Effect of ants on Mus musculus Linnaeus , 1758 ( Rodentia : Muridae ) carcasses decomposition : a preliminary study in an Atlantic Forest fragment

Different ant species are frequently associated with carcasses during decomposition and they are present from the early until the late postmortem phase. Despite the common presence of Formicidae on carcasses, their role in the decomposition process is generally neglected. Here, we evaluate the effect of ants on Mus musculus decomposition pattern and abundance of associated dipteran fauna in the Neotropical region. We also discussed the role of different ant species in the decomposition process, emphasizing their predatory or necrophagous behaviour. Six carcasses of M. musculus were placed in the field and distributed equally in two treatments: “with ants” and “without ants”. Only in the “without ants” treatment, ants were not allowed to access the carcasses. A total of six ant species were registered on carcasses. The most abundant ant species were Linepithema neotropicum and Pheidole sp., which exhibited an intense predatory behaviour on dipteran larvae, resulting in a delay in the decomposition process. In contrast, Camponotus rufipes workers were observed acting exclusively as necrophagous and their behaviour clearly accelerated the decomposition process on one carcass. Considering the entire decomposition process, the average abundance of adult flies was not different between the two treatments. However, we found a significant difference in the abundance of dipteran larvae between them. Our findings provide a preliminary study about the effect of Neotropical ant species presence on the decomposition pattern. We highlight the contradictory role that different ant species may exhibit on carcasses and how these presented findings could be applied in medico legal reports.


Introduction
Insects are usually the first animals to explore a dead body, using it for oviposition or feeding (Amendt, 2004).They are present in all processes of carcass decomposition (Payne, 1965;Jirón & Cartín, 1981;Byrd & Castner, 2010) and their life cycle pattern and behavioural ecology are considered powerful predictors for estimating the minimum postmortem interval (PMImin) and interpreting the events that occurred during the commission of a crime (Wells & Sperling, 2001;Bonacci et al., 2011).
In forensic entomology, insects classified as necrophagous include species that feed upon the tissues of decomposing bodies during at least one phase of their lives, and as predatory, those which feed on the fauna associated in carcass decomposition.When insects feed both on decaying carrion and also on other insects, they are considered omnivorous (Smith, 1986;Oliveira-Costa, 2011).
Palavras-chave: carcaça, entomologia forense, Formicidae, relatório médico-legal.tially feed on other colonizers, especially on eggs and fly larvae (Moretti & Ribeiro, 2006), and consequently, delay the decomposition process (Early & Goff, 1986).An understanding of the roles of different ant species on carcasses is crucial to interpret crime circumstances and avoid wrong conclusions in forensic studies.
The intrinsic local conditions and specific entomological diversity determine the pattern of carcasses decomposition.It is therefore important develop regional forensic entomological studies to provide a useful and applicable source of information (Carvalho et al., 2017).Despite many forensic entomological studies have been published in Brazil, there are important regional gaps regard how the behaviour of ant species can affect decomposition processes (see Cruz & Vasconcelos, 2000;Andrade-Silva et al., 2015;Maciel et al., 2015;Paula et al., 2016;Sales et al., 2016).Here, we evaluated the role of different ant species on carcass decomposition patterns in an Atlantic Forest fragment, in southeast Brazil.We emphasized the predatory or necrophagous behaviour of ant species and also reported the effect of ants on dipteran abundance associated with carcasses.

Methods
This study was carried out in a disturbed fragment of Atlantic Forest (21°46'48"S, 43°22'24"W), located in the municipality of Juiz de Fora, Minas Gerais state, Brazil.The climate is characterized as warm subtropical -Cwa, according to Köppen (Alvares et al., 2013), with two defined seasons: a hot and rainy summer and a cooler and dryer winter.The average annual rainfall is 1.536 mm and average annual temperature is 19.4°C.The experiments were conducted during the hot season.The temperature ranged between 30.8°C and 37.1°C during the 11 days of experiment, (T mean = 34.3± 1.83°C) and the relative humidity between 48.1% and 69.0%, (mean = 56.5 ± 7.20%).
Six individuals of Mus musculus (Linnaeus, 1758) (Rodentia: Muridae) (weight mean = 31.1 ± 1.3 g) were killed by cervical dislocation and their carcasses immediately were placed in the field.All carcasses were obtained from the vivarium of the Center for Reproductive Biology of Universidade Federal de Juiz de Fora.The procedure was approved by the Ethics Committee on Animal Experimentation of Universidade Federal de Juiz de Fora (Protocol No. 013/2009).
Each carcass was placed inside a cage (15 cm x 26 cm x 10 cm) protected by 1.6 cm wire mesh to prevent access by vertebrate scavengers.Carcasses were distributed equally in two treatments: "with ants" and "without ants".The carcasses were distributed in transect, interspersing the treatments at 10 m intervals.In both treatments, a tray containing water and detergent (9:1) was attached below to prevent access by ants and to allow their colonization by flying insects.In the treatment "with ants", cages were connected to the soil surface by four plastic bridges (7 cm x 10 cm) that allowed ants access the carcass without contact with the solution (Figure 1). .151 The observations started one day after the death of the mice, which corresponded to the first day of decomposition and was performed daily until the carcasses reached the final stage of decomposition.Each carcass was inspected for 15 min, always in the interval between 2 p.m. and 6 p.m.The observations were distributed in 3 repetitions of 5 min each with an interval of 3 min between them.At this time, we registered the abundance of ant workers on carcasses and also ant species feeding behaviour by ad libitum observation.We registered if ant species fed directly on carrion or on larvae over the carcasses, categorizing them as necrophagous or predatory, respectively.In each repetition, a scan sampling was performed to record the abundance of adult dipterans.After observations, ants were sampled through manual collection and transferred to a vial filled with ethanol (70%).In laboratory, ants were sorted to species and identified to the lowest taxonomic level possible using available taxonomic keys (Fernandez, 2003) and with the additional assistance of taxonomists (see Acknowledgments). Adult flies were not collected to avoid interference in their oviposition and consequently, in the process of colonization and decomposition.In order to measure the abundance of dipteran larvae present during the decomposition process, we quantified the larvae found in the carcass, cage area and fallen in the tray at the end of each carcass observation.We classified the stages of carcass decomposition following Payne (1965) as: fresh, bloated, active decay, advanced decay, dry stages.
To compare the abundance of adult flies between the two treatments, we pooled the three abundance registered data and conducted generalized linear mixed models (GLMMs) with quasi-Poisson error distribution.We treated individual carcasses as a random effect in the analysis to control for pseudo-replication stemming from our repeated measures design.Further, the Student t-test was used to compare the abundance of dipterans larvae between the two treatments.All analyses were conducted using the software R (R Core Team, 2016) with model checked for the distribution of errors and over-dispersion in the data.We carried out the GLMM using the lme4 v 1.1-11 package (Bates et al., 2015).

Results
Decomposition time at the treatment "with ants" varied between 4 and 11 days among carcasses, while the three carcasses of the treatment "without ants" reached the dry stage in the fifth day of exposure (Table 1).In total, six ant species were observed on the carcasses (Table 1).
Table 1.The decomposition time until dry stage, ant species with their respective abundance and feeding behaviour, and total dipteran larvae abundance observed in each carcass.Workers of Camponotus rufipes (Fabricius) were registered from the 1 st until the 5 th day of decomposition (Figure 2), feeding on mice carrion (see Table 1) and thus acting strictly as necrophagous ants.These ants were especially abundant on the faster decomposed carcass (C4), at the fresh and bloated stage.This carcass had the skull totally exposed with limbs and tail dilacerated at the 2 nd day of observation (Figure 3), reaching the dry stage at the 4 th day of decomposition.

Total
On the other hand, Linepithema neotropicum Wild and Pheidole sp. were especially abundant on the other two carcasses, which took a longer time to reach the dry stage (Table 1).Both ant species were observed on almost all days of decomposition time (Figure 2), strictly preying on dipteran larvae (Figure 4; Figure 5).Similar to C. rufipes, Camponotus scipio Forel was abundant only during the fresh and bloated stages (Figure 2) and only removed dipteran larvae from the carcass, thus not exhibiting the necrophagous behaviour.Only one individual of Acromyrmex subterraneus (Forel) and Camponotus sp. were observed transporting a small piece of mice carrion with their mandibles.
In all six exposed carcasses, the decay can be considered accelerated.Dipteran eggs  were registered at the first day of observation and larvae, at the second day.Considering the whole decomposition time, the average abundance of adult flies was not different (χ 2 = 1.272; p = 0.54) between the treatments "with ants" (977 ± 227) and "without ants" (1835 ± 181; see Figure 6).However, we found a significant difference in abundance of dipteran larvae between the two treatments (t = -5.7294;df = 4; p = 0.004; Figure 7).

Discussion
The present study highlights the species-specific consequences of ant presence to carcasses decomposition patterns.On carcasses which ant access were allowed, we observed large differences in the time of decomposition among them: only four days when ant species fed directly on carrion, and 10 or 11 days when species predated intensively on dipteran larvae.The short decomposition time registered for C4 carcass was probably related by the high necrophagous behaviour of C. rufipes registered at the former decomposition stages.Exhibiting a contrasting behaviour, L. neotropicum and Pheidole sp. were probably responsible for the decomposition delay of the other carcasses with ant access.Their occurrence affected the sarcosaprophagous community due to intense removal of immature stages of dipterans from the carcasses.In contrast to what had been reported for other ant species (Wells & Greenberg, 1994;Archer & Elgar, 2003;Chen et al., 2014;Sales et al., 2016), we did not observe ants preying on adult flies or exhibiting aggressive behaviour against them.
A delay in the decomposition time registered for two carcasses where L. neotropicum and Pheidole sp. were highly abundant suggested an important role of these species to the decomposition pattern, a result of their predatory behaviour.Besides L. neotropicum is widespread in Neotropical region, the present study compris-.157es the first record of this species colonizing carcasses in Brazil.Pheidole Westwood is one of the most hyperdiverse and prevalent ant genera in the world (Wilson, 1976;Pie & Traniello, 2007) and there is a wide range of feeding preferences and habits among species that belong to this taxon (Fowler et al., 1991).In the neotropics, it was reported that workers of Pheidole radoszkowskii Mayr mainly exhibited necrophagous behaviour but were also observed preying on dipteran larvae (Sales et al., 2016).Thus, the inference about the role of Pheidole in the rate of decomposition is applicable only when the particular species behaviour and ecology is known.
The occurrence of C. rufipes in forensic studies in Neotropical region was already reported (Luederwaldt, 1911;Moura et al., 1997;Souza et al., 2008;Andrade-Silva et al., 2015) but their particular role to the decomposition process was ignored.Our study showed a clear contribution of C. rufipes in accelerating the decomposition time.Also, they were able to cause lacerations on carrion due their intense necrophagous behaviour, which should be taken into account in medicolegal reports for a proper distinguish between post and ante-mortem injuries.
The present study is a preliminary description about the role of Neotropical ant species on carcasses decomposition pattern.We demonstrate that carcasses where Linepithema neotropicum and an unidentified Pheidole species were abundant presented a delay in their decomposition time, probably in function of the predatory behaviour of these species.In contrast, carcass where Camponotus rufipes were abundant showed an accelerated decomposition time, especially due to C. rufipes acting as necrophage.We recommend that Formicidae should not be a neglected taxa in forensic conclusions and we point out the need to expand the knowledge about ant species-specific effect on carcasses and applying it to medicolegal reports.

Figure 1 .
Figure 1.Cage used in "with ants" treatment.Trays were filled with water and detergent and plastic bridges were installed to allow ant access to carcass after cage placement.

Figure 2 .
Figure 2. Mean abundance log-transformed of ant species that colonized M. musculus carcasses over the 11 days of exposure.Ant abundance data were log-transformed for plotting and provide a better visualization.Bars represent standard errors.

Figure 3 .
Figure 3. Colonization of M. musculus carcass (C4): a) First day of decomposition, with presence of Camponotus rufipes workers; b) Second day of decomposition, with the evidence of the skull totally exposed.

Figure 6 .
Figure6.Mean abundance log-transformed of ants and dipteran adults that colonized M. musculus carcasses in the treatments "with ants" and "without ants".Air temperature and humidity over the 11 days of carcasses exposure are also presented.Mean abundance data were log-transformed for plotting and provide a better visualization.Bars represent standard errors.

Figure 7 .
Figure 7. Variation in dipteran larvae abundance among treatments.Bars represent standard errors.