Evidences that human disturbance simplify the ant fauna associated a Stachytarpheta glabra Cham . ( Verbenaceae ) compromising the benefits of ant-plant mutualism

Interaction among species, like ants and plants through extrafloral nectaries (EFNs), are important components of ecological communities’ evolution. However, the effect of human disturbance on such specific interactions and its ecological consequences is poorly understood. This study evaluated the outcomes of mutualism between ants and the EFN-bearing plant Stachytarpheta glabra under anthropogenic disturbance. We compared the arthropod fauna composition between two groups of twenty plant individuals, one in an area disturbed by human activities and one in a preserved area. We also check the plant investment in herbivory defense and the consequential leaf damage by herbivore. Our results indicate that such disturbances cause simplification of the associated fauna and lack of proper ant mutualist. This led to four times more herbivory on plants of disturbed areas, despite the equal amount of EFN and ant visitors and low abundance of herbivores. The high pressure of herbivory may difficult the re-establishment of S. glabra, an important pioneer species in ferruginous fields, therefore it may affect resilience of this fragile ecological community.


Introduction
Some plant species have adaptations to promote and facilitate interaction with ants: food bodies, domatia and extrafloral nectaries (hereafter EFNs;Webber et al., 2007).In this interactions, plants provide food or shelter to ants as a reward for protection against herbivores (Bentley, 1977;Del-Claro et al., 1996;Rosumek et al., 2009).As a result, these plants may have reduced herbivory on leaves, buds and flowers and a high reproductive success (Almeida and Figueiredo, 2003;Heil et al., 2004;Rico-Gray and Oliveira, 2007;Rosumek et al., 2009).However, the outcomes of these interactions may change over time, seasonality, identity and number of species involved and level of environmental disturbance (Del-Claro and Oliveira, 1999Oliveira, , 2000;;Lach et al., 2010;Rico-Gray et al., 2004).
Environmental disturbances affect ecological communities in many ways, specially decreasing diversity of species and eroding interaction networks (Del-Claro and Torezan-Silingardi, 2009;Naeem, 1998;Philpott et al., 2010;Pimm, 1984).This may change the direction and strength of ecological interactions (Piovia-Scott, 2011).As a consequence, changes in the outcomes of these interactions may have further effects on the entire ecosystem (Margulis and Fester, 1991;Six et al., 2011).However, relatively little attention has been paid to its effects on mutualisms compared with other interaction like predation and competition (Begon et al., 1996;Six et al., 2011).
Environmental perturbations may cause loss of photosynthetic tissue and lead to the plant regrowth.At this stage the plant is susceptible to herbivore attack due to low physical and chemical protection of young leaves and their nutritional quality (Spiller and Agrawal, 2003).Therefore myrmecophilous plants tend to invest more in production of ant attractors and enhance their protection (Agrawal and Rutter, 1998;Heil et al., 2001;Mondor and Addicott, 2003;Ness, 2003;Wäckers et al., 2001).In a study on the effect of fire on ant-plant interaction through EFNs, Alves-Silva (2011) demonstrated that Banisteriopsis malifolia produces greater amount of nectar in post-fire regrowth.This leads to increase in the number of ants visiting and protecting the plant against herbivores.This author has once again demonstrated this phenomenon in a different year for the same system Banisteriopsis malifoliaant (Alves-Silva and Del-Claro, 2013).The breaking and pruning of plants, such as caused by windstorms, can lead to increased nectar production in young leaves and consequently the amount of ants (Piovia-Scott, 2011).
Stachytarpheta glabra Cham.(Verbenaceae) is a rupestrian shrub, which has leaves carrying many EFNs, a food source for many arthropods, mainly ants (Antonini et al., 2005;Vincent et al., 2002).This plant is endemic of Serra do Espinhaço, the largest Brazilian mountain system in the south eastern, rich in endemism and sensitive to human impacts (Drummond et al., 2005;Giulietti et al., 2005;Jacobi and Carmo, 2008).S. glabra grows only in the ironstone outcrops vegetation, known as Cangas (Viana-Silva and Jacobi, 2012;Viana and Lombardi, 2007), a ecosystem exclusively of this region.These habitats are among the most endangered and little studied in the Neotropical region and the potential recovery of degraded areas is unknown (Jacobi et al., 2008;Viana and Lombardi, 2007).In this abiotically constrained environment, the ant fauna and their behavior are poorly known even though being important on multitrophic interactions with plants and herbivores (Byk and Del-Claro, 2011;Fagundes et al., 2013;Romero, 2002;Sousa-Souto et al., 2008).
In this study, we evaluated the effects of human impacts on the interaction between ants and S. glabra and the richness and abundance of guilds of arthropods associated with S. glabra.We assume that the human disturbance will favor the ant-plant interaction through increased production of EFNs and attracting ants and reducing the amount of herbivores and herbivory on S. glabra (Alves-Silva and Del-Claro, 2013;Piovia-Scott, 2011).We tested the hypothesis that environmental degradation leads to lower abundance and richness, and different composition of arthropods associated with S. glabra (hypothesis 1).The plant produces higher number of EFNs in disturbed area (2), attracting a larger number of individuals and species of ants (3).This leads to smaller amount of herbivores and plant herbivory in the disturbed area (4).Other feeding guilds of arthropods, such as predator spiders, are only affected by environmental degradation and have lower abundance and richness in plants of disturbed area (5).

Material and Methods
The study was conducted from June to July of 2010 in a fragment of ferruginous rupestrian field with 20 ha, at 1.250m high, located in Ouro Preto,Brazil (20°18'45"S,43°33'45" W).The soil consists of a concretion of iron sesquioxide forming hard substrates that are partially penetrable by plants (Antonini et al., 2005;Rosière and Chemale, 2000).The vegetation is dominated by shrub and tree species, especially Veloziaceae, Asteraceae and Malpighiaceae (Jacobi and Antonini, 2008).We selected two neighboring fragments of vegetation: one preserved and one disturbed.The disturbance consists in annual fires and disposal of exogenous soil, in the midst of which the populations grows.In addition, the area was invaded associada e a ausência de uma formiga mutualista anti-herbívoros.Isto leva a quatro vezes mais herbivoria nas plantas da área impactada, a despeito da mesma quantidade encontrada para NEF e formigas visitantes.A grande pressão de herbivoria pode dificultar o reestabelecimento de S. glabra, uma espécie pioneira importante nos campos ferruginosos, dessa forma, afetando a resiliência dessa comunidade ecológica ameaçada.
In each fragment, we selected 20 plants with similar architecture (height, branches, and crown diameter) and minimum distance of 5m from each other (Nahas and Del-Claro, 2011), which we consider as independent samples.We collected the arthropods with an entomological umbrella (Ribeiro et al., 2005).Each plant was hit 10 times with a 1m of length stick and the specimens that fell on the umbrella were collected.Insect species were identified at the Laboratory of Evolutionary Ecology of Canopy Insects and Natural Succession, at the Federal University of Ouro Preto (UFOP), and spiders in the Laboratory of Arachnology, at the Federal University of Minas Gerais (UFMG).Species vouchers were deposited in both collections.The species were separated into feeding guilds according to the literature of the taxon (Gibbs et al., 2010;Marinoni et al., 2001;Nahas et al., 2012).
In order to quantify leaf area lost by herbivory, we randomly collect 25 leaves of each plant.The leaves were scanned and the damage area quantified using the ImageJ software.To assess plant defenses against herbivory, we measured the specific leaf area (SLA; physical defense caused by sclerophilly) and number of active extrafloral nectaries (EFNs; biotic defense in form of ant attractor).For the first variable, the leaves were dried, weighed and measured and the values of SLA per plant were obtained according to the protocol proposed by Cornelissen et al. (2003).For the second variable, all nectaries of each leaf were counted, and also obtained the average per plant.

Data analysis
All variables were transformed by square root function for account data, or logarithm function for measure data.Normality and homoscedasticity tested using Lilliefors and Lavene analysis, respectively.The mean richness and abundance of all arthropods (dependent variables) were compared between areas (independent variable) using t-test for independent samples (plant individuals) We compared the composition of arthropods between the areas with Analysis of Similarity using Bray-Curtis Index (ANOSIM; Clarke, 1993) followed by Non-Metric Multidimensional Scaling (NMDS).The number of NEFs per leaf were compared between areas (independent) using t-test for independent samples.The difference in composition of the ant fauna (dependent: species) among areas (independent), and abundance of species (dependent) were tested by Permutation Multivariate Analysis of Variance (NPMANOVA; Anderson, 2001) and the contribution of each species to the difference obtained by SIMPER.The leaf area, specific leaf area, abundance and richness of herbivores, and the leaf area lost (all dependent) were also compared between areas (independent) using t-test for independent samples.Finally, the mean abundance of spiders and other arthropods (dependent) were compared between areas (independent) with Mann-Whitney test.Parametric tests were run with logarithmic transformed data but original data is show in figures and table in form of mean and confidence interval or median and quartiles according with data distribution.

Discussion
As far we know, this is the only study that evaluates the effects of human disturbance ant-plant interactions on Cangas in Brazil.Our study shows that human disturbance, such as urbanization, fire and exotic species, can not only simplify the fauna associated with a plant as mutualistic interactions also harm through loss of key species of the interaction.All the diversity associated with S. glabra proved being vulnerable to human disturbance.Studies of the effects of human disturbance on interactions are scarce but some can be found to community structure.
Urbanization itself can lead to critical changes of arthropod community and affect neighboring native areas, either through the invasion of exotic species or reduction of species richness (McIntyre, 2000).The loss or alteration of the abundance and distribution of species alters ecosystem processes, ecological interactions, ecosystem resilience to changes and consequently the environmental services (Chapin et al., 2000).The entry of invasive species is an important consequence of human alteration of the environment and leads to reduced local diversity (King and Tschinkel, 2008).However, the conservation of interactions, such as mutualisms by seed-dispersers or pollinators, can reduce or even prevent the entry of alien species (Christian, 2001;Kearns et al., 1998).
Environmental disturbances directly reduce plant diversity in a community through changes in environmental conditions such as physical and chemical changes of soil,   destruction of habitats and species removal (McIntyre and Lavorel, 1994).Indirectly, this leads to reduced arthropod diversity through loss of resources and microhabitats (Bolger et al., 2000).For ants is even more serious because their diversity is directly related to heterogeneity and plant richness and diversity of habitats and resources (Endringer et al., 2007;Ribas et al., 2003;Schmidt et al., 2007), since most of the ants eating foods derived from plants (Blüthgen et al., 2003).This loss of species leads to breakage of ecological interactions and compromising of community through disruption of trophic webs, facilitating entry of diseases and exotic species, weakening interactions involving plants and animals (mutualisms and herbivory) and extinguishing top predators (Tylianakis et al., 2008).The loss of keystones species, whose interactions are very important to the community, it is even more detrimental to the maintenance of functionality of the community (Mills et al., 1993).Anthropogenic disturbances cause a reduction in diversity of arthropods associated with vegetation and changes the composition of species (Floren and Linsenmair, 2001).These authors found a reduction in the number of ants along with an increase in the number of herbivores (Lepidoptera larvae) after removal of vegetation.Lopes and Vasconcelos (2011) have shown that disruption of fire increase the amount of herbivores and damage caused by them in the forest.Basset et al. (2008), showed that disturbed forest areas with gardens, planting or removal of vegetation, have low arthropod diversity in vegetation, but high abundance of herbivores.In the Cerrado, habitat of S. glabra, fire is a major disturbance, leading to reduced abundance and richness of arthropods (Vasconcelos et al., 2009).
Stachytarpheta glabra produces the same amount of EFNs regardless of the presence of disturbances.Thus, plants of both the areas were visited by the same amount of ants.In the Cerrado, Banisteopsis campestris showed higher nectar production in post-fire regrowth, leading to greater amount of ants visiting and lower herbivory (Alves-Silva and Del-Claro, 2013).But the S. glabra from the disturbed area showed higher amount of herbivory, even with a lower abundance of herbivores.So, something beyond the amount of visiting ants determined the high degree of herbivory suffered by S. glabra.
Our study shows that S. glabra in disturbed areas present an inefficient protector and a consequent higher herbivory level.In the disturbed area, we found reduced abundance and richness of herbivore predators along with a relatively lower abundance of aggressive ants visiting S. glabra, resulting in higher herbivory damage than in a preserved area, and despite a similar investment in defense against herbivores (density of EFNs and leaf sclerophylly).Lopes and Vasconcelos (2011) found a greater amount of herbivory in disturbed areas, but detained their discussion to the supposed low amount of defenses in leaf regrowth of plants in disturbed areas poor in nutrients and oblivious to the possible low amount of predators.
Each ant species provides a particular degree of protection of plants, and many species together are little or no effective in protection of plants (Oliveira and Del-Claro, 2005;Rosumek et al., 2009).Ant species which are small, fragile and non-aggressive usually play the role of nectar robbers providing no benefit to the plant or another mutualistic partner (Byk andDel-Claro, 2010, 2011;Oliveira and Del-Claro, 2005).Thus, despite the similar number of ants collecting nectar, the identities of the species and its relative abundances may explain the observed differences between the areas to leaf damages.
The amount and identity of ant species visiting a plant myrmecophytic interaction is determined by the amount and quality of extrafloral nectar (Lange et al., 2013;Rico-Gray and Oliveira, 2007).However, investment of the plant in attracting ants can be influenced by the plant species, the phenology and ontogeny of the plant and EFNs and by external factors such as changes in climate and habitat (Lange et al., 2013;Oliveira and Freitas, 2004;Queiroz et al., 2013).In our study, S. glabra showed no induction or increased biotic defense (EFNs) and physical (sclerophylly) in response to environmental disturbance but still showed higher herbivory in the disturbed area.Thus, the more damage can be the result of inefficient defense provided by the ant species B. heeri, showing a high dependence of protective ants over the non-regulation of constitutive defenses.As the ant fauna available for interaction is a result of the environment quality, the human perturbations indirectly affect the results of the interaction S.glabra-ants-herbivores through the unavailability of species capable of defending the plant.
Studies with ant-plant interactions through ENFs in Cerrado show that C. crassus is the most common and abundant specie in ant-plant interactions involving extrafloral nectaries and exsudate-producer herbivores (Lange et al., 2013;Rico-Gray and Oliveira, 2007;Schoereder et al., 2010).Although B. heeri was present but presents no aggressiveness (Rico-Gray and Oliveira, 2007).Camponotus crassus is considered a dominant arboreal ant and natural sugary secretion consumer (Del-Claro and Oliveira, 1999Oliveira, , 2000;;Fagundes et al., 2012;Goitía and Jaffé, 2009;Rico-Gray and Oliveira, 2007).This means that its large size, aggressiveness, presence of attack mechanisms (namely acidoporo ejector of formic acid), and territorial behavior are key characters that make this species suitable for defense of plants and removing and driving away herbivores and thus reducing herbivory, ensuring the benefits of mutualism (Fagundes et al., 2012;Freitas and Oliveira, 1992).
The high rate of herbivory on S. glabra in disturbed area cannot be explained by the abundance of herbivores, which in turn may be a response to disturbances (Ayres and Lombardero, 2000;Huntly, 1991).Even in low numbers, the presence of one or a few species of herbivores can lead to higher leaf damage (Coley and Barone, 1996), since even low abundance in a herbivore is able to spend more time in the plant and feeding access a large number of plants when it finds its natural enemy, increasing your damage (Beckerman et al., 1997;Byrnes et al., 2006).However, no species of herbivore collected presented a much higher abundance in the disturbed area.Thus, the lack of aggressive ant species also helps to explain the high rate of herbivory.Furthermore, the low abundance of predators and parasitoids of plant-disturbed area may promote the activity of herbivorous (Beckerman et al., 1997;Nahas et al., 2012).Spiders are important predators associated vegetation, and its absence may lead to high rates of herbivory on the plant, especially when associated with low abundance of ants (Nahas et al., 2012).
Stachytarpheta glabra has its trade-off in defense moved to biotic defenses indicated by the particular high production of EFNs.As the human disturbance cause a simplification of the ant fauna (Costa et al., 2010), the diversity of herbivore enemies is reduced, facilitating access and consequently high herbivory damages (McEvoy et al., 1993).S. glabra may be considered as a key-specie for the dynamics and maintenance of ferruginous rupestrian fields (Jacobi and Antonini, 2008;Jacobi et al., 2008).As successional early species, S. glabra modifies the environment making it more suitable for forward species in succession process (Bustamante-Sànchez et al., 2011;Connell and Slatyer, 1977;Padilla and Pugnaire, 2006;Whittaker, 1975).However, its establishment and proper development in the post-disturbance environment depends severely on the arthropod fauna associated (Brown, 1985;Brown and Gange, 1992;De Deyn et al., 2003).The present data adds a new level of concern, the interactions diversity, about the conservation and urgent need of protection and management of this ecosystem which is one of the most threaten.

Figure 3 .
Figure 3.The plants have the same leaf area (a) and specific leaf area (b), but the preserved area was higher abundance and richness of herbivores (c and d).It was the disturbed area that suffered the greatest damage by herbivory (leaf area lost, e: t test: t = 3.23, df = 38, p <0:01).The bars indicate standard deviation.

Table 1 .
List of species and its abundances in both disturbed and preserved areas.

Table 2 .
Comparative data of abundance and richness for arthropod groups in both disturbed and preserved areas.