Santa Virgínia (lat. 23°24.00'S to 23°17.00'S, long. 45°03.00'W to 45°11.00'W) is located in the Northern of PESM (Fig. 1) covering an area of 17, 000 hectares (Instituto Florestal 2006), and altitudes ranging from 870 to 1, 100 meters (Tabarelli and Mantovani 1999). The vegetation is defined as a dense montane humid forest (‘Floresta Ombrófila Densa Montana’) (Veloso et al. 1991) and the annual precipitation is about 2200 mm. The annual mean temperature varies from 18°C to 22°C.

Small mammals were sampled by commercial live-traps (Sherman and Tomahawk-like traps) and pitfall-traps. In July 2008, a pilot experiment was performed from one to three nights, with a total sampling effort of 300 live-traps/night. From September 2008 to September 2009, field survey was carried out bimonthly during five consecutive nights. During this period, we set up six grids with 30 live-traps per grid and 12 transects of pitfall-traps. Live-traps were arranged in a 0.6 ha grids (60 × 100 m each) with 24 trap stations spaced every 20 meters. Each trap station received one Sherman of different size, randomly set (small, 25 × 7.5 × 9.5 cm; medium, 30 × 7.5 × 9.5 cm; large, 37.5 × 10 × 12 cm; H.B. Sherman Trap®, Inc., Tallahassee, Florida, USA). We also set randomly a Tomahawk-like trap (45 × 16 × 16 cm; Rosaminas Serviço Engenharia e Comércio Ltda. Piraúba, Minas Gerais, Brazil) at six trapping stations. Overall, we had 6300 live-trap/night.

The 12 transects of pitfall-traps were pairwise 30 meters apart, from November 2008 to September 2009. Each transect received four plastic buckets (60L, 40 cm top diameter, 35 cm bottom diameter, and 56 cm depth) buried with the rim at ground level, spaced every 10 meters each. The buckets on each line were connected with a 0.5 meters tall plastic drift fence that extended an additional 10 meters at each end, totaling 50 meters of fence. In total, we used 48 buckets, resulting in 1, 440 pitfall-traps/night.

Different sizes and models of traps were used to optimize the sampling, aiming to reduce the selectivity based on body size and/or habits of the animals. Attractive baits (mashed bananas, peanut butter, bacon and corn meal) were placed in both kinds of traps. All traps were checked daily, preferably on the first hours in the morning.

Trapping and handling were carried out under ICMBio licence (number 14428-2) of Instituto Chico Mendes de Conservação da Biodiversidade.

Animals were euthanized according to the protocol of the “Animal experimentation ethics” (Carpenter et al. 1996) and under permission of Instituto Butantan Ethics Committee (242/05). The skins, skulls and partial skeletons were deposited in the Museu de Zoologia da Universidade de São Paulo (MZUSP) (still without MZUSP number), Museu Nacional da Universidade Federal do Rio de Janeiro (MN) and Coleção de Mamíferos da Universidade Federal do Espírito Santo (UFES) (Table 1).

The nomenclature used in this work follows Gardner (2005), Musser and Carleton (2005), Weksler et al. (2006) and Percequillo et al. (2011). External morphologic traits of marsupials were compared with voucher specimens preserved at MZUSP.

Metaphases were obtained from bone marrow and spleen after in vivo injection of a 0.1% colchicine solution (1mL/100g of weight). Cells were suspended in 0.075M KCl solution for 20 minutes at 37°C and fixed in three washes of methanol: acetic acid (3:1). GTG and CBG-banding were performed according to Seabright (1971) and Sumner (1972), respectively. At least 20 metaphases per individual were analyzed to define the diploid number (2n) and fundamental number of autosome arms (FNa). Chromosomes were measured using the program ImageJ version 1.46 (Rasband 2011) to establish the fundamental number, according to Levan et al. (1964). Karyotypes were set up according to the literature, when available.

Specimen identification was carried out through a comparison of our data with previous cytogenetic information, external morphological characteristics, and geographic distribution (see Table 1 references).

Eight individuals were collected, although only one male had been cytogenetically studied. Morphological data and geographic distribution comparisons allow us to identify all as Monodelphis scalops. The morphological traits of these individuals are similar to voucher specimens of Monodelphis scalops preserved at MZUSP under catalogue numbers 1528, 30702, 30712 and 30757. This species has also been reported in São Paulo state, Brazil (Gardner 2005), agreeing to our collecting site (Fig. 1).

Here we present, for the first time, the karyotype of Monodelphis scalops. The karyotype of a male showed 2n=18, FNa=30. Pair 1 is a large submetacentric, pair 2 is a medium metacentric, pairs 3, 4 and 6 are medium subtelocentric, pair 5 is a medium acrocentric and pairs 7 and 8 are medium submetacentric. X chromosome is a small subtelocentric, and the Y is a minute acrocentric (Fig. 2). The short arm of pairs 4 and 6 are difficult to see depending on the condensation of the chromosome and so it was necessary to analyze and measure more than 30 metaphases to define their morphology.

Cytogenetic data helped us to report for first time the presence of Akodon montensis, and Brucepattersonius soricinus in PESM. Cytogenetic information of these species are shown in Fig. 3, Table 1. Briefly, Akodon montensis showed 2n=24, 25 (24+1B), FNa=42 and one individual showed a heteromorphic X chromosome with an enlarged short arm. We also detected one small supernumerary submetacentric (B) in three out of nine individuals analyzed (Fig. 3a).

Brucepattersonius soricinus had 2n=52, FNa=52 (Fig. 3b) and this is the first time that banding-pattern is presented in this species. The CBG-banding pattern in the female specimen showed rather pronounced amount of pericentromeric heterochromatin in all chromosomes (Fig. 3c). GTG-banding allowed the identification of all autosomic pairs and X chromosomes (Fig. 3d).

The remaining species studied in this work have already been recorded in PESM and their karyotypes are in accordance to the literature. Karyotype information of all species analyzed and the chromosomal variability found in this work is shown in Table 1 and Figs 4–7.

Seven out of the 13 rodent species showed species-specific karyotypes: Akodon montensis, Drymoreomys albimaculatus, Oligoryzomys nigripes, Sooretamys angouya, Calomys tener, Juliomys pictipes and Trinomys iheringi (grey cells in Table 1). The identification of the remaining species (Blarinomys breviceps, Brucepattersonius soricinus, Thaptomys nigrita, Euryoryzomys russatus, Nectomys squamipes, and Rhipidomys itoan) required additional morphological and molecular investigation and geographic distribution information (Table 1).

Marsupials presented conserved diploid numbers of 14, 18 and 22 and were identified here by external morphological comparisons.

We proved the cytogenetic analyses as a taxonomic tool, since 7 out of 13 rodent species present species-specific karyotypes (53.8%). Besides, we identified 94% of all species, when cytogenetic data were combined with information of external morphology and geographical distribution (Table 1).

Cryptic species are relatively common in some Neotropical rodent groups and cytogenetic information was indispensable for identifying such species. For instance, Akodon montensis is morphologically indistinguishable from Akodon cursor (Winge, 1887) and both species occur in sympatry in the Atlantic Forest (Christoff et al. 2000). In addition, the occurrence of Akodon cursor previously recorded in Santa Virgínia/PESM (Instituto Florestal 2006) was doubtful till this study, as we proved the occurrence of Akodon montensis by karyotypic analysis.

Another cryptic species case occurs in the genus Thaptomys. Thaptomys sp. (2n=50) and Thaptomys nigrita (2n=52) are morphologically identical, so the karyotypes are the diagnostic information to distinguish both species (Ventura et al. 2004, 2010).

By contrast, Thaptomys nigrita and Brucepattersonius soricinus present very similar karyotypes (2n=52, FNa=52) however their identification can be safely done at the level of genera by external morphological characters. An accurate observation on the karyotypes of Brucepattersonius soricinus and Thaptomys nigrita showed that the pair 1 of Thaptomys nigrita is the largest of the chromosome set (Fig. 5a) meanwhile Brucepattersonius soricinus has the pair 1 similar in size to the others of the set (Figs 3b–d). We also noticed differences regarding sex chromosome morphologies of both species (Table 1). This feature could be a diagnostic tool to differentiate each karyotype, but additional cytogenetic studies (including comparative and molecular cytogenetic data) are needed to support these first observations.

Blarinomys breviceps presents a peculiar karyotype and it could not be considered species-specific due to the great variability in 2n and FNa (Geise et al. 2008, Ventura et al. 2012). Moreover, Ventura et al. (2012) suggested the existence of more species for the monotypic genus Blarinomys in Atlantic Forest since molecular phylogenetic analyses showed two geographically distinct lineages.

Euryoryzomys russatus does not have species-specific karyotype also. Euryoryzomys emmonsae Musser, Carleton, Brothers and Gardner, 1998, and Euryoryzomys nitidus (Thomas, 1884) share the same 2n=80, NFa=86 (Bonvicino and Geise 2006). However, when cytogenetic information is combined with morphologic and geographic distribution data, Euryoryzomys russatus can be confirmed.

Concerning Nectomys squamipes, it is not possible to affirm that this species possess species-specific karyotype with classical cytogenetic data because, when compared to Holochilus brasiliensis (Desmarest 1819), both karyotypes are identical (Yonenaga-Yassuda et al. 1987). Nevertheless, the association of cytogenetic, geographic distribution and external morphological characters allows the recognition of Nectomys squamipes as occurring at PESM (Bonvicino et al. 2008). Nectomys squamipes was considered for years as a carrier of two basic distinct karyotypes: 2n=56 (1 to 3Bs) and 2n=52 (1 to 3Bs), and only after crossings in laboratory, Bonvicino et al. (1996) noticed that two different species could be diagnosed - Nectomys squamipes (2n=56) and Nectomys rattus (Pelzeln, 1883), (2n=52).

The karyotype of Rhipidomys itoan presented here (2n=44, FNa=50 Fig. 4c) is the same one as described by Zanchin et al. (1992) and Silva and Yonenaga-Yassuda (1999). Pinheiro and Geise (2008) also found an identical karyotype for a species referred as Rhipidomys sp., trapped in Picinguaba (PESM), and De Vivo et al. (2011) reported an undescribed species of Rhipidomys that occurs at the Parque Estadual da Serra do Mar. Recently, two new species from Atlantic Forest were described: Rhipidomys tribei Costa, Geise, Pereira and Costa, 2011 and Rhipidomys itoan; and the latter presented 2n=44, FNa=48, 49, 50 (Costa et al. 2011). Santa Virgínia is embedded in the geographical distribution described for this species and molecular analyzes confirmed that this sample belongs to Rhipidomys itoan species. Nevertheless, we do not consider this karyotype species-specific.

Finally, cytogenetic analysis was useful in identifying Thaptomys iheringi as two species – Thaptomys iheringi and Thaptomys dimidiatus (Günther, 1876) - occur in Atlantic Forest. Despite the regular chromosome set of Thaptomys iheringi (not considering B chromosomes) is identical to the one described for the species Thaptomys dimidiatus (2n=60, FNa=116) by Pessoa et al. (2005), the presence of at least one B and the morphology of Y chromosome in Thaptomys iheringi represent good characters to diagnose the species.

Mammals have remarkable diversity in species karyotypes, and rodents exhibit noteworthy variability of diploid chromosome number (O’Brien et al. 2006, Romanenko et al. 2012). For instance, in this work, diploid numbers of rodents ranged from 24 in Akodon montensis to 80 in Euryoryzomys russatus.

The chromosome variation observed here is due to the presence of supernumerary chromosomes (B chromosomes), sex chromosome heteromorphism and/or polymorphism, as well as autosomal polymorphisms. This chromosome variability does not cause a problem in characterizing the species, except in the case of Thaptomys iheringi, in which the presence of at least one B chromosome is sufficient to confirm its identity.

Structural rearrangements may explain much of the observed karyotype diversity in rodents. In this regard, Robertsonian fusions/fissions (whole-arm translocations) and pericentric inversions, have long been considered the predominant rearrangements in natural populations of rodents (Patton and Sherwood 1983). Nevertheless, studies with more refined techniques such as fluorescent in situ hybridization and chromosome painting demonstrate that tandem fusions, reciprocal translocations, and paracentric inversions are much more common than previously thought (Hass et al. 2008, Ventura et al. 2009, Romanenko et al. 2012).

Our data showed two species with pericentric inversion rearrangements, Oligoryzomys nigripes and Rhipidomys itoan. Oligoryzomys nigripes showed variation in autosomal pair 3 (Fig. 4b) but this rearrangement had also been reported in pairs 2, 4 and 8, which places this species as one of the most polymorphic within Neotropical rodents (Paresque et al. 2007). The genus Rhipidomys frequently shows 2n=44, except for the 2n=50 reported by Silva and Yonenaga-Yassuda (1999) from Amazonas, in contrast with differences in the FNa (Zanchin et al. 1992, Costa et al. 2011). The variation of FNa, which represents the commonest chromosome change observed for the genus, may be a consequence of pericentric inversion events.

Karyotype diversity is also enhanced in mammals due to the presence of B chromosomes. B chromosomes are extra elements found in the karyotypes of many eukaryotic species. Their functions and molecular composition remain obscure but, apparently in mammals, these chromosomes neither promote phenotypic alterations nor affect fitness of individuals (Jones and Rees 1982, Trifonov et al. 2010). B chromosomes are known in nine Brazilian rodent species (Silva and Yonenaga-Yassuda 2004, Ventura et al. 2012). Herein, we found B chromosomes in four out of 13 species of rodents (30, 76%, i.e. almost a third of the total): Akodon montensis, Brucepattersonius breviceps, Nectomys squamipes and Thaptomys iheringi. Silva and Yonenaga-Yassuda (2004) found B chromosomes in Sooretamys angouya (referred at that time as Oryzomys angouya), however, in our sample, B chromosomes were not observed for this species (Fig. 5c).

Sex chromosome heteromorphisms/polymorphisms were found in Akodon montensis and Oligoryzomys nigripes, and the variation is due to addition/deletion of constitutive heterochromatin, as described by Kasahara and Yonenaga-Yassuda (1982) and Paresque et al. (2007), respectively.

Cytogenetic data exposed three diploid numbers for the family Didelphidae: 2n=14, 18 and 22 (Reig et al. 1977, Carvalho et al. 2002). As the karyotypes of American marsupials are conserved, cytogenetic analyses cannot be considered as a diagnostic tool to identify species. However, differences in banding patterns could help in the characterization of some taxa, for instance, Marmosops incanus (Svartman 2009).

In the present paper we report for the first time the karyotype of Monodelphis scalops which is similar to the one described for Monodelphis kunsi Pine, 1975 and Monodelphis brevicauda (Erxleben, 1777) by Carvalho et al. (2002), except for the morphology of the sex chromosomes (Fig. 2). Besides, Monodelphis scalops karyotype differs from Monodelphis rubida (Thomas, 1899) (2n=18, FNa=32) (Pereira et al. 2008) due to the presence of one acrocentric pair (#5) instead of a biarmed pair (Fig. 2).