The reallocation of the ant species Dinoponera lucida Emery ( formicidae : ponerinae ) population increasing Its local genetic diversity

The aim of the current study is to describe the genetic consequences of the reallocation process in order to preserve an entire Dinoponera lucida Emery, 1901 population. The sample collection and the mitochondrial data analysis were conducted before and after the reallocation of nests in two conservation unities in Espirito Santo State, Southeastern Brazil. The data analysis showed that the genetic variability inside the forest fragmentation has considerably increased above the natural levels. The importance of the present study relies on information and notifications about the herein studied populations, which will be provided to future studies.


Introduction
The Atlantic Forest is one of the richest biomes in the world and one of the 25 biodiversity hotspots (Myers et al., 1999), but most of its remaining biome is distributed in small forest fragments (Ribeiro et al., 2009).This habitat fragmentation process may represent a constant threat to endemic species (Brooks et al., 2002), such as the giant ant Dinoponera lucida Emery 1901, whose geographic distribution is limited to Bahia and Espírito Santo States, in Northeastern and Southeastern Brazil, respectively (Campiolo & Delabie, created through the fission of bigger colonies.Thus, nests are established in an aggregated spatial distribution pattern inside the forest fragment (Mariano et al., 2008;Peixoto et al., 2010), and are probably genetically related to each other (Araújo, 1994).Such distribution pattern increasingly limits the species' dispersal ability.
Habitat fragmentation causes the isolation of populations and increases endogamy, which results in low genetic variability inside inbreeding groups and high genetic divergence between isolated groups (Packer & Owen, 2001).In addition, populations with dispersion limitations, such as D. lucida, are under high extinction risks due to stochastic events (Burkey, 1989).Although the process that causes stochastic events may naturally occur, the anthropic exploration accelerates its natural course and leads to the extinction of populations haplotypes.Because of the high forest fragmentation, not just entire forest fragments disappear, but a habitat might become no longer suitable to host highly adapted populations (Fahrig, 1997;Brooks et al., 2002).Endemic species are even more threatened under these circumstances, and this is why the species D. lucida requires ultimate population genetic studies.
The ecological theory features fragmented populations as a set of subpopulations form a metapopulation.These metapopulations experience extinction and recolonization events within subpopulations, and they persist in landscapes depending on the dynamic flow of the individuals.From a genetic point of view, small subpopulations are prone to extinction because of endogamy.However, individuals from other large subpopulations can migrate to a small subpopulation, and it could increase the genetic diversity in and avoid the extinction of this small subpopulation.Such effect is known as rescue effect (Begon, 2006); the large populations are called "donor" or "source", and the small ones, "sink" or "receiver".
As it was previously mentioned, D. lucida populations have special dispersion limitations.In the framework of fragmentation and metapopulation theory, it is important understanding how the migration of individuals from a donor (or source) population affects the genetics of a receiver (or sink) population.Therefore, a rescue and reallocation process was designed using an entire D. lucida population.The experiment took place in 2009, when a seaport construction site was set in an Atlantic Forest fragment.We tested the hypothesis that the reallocation of an exogenous D. lucida population would increase the genetic variability of populations located in isolated forest fragments.

Material and Methods
The forest fragments that have donated and received the reallocated population are located in Aracruz County, Espírito Santo State, Southeastern Brazil.The research team visited conservation units in Aracruz County aiming to assess the viability of reallocating the rescued nests to these units, before moving the population in.The criteria for selection were fragments under environmental protection policies, the shortest distance from the original area as possible, and the occurrence of the species  Back to 2009, at the time the procedure was put in place, one sample was collected from each nest for genetic analysis, as well as samples from each of the 18 AWDMNP and 17 DVFMNP nests of the native populations in the CU, in order to gather genetic data of the D. lucida populations before the reallocation.The research team collected new samples from all monitored nests in the receiving fragments in 2011.The DNA extraction was conducted according to the protocol by Waldschmidt et al. (1997) in the Insect Molecular Biology Laboratory of Federal University of Viçosa.Sequences of the Cox1-Cox2 mitochondrial gene and intergenic tRNA leu regions were obtained using primers developed by Resende et al. (2010), which were later aligned in the Mega 5.0 software (Tamura et al., 2011).
The haplotype network analysis was conducted in the Network 4.6 software, through the median-joining method (Bandelt et al., 1999).The Analysis of Molecular Variance (AMOVA) was performed using the Arlequin 3.5 (Excoffier & Lischer, 2010) applied to the data collected before and after the reallocation process.The "Among populations" hierarchical level corresponded to the BRTW, AWDMNP and DVFMNP populations in the AMOVA conducted before the reallocation.The same hierarchical level consisted of AWDMNP and DVFMNP populations in the AMOVA conducted after the reallocation process, since the BRTW population was reallocated to those both receivers.The AMOVA was applied to three hierarchical levels in order to take the potential subpopulations into account, because previous reviews refer to aggregate distribution patterns within a single forest fragment.

Results
Four haplotypes were identified in the data set: one in each park (named H1 in AWDMNP; H4 in DVFMNP) and two in the reallocating population (H2 and H3, both from BRTW) (Fig 2).As it was expected, the AMOVA conducted before the reallocation process has shown high genetic variance between populations from different forest fragments (83%) when the three populations, namely: BRTW, AWDMNP and DVFMNP, where studied as a single one (Table 1).
The second AMOVA has shown that the genetic variance between AWDMNP and DVFMNP has decreased to 55% after the reallocation, whereas the diversity within populations and subpopulations has increased.
It is worth mentioning that after the five-year monitoring, 82% of the reallocated nests has survived in the receiver fragments (Ferreira et al., in prep).

Discussion
The main aim of the present study was to describe the genetic consequences of the reallocation process involving an entire D. lucida population.Instead of collecting genetic samples from all populations and simulate the genetic diversity in order to predict the reallocation impacts, it was made the choice for effectively performing the reallocation procedure and monitoring its effects for the following five years.It was demanding to put this intervention in place in order to save BRTW haplotypes from extinction.
The haplotype network analysis showed that the haplotypes H2 and H3, both from the BRTW fragment, were more similar to each other than the H1 and H4, from AWDMNP and DVFMNP, respectively.Such finding may be justified by the geographic distance, because H2 and H3 come from the same area, and it suggests the existence of recent divergence.D. lucida populations living in a certain forest fragment often have one or few haplotypes (Resende et al., 2010).There is high genetic diversity from population to population, as well as low variability inside a single population.
Since it is common finding one or few haplotypes in a single population, it is possible inferring that the smaller

VS DF SS E(MS) %Variation
Among the area, the lower the genetic diversity of D. lucida.The genetic diversity within the populations and subpopulations of both receivers in the CU has increased after the donor population was reallocated, whereas the variation between AWDMNP and DVFMNP populations decreased, due to the homogenization of nests from the BRTW population (Table 2).Thus, the haplotypic diversity in the receiver populations in the present study is higher than natural.It means that the herein studied hypothesis was accepted, because of the high survival rate presented by the reallocated population.The reallocation effectively increased the genetic diversity of subsequent generations.It is a satisfactory result, since the literature shows that the rate of successful reallocation, repatriation and translocation processes is low (Dodd & Seigel, 1991).
variation may enable enhancing the population's survival, thus simulating a rescue effect.This result may be representative for the entire D. lucida species.In addition, the previously conducted search for suitable and similar environments may have prevented or minimized adaptation issues.Genetic studies about reallocation and translocation process are rare, mainly studies of this nature involving groups of invertebrates (Sherley et al., 2010).Yet, these studies are essential to assess the persistence of the introduced population (Armstrong & Seddon, 2008).The reallocation program appears to have succeeded after five years monitoring the studied populations.Yet, the herein adopted reallocation procedures shall not be replicated if the basic information provided in the current study are not followed.This is the first time a reallocation program involving an invertebrate population is conducted based on a genetic data analysis associated with a middle time monitoring.It is strongly recommended that future studies do not ignore the importance of the herein presented information.
. The Barra do Riacho Terminal Waterway (BRTW, -19 o 50.548', -40 o 03.827') was the donor forest fragment supposed to be deforested, and it comprised 11 hectares.The conservation units selected to host the nests were: David Victor Farina Municipal Natural Park (DVFMNP, -19 o 55.850', -40 o 07.774', 44 hectares) and Aricanga Waldemar Devens Municipal Natural Park (AWDMNP, -19 o 48.827', -40 o 19.959', 515 hectares).These fragments belong to a highly fragmented landscape, and the two CU fragments are located approximately 25 kilometers from each other (in a straight line) (Fig 1).All the 19 nest found in the BRTW forest fragment were open, and every individual (eggs, young and adults forms) were removed and reallocated to artificial nests constructed in two Conservation Units (CU), according to the protocol designed byFerreira et al. (in prep).The reallocated population was monitored throughout the following five years in order to observe its behavioral and ecological reactions.

Table 1 .
Analysis of Molecular Variance (AMOVA) applied to three hierarchical levels on data set before reallocation.VS: Variation Source; DF: degrees of freedom; SS: sum of squares; E (MS): Estimate medium squares -variance components.

Table 2 .
Analysis of Molecular Variance (AMOVA) applied to three hierarchical levels on data set after reallocation.VS: Variation Source; DF: degrees of freedom; SS: sum of squares; E(MS): Estimate medium squares -variance components.