Acromyrmex phorids in Atlantic forest: following the footsteps of Thomas Borgmeier

ELIZALDE LUCIANA; QUEIROZ, JARBAS M

Simposio; Simposio de Mirmecologia; 2013

INTRODUCTION Phorids (Diptera: Phoridae) are one of the most important families of insects because of the large number of species and the diversity of life forms (Disney 1994). Traditionally, they were considered saprophytic insects, however, one of the most common habits in this family is the parasitoid one (Disney 1994). In fact, there are at least 36 genera of phorids that are parasitoids, nearly 72% of which have infrequent hosts for the parasitoid life habit: adult ants (Disney 1994). Associated with leaf-cutting ants there are two types of phorids, those species who spend most of their life cycle within the nest and those with females that are parasitoids. The first group of phorids is assumed to feed on the abundant resources that are within the nests of leaf-cutting ants (i.e. fungus culture). These myrmecophile phorids may be important from an applied perspective since they may transmit pathogens among ant colonies. However, very little is known about their biology. The other group, those who are parasitoids of leaf-cutting ants, are known since 1900 but scientific studies began in 1923 with the pioneering work done by Borgmeier in Brazil (Borgmeier 1931; Borgmeier 1929; Borgmeier 1928). This author described three of the six genera that parasitize leaf-cutting ants. These parasitoid females search for worker ant hosts to oviposit in foraging trails, cutting sites and external refuse dumps (Elizalde and Folgarait 2011). Once they selected a host ant, they deposit an egg within the body, in different places according to the species, using a sclerotized ovipositor. A larva emerges from this egg that eventually kills the host just before pupation (Elizalde and Folgarait 2011; Tonhasca et al. 2001; Tonhasca 1996). The pupa may be formed inside the ant's body, either in the head or thorax (reviewed in Elizalde & Folgarait 2011), or can leave the ant body to pupate outside the host. In the case of the species that pupate outside the host, more than one can be formed by ant larva host (Elizalde and Folgarait 2011; Erthal and Tonhasca 2000). The life cycle is detailed in Figure 1. Figure 1. Generalized life-history of a phorid parasitoid of leaf-cutting ants. While Borgmeier?s work on phorid parasitoids of leaf-cutting ants was taxonomic, his research was motivated by the potential to use these parasitoids for biological control of leaf-cutting ants (Borgmeier 1931), which are a major agriculture in Brazil and throughout its distribution (Della Lucia 2003). Borgmeier knew that to use enemies as biological control it was necessary to know the biology of natural enemies in detail, but the first step was to study their diversity and describe new species. Despite these ideas clearly expressed in his publications, research on behavior and efficiency of these parasitoids as natural control agents did not start until the 90?s, and took place mainly in Brazil. All these studies focused on parasitoid species attacking Atta ants (Bragança et al. 2009; Bragança et al. 2003; Bragança et al. 2002; Bragança and Medeiros 2006; Erthal and Tonhasca 2000; Tonhasca et al. 2001; Tonhasca 1996), but parasitoids interacting with Acromyrmex are almost unknown in Brazil, except for the pioneering work done by Borgmeier. Interestingly, an important number of phorid species associated with leaf-cutting ants that were described by Borgmeier parasitize ants of the genus Acromyrmex. Borgmeier describes many phorid species associated with leaf-cutting ants from south-east Brazil, mainly from Rio de Janeiro State. There, the Atlantic Forest is severely affected by habitat fragmentation, and what remain exists as small?medium sized fragments surrounded by a deforested matrix. This matrix is used as pasture for cattle ranging, but once was used for coffee plantations (Francelino et al. 2012). Forest remnants represent 15.7% of the vegetation cover of the area, and is mostly represented by secondary lowland and submontane semi-deciduous forests (Oliveira-Filho & Fontes 2000; Francelino et al. 2012). Several studies have shown the negative effect that habitat fragmentation has on parasitoids (reviewed in Tscharntke et al. 2002). Phorid parasitoids are expected to be more susceptible than their leaf-cutting ant hosts to habitat fragmentation because of their narrow niche-breadth (Cagnolo et al. 2009; Henle et al. 2004) and their small body size (Elizalde & Folgarait 2011). Body size correlates positively with dispersal abilities (Blackburn et al. 1999) and adult phorid parasitoids do not have special long-dispersal means, thus making dispersal throughout a hostile matrix a risky task for phorids. In fact, there is some evidence that phorid parasitoids of Atta leaf-cutting ants are negatively affected by habitat fragmentation (Almeida et al. 2008; Rao 2000). Here, we evaluate whether habitat fragmentation is a negative factor affecting the interactions between phorid parasitoids and their Acromyrmex leaf-cutting ant hosts. We first described the quantitative interactions among Acromyrmex ants and their parasitoids in Atlantic Forest, and compare our findings with Borgmeier?s work for these phorids in the area. Then, we compared parasitoid species richness, abundance and proportion of nests with parasitoids in fragmented and well-preserved forests. Thus, the focus of this work is on Acromyrmex ant hosts, which do not share phorid parasitoid species with Atta (Elizalde & Folgarait 2011). METHODS The study was conducted in two areas within the Atlantic Forest of Rio de Janeiro state, chosen based on their different degree of fragmentation: the Vassouras-Barra do Piraí area was selected as the highly fragmented forest, and the continuous forest is located in a NGO reserve, Reserva Ecológica de Guapiaçu?REGUA. Eleven fragments were selected in the fragmented area, with a range of size variation of 0.1?10.6 km2 and spanning over an area of 726 km2. Two areas of the continuous forest were selected, separated by 2 km and covering an area of 9 km2. Walking along transects of known length we counted all Acromyrmex nests and collected phorids attacking over ants, to later identify them in the laboratory. Later, we obtained ant species density by dividing the number of ant species in each transect by its area. We evaluate the effect of forest fragmentation over parasitoid species richness, phorid abundance over each nest, and the proportion of nests with parasitoids. RESULTS AND DISCUSSION Acromyrmex-parasitoid interactions. We found five Acromyrmex species. Acromyrmex niger was by far the most abundant species (Table 1). The other species collected were A. subterraneus, with three subspecies (A. s. brunneus, A. s. molestans, A. s. subterraneus), A. coronatus, A. aspersus, and A. disciger (Table 1). Seven species of phorid parasitoids were found, with five of them attacking exclusively A. niger (Table 2). Only one phorid species was attacking A. disciger and another was attacking A. coronatus only (Table 2). None of the subspecies of A. subterraneus nor A. aspersus were attacked by phorids. The parasitoids Apocephalus lutheihalteratus and Myrmosicarius catharinensis were the most abundant species, and they used A. niger as host. The other parasitoid species had very low abundance (Table 2). Only five nests from A. niger in the continuous forest had more than one phorid species attacking at the same time (Ap. lutheihalteratus and M. catharinensis found together in four nests and Ap. lutheihalteratus and Neodorhniphora similis in the other nest). Surprisingly, we collected all phorid species attacking Acromyrmex that Borgmeier reported in the region (Borgmeier 1928, 1929, 1931). This suggests that our sampling was enough, but also that parasitoid communities remain similar even after 70 years. The only exception was Ap. lamellatus, a phorid species that we did not collect and Borgmeier recorded it in association with A. niger (Borgmeier 1926). In addition, it was notable that few nests had two phorid species attacking ants at the same time, and this only occurred in the continuous forest. Effect of forest fragmentation on ant-phorid interactions. Ant communities in the fragmented and continuous forests were similar. Only A. disciger and A. s. molestans were not collected in the continuous forest (Table 1). All Acromyrmex species were present in the fragmented forest, when all fragments were pooled. In fact, ant species density was higher in the fragmented forest (U = 20, P = 0.03). However, nest density of all Acromyrmex species did not differ between forest types (U = 18, P = 0.12) nor for A. niger (U = 12, P = 0.75). Only 37.5% of the fragments with A. niger had phorids attacking it. Rarefied phorid species richness in continuous forest was 4.81 (lower 95% confidence interval = 3.51), higher than the three species gathered over A. niger in the fragmented forest. In addition, parasitoid abundance over nests was 12 times higher than that found in fragmented forest (Kruskal-Wallis 21 = 12.9, P = 0.0003). Continuous forests also had a 4 times higher proportion of A. niger nests with parasitoids than fragmented forest (21 = 7.06, P = 0.008). In fact, phorids attacking ants were collected only in 3 from the 8 forest fragments sampled, where A. niger was present. Only two phorid species were abundant enough for comparisons at phorid species level, Ap. lutheihalteratus and M. catharinensis. The first species was not present in the fragmented forest, and in the continuous forest 17.3% of the A. niger nests sampled had this species attacking, with most of these nests having only one Ap. lutheihalteratus attacking (except for two nests). Meanwhile, M. catharinensis in the continuous forest was present in 26.9% of the A. niger nests sampled, and typically two phorids were attacking per trail (two phorids in median, range 1?6); however, in the fragmented forest only 5.9% of the nests had phorids of this species and in all cases only one individual was attacking ants. These results show that habitat fragmentation had a negative effect on phorid parasitoids of Acromyrmex, mainly affecting abundance. However, their ant hosts were not noticeably affected by forest fragmentation. The effect on phorids was particularly evident for the number of adult parasitoids attacking ants and the proportion of nests with parasitoids. Although species richness was also lower in the fragmented forest, the difference was due to two species, one of them with low abundance (Myrmosicarius simplex). Thus, our results add to the body of knowledge that posits parasitoids as very susceptible animals to this type disturbance (Cagnolo et al. 2009; Kruess and Tscharntke 2000). In conclusion, our work showed that even when phorid abundance is considerably reduced in forest fragments, possibly because of their inability to reach fragments, phorids of some species are able to live there. Because agricultural activities are among the main drivers of habitat fragmentation (Foley et al. 2005), it is necessary to understand how leaf-cutting ants? natural enemies such as phorids are affected by forest fragmentation. An important management tool to control pests in agricultural croplands is to leave fragments of native forest around crops, which function as refuges for natural enemies (Bragança et al. 1998; Zanetti et al. 2000). Results from this work suggest that these fragments would not represent important refuges for these phorid parasitoids. However, since some parasitoids were able to live on fragments, more studies on the population dynamics of these phorids on fragmented landscapes are necessary to understand if and how forest fragments could function as reservoirs for parasitoids of leaf-cutting ants. We are working on this line, following Borgmeier?s footsteps and ideas about the importance of Acromyrmex phorids, with an effort to reach his proposed goal: evaluate carefully whether these are proper tools to control leaf-cutting ants.