Despite the importance and increasing knowledge of ecological networks, sampling effort and intrapopulation variation has been widely overlooked. Using continuous daily sampling of ants visiting three plant species in the Brazilian Neotropical savanna, we evaluated for the first time the topological structure over 24 h and species-area relationships (based on the number of extrafloral nectaries available) in individual-based ant-plant networks. We observed that diurnal and nocturnal ant-plant networks exhibited the same pattern of interactions: a nested and non-modular pattern and an average level of network specialization. Despite the high similarity in the ants’ composition between the two collection periods, ant species found in the central core of highly interacting species totally changed between diurnal and nocturnal sampling for all plant species. In other words, this “night-turnover” suggests that the ecological dynamics of these ant-plant interactions can be temporally partitioned (day and night) at a small spatial scale. Thus, it is possible that in some cases processes shaping mutualistic networks formed by protective ants and plants may be underestimated by diurnal sampling alone. Moreover, we did not observe any effect of the number of extrafloral nectaries on ant richness and their foraging on such plants in any of the studied ant-plant networks. We hypothesize that competitively superior ants could monopolize individual plants and allow the coexistence of only a few other ant species, however, other alternative hypotheses are also discussed. Thus, sampling period and species-area relationship produces basic information that increases our confidence in how individual-based ant-plant networks are structured, and the need to consider nocturnal records in ant-plant network sampling design so as to decrease inappropriate inferences.
Citation: Dáttilo W, Fagundes R, Gurka CAQ, Silva MSA, Vieira MCL, Izzo TJ, et al. (2014) Individual-Based Ant-Plant Networks: Diurnal-Nocturnal Structure and Species-Area Relationship. PLoS ONE 9(6): e99838. doi:10.1371/journal.pone.0099838
Editor: Stephen C. Pratt, Arizona State University, United States of America
Received: March 12, 2014; Accepted: May 16, 2014; Published: June 11, 2014
Copyright: © 2014 Dáttilo et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The authors confirm that all data underlying the findings are fully available without restriction. All data are included within the manuscript.
Funding: These authors have no support or funding to report.
Competing interests: The authors have declared that no competing interests exist.
Ants and plants can interact in different ways, from facultative to highly specialized relationships . For instance, extrafloral nectar-mediated ant–plant mutualisms are among the most remarkable ecological interactions in terrestrial ecosystems . In this type of ant-plant interaction, plants with extrafloral nectaries (EFN-bearing plants) produce a liquid rich in carbohydrates and amino acids, which attracts different ant species . In exchange for food, some ants can protect the host plant against potential herbivores . At the community level, different ant and plant species can interact with each other generating complex ecological networks; where plant and ant species are depicted as nodes and their interactions are depicted by links , , .
Focused on the structure of ant-plant ecological networks, some studies have found some non-random patterns of interactions around the world as, for example, the nested pattern , , . This pattern indicates that within an ant-plant network there is a core of generalist species (those with the most interactions), which interact among themselves, and specialists species (those with fewer interactions) also interacting with the generalist species in cohesive subgroups . Another characteristic of these facultative ant-plant networks is that they do not exhibit a modular pattern of interactions, since there are no groups of ants specialized in feeding on a given group of plants , ; as previously demonstrated in symbiotic networks involving ants and myrmecophytes . However, despite the importance and increasing knowledge of ant-plant networks at the community level, no study has evaluated how intrapopulation variation in plants can contribute to the organization of the ant species associated to individual EFN-bearing plants. As a single EFN-bearing plant can be associated to several ant species in a predictable way, we can also use a network approach to evaluate the structure of such individual-based ant-plant networks.
We already know that within an EFN-bearing plant population the richness of associated ants differs among plants , ,  One of the main factors explaining this intrapopulation variation is based on the difference in the reward offered (quantity and quality of nectar), where individuals with better rewards would be most visited by ants , . On the other hand, nectar quantity and quality can vary along the day, and on the same plant , , , which means that an individual plant may be not considered to be a good resource throughout the day. Therefore, this variation in nectar production can influence the daily rate of ant foraging , and colony growth . In addition, foraging of some ant species, including those that forage on EFN-bearing plants, can be strictly diurnal, nocturnal, or both , , Therefore, there are good reasons to expect temporal turnover of species composition of ant assemblages and intrapopulation variation in ant-plant networks.
An individual EFN-bearing plant can be viewed as an “island of resources” for ants and compared with the classical model of species-area relationship proposed by MacArthur and Wilson . In ant-plant interactions, it is possible to expect an increase of ant richness throughout plant ontogeny, mainly due to higher availability of food resources for ants (e.g., number of extrafloral nectaries-EFNs) , , . In fact, bigger and older plants can increase their nutrient-acquiring and photosynthetic capacity and invest in biotic defense via extrafloral nectaries . The increase in the richness of ant occurs mainly because bigger plant individuals with higher number of EFNs would allow for a greater spatial segregation in the use of their food resources, which increases ant coexistence . Thus, the difference in ants’ presence on plants from the same population is linked not only by available food resources, but also by the foraging time per ant species and species-area relationships. This intrapopulation variation can directly affect the fitness of each individual plant  since it is expected that, in general, the most visited individuals can be better protected against herbivores .
Here, using both diurnal and nocturnal sampling, we evaluated the network structure and species-area relationships in individual-based ant-plant networks. Specifically, due to nectar production features and ant foraging shifts between day and night, we hypothesized that there would be changes in the composition and in the specific positions of ants within the network when diurnal and nocturnal networks were compared. Moreover, based on species-area relationship, we also postulated that plant individuals with higher number of EFNs had a greater richness of associated ants (i.e., greater degree within a nested network). In order to test our hypotheses, we conducted diurnal and nocturnal sampling of ants interacting with three plant species in different regions in the Brazilian Neotropical savanna.
Materials and Methods
All necessary permits were obtained for the described field studies. All research was conducted with the approval of the Brazilian Chico Mendes Institute for Biodiversity Conservation (ICMBIO/SISBIO, Permit Number 12427/2007) and the Minas Gerais Forest Institute (IF/MG, Permit Number 064/11).
Study Area and Species Studied
We sampled individual-based ant-plant networks in three different localities of the Brazilian Neotropical savanna (Cerrado biome). In each field site we observed and annotated the interactions between individuals of one EFN-bearing plant species and their associated ants. As the plants’ composition changes geographically, is not logistically possible to access the same species in all sites, so we choose the three most common EFN-bearing plant species in each site: Chamaecrista mucronata (Leguminosae-Caesalpinioideae), Stachytarpheta glabra (Verbenaceae), and Qualea grandiflora (Vochysiaceae) (Figure 1).
Ant foragers shown are: A) Pseudomyrmex pallidus (Pseudomyrmecinae), B) Camponotus crassus (Formicinae), and C) Ectatomma tuberculatum (Ectatomminae). Arrows indicate the position of the extrafloral nectaries.
We quantified the interactions between ants and S. glabra in July of 2011 and C. mucronata in September 2011 at Parque Estadual do Itacolomi (20°26’S and 43°30’W, for S. glabra and 20°24′S and 43°30′O, for C. mucronata), located in Ouro Preto, in the center of State of Minas Gerais, Brazil. The regional climate, according to Köppen classification is tropical savanna (Cwa) with a mean annual temperature of 21°C and 1700 mm of precipitation. The park covers an area of ca.7000 ha with different savanna physiognomies (44%) . S. glabra is the most common species in the field site, with mean height of 1.5 to 2 m with very small extrafloral nectaries spreading on the upper leaf surface (Antonini et al. 2005). C. mucronata is one of the most common species of campos rupestres (rupestrian fields). C. mucronata is a highly branched shrub with average height of 0.8–1 m, presenting compound leaves with 6–8 leaflets, and one nectary per leaf . Rupestrian fields are characterized by tortuous trees and shrubs immersed in open fields of grasses and rocky outcrop .
We sampled the interactions between ants and individuals of Q. grandiflora in November 2013 at Estação Ecológica Serra das Araras (EESA) (15°38′S and 57°12′W), located in the municipality of Porto Estrela, state of Mato Grosso, Brazil. The regional climate, according to Köppen classification is tropical savanna (Cwa) with a mean annual temperature of 28°C and 1400 mm of precipitation. The reserve area covers 28,700 ha of continuous forest with different savanna physiognomies . Q. grandiflora is a typical and abundant woody tree of the Brazilian Neotropical savanna which reaches up to 15 m tall, bearing paired oval and elevated extrafloral nectaries on the stem next to the insertion of the leaves, and on the bud pedicels .
We assessed the interactions between ants and plants along one transect of 300 m × 10 m for C. mucronata and S. glabra and 500 m × 10 m for Q. grandiflora. Within each transect in each of three different sites, we checked all individuals of the three species considered and recorded all ants collecting nectar from their EFNs. We used a ladder to sample branches of a few larger individuals. We marked all sampled individuals during the day to be reviewed at night. Ant-plant interactions were sampled in two time intervals: Diurnal sampling: 08∶00–12∶00, and Nocturnal sampling: 20∶00–23∶00. However, the number of plants sampled could vary among samplings, since not all plants sampled in the day had ants at night, and vice versa. Moreover, no plants with hemipterans or any other visible liquid-resource were included in our sampling. This restriction was made in order to avoid ant aggregation caused by attraction to another kind of food resource. We also recorded the maximum height of each individual (ground to canopy) at the time of each collection. In this study, we used the number of EFNs rather than plant height to evaluate species-area relationships, mainly because the number of EFNs would make a more robust indicator of resource availability for ants. For this, we estimated the number of EFNs per plant based on the method proposed by Blüthgen et al. , which consists in counting the number of EFNs of three random branches, extracting the mean value, and then multiplying by the number of total branches in the plant. This procedure was done for each individual of C. mucronata and Q. grandiflora. In cases where plants had a variable number of EFNs per leaf (i.e., S. glabra), we calculated the mean number of EFNs using ten random leaves per branch (three branches), and then we multiplied by the number of total branches in the plant.
Ant specimens were identified at the lowest possible taxonomic level with the assistance of identification keys available in the literature and by morphological comparisons of ant species deposited in our reference collections. Ant vouchers were deposited in the: Coleção Entomológica do Departamento de Biodiversidade Evolução e Meio Ambiente da Universidade Federal de Ouro Preto for ants associated to C. mucronata and S. glabra, and Coleção Entomológica of Universidade Federal de Mato Grosso (CEMT) for ants associated to Q. grandiflora.
Network Analysis and Statistics
We used each plant species and associated ants as independent ant-plant networks. Each individual-based ant-plant network was built by an adjacency matrix A, where aij = number of interaction from an individual plant j by the ant species i, and zero otherwise . For each plant species (Chamaecrista mucronata, Stachytarpheta glabra, and Qualea grandiflora), we build interaction ant-plant networks according to the sampling period (diurnal, nocturnal, or both together), totaling nine networks.
We first evaluated the specialization for each of the nine ant-plant networks using an index extremely robust to changes in sampling intensity and the number of interacting species called: H2’ , . In this index, extreme generalization of an ecological network is H2’ = 0 and extreme specialization is H2’ = 1. Then we performed a second approach that involves the search for non-random patterns of interactions commonly found in ant-plant networks studied at the community level. Specifically, we evaluated if selective ant species would visit only a subset of plant individuals visited by the generalist ant species (i.e., nested pattern of ant-plant interactions). For this, we estimated nestedness using the NODF-metric  in the ANINHADO program, a program developed by Guimarães and Guimarães  to perform rapid and automatic calculation of nestedness in bipartite matrices. The values of this metric range from 0 (non-nested) to 100 (perfectly nested). NODF-values are less prone to Type I statistical error when compared to other nestedness indices . Moreover, we tested whether within each ant-plant network there were groups of ant species strongly associated with a particular set of individual plants, as expected in a modular network. For this we used the modularity index (M) based on Simulated Annealing (SA) (range 0–1) ,  using the software MODULAR . This index ranges from 0, no subgroups, to 1, totally separated subgroups . We generated random matrices to test the significance of modularity and nestedness according to the Null Model II (CE)  using functions within the software MODULAR and ANINHADO (n = 1000 randomizations for each network). In this null model, the probability of occurrence of an interaction is proportional to the number of interactions of both ant species and plant individuals . We used these network descriptors and null model because they provide a way to characterize the organization of these networks and allows direct comparison with previous works on ant-plant networks. We used the recent formula proposed by Dáttilo et al.  to describe ant species as peripheral (selective species, those with fewer interactions) or generalist core (generalist species, those with the most interactions) components of the networks: Gc = (ki − kmean)/σk, where ki = mean number of links for a given ant species, kmean = mean number of links for all ant species in the network, and σk = standard deviation of the number of links for ant species. Gc >1 are ant species of the generalist core, and Gc <1 are peripheral ant species. This categorization enabled us to evaluate the temporal turnover in the specific positions of ants within each network (e.g., shifting from peripheral to generalist core between diurnal and nocturnal networks). Moreover, we computed Jaccard’s similarity index (JSI) for each network in order to explore the turnover in ant composition between diurnal and nocturnal networks (see ). Jaccard’s similarity index between D (diurnal) and N (nocturnal) network was computed as follows: JSI (DN) = A/(A+B+C), where A is the number of ant species shared between the two sampling time periods, B is the number of ant species present only in the first period, and C is the number of ant species present only in the second period.
Finally, to test our hypothesis that the number of EFNs could increase the richness of associated ants (i.e., greater degree within a nested network), for each individual-based ant-plant network studied, we used Simple Linear Regressions (SLR) with the number of EFNs as predictor variable and ant richness as a dependent variable. However, before performing SLR analysis, we used a one-sample Student t-test to evaluate the statistical variation of our predictor variable (number of EFNs) around their own mean for all plant species (all p-values <0.05). This analysis is especially important to assess whether the number of EFNs varied enough in our individual plants to be able to influence the richness of associated ants. We did SLR analysis and Student’s t-tests using the vegan package  in the R–software version 184.108.40.206 .
We recorded 32 ant species representing 13 genera and six subfamilies (Appendix S1) interacting with the three studied plant species.
In the diurnal individual-based network of Chamaechrista mucronata, we recorded 11 ant species (two exclusive) foraging on nectaries of 24 individuals, and 11 ant species (two exclusive) on 26 individuals in the nocturnal network (Figure 2A) (n = 26 individuals were sampled in both periods). The total ant richness (the two periods of sampling) was 13 species. Mean ant richness per plant individual was similar in the diurnal (Mean ± SD: 2.12±1.03) and nocturnal (2.03±1.08) networks. However, when the two sampling periods were combined, there was an increase in the mean ant richness per individual (3.14±1.61). In addition, despite the high similarity in ant composition between diurnal and nocturnal networks (Jaccard similarity index = 0.866), the ant assemblage present in the generalist core (i.e., those with the most links) totally changed between the two sampling periods. The ant species Camponotus crassus (Formicinae) and Pseudomyrmex gracillis (Pseudomyrmecinae) were part of the generalist core in the diurnal networks (18.1% of the total ant species sampled), while only Camponotus melanoticus (Formicinae) was part of the generalist core in the nocturnal networks (9.09% of the total ant species sampled). Whereas, three ant species were part of the generalist core including both diurnal and nocturnal sampling: C. crassus, C. melanoticus, and P. gracillis (23% of the total ant species sampled). In addition, we observed an average level of network specialization in both diurnal and nocturnal networks (Diurnal: H2’ = 0.519. Nocturnal: H2’ = 0.499). However, the level of network specialization decreased when we included both diurnal and nocturnal data (H2’ = 0.409) (Table 1).
For each plant species we built ant–plant networks using diurnal, nocturnal, and both together sampling. Within each network, node represents one ant species (left) or plant individual (right), and lines represent ant–plant interactions. The nodes are arranged according their position in the nestedness ranking. Rectangle height is proportional to the number of interactions of each species. Asterisks (*) and red rectangles denote ant species that were present in the generalist core. Species codes are in Appendix S1.
In the diurnal individual-based network of Stachytarpheta glabra, we recorded nine ant species (two exclusive) foraging on nectaries of 41 individuals, and 13 ant species (six exclusive) on 41 individuals in the nocturnal network (Figure 2B) (n = 41 individuals were sampled in both periods). The total ant richness (two sampling periods) was 15 species. Mean ant richness per plant individual was similar in the diurnal (Mean ± SD: 1.85±0.89) and nocturnal (2.21±1.51) networks. However, when the two sampling periods were combined, there was an increase in the mean ant richness per individual (3.02±1.35). In addition, despite the high similarity in ant composition between diurnal and nocturnal networks (Jaccard similarity index = 0.722), the ant assemblage present in the generalist core totally changed between the two sampling periods. The ant species C. crassus and Pseudomyrmex pallidus (Pseudomyrmecinae) were part of the generalist core in the diurnal networks (22.2% of the total ant species sampled), while C. melanoticus and Camponotus novogranadensis (Formicinae) were part of the generalist core in the nocturnal networks (15.3% of the total ant species sampled). Whereas, three ant species were part of the generalist core including both diurnal and nocturnal sampling: C. crassus, C. melanoticus, and C. novogranadensis (23% of the total ant species sampled). In addition, we observed an average level of network specialization in both diurnal and nocturnal networks (Diurnal: H2’ = 0.494. Nocturnal: H2’ = 0.518). However, the level of network specialization decreased when we included both diurnal and nocturnal data (H2’ = 0.401) (Table 1).
In the diurnal individual-based network of Qualea grandiflora, we recorded 13 ant species (no exclusive) foraging on nectaries of 31 individuals, and 15 ant species (two exclusive) on 30 individuals in the nocturnal network (Figure 2C) (n = 31 individuals were sampled in both periods). The total ant richness (two sampling periods) was 15 species. Mean ant richness per plant individual was similar in the diurnal (Mean ± SD: 1.87±0.84) and nocturnal (2.22±1.1) networks. However, when the two sampling periods were combined, there was an increase in the mean ant richness per individual (2.54±0.99). In addition, despite the high similarity in ant composition between diurnal and nocturnal networks (Jaccard similarity index = 0.882), the ant assemblage present in the generalist core totally changed between the two sampling periods. The ant species C. crassus and P. pallidus were part of the generalist core in the diurnal networks (15.3% of the total ant species sampled), while only the giant tropical ant Paraponera clavata (Ponerinae) was part of the generalist core in the nocturnal networks (6.6% of the total ant species sampled). Whereas, two ant species were part of the generalist core including both diurnal and nocturnal sampling: C. crassus and P. clavata (13.3% of the total ant species sampled). In addition, we observed an average level of network specialization in both diurnal and nocturnal networks (Diurnal: H2’ = 0.509. Nocturnal: H2’ = 0.499). However, the level of network specialization decreased when we included both diurnal and nocturnal data (H2’ = 0.424) (Table 1).
Evaluating non-random patterns of ant-plant interaction within each network, we observed that all individual-based networks for the three plant species studied, exhibited a significantly nested network topology (NODF values ranging from 27.84 to 58.62. all p-values <0.05), suggesting that the interactions recorded for plant individuals scarcely visited by ants are a cohesive subset of the interactions found on the most visited individuals. In addition, no network was significantly modular when compared with the neutral patterns of ant-plant interactions (null models) (M values ranging from 0.293 to 0.421. all p-values >0.05) (Table 1), and therefore, there is no group of ant species that feed specifically on a particular group of plant individuals. Finally, when we tested the hypothesis of species-area relationship, we observed that increasing the number of EFNs does not cause an increase in ant richness (i.e., number of links) foraging on such individuals in any of the studied networks (Simple Linear Regression: all p-values >0.05).
Using an intrapopulation approach and three plant species, we evaluated the structure of individual-based ant-plant networks in both diurnal and nocturnal sampling in the Brazilian Neotropical savanna. Our results show that diurnal ant-plant networks exhibited the same pattern of interactions when compared with nocturnal networks. However, despite the high similarity in ant composition between the two sampling periods, ant composition in the generalist core totally changes between diurnal and nocturnal sampling for all plant species studied. Whereas, we did not observe any effect of the number of EFNs on ant richness foraging on plants in any of the studied ant-plant network.
Studies dealing with ecological interaction networks at the community level are steadily growing in the literature , , , although only a few address intrapopulation variations within ecological networks , , . Furthermore, there is an on-going discussion on the standardization of sampling efforts  and, in some cases, additional nocturnal sampling . Here we show that individual-based ant-plant networks exhibited the same patterns of interactions found in networks involving ants and EFN-bearing plants studied at the ecological community level: a nested and non-modular pattern of ant-plant interactions and an average level of network specialization , , , , . Interestingly, these non-random patterns were stable in all our intrapopulation ant-plant networks regardless of the sampling period (diurnal, nocturnal, or both periods together).
At the community level, we know that different factors can structure both nested and non-modular patterns into ant-plant networks, for example: temperature, precipitation, soil pH, canopy openness, nectar features and phenology, abundance, body size, and dominance hierarchy among ants , , , , , , , , . So, why do individual-based ant-plant networks exhibit similar patterns of ant-plant interactions when compared to ant-plant networks evaluated at the community level? We observed that species of the genus Camponotus were the most extreme generalist ant species in our networks, followed by Pseudomyrmex and Paraponera. Except for Paraponera, all other genera are particularly abundant in the Brazilian Neotropical savanna . Moreover, Camponotus, Pseudomyrmex and Paraponera are strongly associated with EFN-bearing plants , . Such tree-dwelling genera are highly aggressive, occupy large territories, have mutually exclusive patterns of distribution within an environment, and use diversified strategies to monopolize this highly nutritive resource , , , . Therefore, based on our knowledge of natural history in the field and literature (see information above), we hypothesize that dominance hierarchy could really be the main mechanism that structures the nested pattern in our ant-plant networks . Competitive dominance can occur by: 1) numerical dominance, as observed for Camponotus ; or 2) aggressive displacement of competitors by solitary foragers, as in P. clavata (pers. obs.). In both cases, dominant ant species would visit most plants in an environment (i.e., generalist core components) while submissive and subordinate ant species co-occur with dominant species only in a few plants , , ,  Thus, this pattern of ant-plant interactions would generate the cohesive subgroups expected in nested intrapopulation networks. Additionally, the lack of a modular pattern in our ant-plant networks may be explained because the resource offered by EFN-bearing plants is seasonal over space-time, ant species do not exhibit “fidelity” of foraging on the same group of individual plants , , . Therefore, when a plant does not secrete nectar, the ants can use other resources available on the foliage , and therefore interactions tend to be more generalized .
When we included both diurnal and nocturnal sampling together, we observed a decrease in network specialization for all plant species. This was possibly due to the increase in the mean richness of associated ants per individual plant caused by the entrance of new highly generalized ant species. Using both diurnal and nocturnal sampling (greater sampling effort), we should expect an increase in the number of species and interactions  and, consequently, an increase in the probability to find unique interactions and ant species foraging in only particular period (i.e., ‘forbidden links’) , . The record of these forbidden links can exert a strong effect on the structure of an ecological network , , , mainly because it reduces potential sampling bias, as for example, the record of only spatially abundant species. We know that ant foraging can be strictly diurnal, nocturnal, or both , including ants that feed on EFN-bearing plants , . The temporal partitioning in the daily rhythm of ant foraging on extrafloral nectaries of all three plant species studied by us could directly reduce competitive interactions among sympatric ant species , , and promote coexistence of greater number of ant species foraging on the same plant individual. Therefore, when recording ant-plant interactions in different sampling periods, we can increase our understanding of the true role of ant species within an ecosystem.
In our study, some ant species changed their role in network structure, shifting from peripheral to generalist core between diurnal and nocturnal sampling. In other words, this “night-shift” indicates that the ecological dynamics of these ant-plant interactions can be temporally partitioned (day and night) at a small spatial scale. For example, in sites with high temperature and low humidity during the day (e.g., Brazilian Neotropical savanna and deserts) due to ecophysiological limitations there is a strong difference in the composition and abundance of ants and herbivores over a 24 h period , . We believe this temporal turnover of ant composition in the same individual plant could offer a better protection against a wider range of herbivores, since the protection ability of the host plant can vary among associated ant species . In other words, it is possible that in some cases, processes shaping mutualistic networks formed by protective ants and plants may be underestimated by diurnal sampling alone, possibly because most herbivore activity is also during the night, when weather conditions are milder, that is, temperature falls to 20–22°C and humidity increases to 0–60%.
Despite a significant variation in the number of EFNs, our hypothesis that plant individuals with higher number of EFNs have greater richness of associated ants was not supported. At least two alternative and non-mutually exclusive hypotheses could explain this pattern: 1) due to extrafloral nectar being a highly nutritive and predictable food resource, competitively superior ants could dominate this resource and allow the coexistence of only a few other ant species , , and 2) due to the limited number of workers in an ant colony, it is possible that the dominant species can only monopolize low trees and with low or hardly any canopy connectivity (i.e., resource access) . In the first case, ant richness would not increase during plant ontogeny (number of EFNs), because competitively superior and territorial ground-nesting ant species could limit the access to the EFN-bearing plants by submissive ant species , . This pattern can be stronger in environments where plants are relatively low and with few “active” nectaries. Therefore, most of EFN-bearing plants in these environments could support only a few ant species. In the second case, plants are relatively low in the Brazilian Neotropical savanna, for instance, in our study the largest tree was 7.43 m tall. However, in environments where tree height may reach up to 50 m (e.g., some tropical rainforests) , it is expected that different ant species can coexist on the same individual . This is because plant growth leads to gradual increase in the number of microhabitats for nesting sites , and in the abundance of available food resources: liquids secretions (nectaries and honeydew-producing hemipterans) ,  and herbivore insects (important resource for arboreal predatory ants) .
Finally, there still remains a degree of uncertainty about the true role of species-area relationships in ecological ant-plant networks, and thus future studies should evaluate the effect of plant height on ant-plant networks in structurally more complex environments.
In summary, we demonstrate that (1) our individual-based ant-plant networks exhibited similar patterns of interactions when compared with ant-plant networks studied at the community level, suggesting that ant-plant interactions are much more dynamic than expected a priori, since ant species can change their role in network structure between diurnal and nocturnal period; (2) although plants with higher numbers of EFNs do not exhibit greater richness of foraging ants, our study produces basic information that increases our confidence in how individual-based ant-plant networks are structured, and the need to consider nocturnal records in ant-plant network sampling design so as to decrease inappropriate inferences; and (3) even though there is still much to do in order to have a better knowledge of the patterns and processes related to intrapopulation ant-plant networks, mainly to the true role of ant dominance hierarchy and availability of food resources in such networks, many things can be done when our suggestions are considered.
Species codes of ants recorded in two time intervals (D = diurnal. N = nocturnal) foraging on individuals of (C) Chamaecrista mucronata (Leguminosae–Caesalpinioideae), (S) Stachytarpheta glabra (Verbenaceae), and (Q) Qualea grandiflora (Vochysiaceae) in the Brazilian Neotropical Savanna. Please see Figure 1 and text for more information.
Sérvio P. Ribeiro has provided useful information on species-area relationships. The authors also wish to thank Lucas Kaminski and Denise Lange for support in the identification of Qualea grandiflora and to Gustavo Araújo for his help in the fieldwork. WD was supported by the CNPq (no. 237339/2012-9) and CONACYT (no. 489746). MCLV was supported by CNPq. RF, CAQG and MSAS were supported by CAPES. KDC has a researcher fellowship from CNPq.
Conceived and designed the experiments: WD RF TJI CDC KDC VRG. Performed the experiments: WD RF CAQG MSAS MCLV. Analyzed the data: WD. Contributed reagents/materials/analysis tools: WD TJI CDC KDC VRG. Contributed to the writing of the manuscript: WD RF CAQG MSAS MCLV TJI CDC KDC VRG.
- 1. Rico–Gray V, Oliveira PS (2007) The ecology and evolution of ant–plant interactions. Chicago: University of Chicago Press. 320 p.
- 2. Rosumek FB, Silveira FAO, Neves FS, Barbosa NPD, Diniz L, et al. (2009) Ants on plants: a meta-analysis of the role of ants as plant biotic defenses. Oecologia 160: 537–549. doi: 10.1007/s00442-009-1309-x
- 3. Koptur S, Rico–Gray V, Palacios–Rios M (1998) Ant protection of the nectaried fern Polypodium plebeium in central Mexico. Am J Bot 85: 736–739. doi: 10.2307/2446544
- 4. Guimaraes PR, Rico–Gray V, dos Reis SF, Thompson JN (2006) Asymmetries in specialization in ant–plant mutualistic networks. Proc R Soc Lond B Biol Sci 273: 2041–2047. doi: 10.1098/rspb.2006.3548
- 5. Chamberlain SA, Kilpatrick JR, Holland JN (2010) Do extrafloral nectar resources, species abundances, and body sizes contribute to the structure of ant–plant mutualistic networks?. Oecologia 164: 741–750. doi: 10.1007/s00442-010-1673-6
- 6. Dáttilo W, Guimarães PR, Izzo TJ (2013a) Spatial structure of ant–plant mutualistic networks. Oikos 122: 1643–1648. doi: 10.1111/j.1600-0706.2013.00562.x
- 7. Dáttilo W (2012) Different tolerances of symbiotic and nonsymbiotic ant–plant networks to species extinctions. Net Biol 2: 127–138.
- 8. Lange D, Dáttilo W, Del–Claro K (2013) Influence of extrafloral nectary phenology on ant–plant mutualistic networks in a Neotropical Savanna. Ecol Entomol 38: 463–469. doi: 10.1111/een.12036
- 9. Bascompte J, Jordano P, Melián CJ, Olesen JM (2003) The nested assembly of plant–animal mutualistic networks. Proc Nat Acad Scienc USA 100: 9383–87. doi: 10.1073/pnas.1633576100
- 10. Díaz–Castelazo C, Sanchez–Galván IR, Guimaraes PR, Raimundo RLG, Rico–Gray V (2013) Long–term temporal variation in the organization of an ant–plant network. Ann Bot 111: 1285–1293. doi: 10.1093/aob/mct071
- 11. Dáttilo W, Marquitti FMD, Guimaraes PR, Izzo TJ (2014a) The structure of ant–plant ecological networks: is abundance enough?. Ecology 95: 475–485. doi: 10.1890/12-1647.1
- 12. Dáttilo W, Izzo TJ, Vasconcelos HL, Rico–Gray V (2013b) Strength of the modular pattern in Amazonian symbiotic ant–plant networks. Arthropod–Plant Interac 7: 455–461. doi: 10.1007/s11829-013-9256-1
- 13. De la Fuente MAS, Marquis RJ (1999) The role of ant–tended extrafloral nectaries in the protection and benefit of a Neotropical rainforest tree. Oecologia 118: 192–202. doi: 10.1007/s004420050718
- 14. Oliveira PS, Rico–Gray V, Díaz–Castelazo C, Castillo–Guevara C (1999) Interaction between ants, extrafloral nectaries and insect herbivores in Neotropical coastal sand dunes: herbivore deterrence by visiting ants increases fruit set in Opuntia stricta (Cactaceae). Funct Ecol 13: 623–631. doi: 10.1046/j.1365-2435.1999.00360.x
- 15. Rios RS, Marquis RJ, Flunker JC (2008) Population variation in plant traits associated with ant attraction and herbivory in Chamaecrista fasciculate (Fabaceae). Oecologia 156: 577–588. doi: 10.1007/s00442-008-1024-z
- 16. Baker–Méio B, Marquis RJ (2012) Context–dependent benefits from ant–plant mutualism in three sympatric varieties of Chamaecrista desvauxii. J Ecol 100: 242–252. doi: 10.1111/j.1365-2745.2011.01892.x
- 17. Alves–Silva E, Del–Claro K (2013) Effect of post–fire resprouting on leaf fluctuating asymmetry, extrafloral nectar quality, and ant–plant–herbivore interactions. Naturwissenschaften 100: 525–532. doi: 10.1007/s00114-013-1048-z
- 18. Schupp EW, Feener DH (1991) Phylogeny, lifeform, and habitat dependence of ant–defended plants in a Panamanian forest. In: Cutler DF, Huxler CR, editors. Ant–Plant Interactions. Oxford: Oxford University Press. 175–197.
- 19. Heil M, Fiala B, Baumann B, Linsenmair KE (2000) Temporal, spatial and biotic variations in extrafloral nectar secretion by Macaranga tanarius. Funct Ecol 14: 749–75. doi: 10.1046/j.1365-2435.2000.00480.x
- 20. Falcão JCF, Dáttilo W, Izzo TJ (2014) Temporal variation in extrafloral nectar secretion in different ontogenic stages of the fruits of Alibertia verrucosa S. Moore (Rubiaceae) in a Neotropical savanna. J Plant Interac 9: 137–142. doi: 10.1080/17429145.2013.782513
- 21. Byk J, Del–Claro K (2011) Ant-plant interaction in the Neotropical savanna: direct beneficial effects of extrafloral nectar on ant colony fitness. Popul Ecol 53: 327–332. doi: 10.1007/s10144-010-0240-7
- 22. Hölldobler B, Wilson EO (1990) The ants. Cambridge: Harvard University Press. 732 p.
- 23. Díaz–Castelazo C, Rico–Gray V, Oliveira PS, Cuautle M (2004) Extrafloral nectary–mediated ant–plant interactions in the coastal vegetation of Veracruz, Mexico: Richness, occurrence, seasonality and ant foraging patterns. Ecoscience 11: 472–481.
- 24. MacArthur RH, Wilson EO (1967) The Theory of Island Biogeography. Princeton: Princeton University Press. 203 p.
- 25. Blüthgen N, Verhaagh M, Goitía W, Jaffé K, Morawetz W, et al. (2000) How plants shape the ant community in the Amazonian rainforest canopy: the key role of extrafloral nectaries and homopteran honeydew. Oecologia 125: 229–240. doi: 10.1007/s004420000449
- 26. Davidson DW, Cook SC, Snelling RR, Chua TH (2003) Explaining the abundance of ants in lowland tropical rainforest canopies. Science 300: 969–972. doi: 10.1126/science.1082074
- 27. Campos RI, Vasconcelos HL, Ribeiro SP, Neves FS, Soares JP (2006) Relationship between tree size and insect assemblages associated with Anadenanthera macrocarpa. Ecography 29: 442–450. doi: 10.1111/j.2006.0906-7590.04520.x
- 28. Villamil N, Márquez-Guzmán J, Boege K (2013) Understanding ontogenetic trajectories of indirect defence: ecological and anatomical constraints in the production of extrafloral nectaries. Ann Bot 112: 701–709. doi: 10.1093/aob/mct005
- 29. Gómez JM, Perfectti F (2012) Fitness consequences of centrality in mutualistic individual–based networks. Proc R Soc Lond B Biol Sci 279: 1754–1760. doi: 10.1098/rspb.2011.2244
- 30. Heil M, McKey D (2003) Protective ant–plant interactions as model systems in ecological and evolutionary research. Annu Rev Ecol Syst 34: 425–453.
- 31. Fujaco MAG, Leite MGP, Messias MCTB (2010) Multitemporal analysis of land use changes in the Itacolomi State Park (MG) by geoprocessing techniques. Rev Esc Minas 63: 695–701. doi: 10.1590/s0370-44672010000400016
- 32. Dutra VF, Garcia FVP, Lima HC (2008) Caesalpinioideae (Leguminosae) nos Campos Rupestres do Parque Estadual do Itacolomi, MG, Brasil. Acta Bot Bras 22: 547–558. doi: 10.1590/s0102-33062008000200021
- 33. Zardo DC, Carneiro AP, Lima LG, Santos–Filho M (2010) Comunidade de artrópodes associada à serrapilheira de cerrado e mata de galeria, na Estação Ecológica Serra das Araras – Mato Grosso, Brasil. Rev Uniara 13: 105–113.
- 34. Costa FMCB, Oliveira–Filho AT, Oliveira PS (1992) The role of extrafloral nectaries in Qualea grandiflora (Vochysiaceae) in limiting herbivory: an experiment of ant protection in cerrado vegetation. Ecol Entomol 17: 363–365. doi: 10.1111/j.1365-2311.1992.tb01071.x
- 35. Blüthgen N, Menzel F, Blüthgen N (2006) Measuring specialization in species interaction networks. BMC Ecol 6: 12–18. doi: 10.1186/1472-6785-6-9
- 36. Blüthgen N, Menzel F, Hovestadt T, Fiala B, Blüthgen N (2007) Specialization, constraints, and conflicting interests in mutualistic networks. Curr Biol 17: 341–346. doi: 10.1016/j.cub.2006.12.039
- 37. Almeida–Neto M, Guimaraes P, Guimaraes PR, Loyola RD, Urlich W (2008) A consistent metric for nestedness analysis in ecological systems: reconciling concept and measurement. Oikos 117: 1227–1239. doi: 10.1111/j.0030-1299.2008.16644.x
- 38. Guimarães–Jr PR, Guimarães PR (2006) Improving the analyses of nestedness for large sets of matrices. Environ Modell Softw 21: 1512–1513. doi: 10.1016/j.envsoft.2006.04.002
- 39. Guimerà R, Sales–Pardo M, Amaral LAN (2004) Modularity from fluctuations in random graphs and complex networks. Phys Rev E 70: 025101. doi: 10.1103/physreve.70.025101
- 40. Guimerà R, Amaral LAN (2005) Functional cartography of complex metabolic networks. Nature 433: 895–900. doi: 10.1038/nature03288
- 41. Marquitti FMD, Guimaraes PR, Pires MM, Bittencourt LF (2014) MODULAR: Software for the Autonomous Computation of Modularity in Large Network Sets. Ecography 37: 221–224. doi: 10.1111/j.1600-0587.2013.00506.x
- 42. Olesen JM, Bascompte J, Dupont YK, Jordano P (2007) The modularity of pollination networks. Proc Natl Acad Sci USA 104: 19891–19896. doi: 10.1073/pnas.0706375104
- 43. Díaz–Castelazo C, Sanchez–Galván IR, Guimaraes PR, Raimundo RLG, Rico–Gray V (2013) Long–term temporal variation in the organization of an ant–plant network. Ann Bot 111: 1285–1293. doi: 10.1093/aob/mct071
- 44. Oksanen J, Kindt R, Legendre P, O’Hara RB (2007) Vegan: Community Ecology Package, Version 1.8. Available: http://cran.r-project.org. Accessed 2014 Jan 15.
- 45. R Development Core Team (2012) R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria.
- 46. Ings TC, Montoya JM, Bascompte J, Blüthgen N, Brown L, et al. (2009) Review: Ecological networks–beyond food webs. J Anim Ecol 78: 253–269. doi: 10.1111/j.1365-2656.2008.01460.x
- 47. Vázquez D, Blüthgen N, Cagnolo L, Chacoff NP (2009) Uniting pattern and process in plant–animal mutualistic networks: a review. Ann Bot 103: 1445–1457. doi: 10.1093/aob/mcp057
- 48. Hagen M, Kissiling WD, Rasmussen C, Aguiar MAM, Brown DW, et al. (2012) Biodiversity, species interactions and ecological networks in a fragmented world. Adv Ecol Res 46: 89–210.
- 49. Araújo MS, Martins EG, Cruz LD, Fernandes FR, Linhares AX, et al. (2010) Nested diets: a novel pattern of individual–level resource use. Oikos 119: 81–88. doi: 10.1111/j.1600-0706.2009.17624.x
- 50. Cantor M, Pires MM, Longo GG, Guimaraes PR, Setz EZF (2013) Individual variation in resource use by opossums leading to nested fruit consumption. Oikos 122: 1085–1093. doi: 10.1111/j.1600-0706.2012.00070.x
- 51. Dáttilo W, Serio–Silva JC, Chapman C, Rico–Gray V (2014) Highly nested diets in intrapopulation tropical monkey–resource food webs. Am J Primatol (in press). doi:10.1002/ajp.22261.
- 52. Nielsen A, Bascompte J (2007) Ecological networks, nestedness and sampling effort. J Ecol 95: 1134–1141. doi: 10.1111/j.1365-2745.2007.01271.x
- 53. Devoto M, Bailey S, Memmott J (2011) The “night shift”: nocturnal pollen–transport networks in a boreal pine forest. Ecol Entomol 36: 25–35. doi: 10.1111/j.1365-2311.2010.01247.x
- 54. Dáttilo W (2012) Different tolerances of symbiotic and nonsymbiotic ant–plant networks to species extinctions. Net Biol 2: 127–138.
- 55. Chamberlain SA, Holland JN (2009) Body size predicts degree in ant–plant mutualistic networks. Func Ecol 23: 196–202. doi: 10.1111/j.1365-2435.2008.01472.x
- 56. Chamberlain SA, Kilpatrick JR, Holland JN (2010) Do extrafloral nectar resources, species abundances, and body sizes contribute to the structure of ant–plant mutualistic networks?. Oecologia 164: 741–750. doi: 10.1007/s00442-010-1673-6
- 57. Rico–Gray V, Díaz–Castelazo C, Ramirez–Hernandez A, Guimarães PR, Holland JN (2012) Abiotic factors shape temporal variation in the structure of an ant–plant network. Arthropod–Plant Interact 6: 289–295. doi: 10.1007/s11829-011-9170-3
- 58. Dáttilo W, Rico–Gray V, Rodrigues DJ, Izzo TJ (2013) Soil and vegetation features determine the nested pattern of ant–plant networks in a tropical rainforest. Ecol Entomol 38: 374–380. doi: 10.1111/een.12029
- 59. Dáttilo W, Sánchez–Galván I, Lange D, Del–Claro K, Rico–Gray V (2014) Importance of interaction frequency in analysis of ant–plant networks in tropical environments. J Trop Ecol 30: 165–168. doi: 10.1017/s0266467413000813
- 60. Dáttilo W, Díaz-Castelazo C, Rico-Gray V (2014) Ant dominance hierarchy determines the nested pattern in ant-plant networks. Biol J Linn Soc (in press).
- 61. Santos GMM, Dáttilo W, Presley SJ (2014) The seasonal dynamic of ant-flower networks in a semi-arid tropical environment. Ecol Entomol (in press)
- 62. Silvestre R, Brandão CRF (2001) Formigas (Hymenoptera, Formicidae) atraídas a iscas em uma ilha de cerrado no município de Cajuru, estado de São Paulo, Brasil. Rev Bras Entomol 44: 71–77.
- 63. Oliveira PS, Brandão CRF (1991) The ant community associated with extrafloral nectaries in the Brazilian cerrados. In: Cutler DF, Huxler CR, editors. Ant–Plant Interactions. Oxford: Oxford University Press. 198–212.
- 64. Davidson DW, Cook SC, Snelling RR (2004) Liquid–feeding performances of ants (Hymenoptera: Formicidae): ecological and evolutionary implications–external site. Oecologia 139: 255–266. doi: 10.1007/s00442-004-1508-4
- 65. Del–Claro K, Oliveira PS (1999) Ant-Homoptera Interactions in a Neotropical Savanna: The Honeydew–Producing Treehopper, Guayaquila xiphias (Membracidae), and its Associated Ant Fauna on Didymopanax vinosum (Araliaceae). Biotropica 31: 135–144. doi: 10.2307/2663967
- 66. Livingston GF, Philpott SM (2010) A metacommmunity approach to co–occurrence patterns and the core–satellite hypothesis in a community of tropical arboreal ants. Ecol Res 25: 1129–1140. doi: 10.1007/s11284-010-0738-7
- 67. Schoereder JH, Sobrinho TG, Madureira MS, Ribas CR, Oliveira PS (2010.) The arboreal ant community visiting extrafloral nectaries in the Neotropical cerrado savanna. Terr Arthropod Rev 3: 3–27. doi: 10.1163/187498310x487785
- 68. Blüthgen N, Fiedler K (2004) Competition for composition: lessons from nectar–feeding ant communities. Ecology 85: 1479–1485. doi: 10.1890/03-0430
- 69. Blüthgen N, Fiedler K (2004) Preferences for sugars and amino acids and their conditionality in a diverse nectar–feeding ant community. J Anim Ecol 73: 155–166. doi: 10.1111/j.1365-2656.2004.00789.x
- 70. Blüthgen N, Stork NE (2007) Ant mosaics in a tropical rainforest in Australia and elsewhere: a critical review. Austral Ecol 32: 93–104. doi: 10.1111/j.1442-9993.2007.01744.x
- 71. Rico–Gray V, Morais HC (2006) Efecto de una fuente de alimento experimental sobre una asociación hormiga–hemíptero. Acta Zool Mex 22: 23–28.
- 72. Olesen JM, Bascompte J, Dupont YL, Elberling H, Rasmussen C, et al. (2011) Missing and forbidden links in mutualistic networks. Proc R Soc Lond B Biol Sci 278: 725–732. doi: 10.1098/rspb.2010.1371
- 73. Jordano P, Bascompte J, Olesen JM (2003) Invariant properties in coevolutionary networks of plant–animal interactions. Ecol Lett 6: 69–81. doi: 10.1046/j.1461-0248.2003.00403.x
- 74. Vázquez DP (2005) Degree distribution in plant–animal mutualistic networks: forbidden links or random interactions? Oikos 108: 421–426. doi: 10.1111/j.0030-1299.2005.13619.x
- 75. Baldock KC, Memmott J, Ruiz-Guajardo JC, Roze D, Stone GN (2011) Daily temporal structure in African savanna flower visitation networks and consequences for network sampling. Ecology 92: 687–698. doi: 10.1890/10-1110.1
- 76. Oliveira PS, Rico–Gray V, Díaz–Castelazo C, Castillo–Guevara C (1999) Interaction between ants, extrafloral nectaries and insect herbivores in Neotropical coastal sand dunes: herbivore deterrence by visiting ants increases fruit set in Opuntia stricta (Cactaceae). Funct Ecol 13: 623–631. doi: 10.1046/j.1365-2435.1999.00360.x
- 77. Strauss SY, Stanton ML, Emery NC, Bradley CA, Carleton C, et al. (2009) Cryptic seedling herbivory by nocturnal introduced generalists impacts survival, performance of native and exotic plants. Ecology 90: 419–429. doi: 10.1890/07-1533.1
- 78. Byk J, Del–Claro K (2010) Nectar and pollen–gathering Cephalotes ants provide no protection against herbivory: a new manipulative experiment to test ant protective capabilities. Acta Ethol 13: 33–38. doi: 10.1007/s10211-010-0071-8
- 79. Powell S, Costa AN, Lopes CT, Vasconcelos HL (2011) Canopy connectivity and the availability of diverse nesting resources affect species coexistence in arboreal ants. J Anim Ecol 80: 352–360. doi: 10.1111/j.1365-2656.2010.01779.x
- 80. Milliken W (1998) Structure and composition of one hectare of Central Amazonian terra firme forest. Biotropica 30: 530–537. doi: 10.1111/j.1744-7429.1998.tb00093.x
- 81. Djieto–Lordon C, Dejean A, Gibernau M, Hossaert–McKey M, McKey MD (2004) Symbiotic mutualism with a community of opportunistic ants: protection, competition, and ant occupancy of the myrmecophyte Barteria nigritana (Passifloraceae). Acta Oecol 26: 109–116. doi: 10.1016/j.actao.2004.03.007
- 82. Floren A, Biun A, Linsenmair EK (2002) Arboreal ants as key predators in tropical lowland rainforest trees. Oecologia 131: 137–144. doi: 10.1007/s00442-002-0874-z