66).

The nest density recommended to control colonies (EIL) is far beyond the density at which Atta colonies occur. Previous studies conducted by Jonkman (7) using aerial photography to survey the density of A. vollenweideri nests over an area of 80,000 km² in Paraguay showed that only 10% of the area contained living nests. This is in line with our results, which show that a small portion of the surveyed area had colonies. We recognize that calculating EIL based on consumption data from the bibliography has a strong bias. However, all species of Atta are expected to forage annual amounts within a similar range. First, most species possess nests of different shapes but similar sizes (67, 68), which results in similar colony sizes and consumption (69, 70). Second, fungus gardens cultivated by all species are highly conserved across species at a continental scale (3, 71) and, therefore, are expected to decay foraged material in a similar way. We believe this common EIL is an underestimation because we consider toxic baits as the standard control measure, following the established rationale for all leaf-cutting ants (72). Nevertheless, it is known that grass-cutting ants of both genera, Atta and Acromyrmex, do not always accept baits (5). Species not accepting baits should be controlled with thermal fog, which is not just more costly but also extremely harmful to the environment compared with baits (5). Although we consider this common EIL to be robust, we believe it is an underestimation. Our main goal was to obtain an index that would allow us to carry out an overall approximation for the genus, not species- or location-specific comparisons. In summary, the EIL obtained during our work falls in the range of the overall median for all studies summarized in Table 1 , which is 5.42 nests/ha ( Figure 4 ). To date, average reported nest densities for Atta species fall mostly below this average EIL, in both natural and modified habitats (73–75).

In recent decades, it was mostly assumed that Atta was a severe pest of extensive livestock production. Given our work, it could be said there is no conclusive evidence for Atta to be considered a severe pest. In addition, we should consider the added negative effect of controlling a species without numerical evidence supporting such a decision. Several Atta species have been confirmed as key engineer species (76). An irrational eradication from areas aimed at extensive livestock production would preclude extremely important ecosystem processes related to the presence of Atta nests (9, 77, 78). Our data, similar to the previously published works, suggest that Atta species, despite being conspicuous herbivores in their habitats, should not always be considered a pest of livestock production in rangelands. This differs from past and current policies concerning grass-cutting ant control in the countries where A. vollenweideri is widely distributed. For instance, decades ago, Argentina, Paraguay, and Uruguay promoted and made the control of A. vollenweideri mandatory through laws and their subsequent regulatory acts (79–81), which are still in use. Now, and based on the current evidence, new regulations are needed. These should include the concept that a species is a pest in a specific production system at a given population level, i.e., the critical economic injury level, as discussed in this study, for livestock production in rangelands. It is now time to apply the methods already developed for estimating the population density of Atta nests and using them as a decision criterion in rational pest management practices (1, 34, 35).

Results show that A. vollenweideri consumed approximately 276 kg dry weight/ha/year of plants, the most foraged being grasses (70%) but also cutting herbs (25%) and trees (5%). This consumption represents 7% of the pasture demanded to raise a calf according to the management regime applied by farmers. Our calculated EIL was 5.85 nests/ha, which falls in the range of previous works. Colonies were absent in 93.6% of a surveyed area of 15,000 ha, while their density was below the EIL in 6.2% of the area and surpassed the EIL in only 0.2%.

These results question the perception that Atta leaf-cutting ants are a pest of livestock production. Although ants consume a small percentage of cattle’s demand, evidence that ants and cattle are competing in the few cases that density surpasses the EIL is arguable. In the countries where A. vollenweideri is present, decision makers have promulgated several acts making its control mandatory. It is time to revisit these regulations and the pest status of A. vollenweideri by including the use of EIL as a control criterion.

The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding author.

Conceptualization: JS and MB. Funding acquisition: JS and MB. Data acquisition: JS. Analysis: JS and MB. First draft writing: MB. Writing review and editing: MB. All authors contributed to the article and approved the submitted version.