Among caviomorph rodents, social organization ranges from solitary living to social forms, in which individuals interact frequently, overlap their range areas, and share resting locations (e.g. ^{Ebensperger et al. 2004}; ^{Ebensperger & Hayes 2008}). These rodents include species that differ morphologically and physiologically and use a great diversity of habitats (^{Verzi et al. 2015}) which seems to covary with groupliving (^{Lacey & Ebensperger 2007}).

Thus, caviomorphs offer unique opportunities to examine factors associated with variation in social organization across and within species (e.g. ^{Maher & Burger 2011}). Within caviomorphs, the octodontids (Octodontidae) comprises 16 extant species grouped in six living genera with surface-dwelling (Octomys), fossorial (Octodontomys, Octodon, Tympanoctomys), completely subterranean (Spalacopus) and semisubterranean (Aconaemys) habits (^{Ojeda et al. 1996}; ^{Gallardo et al. 2007}; ^{Lessa et al. 2008}).

Aconaemys or rock rats used open areas and forests, on both eastern and western slopes of the Andes (^{Pearson 1984}; ^{Gallardo & Mondaca 2002}).

Three species are currently recognized: A. fuscus, A. sagei, and A. porteri (^{Gallardo & Mondaca 2002}). Aconaemys fuscus inhabits highland forests of Araucaria araucana (Araucariaceae) and sandy areas (^{Muñoz-Pedreros 2000}), while A. sagei inhabits ungrazed bunchgrass and Nothofagus forest with bamboo cover (^{Pearson 1984}). Aconaemys porteri is distributed from Volcán Villarrica to Puyehue in Chile, and from both Parque Nacional Lanín and Parque Nacional Nahuel Huapi in Argentina (^{Pearson 1984}; ¹⁹⁹⁵; ^{Gallardo & Reise 1992}; ^{Gallardo & Mondaca 2002}). In these areas, Porter’s rock rats inhabit dense bamboo and southern beech rainforests (^{Pearson 1983}; ¹⁹⁸⁴; ^{Gallardo & Reise 1992}).

Evaluating the extent of the social behavior of Aconaemys remains critical to determine whether ecological conditions and species-specific attributes drove the evolution of group living in the octodontids. A recent comparative analysis suggested that social behavior in caviomorphs was gained and lost repeatedly, perhaps originating from an ancestral species that was socially flexible, and where the loss of group-living has been associated to the use of habitats with high plant cover (^{Sobrero et al. 2014a}). Moreover, robust knowledge about multiple ecological factors (distribution of resources such as food, predation risk, and soil conditions associated with digging burrows or nesting sites) surely will contribute to build a theory of sociobiology that is closer and more consistent with the diversity of mammalian social behavior (^{Tang-Martínez 2003}; ^{Hayes et al. 2007}). Taken together, establishing how group-living varies across species is critical for comparative studies to examine the origin and the adaptive value of this behavior (^{Blumstein & Armitage 1998}; ^{Ebensperger 2001}; ^{Ebensperger & Blumstein 2006}). The scarce available and anecdotal evidence suggests that Porter’s rock rats lives in small groups in communal burrow systems and both diurnal and nocturnal activity has been recorded (^{Pearson 1983}; ^{Verzi et al. 2015}). Also, in captivity, individuals displayed a high tolerance for conspecifics (^{Verzi et al. 2015}); however it has not been tested in natural populations and laboratory conditions can modify behavior of captive individuals (^{Calisi & Bentley 2009}).

Our aim in this research note is to report data of the activity patterns and social behavior of freeliving A. porteri. The study site was located at the Fundo San Pablo de Tregua (39°35’S, 72°05’W), Panguipulli, Chile.

San Pablo de Tregua is characterized by a dry and short summer where mean annual precipitation is around 5000 mm, and ambient temperature ranges from 5°C to 20°C (^{Alvarez & Lara 2008}; ^{Guevara et al. 2015}). The site consisted of Valdivian rainforest where dominant tree species were coigüe (Nothofagus dombeyi), raulí (N. alpina), and tepa (Laureliopsis philippiana) (^{Schlegel & Donoso 2008}). Animals were captured using a combination of 172 (14 x 14 x 40-cm) Tomahawk (model 201; Tomahawk Live Trap Company, Hazelhurst, Wisconsin) and leg-hold traps during 8 days in January, 2011. Qualitatively, putative nest sites used by A. porteri are structurally similar to burrows and resting locations of O. degus (^{Fulk 1976}; ^{Lessa et al. 2008}), consisting mainly of oblique tunnels connecting the surface to the nest, with several branches and active openings (4-5) associated to patches of perennial bamboo (Chusquea quila), covering part of the burrow system. We defined rats burrow as areas of 1.5 m² covered by vegetation, in which we found signs of A. porteri presence (i.e., feces) and where radiocollared individuals were found during trapping and telemetry periods. Ground mounds were sometimes found outside openings and fresh feces in burrows openings allow us to determine if each burrow system was active.

Moreover, in two occasions rats’ vocalizations were listened (but not recorded and analyzed) during trapping. We placed traps near burrow openings and inside patches with high bamboo cover and baited them with rolled oats, cereals and sunflower seeds. Traps were opened 08:00h and closed 00:00h and checked every hour. During each capture, we recorded sex, body mass and reproductive status, and each animal was marked with an ear tag (Monel 10051; National Band and Tag Co., Newport, Ky., USA). Low number of rats represents the species at a normal density, based on knowledge about population ecology of other caviomorph rodents (e.g. ^{Cassini 1991}; ^{Ebensperger et al. 2008}). All adult-sized individuals (N = 5) were fitted with a radiocollar weighing 7–9 g (BR radiocollars; AVM Instrument Co., Colfax, California) with unique pulse frequencies. Weight of radio-collars represented about 4–5% of body adult weight (e.g. ^{Ebensperger et al. 2008}). At the end of our study all radiocollared animals were recaptured and radiocollars were removed (^{Ebensperger et al. 2004}; ²⁰¹²).

During 5 days and nights it was performed homing technique every two hours, to determine resting locations or putative nest places, using LA 12-Q receiver (for radiocollars tuned to 150.000–151.999 MHz frequency; AVM Instrument Co., Colfax, California) and a handheld 3-element yagi antenna (AVM instrument Co., Colfax, California). Once located, the position of each animal was marked with flagging material coded for individuals. Each radiofix location was then referenced twice with a Garmin portable global positioning system (Garmin International Inc., Olathe, Kansas), precision always was within 5 m. The determination of group composition required the compilation of a symmetric similarity matrix of pairwise association of the resting locations of all adult animals during homing (^{Whitehead 2008}). Social organization was quantified based on the number and sex composition of adult members in a social group (^{Ebensperger & Hayes 2016}). Thus, we conducted a hierarchical cluster analysis of the association matrix in SOCPROG software (^{Whitehead 2009}).

To estimate daily activity patterns and range areas, we recorded locations hourly of all radiocollared animals for 5 days and 4 nights in 2011, at nighttime (21:00-07:00 h) and daytime (07:00-21:00 h). Sunrise occurred at approximately 06:30 h, whereas sunset occurred at 20:30 h. The spatial location of animals was determined using triangulation (^{Kenward 2001}). We used 2 LA 12-Q receivers, each connected to a null peak antenna system (AVM Instrument Co.). Every null peak system had four 7-element yagi antennas. Distance between antenna stations was about 85 m. Bearings from both antenna stations were then transformed into x–y locations with the software Locate II (^{Nams 1990}). We calculated the distance traveled (in m), between successive scans as a measure of aboveground activity of the Porter’s rock rats. The same individuals were monitored throughout consecutive days and nights. As a result, locations recorded at 24-h intervals were not independent of one another. Consequently, and for statistical analyses, we divided the entire data collection period into 5 days and 4-night cycles, defined on the basis of sunrise and sunset at study site. For the daytime portion of each activity cycle, we calculated the mean distance travelled for each radiocollared individual. We used a similar approach for the nighttime portion of the activity cycle. As a result, each radiocollared individual contributed 2 dependent data points to our analysis of activity. We used repeated-measures analysis of variance to examine the effect of activity time (day versus night) on individual activity of males and females.

The range area was determined from locations recorded through triangulation and included animal resting locations. While triangulation is thought to interfere less with the activity of radio-collared animals compared with homing (^{Kenward 2001}; ^{Ebensperger & Blumstein 2006}), the topography and cover type of our study site precluded the use of longrange radio-fixings because of signal bounce. Thus, although we used the homing technique, we previously trained ourselves to locate animals quickly to minimize disrupting their navigation or locomotion behavior (see ^{Ebensperger et al. 2008}). Data points from each individual were mapped using the 95% minimum convex polygon algorithm in Ranges 6 (^{Kenward et al. 2003}). Pairwise estimates of the percent overlap between polygons for different animals also were calculated using Ranges 6. We compared the mean size (in m²) of range areas and percent range overlap by male and female Porter’s rock rats with Mann–Whitney U-tests. We used Wilcoxon matched-pair tests to compare percent overlap in range areas of individuals assigned to the same burrow location associations and percent overlap that these individuals had with individuals assigned to different associations in January 2011. All statistical analyses were calculated using Statistica 7.0 (StatSoft Inc. 1984–2004) and results are reported as mean ± SE.

Animals used 2 to 9 burrow locations, namely where one or more rats were found repeatedly during total radioscans (n=22). We recorded 21 scans in which animals shared burrow or resting locations. Of these, 20 observations involved male–female pairs, and 1 involved 2 males–1 female associations. Distance moved between radioscans was variable through time of day or night (Fig. 1). The distance moved between any two consecutive radio fixes during the night averaged 49 ± 28 m (n = 5, range: 1-249 m), and was 1.4 times larger than distance moved during day (36 ± 25 m, range: 0.5-425 m). However, this difference was not statistically significant (F_8,9 = 0.420, P = 0.535). Five individuals were radio-tracked during the day and night, which provided an average of 98 useful radio fixes per animal. Spatial overlap among radio-collared Porter’s rock rats was relatively low (20 ± 5%) ranging from 3% to 34% (Fig. 2). If data from all individuals are combined, the size of range areas averaged 48 ± 13 m² (n = 5). When sex was examined, females tended to range over larger areas (51 ± 3 m², n = 3) than males (45 ± 39 m², n = 2), a non-statistically significant difference (Mann–Whitney U test, z = 3.0, p > 0.10).

Fig. 1 Mean (± 1 SE) distance moved (m) since previous scan of Porter’s rock rats (Aconaemys porteri) monitored every 1 h for 4 days and 6 nights at San Pablo de Tregua. 

Fig. 2 Range areas (95% minimum convex polygons) of five Porter’s rock rats (Aconaemys porteri) during the activity period. The arrow signals the geographic north. 

Taken together, and based on this short-term study, results suggest that A. porteri are social. However, compared to other octodontids, social behavior of A. porteri seems relatively low, in terms of their low spatial overlap and group size. For example, evidence for Aconaemys suggests that A. fuscus forms social group up to 7 individuals (^{Muñoz-Pedreros 2000}). On the other hand, similar studies, using telemetry techniques in O. degus, revealed range areas over 1 ha in summer with spatial overlap between individuals ranging from 12.5-59.4% (^{Quirici & et al. 2010}). Our results suggest a smaller group size for A. porteri, however future studies are recommended, including larger sample sizes or different populations, to confirm results reported here. Additionally, our results support that A. porteri is active during daytime and nighttime. Continuous pattern of activity in other octodontids (Octodon lunatus) may be the consequence of relatively stable microclimatic or cover conditions (^{Jensen et al. 2003}; ^{Sobrero et al. 2014b}). Plant cover has been shown to be an important factor for the distribution (e.g. ^{Birney et al. 1976}), social group size and space use in caviomorphs, included octodontids (^{Quirici & et al. 2010}; ^{Sobrero et al. 2014b}) and cavies (^{Asher et al. 2004}; ^{Taraborelli 2008}). Compared with range areas of other octodontids like Octodon degus (^{Hayes et al. 2007}), O. lunatus (^{Sobrero et al. 2014b}), and Octodontomys gliroides (^{Rivera et al. 2014}), A. porteri showed small range areas, where polygons matched the spatial distribution of bamboo patches. Finally, these results are consistent with the hypothesis of the evolution of sociality in Octodontids, in which social living would be a derive characteristic evolving relatively recently from solitary-living ancestors. This novel evidence gives new insights into the social behavior of this species and the evolution of sociality across caviomorph rodents.