Introduction

The South American gray fox (Lycalopex griseus) is one of the three canids present in Chile, distributed from Arica to Magellan, where it thrives in a variety of habitats (^{Iriarte and Jaksic 2012}). The ecological aspects of this canid species have been extensively studied (^{Iriarte and Jaksic 2012}), but information still remain scarce with regard to research in anthropized areas. The habitats of this fox species are currently under a progressive transformation across a large part of its range, resulting from the conversion of native forests into commercial crops (^{Echeverria et al. 2008}). Thus, knowing the ecology of this species in disturbed environments is key. Any alteration of the habitat requires species to display ecological flexibility to exploit the available microhabitats (^{Crooks 2002}), including the existing food resources, leading to various effects in the community where it thrives. The objective of this study was to document the feeding habits of L. griseus in an agroecosystem located to the center-south of Chile, characterized by a low biodiversity (^{Benton et al. 2003}). In addition, this study compared the relative contribution of the prey species according to their origin (native vs. exotic), as well as in terms of the relative contribution to total biomass consumed.

Materials and Methods

The study area is located 20 km from Mulchén in the center-south of Chile (-37° 49’ 42’’ S, -0.72° 14’ 51’’ W). This biogeographic zone has a humid climate (Di Castri and Hajek 1974) and was originally covered by deciduous forests (^{Gajardo 1994}); today, this area has been converted into an agroecosystem mainly used to grow lupine (Lupinus spp.) and oat (Avena spp.).

During the spring (southern hemisphere) of 2016 (November/December), the study area was walked in search of feces through a path (forked in some areas) that runs across crops, covering a total of 4 km. Fox scats were recognized based on gross morphology (^{Chame 2003}) as well as on images from camera traps arranged systematically (separated by 500 m) throughout the study area (^{Kays and Slauson 2008}). These made it possible to identify individuals of the species. Scats were collected in paper bags for subsequent laboratory analysis. In the laboratory, scat samples were dried at 60 °C and then shredded manually to obtain the remains of prey. These remains were identified based on skull morphology (^{Pearson 1995}) and the arrangement of cuticle hairs (^{Day 1966}), in addition to morphotypical comparisons vs. reference vouchers in the collection of Ecology and Biodiversity Laboratory at Universidad Católica de Temuco. Birds were identified based on feathers, culmen and nails. Then, the abundance of each food item relative to the total was quantified, thus obtaining the relative frequency of the different dietary groups (^{Rau 2000}). Dietary diversity was estimated using ^{Levins’ index (β; Levins 1968}), which allows determining prey diversity from the frequency of prey consumption and the total prey categories observed. The standard deviation of this index was calculated through the Jackknife method (^{Jaksic and Medel 1987}).

The effect of biomass on the dietary preferences of the gray fox was evaluated using the trophic isocline method (^{Kruuk and DeKock 1980}), where the frequency of prey occurrence is graphically incorporated according to the biomass of each relative to the total consumed (^{Rau 2000}). Trophic isoclines are arranged around percentages that allow setting the importance of a particular prey category in the diet of the predator studied (^{Rau 2000}). In addition, we calculated the prey geometric mean based on the sum of the product of the number of individual preys by the log10 transformation of its weight (^{Muñoz-Pedreros et al. 2018}). Rodent weight figures were obtained from ^{Muñoz-Pedreros and Gil (2009)}, as well as from the birds of ^{Norambuena and Riquelme (2014)}. In the latter group, passeriform birds were considered as a consumption benchmark according to evidence from previous studies (^{Zúñiga et al. 2008}). Arthropods and vegetation were excluded from this analysis, the former due to their poor representation relative to the total, and the latter due to the impossibility of identifying the remains collected. In addition, the prey categories observed were compared according to origin (native vs. exotic; ^{Muñoz-Pedreros and Yáñez 2000}), in terms of richness and abundance.

Results

A total of 73 scat samples were collected, finding 13 prey categories or items. These were dominated by mammals — mainly rodents of the family Cricetidae (4 species) and Muridae (1 species; Table 1), with a combined relative frequency of nearly 50 % (Figure 1). In addition, the remains (incisors and nails), of an Echimyidae, the coypu (Myocastor coypus), which was the prey of largest size. Lagomorphs — an exotic group at this latitude — were also included in this dietary pattern. Birds and arthropods ranked second to mammals in relative frequency of consumption by the gray fox. The trophic diversity was β = 7.43 + 1.93 (excluding plant materials). Native species (mammals) were represented by five species (62.5 % of total), with an abundance of 54 individuals (79.41 %; non-identified rodents were excluded in both cases).

Figure 1 Percentage of prey consumption by Lycalopex griseus in south-central Chile. m. n. i.: unidentified murids. 

The evaluation of the effect of prey biomass on trophic isoclines revealed that rodents exceeded the 50 % isocline despite representing 40 % of total biomass (Figure 2). The coypu and the hare (Lepus europaeus) were assigned to an intermediate isocline (5 %), while the rest of preys analyzed were allocated to the lowest isocline (1 %). The geometric mean for the preys analyzed was 35.35 grams.

Figure 2 Trophic isoclines of preys consumed by Lycalopex griseus in the study area. Bd: Birds; Lep: Lepus europaeus, Myo: Myocastor coypus; Ory: Oryctolagus cuniculus; Rd: Rodents. 

Discussion

The results obtained regarding prey diversity are consistent with those reported in other studies conducted in the center-south of Chile (^{Rau et al. 1995}; ^{Zúñiga et al. 2008}), where small mammals — particularly rodents — constituted the core of the gray fox food spectrum. Abrothrix longipilis, A. olivaceus, Oligoryzomys longicaudatus, and, to a lesser extent, Irenomys tarsalis constituted the core of the gray fox trophic diversity in the study area. These species are characterized by their flexibility and the use of environments with little plant cover, typical of agroecosystems (^{Muñoz-Pedreros and Gil 2009}). The absence of species sensitive to habitat loss /or with highly restricted distributions was evidenced (^{Quintana 2008}). The diversity of rodents observed in this study partially agrees with the records of rodents for this latitude (^{Muñoz-Pedreros 1992}), where the sampling season coincides with their reproductive period (^{González and Murua 1983}). This allowed recording a significant representation of local prey availability. The presence in feces of an exotic species, Rattus norvegicus, which is common in anthropized environments (^{Fernández and Simonetti 2013}), evidences the spatial disruption that characterizes agroecosystems, where this rodent is found in optimum conditions given its ecological versatility (^{Jaksic et al. 2002}). The geometric mean obtained is similar to the average weight of rodents, thus highlighting the importance of this group relative to the total observed. The same is interpreted from trophic isoclines: the positioning of rodents on the 50 % isocline suggests that this group is a key food item within the food spectrum (^{Rau 2000}). The presence of a single coypus specimen in the food spectrum, considering the size of this species relative to that of the gray fox (10 kg; ^{Muñoz-Pedreros and Gil 2009}), suggests that its consumption may have been due to scavenging rather than hunting, which assumes an incidental consumption. This is based on the fact that the gray fox and the coypu are of similar size (^{Iriarte and Jaksic 2012}), which would make it difficult for the fox to hunt it; additionally, there are no records of this prey in the gray fox diet in areas where both coexist (^{Zúñiga et al. 2008}). The higher number of scats collected in this study vs. the figure recorded by ^{Torés (2007)} for the culpeo fox (Lycalopex culpaeus), where a significant representation of the local prey diversity was reported, suggests that the number of samples in this study was sufficient to achieve representativeness of the dietary analysis.

A noteworthy finding is the absence of marsupials such as the Monito del Monte (Dromiciops spp.), in the diet of the gray fox, which contrasts with other studies on this canid in the center-south of Chile (^{Zúñiga et al. 2008}). This fact could be explained by the constraints of the study area to be occupied by this marsupial, due the close relationship of the monito del monte with the native forest, as well as its displacement limitations (^{Fonturbel et al. 2010}). On the other hand, the broad range of movement of the fox across patches of native forest (^{Zúñiga et al. 2009}) would facilitate the consumption of this prey. This suggests a small population size of this marsupial species, which deserves to be further investigated. In this way, the absence of arboreal species in the diet of the gray fox is compensated by the presence of species of terrestrial habits. However, the presence of Irenomys tarsalis in the diet of the fox shows the flexibility of this rodent in the use of microhabitats with low vegetation cover, despite its preference for forested areas (^{Kelt 1993}), which would lead to a lower frequency of capture vs. the rest of the cricetids observed.

The low consumption of lagomorphs according to the trophic isoclines contrasts with the observations for Lycalopex culpaeus, a canid of larger size than the gray fox that has been reported in other latitudes (^{Rubio et al. 2013}; ^{Zúñiga and Fuenzalida 2016}), where the frequency of consumption is considerably higher. This finding leads to setting two types of hypothesis: on the one hand, the assumed high abundance of lagomorpha this group is not supported, despite the success experienced as exotic species at this latitude, which has enabled it to spread rapidly across the Chilean territory (^{Jaksic et al. 2002}). On the other hand, an interference effect by larger predators, such as the puma, in the study area, leading to a spatial overlap (^{Zúñiga et al. 2009}), and ultimately, to competition by exploitation. Both hypotheses should be tested in further studies to investigate which ecological elements affect this canid in anthropized areas, taking into account the variations in prey availability resulting from seasonal effects and changes of land use.

The frequency of bird consumed by L. griseus in the study area is consistent with the records from native forests in the same biogeographic zone (^{Zúñiga et al. 2008}). However, the impossibility to identify the species consumed has restrained the estimation of the trophic spectrum of this canid (underestimation of the Levins’ index; ^{Greene and Jaksic 1983}), considering the diversity of birds in this habitat (^{Zúñiga 2014}). Nonetheless, the low contribution of birds to total prey biomass, as reflected by their inclusion in the 1 % isocline, suggests that this group is a minor component of the gray fox diet, in spite of its frequency of consumption. A similar finding was observed for arthropods, i.e. the amount of arthropod biomass was negligible in relation to the total, despite their frequency of consumption. This fact is even more significant considering arthropod richness at a regional scale (^{Peña 1987}); hence, it should be interpreted with caution.

The plant remains observed differ from what is documented in localities including native forests, where the presence of seeds of several species has been reported (^{Zúñiga et al. 2008}; ^{Morales-Paredes et al. 2015}), suggesting a potential role of the gray fox in seed dispersal. The scarcity of plant remains may indicate a change in the vegetation pattern as habitat fragmentation continues in the study area. Given this scenario, further studies are required that consider the richness of plant species according to their interaction with the local fauna.