Check list of ground-dwelling ants ( Hymenoptera : Formicidae ) of the eastern Acre , Amazon , Brazil

The ant fauna of state of Acre, Brazilian Amazon, is poorly known. The aim of this study was to compile the species sampled in different areas in the State of Acre. An inventory was carried out in pristine forest in the municipality of Xapuri. This list was complemented with the information of a previous inventory carried out in a forest fragment in the municipality of Senador Guiomard and with a list of species deposited at the Entomological Collection of National Institute of Amazonian Research– INPA. The resulting list covered 268 species distributed in 52 genera and nine subfamilies, and records 23 species and four morphospecies for the first time in the state of Acre. Due to the large environmental heterogeneity, future inventories will be crucial to properly describe and understand ant species distribution patterns in southwestern Amazon.


Introduction
Although tropical forests represent only 7% of the planet´s surface, these ecosystems host more than a half of known species of plants and animals (Wilson 1988;May 2010).Paradoxically, a large amount of species found in tropical forests are associated with a relatively low number of inventories, limiting the knowledge about taxonomy and geographic distribution of species (Dourojeanni 1990).This fact can be found in faunistic and floristic studies in the sate of Acre, west of the Brazilian Amazon.
Several studies on plants (Silveira et al. 1997;Goldenberg and Meirelles 2011), mammals (Calouro 1999), birds (Whitaker and Oren 1999;Guilherme 2001;Guilherme and Borges 2011), amphibians (Lima et al. 2007) and some arthropod groups (Morato 2001;Villarreal-Manzanilla and Pinto-Da-Rocha 2006;Oliveira et al. 2009;Carvalho and Esposito 2010) suggests a large diversity and endemism in forested areas of Acre.However, the distribution of terrestrial invertebrates is relatively unknown (Morato et al. 2005).For example, to our knowledge, only one ant survey carried out in Acre state was published until now (Oliveira et al. 2009), although these insects are frequently mentioned as key-organisms to the functioning of tropical forests by changing the physical and chemical soil structure (Folgarait 1998;Moutinho et al. 2003, Sousa-Souto et al. 2007), seed dispersion (Handel andBeattie, 1990, Levey andByrne, 1993) and control of herbivorous insect population (Dyer and Letourneau 1999).
The aim of this study was to compile the occurrences of species sampled at different areas in the state of Acre.We organized the information available from a previous inventory undertaken in a forest fragment in the Abstract: The ant fauna of state of Acre, Brazilian Amazon, is poorly known.The aim of this study was to compile the species sampled in different areas in the State of Acre.An inventory was carried out in pristine forest in the municipality of Xapuri.This list was complemented with the information of a previous inventory carried out in a forest fragment in the municipality of Senador Guiomard and with a list of species deposited at the Entomological Collection of National Institute of Amazonian Research-INPA.The resulting list covered 268 species distributed in 52 genera and nine subfamilies, and records 23 species and four morphospecies for the first time in the state of Acre.Due to the large environmental heterogeneity, future inventories will be crucial to properly describe and understand ant species distribution patterns in southwestern Amazon.Patrícia Nakayama Miranda 1,2* , Marco Antônio Oliveira 3 , Fabricio Beggiato Baccaro 4 , Elder Ferreira Morato 1 and Jacques Hubert Charles Delabie 5,6   Check list of ground-dwelling ants (Hymenoptera: Formicidae) of the eastern Acre, Amazon, Brazil Projeto de Assentamento Agroextrativista (PAE) Chico Mendes (10° 53' S e 68° 21' W) is an area of 24,898 ha (Acre 2006), located in the municipality of Xapuri, Acre, Brazil (Acre 2009).Approximately 84 families are distributed in PAE Chico Mendes.These families carry out the collection of Brazilian nuts (Bertholletia excelsa Kunth), latex (Hevea brasiliensis (Willd.Ex A. Juss) Mull.Arg.) extractivism, and some families develop the sustainable low intensity forest management (5.4m 3 /ha).Approximately 90% of the area is covered by forests (Coopeagro 2001), which the principal vegetation formations are represented by open forest dominated by bamboo, open forest dominated by palms and dense rain forest (Acre 2006).

Sampling
At FEC ants were sampled using pitfall traps installed in nine plots of approximately 1 ha (90 x 80 m).Three plots were established in primary forests; three in approximately 19 years old secondary forests; and three in secondary managed forests (burned, deforested and left for regeneration since 2001).The distance between areas varied of 0.1 to 6.5 km.In each area 20 pitfalls were installed every 25 m.The traps remained in field for seven days.The samples were done from June 2001 to January 2005, totaling 2,400 pitfalls (Oliveira et al. 2009;2011).
Ants were sampled in nine sustainable low intensity forest management areas at PAE Chico Mendes during August and September 2010.Areas were chosen in order to cover the different forests structures found on the settlement project.The distance between areas varied between approximately 4 km and 12 km.In each area, ants were sampled by pitfall traps (9 cm diameter plastic cup) in two 10 ha plots, totaling 18 plots (nine in managed areas and nine plots in unmanaged areas).In each plot, two 200 m-long parallel transects 50 m from each other, were established in the center of the plot.Ten pitfall traps were installed every 20 m along each transect, totalizing 360 traps (20 per plot).The pitfalls had a solution of 70% alcohol and were left open for four days before being removed.
Ants were sorted and identified using taxonomic keys (Fernández 2003;Wilson 2003;Hölldobler and Wilson 1990) and by comparison with specimens deposited at Entomological Collection of INPA, Regional Museum of Universidade Federal de Viçosa and Laboratório de Mirmecologia do Centro de Pesquisas do Cacau (CEPEC/ CEPLAC).The taxa that had not been identified to a species level were named with a unique code, thus the morphotypes sampled in PAE Chico Mendes were indexed with morphotypes sampled in FEC.Voucher specimens of FEC surveys were deposited at Entomological Collection of Universidade Federal do Acre, Regional Museum of Universidade Federal de Viçosa and Centro de Pesquisas do Cacau (CPDC collection, CEPEC/CEPLAC).A full reference collection of PAE Chico Mendes was deposited in Entomology Collection of INPA and Entomological Collection of Universidade Federal do Acre.
The nomenclature followed Bolton et al. (2005), with subsequent amendments available at the Online Catalog of the Ants of the World (AntCat) web site (http://antcat.

Data analyses
Because the sampling units and pitfall trap density were different, species rarefaction curves were used to compare species richness between areas.The rarefaction curves were generated by random samples of species occurrence without replacement in Estimate S 8.20 software (Colwell 1997), using the Coleman method (Coleman 1981).As ants are colonial organisms and often use established pheromone trails, using abundance of individuals may alter diversity estimates (Longino et al. 2002).Therefore, the rarefaction curves were built using the species frequency instead of worker numbers (Colwell and Coddington 1994;Gotelli and Colwell 2001).
The rarefaction curve of species tended to be asymptotic only for FEC inventory (Figure 2).This result was expected because of the larger sampling effort carried out for 3.5 years, which covered different pluviometric regimens (Oliveira et al. 2009(Oliveira et al. , 2011)).Although the rarefaction curve of PAE Chico Mendes did not show signs of stabilization, the increasing of new species ratio was similar from FEC areas (Figure 2).This suggests that, if the same amount of sampling effort was employed, the total number of species expected for PAE Chico Mendes would be similar to the number found at FEC forest systems (276 species).
The inventory undertaken at PAE recorded 22 species and four morphospecies (~11%) not registered in the previous study in FEC.The combined list presented here represents 267 species distributed in nine subfamilies (Table 1).Besides the relative large number of species, 83 morphotypes could not be compared between the inventory collections and were excluded from the list.Approximately 42% of species were sampled at both, PAE Chico Mendes and FEC (Table 1).Pseudomyrmex unicolor Fr.Smith, 1855 was the only species deposited at Coleção de Invertebrados do INPA, which were not sampled in FEC and PAE Chico Mendes.
As in other Amazon areas, the subfamily Myrmicinae showed a higher number of species.In soil and leaf litter ants inventories, this subfamily is normally the most frequent and diverse, probably because it comprises generalist species, specialists predators and fungusgrowing ants which have different nesting habitats, such as rotting wood, soil, leaf litter and on the vegetation (Fowler et al. 1991).The subfamilies that harbor strictly or partially arboreal species, such as Pseudomyrmecinae (Ward 1999) and Dolichoderinae (Shattuck 1992), probably were underestimated due to the sampling methods used in our inventories.For example, among the Pseudomyrmex species sampled, only two (P.termitarius and P. tenuis) are terrestrial (Kempf 1960), while in the genus Dolichoderus, only D. imitator nests in the leaf litter.
Although the ant fauna of the state of Acre is still poorly understood, the number of ant inventories for the Amazon basin growed exponentially in recent decades (Benson and Harada 1988;Majer and Delabie 1994;Vasconcelos andDelabie 2000, Vasconcelos et al. 2001;Vasconcelos et al. 2003;Vasconcelos and Vilhena 2006;Santos et al. 2008;Donoso and Ramón 2009;Harada and Ketelhut 2009;Souza et al. 2012).We expect this study help to stimulate new ant inventories in state of Acre, which is crucial to properly describe and understand the ant species distribution patterns in the southwestern Amazon.

Figure 1 .
Figure 1.Map of the sampling areas and additional information about ant species distribution available for the state of Acre.

Figure 2 .
Figure 2. Ant incidence-based rarefaction curves for two sampling areas in eastern Acre.The dotted lines represent the 95% confidence interval around the average.