Species composition of small non-volant mammals in the Parque Estadual das Fontes do Ipiranga, São Paulo, Brasil

This study provides the first inventory of small non-volant mammals in the Parque Estadual das Fontes do Ipiranga (PEFI), a protected area in the city of São Paulo, Brazil. The data was collected from 2015 to 2017 in 16 research campaigns with a duration of five days each. Four areas with different phytophysiognomies were sampled throughout the dry and rainy seasons. We sampled small mammals using live-capture and pitfall traps. Eleven species were captured, composed of six rodents and five marsupials. The sampling effort involved 5,600 traps/night, there were 527 capture events, and we captured 302 distinct individuals: 174 marsupials and 128 rodents. Recaptures accounted for 42.7% of the total captures. No significant differences were observed in the richness and abundance of small non-volant mammals between the different phytophysiognomies. We also found no significant differences in the richness and abundance of small non-volant mammals between the dry and rainy seasons. The relative abundance (Ar) and constancy index (C) of the species showed that the three most abundant and common species in the PEFI are: Didelphis aurita, Akodon montensis and Oligoryzomys nigripes, which represented 93.7% of the captures. Tomahawk traps accounted for 69% of the total captures, and pitfall traps were responsible for detecting the greatest richness, capturing 81.8% of the species. Comparing the efficiency of the different sampling methods in capturing small mammals in the PEFI, we observed significant differences between both pitfall versus Tomahawk and pitfall versus Sherman in the understory. The results obtained in this study are consistent with the past and current situations of the forest fragment which are in recovery after significant altered by anthropic activity. In light of this scenario of degradation and isolation, a defaunation process affecting the mastofauna is very likely in the PEFI, which favors the establishment and dominance of generalist species. This study could be the basis for further monitoring programs of small non-volant mammals. The data obtained here will also increase knowledge about the diversity of small mammals in urban fragments of the Atlantic Forest and demonstrate the importance of the PEFI for the maintenance of ecologically important species within the largest metropolitan region in Brazil. These species play important biological roles for the maintenance of ecological interactions and for the provision of rare ecosystem services for the anthropic landscape, which is of great value to the city of São Paulo.


Introduction
The Atlantic Forest is one of the most threatened biomes in the world and one of the 25 global biodiversity hotspots (Mittermeier et al. 2004, Ribeiro 2009, Jenkins et al. 2013. It originally covered a total area of 1,315,460 km 2 , spread over 17 Brazilian states (Peres 2010). The Atlantic Forest is also one of the regions with the highest biological richness on the planet (Mittermeier et al. 2006). Its remnants currently comprise only 12.4% of its original coverage and are mostly found in small fragments of less than 250 ha, of which only 9.3% are within protected areas (Hirota 2018). Although extremely degraded, the Atlantic Forest remains remarkably heterogeneous and its wide variety of ecosystems allows the occurrence of diverse plant and animal species (Galindo-Leal & Câmara 2005).
Considered to be the Brazilian biome with the second largest number of species and endemic mammals (MMA 2002, Paglia et al. 2012, the Atlantic Forest is one of the most diverse regions of small mammals in South America (Galindo-Leal & Câmara 2003), comprising 105 species of rodents (Patton et al. 2015) and 30 species of marsupials (Bovendorp et al. 2017). However, the absence of published species inventories of some areas has created a significant knowledge gap concerning the presence and distribution of its taxa (Costa et al. 2005, Brito et al. 2009, Galetti et al. 2009, De Vivo et al. 2011. Bovendorp et al. (2017) compiled information from 136 studies conducted on small non-volant mammals from seven different types of vegetation in the biome and these results enabled the identification of priority areas for future sampling efforts. Despite some advances in research, further studies on the diversity of the Atlantic Forest species are necessary in order to increase both the understanding and direct conservation efforts for biodiversity (Galetti et al. 2009, Ribeiro 2009).
The Parque Estadual das Fontes do Ipiranga -PEFI is among the few, but important, remaining areas of Atlantic Forest effectively protected as Conservation Units in the city of São Paulo ). Its boundaries have been set since 1893, when the PEFI area had approximately 697 ha (Barbosa et al. 2002). Since then, this fragment has been affected by the construction of highways and avenues, the urbanization of neighboring districts and by fires, leading to a decline in its vegetation cover (Peccinini & Pivello 2002). The PEFI vegetation is typical of dense tropical rainforest of the Atlantic hillside (São Paulo 2008), where altitude ranges from 770 to 825 m (Barbosa et al. 2002). The area is typically used for recreation, teaching and research. A number of studies were carried out in the PEFI, and its hydrography, topography, plant physiognomy and climate are well known (Fernandes et al. 2002, Santos 2008, Villagra & Romaniuc-Neto 2010. However, there have been very few studies on the local fauna (Bicudo et al. 2002, Malagoli et al. 2008) until 2013, when some studies on vertebrates began to be carried out (Perrella & Guida 2013, Benedicto 2015, Monticelli & Morais 2015, Lisboa et al. in press, Moraes 2017, Rossi 2017, Monticelli & Antunes 2018, Perrella et al. 2018, Monticelli 2019, Rossi et al. 2020). The aim of these studies has been to increase the knowledge about local fauna and create opportunities for new research.
Studies concerning mammals in the PEFI were mainly conducted on larger species. Thus, despite the advances in research on mammalian fauna, knowledge remains deficient regarding the composition of the small non-volant mammal community. Even though the PEFI is situated along the Atlantic Forest phytophysionomy best studied regarding to composition of small non-volant mammals (Bovendorp et al. 2017, Fiqueiredo et al. 2017, it is precisely located in the Paulista plateau, a regional sampling gap for the group (Figueiredo et al. 2017). Therefore, this study has three objectives: 1) to inventory the small non-volant mammals occurring in the PEFI in order to create a database for future monitoring programs; 2) to compare the success of distinct sampling methods; and 3) to compare richness and abundance of small non-volant mammals throughout seasons and phytophysiognomies.

Study area
The Parque Estadual das Fontes do Ipiranga (PEFI) is located in the municipality of São Paulo (Figure 1) and borders the municipality of Diadema (23º38'08" S and 23º40'18" S and 46º36'48" W and 46º38'00" W) (Fernandes et al. 2002). The PEFI is one of the country's largest and most important remnants of Atlantic Forest inside an urban area (Bicudo et al. 2002) and the third largest fragment of the biome in the city of São Paulo (Rancura & Cerati 2020). It currently comprises a total area of 526.4 ha with 340 ha of biological reserve (Bicudo et al. 2002).
The original area of the PEFI was composed of lands owned by farmers and which underwent a vegetation recovery process after their expropriation by the São Paulo State government, making it an area of secondary forest with little more than 100 years of recovery (Barbosa et al. 2002, Barros et al. 2002. The PEFI phanerogamic flora is composed of 1,159 species from 129 families (Barros et al. 2002, Villagra 2008. The climate in the city of São Paulo is categorized as Cwa (according to Koppen classification), also called humid subtropical, and is marked by a dry winter and a rainy summer. The average temperature and rainfall establishes that the driest and coldest periods are between the months of April and September, whereas the warmest and rainiest periods of the year correspond to the months of October to March (IAG 2017).
The Atlantic Forest fragment closest to the PEFI is the Parque Estadual da Cantareira which is located approximately 20 km away. Such a scenario of isolation associated with the urban pressure generated by the growth of the surrounding cities led to negative impacts on PEFI biodiversity (Gomes et al. 2003, Monticelli & Morais, 2015. Based on the Unit Management Plan directions, this study sampled four different points of the PEFI, namely: Instituto de Botânica 1 (Ibot.1) -an area of dense forest with homogeneous canopy; Instituto de Botânica 2 (Ibot.2) -forest with high-sized heterogeneous canopy; Parque de Ciência e Tecnologia 1 (Cient.1) -forest with sparse homogeneous canopy; Parque de Ciência e Tecnologia 2 (Cient.2) -forest with discontinuous canopy/degraded forest.

Data collection, capture and tagging
Data collection took place from 2015 to 2017, for a total of 16 monthly research campaigns with a duration of five days each. In all campaigns, the four aforementioned areas (Ibot.1, Ibot.2, Cient.1 and Cient.2) were sampled. Every area was sampled four times, for two campaigns in the rainy season and two in the dry season.
In order to capture the small non-volant mammals, we employed two types of live capture traps: a box-trap (Sherman, size 30 x 7.5 x 9 cm) and a cage-trap (Tomahawk, size 45 x 20 x 20 cm). Traps were baited with a mixture of sardines, cornmeal, bananas, peanut butter and pineapple essence. The mixture was replenished daily.
Two parallel 100-meter lines were established in each area, 30 m apart from each other. Each line had 10 capture stations equidistant at every 10 m. In every station, three live traps were placed spread along the ground (named "forest layer 1") and at about two meters high (named "forest layer 2"), attached to understory branches, totaling 60 live traps per sample site.
Pitfall traps were also arranged in the sampled points, but at 100 m apart and parallel to the live trap lines. On each site, 10 buckets were installed, 10 m apart from each other, buried up to the soil level (named "forest layer 0"), connected by a guide fence made with black 80-centimeter high plastic canvas and supported by wooden poles and metal staples. Inside each bucket, a piece of expanded polystyrene foam was set down to avoid animal drowning in case of flooding.
Upon the animal capture, the following procedures were executed: taxonomic identification, individual tagging with a numbered metallic earring (Ear tags, National Band and Tag Company, USA), collecting of feces and ectoparasites, weighing, body measuring, sex and reproductive condition recording and subsequent release at the capture site.
The species identity was determined following Gardner (2007) and Patton et al. (2015). Taxonomic identifications of representatives of genera with cryptic species-diversity were performed based on cytogenetic analyses carried out in the Special Laboratory of Ecology and Evolution of the Instituto Butantan.
As there are no previous studies concerning small non-volant mammals from the PEFI, testimony specimens of all the species captured were collected. The specimens were deposited in the mammal collection of the Museu de Zoologia da Universidade de São Paulo and prepared according to the guidelines established by the institution.
The capture success was determined by the total number of captures multiplied by 100 and divided by the capture effort (traps per night). All procedures were authorized by the pertaining environmental agencies under license SISBIO no. 45520 and SISGEN no. AE48610.

Richness, abundance and seasonality
The species richness was assessed through the non-parametric estimator Jackknife 1 (Burnham & Overton 1979), using the software EstimateS 9.0 (Colwell 2013). The analysis was performed with 100 randomizations and the days as the sample unit.
In order to compare the richness and abundance among the different phytophysiognomies sampled, we employed a Kruskal-Wallis test. The capture data was also used to evaluate putative differences in richness and abundance of small mammals between the dry and rainy seasons. Finally, we applied a paired t-test for parametric data and the Mann-Whitney test for nonparametric data. In all cases, the significance level adopted was 5% (p < 0.05).
The relative abundance (Ar) of species was determined by the number of individuals of each species captured multiplied by 100 and divided by the total number of individuals captured. The constancy index (C), which allows species to be grouped into categories based on its capture frequency, was established as: common species -present in more than 50% of the samples; relatively common species -present in 25 to 50% of the samples; and rare species -present in less than 25% of the samples (Dajoz 1983).

Capture method assessment
The different methods implemented for the capture of small mammals in the PEFI were compared to assess their efficacy at the study site, using the Kruskal-Wallis test and the Dunn post-test.

Data collection, capture and tagging
During this study, 11 species of small non-volant mammals were captured, six of which were rodents belonging to the Cricetidae and Caviidae families and five were marsupials belonging to the Didelphidae family ( Figure 2). From a 5,600 traps/night sampling effort, 527 capture events occurred, represented by 302 individuals (174 marsupials and 128 rodents; Table 1). Recaptures accounted for 42.7% of the total capture events. The success rate of small mammal capture was 9.41%, being 5.55% in the dry season and 3.85% in the rainy season. The vegetation formation with the highest number of captures was Cient.1, followed by Cient.2, Ibot.2 and Ibot.1 (Table 1).

Richness, abundance and seasonality
The estimated richness of small mammals calculated was close to the empirical results, indicating 12.04 species occurring in the PEFI. (Figure 3). There was no significant difference in the small mammal richness among the different sampled areas (H = 3.779; p = 0.286), nor in the abundance among the different sampled physiognomies (H = 0.602; p = 0.895).
Nine species were captured during the dry period, five rodents and four marsupials, and six species during the rainy season, four rodents and two marsupials. Although more captures occurred during the dry months (n = 167) compared to the rainy months (n = 135), we found no significant difference in the total richness of small mammals (rodents and marsupials) between the two periods (t = -1.275; p = 0.211). Similarly, there was no significant difference in the richness of rodents (t = 2.507; p = 0.120) and marsupials (t = -0.547; p = 0.340) separately between the seasons. The total abundance of small mammals did not vary significantly between the dry and rainy periods (t = 1.999; p = 0.147), nor did it for rodents (u = 0.480; p = 0.315) and marsupials (u = 0.626; p = 0.265) separately.
In addition to the species of small non-volant mammals, field activities led to the record of four other non-target native taxa, namely: brown howler monkey (Alouatta guariba Humboldt, 1812), orange dwarf porcupine (Coendou spinosus F. Cuvier, 1823), nine-banded armadillo (Dasypus novemcinctus Linnaeus, 1758) and three-toed sloth (Bradypus variegatus Schinz, 1825). Additionally, there were unintentional captures of alien species, such as the common marmoset (Callithrix jacchus Linnaeus, 1758), the black-tufted marmoset (Callithrix penicillata É. Geoffroy, 1812) and domestic animals, such as the common household cat (Felis catus Linnaeus, 1758). Taxonomic identification of non-target species followed Abreu et al. (2020).  Table 2. Species of small non-volant mammals captured in the Parque Estadual das Fontes do Ipiranga and their respective values of relative abundance (Ar) and constancy index (C). Ptf = pitfall traps; sh = Sherman; t = Tomahawk method were D. aurita, the most abundant species in the PEFI. With a sampling effort of 800 traps/night and accounting for only 16% of the total capture events, the pitfall trap was not the most effective method of capture, but it was responsible for detecting nine of the 11 species of small mammals. Despite its lower sampling effort and the absence of bait, this method proved to be the most efficient and was essential for the sampling of small mammals in the area. Box traps (Shermans) were the least successful in terms of capture events. With a 2,400 traps/night sampling effort, the same as the Tomahawk, they captured five of the 11 species and their capture events represented only 15% of the total. Concerning the captures in the different strata of the forest, pitfall traps (0) presented the most successful capture rate of 13.65%, followed by live traps (Sherman and Tomahawk) located in the forest substrate (1), with a  (2), with a rate of 3.66%. Assessing the efficiency of the methods for capturing different species of the PEFI, we found a significant difference between the methods (H = 30.24; p = 0.0001). Post-test analysis showed a significant difference between pitfall versus Tomahawk from forest stratum 2 (p = 0.05) and pitfall versus Sherman from forest stratum 2 (p = 0.05). We found no significant difference between pitfall, Sherman (1) and Tomahawk (1). There was also no significant difference between Sherman (1), Sherman (2) and Tomahawk (2) methods.

Discussion
Environmental characteristics such as vegetation type, primary production and terrain directly impact the mammalian community present in a given area (Peres 2000, Haugaasen & Peres 2005, Galetti et al. 2009). Anthropic actions, such as habitat suppression and fragmentation, also affect the permanence of mammal populations in different environments (Chiarello 1999, Cullen-Junior et al. 2000, Peres 2000, Galetti et al. 2009, Brocardo & Cândido-Junior 2012. These factors in association with the size of the remaining natural area may determine the richness of mammal species (Chiarello 1999), as the absence of large protected areas has been directly related to the decrease in species, especially those of larger size (Chiarello 2000, Gurd et al. 2001, Ceballos et al. 2005, Cardillo et al. 2005, Jorge et al. 2013. Large remnants of the Atlantic Forest are related to the viable maintenance capacity of several species of mammals (Chiarello 1999, Cullen-Junior et al. 2000. Contrarily, the PEFI is an example of the loss of vegetation cover and fragmentation which, due to its isolation as a small area inside the anthropic landscape of São Paulo, the most populous city in Brasil, presents a low richness of small nonvolant mammals when compared with other areas of the Atlantic Forest found nearby and better preserved: 32 species found in the Estação Ecológica do Bananal (Abreu-Junior & Percequillo 2019); 23 in the Reserva Florestal do Morro Grande ; and 21 in the Parque Estadual Carlos Botelho ). However, the number of species recorded in the PEFI is still compatible with the majority of the studies reported for the Atlantic Forest biome (Figueiredo et al. 2017). These results were expected, considering that a substantial reduction in mammal richness has been reported for small fragments of the Southern Atlantic Forest (Abreu-Junior & Köhler 2009, Brocardo & Cândido-Junior 2012 in Southeastern (Chiarello 1999, Briani et al. 2001, Pardini et al. 2005) and Northeastern Brazil (Silva Junior & Pontes 2008. In light of its loss of vegetation cover and isolation, it is possible that a defaunation process is occurring in the PEFI, affecting not only the small mammals, but mainly other mammals which are larger and/ or have greater habitat requirements. In the PEFI no invasive rodents were captured, which is surprising considering all its anthropogenic modifications. According to Bovendorp et al. (2017), 24% of the Atlantic Forest fragments have at least one species of invasive rodent.
The forest types present in areas Ibot.1 and Ibot.2 are considered similar to each other according to Bicudo et al. (2002) and they show a relatively larger number of arboreal individuals compared to the other areas, with distinct aggregation patterns and dense understory. Fewer species were found in these areas. The Cient.1 area is comparable to Ibot.1 and 2 (Bicudo et al. 2002) and presented a similar number of species. These three areas are considered to be less degraded than Cient.2 (Bicudo et al. 2002).
Although we did not find significant differences in small mammal richness in the PEFI among four sampled areas, the Cient.2 had the highest number of species, where eight of the 11 species were captured. Of these eight, six were terrestrial or semi-fossorial (C. aperea, T. nigrita, A. montensis, O. nigripes, M. americana and M. iheringi). This area is comprised of continuous forest, degraded by fires that occurred in the late 1990s and early 2000s (Peccinini & Pivello 2002).  suggest that younger or more altered forests, as found in Cient.2, lead to greater biomass production and, consequently, greater availability of fruits and arthropods which are the main items consumed by small non-volant mammals. In addition, this variety of phytophysiognomy has a relatively more opened canopy with a denser understory which favors the proliferation of terrestrial or understory species and minimizes the chance of occupation by forest canopy species .
The Ibot.1 site was the only one in which C. philander, an arboreal species commonly found in canopies but also in the understory (Delciellos et al. 2006), was captured. The exclusive capture at this site may be related to the fact that the species is described as arboreal of medium to high canopies (Aragona & Marinho-Filho 2009) and the site presents a large concentration of 4 to 6 m tall trees, due to its absence of recent major impacts.
Communities of small non-volant mammals are long considered to be generally composed of two or three dominant species and other species tend to be in greater rarity (Fleming 1975), a pattern largely considered to be a response to adaptive flexibility of dominant species. The results found in this study are in accordance with this expected pattern for a community of small non-volant mammals in degraded/ recovering areas of the Atlantic Forest (Pardini et al. 2005, Puttker et al. 2008. Other characteristics of the PEFI such as isolation, secondary forest and urban surroundings might also promote this scenario of species occupation. These elements favored the establishment and predominance of three generalist species in the PEFI: D. aurita, A. montensis and O. nigripes, which accounted for a combined 93.7% of total captures. Seven of the 11 species captured in the PEFI are among the 22 species suggested as hyper-dominant in the Atlantic Forest (Bovendorp et al. 2017).
The process of forest regeneration and the impacts it suffers, such as fires, cause environmental changes that may reflect on the structure of the small mammal community, as observed by Oliveira (1995), creating new species dynamics to be later assessed. Pinotti (2010) suggests that both structural characteristics of the forest (biomass and depth of leaf litter, branch volume and number) as well as food availability (arthropod biomass in the soil, richness of fruiting plants and number of individuals fruiting in the understory) are liable to change as a result of the forest regeneration process and are strongly linked to favoring or disadvantaging species, either specialists or generalists. Generalist species such as D. aurita, A. montensis and O. nigripes also benefit from a greater availability of food resources found in areas at an earlier stage of regeneration. Specialized forest species have a higher occupancy capacity in more mature forests, where these resources are scarcer (Pinotti 2010). However, it is important to note that the absence of connectivity between the PEFI and other forest fragments prevents the recolonization of the area by other species, which could potentially increase the richness of the rodent and marsupial community (Pardini et al. 2005). As D. aurita was the most abundant species in the PEFI, the low richness found may have been influenced by the presence of this species. Fonseca & Robinson (1990) suggested that the increase in the density of D. aurita, due to the absence or small abundance of predators, for example, could be related to the low richness of small terrestrial mammals in smaller fragments, either due to competition for resources or even predation (Graipel et al. 2003).
The absence of PEFI predators due to reduced vegetation cover and all the effects of isolation inside the anthropic landscape may also be related to the rates of recapture found in this study. The high recapture value (42.7%) suggests that the areas sampled were not occupied by new individuals during the study period. Such a low turnover of individuals may be related to the absence of medium and large predators, which play an important role in the dynamics and structure of the mammal community (Fonseca 1988). For the PEFI, there are no records of native carnivorous mammals and this absence can decrease the rate of predation, favor the permanence of individuals in the same area for a longer time and, consequently, decrease the turnover of individuals between adjacent territories thus resulting in spatial stability, a fact also observed by Lessa et al. (1999).
Regarding the species found in the different strata, the trapping method alone is not sufficient to estimate the use of vertical space by small mammals (Delciellos et al. 2006, Preveddello et al. 2008). The finding of species described as terrestrial (A. montensis) or arboreal (J. pictipes) (Paglia et al. 2012) on ground and understory may be related to physical characteristics of the area, such as connectivity due to the presence of lianas and fallen trunks, seasonal variations in food availability and interspecific competition (Begon et al. 2006, Lambert et al. 2006, Hannibal & Caceres 2010. Didelphis aurita and O. nigripes are considered scansorial and can be found on the ground and understory (Paglia et al. 2012). Although young individuals of D. aurita have been previously reported to be more prevalent in the upper strata (Prevedello et al. 2008), in this study, 62.83% (n = 71) of young individuals of D. aurita were captured on the ground and 37.17% (n = 42) in understory.
Whereas not significantly different, the capture success rate of this study was greater in the dry season (5.55%) than in the rainy season (3.85%). This can possibly be explained by the variation throughout the year in the availability of food resources. Higher rainfall is related to a greater supply of arthropods (Janzen 1973, Charles-Dominique 1983, Wolda 1993, Santos-Filho et al. 2008) and fruits in the environment (Foster 1982, Charles-Dominique 1983, Julien-Laferrière & Atramentowicz 1990, Bergallo & Magnusson 1999, 2002, Santos-Filho et al. 2008). This makes the animals more likely to find food while moving less (Stallings 1988), thus decreasing the efficiency of the baits and the probability of capturing small mammals (MacClearn et al. 1994). A higher relative rate of capture in the season of lower rainfall has previously been observed in other studies with determined seasonality (Mello 1980, O' Connell 1989, MacClearn et al. 1994, Vieira 2002, Alho 2003, Santos-Filho et al. 2008). Among the five species of marsupials captured, three were captured exclusively in the dry period (M. iheringi, C. philander and G. microtarsus), one was captured in both periods, but more often in the dry period (D. aurita), and one species was captured only in the rainy season (M. americana). For rodents, among the six species captured during the study, one was exclusively captured during the rainy season (B. breviceps), two during the dry season (T. nigita and C. aperea) and three were captured in both periods (A. montensis, O. nigipes and J. pictipes). A. montensis was the most frequently captured species in the dry season and O. nigripes in the rainy season.
As there is no previous research on small non-volant mammals from the PEFI, this novel study can be a starting point for monitoring programs of this group, aiming to evaluate possible local ecological changes and advances in the defaunation process.
The different capture methods used in this study complemented each other, as they were responsible for the capture of different species, showing the importance of the use of diverse techniques to capture small non-volant mammals (Santos-Filho et al. 2006, Caceres et al. 2011, Bovendorp et al. 2017. Even though baits are not used in pitfall traps, they have previously been associated with higher capture success rates (Hice & Schmidly 2002, Santos-Filho et al. 2006. In this study, the method with the greatest success in capturing small mammals was the use of cage-type traps (Tomahawk). The difference in the capture success rate of Tomahawks versus Shermans was probably due to the size of the traps. Since the most abundant animal species captured in this study reached around 2 kg, this size was probably incompatible with the Shermans culminating in a lower capture rate using this method.
Considering the great influence marsupials and small rodents have on forest dynamics, as well as their role as habitat quality indicators , the importance of knowing their community formation cannot be underestimated, as its understanding may direct conservation efforts of the forests and species.
Of the 11 species captured in this study, none were included in the national list of threatened species (ICMBio 2018), but six (54.55%) are endemic to the Atlantic Forest: B. breviceps, J. pictipes, T. nigrita, D. aurita, M. iheringi and G. microtarsus (Paglia et al. 2012). In the state of São Paulo, M. iheringi is on the list of threatened species in the "vulnerable" category, and five of the species captured are classified as "almost threatened" (B. breviceps, J. pictipes, T. nigrita, M. americana and G. microtarsus) (São Paulo 2018).
The results of this study add knowledge about the biodiversity of the Atlantic Forest in an urban fragment in São Paulo. Although species richness is not high and generalist species are predominant, the PEFI is important for the maintenance of different animal and plant species, which in turn, play important biological roles for the maintenance of ecological interactions. This demonstrates the great value of the PEFI to the city of São Paulo, due to its provision of rare ecosystem services for the anthropic landscape of the city.