New record and distribution extension of the rare Atlantic Forest endemic Abrawayaomys ruschii Cunha & Cruz, 1979 (Rodentia, Sigmodontinae)

We recorded three individuals of Abrawayaomys ruschii Cunha & Cruz, 1979 in Chapeco, Santa Catarina state, expanding the known geographic distribution of the species in approximately 370 km west from its nearest locality. It is the second record of the species in this type of physiognomy, characterized by the transition of Seasonal Deciduous and Mixed Ombrophilous forests. Additionally, this is the closest record, about 200 km east, from the recently described Abrawayaomys chebezi Pardinas, Teta & D’Elia, 2009.

The genus Abrawayaomys Cunha & Cruz, 1979 is one of the least known among the living sigmodontines (D'Elía et al. 2007;Pardiñas et al. 2009). It can be distinguished by presenting a peculiar spiky fur and unusual craniodental morphology (Pardinãs et al. 2009). Currently, this genus is composed by only two species: Abrawayaomys ruschii Cunha & Cruz, 1979 and Abrawayaomys chebezi Pardiñas, Teta & D'Elía, 2009. Abrawayaomys spp. are found in the Atlantic Forest of Argentina and Brazil, but there are few confirmed records for these species. Abrawayaomys chebezi had its description based on a single individual preserved as skin and skull, with three craniodental remains recovered from Tyto alba owl pellets constituting additional distributional records, comprising four distinct localities in the Misiones province, Argentina (Pardinãs et al. 2009); and A. ruschii that presents a slightly higher number of specimens collected in south and southeastern Brazil in the states of Espírito Santo, Minas Gerais, Paraná, Santa Catarina, and Rio de Janeiro (Pereira et al. 2008;Cherem et al. 2011;Passamani et al. 2011;Ventura et al. 2013;Cerboncini et al. 2014). A. ruschii is currently classified as Least Concern (LC) according to IUCN Red list of Threatened Species, however, there is little information on its occurrence extension, conservation status, threat degree and ecological requirements (Pardiñas et al. 2008).
In this article we expand the known distributional range of A. ruschii in approximately 370 linear km west of its nearest locality previously recorded at the State Park of Serra do Tabuleiro, on the coast of Santa Catarina state (Cherem et al. 2011). The new record is located in the city of Chapecó, also in Santa Catarina state, southern Brazil (27°08′32″ S, 052°42′59″ W), in a fragment of native forest with approximately 200 hectares ( Figure 1). The sample site is located in a transitional region between the phytophysiognomies of Seasonal Deciduous and Mixed Ombrophilous forests (known as forest with Araucaria), inside the Atlantic Forest domain, at an altitude of about 600 m (IBGE 2011). The climate is defined as Cfb type according to Köppen climate classification (Peel et al. 2007), characterized by humid mesothermic with annual average temperatures ranging between 18°C and 19°C and annual average rainfall of 1,800 mm. Rainfall is equally distributed throughout the year. The area is a private property of a food industry (Cooperativa Central Oeste Catarinense-AURORA), and it is not protected as a conservation unit. Historically, the study area has been affected by selective removal of wood and use of the land for cellulose monoculture. Part of the area is now in regeneration process, but is still subject to disturbances, especially illegal hunting (Maestri et al. 2014).
The sampling was conducted over a period of 13 months (August 2011 to September 2012). Interception and fall traps (pitfall traps) were installed in five distinct sampling sites, distant at least 400 m between each other. At each sampling site a pitfall trap station was installed; each formed a Y-shape, and consisted of 10 buckets with capacity of 100 L each, 10 m apart from each other, and interconnected by a guide-fence 50 cm high. The buckets remained open for 10 days per month over the sampled period, comprising a total sampling effort of 6,500 bucket-nights. The traps were daily reviewed during the sampling periods. Two adult females and an adult male (age classes 3-5; see the toothwear criteria established by Patterson 1992) Table 1. The identification at genus level was accomplished based on external morphological characters (Bonvicino et al. 2008) easily recognizable (average head and body size, dorsal and lateral fur composed of hardened and awn-shaped guard-hairs, and in general, with gray base and black or yellow extremity) ( Figure  2). The identification at specific level considered some diagnostic external and craniodental characters cited by Pardiñas et al. (2009) on the description of A. chebezi, and recognizable through specimen comparisons with the original description (Pardiñas et al. 2009) and with other specimens from the two species analyzed. Specimens examined in this study were: MCNU3630, MCNU3629, and MCNU3628, two specimens of A. ruschii deposited in Museu Nacional, Rio de Janeiro (MN), MN23075 from Reserva Biológica de Forno Grande, Espírito Santo (locality 1, Figure 1) and MN73415 from São Sebastião das Águas Claras, Minas Gerais (locality 5, Figure 1), besides the photographs of A. chebezi (locality 10, Figure 1) in Pardiñas et al. (2009). The dorsal and ventral fur color of the individuals collected in this study is composed by a yellow-gray tonality, resembling individuals of A. ruschii, while the individuals of A. chebezi have a brown tonality. Furthermore, our samples and specimens of A. ruschii share a tuft of hairs on the tip of the tail (Figure 3) which is long, hairy, and whitish compared to the shorter and more brownish hairs in A. chebezi. Also, the orientation of the upper incisor (defined by the position of the cutting edge of the incisor relative to the vertical plane of the incisors) is slightly opisthodont (88°) in both new specimens and A. ruschii analyzed, and is proodont (95°) in A. chebezi (Pardiñas et al. 2009). In the quantitative analysis of the skull, our samples and specimens of A. ruschii present a larger incisive foramen width, smaller rostral length, larger zygomatic plate width, larger mandible length, and smaller mandible height, when compared to A. chebezi (Table 1). Apparently, there is a considerable amount of phenotypic variation in the genus, and the scarcity of available specimens prevents a full evaluation of the variation at intra and interspecific levels (Pardiñas et al. 2009). Based on the current knowledge, the analyzed morphological characteristics seem to us as reliable diagnostic characters to identify the captured individuals as A. ruschii. Brazil. Cranial measurements were taken using digital photography. Also are presented the measures of the two holotypes: A. ruschii (MN23075) and A. chebezi (MACN20253), the latter measurements taken from Pardiñas et al. (2009). Age class following toothwear classes proposed by Patterson (1992 Otherwise, there are some apparently morphological idiosyncrasies in the new collected specimens, which differ from both A. ruschii and A. chebezi. For instance, the condylobasal length is greater in the new specimens than in the others, even considering the smallest individual captured by us (MCNU 3628), as also is the least interorbital breadth. Such features could be related to the geographic distance from these specimens to the others examined. However, other characteristics showed appreciable variation following size of individuals, as the zygomatic breadth, the incisive foramina length, and the bullar length less tube. In these high variable features, the new specimens could be closer related to A. ruschii or A. chebezi depending on the size of the individuals analyzed. These features show how variable morphological characters can be inside a single population, and this variation can potentially be the same in other regions if we were able to explore more than one or two individuals from a single location. In this way, we advocate that the quantitative differences are highly variable between specimens of A. ruschii, and for now we could not detect if there is a geographical pattern explaining this variation.
The species A. ruschii appears to be a naturally rare species, because at the nine localities where it was sampled, including this study, it has always been found at low abundance. A great capture effort may be required to increase the probability of surveying individuals from this species. As an example, in our study, we collected just three individuals with an effort of 6,500 trap-nights, resulting in a capture success of 0.046%. In addition, little is known about the ecology of A. ruschii, and this new record represents not only an increase in geographical distribution, but it is also the second record inside the interior Atlantic Forest (see Cerboncini et al. 2014), where forest phytophysiognomy is different from Atlantic Forest stricto sensu, increasing the habitat range occupied by this species.
It is remarkable that the new record of A. ruschii is geographically closer to records from A. chebezi (ca. 200 km from the region of Misiones, Argentina), than from records of A. ruschii (ca. 370 km) (Figure 1). This may suggest that there is a continuum in the geographic distribution of both species, and perhaps even sympatric areas among them may occur. Inventories on the intermediate region between the new record and that of A. chebezi (east of Misiones and extreme western of Santa Catarina state) would help to reveal the distribution limits of each species. No geographic barrier is evident among these localities. Also, considering the great morphological variation within A. ruschii specimens (as evidenced in the three specimens here examined), it is necessary that a molecular analysis is done taking into account all specimens available to date, in order to better understand this variation and to help delimit the species in the genus. Future studies Finally, an addendum on the method, considering the known difficulties on the capture of this species: possibly, the great sampling effort employed on this area, in addition to the methodology of pitfall traps (6,500 pitfall-nights), contributed to the capture of A. ruschii. The use of interception and fall traps allows a broader sampling of the community (captures more than one individual per night), is less selective (does not have bait), and often capture a greater species richness (Umetsu et al. 2006;Cáceres et al. 2011). So our results are in accordance with previous studies (Passamani et al. 2011), and the use of this capture method combined with long-term surveys should be encouraged in small mammal studies.