Terrestrial fauna of the largest granitic cave from Southern Hemisphere , southeastern Brazil : A neglected habitat

Studies focusing cave fauna on granitic caves are relatively rare when compared to carbonatic ones and no one considered replicas in the sampling to test fauna distribution patterns. We describe the terrestrial fauna of Riacho Subterrâneo cave through four sampling occasions (replicas) in different seasons. We analyzed seasonality and substrate preference of terrestrial invertebrates and discussed the importance of this neglected habitat as a refuge for fauna. Furthermore, we stressed the importance of the replicas in order to detect subterranean biodiversity patterns. The cave represents the greatest richness considering igneous rocks in Brazil (199 taxa) and has an important role as refuge for epigean fauna, besides the maintenance of troglophilic and trogloxene populations.


Introduction
In the worldwide literature, studies on granitic caves are relatively rare when compared to carbonatic ones (Juberthie, 2000).Usually, granitic (an igneous rock) caves are small, may vary in shape (Twidale and Bourne, 2008) and their final structure is shaped mainly by the power of water (Romaní et al., 2010).
Maria Elina Bichuette, André R. Nascimento, Diego M. von Schimonsky, Jonas E. Gallão, Leonardo P.A. Resende, Tamires Zepon Volume 12 number 2  may -august 2017 They are often formed by large agglomerates of blocks with many openings to surface.In this context, granitic caves can offer microhabitats for the establishment of communities with surface faunistic components (Romaní et al., 2010).Granitic caves are formed by marine/fluvial erosion or talus deposits, with possibility of chemical dissolution (Finlayson, 1986).Granitic and gnaissic caves represent about 3.4% of the Brazilian recorded caves (216 caves) (Cadastro Nacional de Cavidades, 2016).Besides the small sizes compared to other lithologies (e.g., limestone), granitic caves are capable to harbor complexes biological systems (Willems et al., 2002).
Studies comparing granitic caves with other lithologies regarding fauna composition are rare in Brazil (e.g., Gnaspini and Trajano, 1994;Souza-Silva et al., 2011;Bernardi et al., 2012), most conducted in small caves (less than 190 m) and without replicas along annual cycles, which bring limited data to propose faunistic patterns and protection policies.
Subterranean fauna is classified by three categories, associated to the different habitats in according to its ecological and/or evolutionary characteristics.In this context, the troglobites are the most specialized and restricted to subterranean habitats, showing autapomorphies related to their isolation (the classical troglomorphisms are reduction until absence of eyes and body melanic pigmentation sensu Christiansen, 2005); troglophiles are facultative species with established populations inside and outside caves and trogloxenes are those animals utilizing the subterranean habitats as shelters and completing their life cycle outside caves (modified from Holsinger and Culver, 1988).
The topography and mapping, mandatory in environmental studies of relevance which is based on Decree 6640 (Brasil, 2008), are not easy to perform in granitic caves.As consequence, these habitats are frequently neglected considering fauna and physical attributes.Indeed, the demand for extraction of igneous rocks for economic purposes in Brazil is high, which represents a threat to this habitat.
We describe herein the subterranean biodiversity of terrestrial invertebrates in the largest granitic cave of the Southern Hemisphere, the Riacho Subterrâneo cave, testing the influence of seasonality in the taxa distribution and the substrate preference.

Study area
Riacho Subterrâneo cave is inserted in the Itu's postorogenic granite suite (Martins, 2011), in the state of São Paulo, southeastern Brazil (Figure 1) in an altitude of 583 m.The climate is classified as wet and sub warm (Nimer, 1989) with the rainy and warm season between October and March and dry and cold season from April to Septem-ber (Nimer, 1989).The average temperature is ca.15°C and the annual precipitation between 1,000 and 1,500 mm (Instituto Nacional de Meteorologia, 2014).This is the largest cave in igneous rock of the Southern Hemisphere with approximately 1,500 meters of development in mapped passageways (Igual, 2011) (Figures 1  and 2).Most recorded granitic caves are small and Riacho Subterrâneo cave did not follow this pattern comparing to the recorded in Brazilian databases (e.g., CNC/SBE).The innumerable entrances of Riacho Subterrâneo cave (Figure 3A) implies in potential colonization ways for the surface fauna to the cave habitat besides food input for the established cave communities.
The vegetal physiognomies are composed mainly by a Semidecidual Seasonal Forest, interspersed by Atlantic Rainforest and Cerrado (Savannah-like vegetation) (Kronka et al., 2005).The surroundings of the cave are hardly impacted by anthropogenic activities (plantations, pasture and housings) (Figure 3B) and, in 2010, we observed a drastic fire in its surroundings (Figure 3C), representing a punctual disturbance.

Samplings
Four samplings were conducted in total.The first one occurred in October 2010, beginning of rainy season; in that occasion, the cave surrounding vegetation was drastically burnt (Figure 3C).The second and third samplings were in August 2012 (dry season) and November 2012 (rainy season).The fourth sampling occurred in March 2013 (end of rainy season), when the cave surrounding was in an advanced regeneration (vegetation) (M.E.Bichuette, pers.obs.) (Table 1).
The cave possesses as main substrates: huge rock blocks (rocky substrate), sediment degraded from the cement among the blocks (unconsolidated), innumerable roots crossing the openings in the ceiling (Figure 2); litter carried from the epigean environment (Figures 3D-E) and few guano piles, and, also, a mix of them (details in the next paragraph).The cave is predominated by twilight and entrance zones, with aphotic zones.There is a small permanent drainage crossing the cave, which has communication with the epigean drainage.Resources (food input) are carried by the drainage and rains through the ceilings openings and also by the penetrating roots.
Samplings were carried out by direct search qualitative method, i.e., to search for fauna in potential areas for their occurrence (Weinstein and Slaney, 1995;Bichuette et al., 2015), in the available substrates: rocky, rocky with litter, unconsolidated sediment (sand and clay), litter, unconsolidated sediment and litter.In total, we sampled 110 hours by each collector in all samplings, since three people along 3-4 days per occasion formed the team.For analysis purposes and considering the main components, we grouped Terrestrial fauna of the largest granitic cave from Southern Hemisphere, southeastern Brazil: A neglected habitat the substrates in seven categories: Rocky (RO); Unconsolidated (UN); Litter (LI); Unconsolidated and Rocky (URO); Unconsolidated and Litter (ULI); Rocky and Litter (RLI) and a mix of Rocky, Unconsolidated and Litter (RULI) (see details at Figures 2 and 3).
The same team conducted epigean collections in the surroundings of the cave in the same occasions of the cave fauna collections.Based on this information and on literature data, a proposition of classification of the cave fauna status was proposed (troglobitic, troglophilic or trogloxene animals).
Individuals were fixed in loco in formalina 4% (Oligochaeta and Turbellaria), ethanol 50% (Diplopoda) and ethanol 70% (remaining taxa).Juvenile individuals that might not have the identification compared to adult forms were not considered in the species counts and analysis.The vouchers are deposited in the reference collections of the Instituto Butantan (IBSP), Laboratório de Estudos Subterrâneos da Universidade Federal de São Carlos (LES/UFSCar) and Museu de Zoologia da Universidade de São Paulo (MZUSP).
Because the taxonomic impediment for some groups and the needs of robust classification to apply the ecological analysis, we use parataxonomy, grouping similar individuals, based on features of external morphology (sensu Majka and Bondrup-Nielsen, 2006).To avoid cascade errors, we use classical literature (Adis, 2002;Brusca and Brusca, 2003;Borror et al., 1989, Rafael et al., 2012) allied to specialists confirmation.

Data analysis
To verify possible seasonality in the distribution of taxa, graphics were constructed for every collection occasion and a Kruskal-Wallis nonparametric test (Kruskal and Wallis, 1952) followed by Mann-Whitney pairwise comparisons with abundance data was realized to verify significant differences among the seasons/occasions.To analyze the sampling effort and sampling efficiency Mao-tau sample-based rarefaction curves and Jackknife 1 and Chao 2 estimators was applied (Colwell et al., 2004).These es- timators consider, in the samplings, the uniques (Jacknife 1) and uniques and duplicates (Chao 2).Singletons and doubletons were also calculated.
Box-Plots diagrams with richness and abundance per substrate were constructed to verify possible preferences per substrates.A Kruskal-Wallis nonparametric test (Kruskal and Wallis, 1952) followed by Mann-Whitney pairwise comparisons with richness data was realized to verify significant differences among the substrates.Besides, a nonparametric multivariate analysis of variance (NP-MANOVA) using Bray-Curtis distance and 9999 permuts (Anderson, 2001) was applied to verify if the fauna composition differs among substrates.These analyses were performed in Estimates (v 9.1) (Colwell, 2013) and PAST (v 3.07) (Hammer et al., 2001) softwares.
The total observed abundance was 472 individuals.Arachnida and Insecta (Figure 5) were the most abundant groups (Figure 6) totaling 221 and 180 individuals respectively (85% of relative abundance), considering the four sampling occasions.Arachnida was predominant in the first sampling occasion and Insecta in the third occasion (Figure 6).The abundance differed significantly among sampling occasions (Kruskal-Wallis, H=35.7, df = 3, p=8.661 - ) and the Mann-Whitney pairwise comparisons indicated that there was significant difference between the first and the second samplings occasions (p=2.523 - ), the first and the fourth (p=0.005654), the second and the third (p=1.076-7 ), the second and the fourth (p=0.03886) and the third and the fourth (p=0.0008667).

Month
In according to the rarefaction curve, the sampling effort (four occasions) was not enough to access the total richness (Figure 7).The estimators for this study were: Jackknife 1 (322.75)and Chao 2 (621.81).Singletons and Terrestrial fauna of the largest granitic cave from Southern Hemisphere, southeastern Brazil: A neglected habitat doubletons means are also showed in Figure 7, and represent the occurrence of only one (single) or two (double) individuals in all samplings.
The substrates with higher richness values were rocky (RO), followed by rocky and litter (RLI), rocky, unconsolidated and litter (RULI) and unconsolidated (UN); meanwhile substrates formed by litter (LI) showed lower richness values (Figure 8A).Similar results were observed for abundance: rocky substrates were the most abundant (RO), followed by rocky and litter (RLI) and unconsoli-dated (UN); meanwhile litter substrates (LI) showed the lowest abundance values (Figure 8B).
The Kruskal-Wallis test (H= 41.6, df=6, p= 0.0002201) and the Mann-Whitney pairwise comparisons showed a significant difference on fauna richness among the substrates (see Table 2).Besides, according to NPMANOVA analysis there is a significant difference on fauna composition among the substrates (p= 0.019) and the post-hoc tests show that rocky substrate (RO) was the most distinct in relation to the other substrates: RO versus LI (p=0.0301),   3).

Discussion
Some of the taxonomic groups recorded in the Riacho Subterrâneo cave were also recorded in other studies focusing granitic caves (Gnaspini and Trajano, 1994;Bernardi et al., 2012) and are common in Brazilian caves of other lithologies (Pinto-da-Rocha, 1995;Trajano and Bichuette, 2010).The high number of accidental records in all sampling occasions (61), such as Chrysomelidae and Lampyridae coleopterans; Braconidae, Ichneumonidae and Formicidae hymenopterans; Pulmonata gastropods; Bothriuridae scorpions, is related to the granitic cave morphology, which has many contacts with the surface and many routes for access of epigean fauna.In this case, the cave can act as a refuge in the dry season or after any unexpected disturbance in the surface environment (as fire or deforestation).The number of troglophilic species was also high (82) and this result is a pattern of many Brazilian caves (see Trajano and Bichuette, 2010).The troglophilic taxa are similar to that observed for Brazilian cave fauna in general, considering high (e.g., order or family) or low (e.g., genus or species) taxonomic levels (e.g., Endecous sp. and Eidmanacris alboannulatus PIZA 1960 orthopterans; Staphylinidae coleopterans; Drosophilidae, Mycetophilidae and Tipulidae dipterans; Ctenidae, Pholcidae, Sicariidae, Theridiidae and Theridiosomatidae spiders).For the trogloxene category, the three records follow the pattern observed for caves from Rainforest of state of São Paulo, with records of Goniosomatinae opilionids (Acutisoma hamatum (ROEWER 1928) and Mitogoniella sp.) and the ctenid spider Enoploctenus cyclothorax (BERT-KAU 1880).These results emphasize the importance of the cave for maintenance of populations of troglophilic and trogloxene populations, besides the refuge for epigean fauna (Appendix 1).Despite many records of troglomorphic groups (13) it was not possible to affirm their troglobitic status without a wider epigean surveys since these taxa live in microhabitats ecologically similar (e.g., deep soil) to the subterranean habitats and are typical edaphic fauna (Trajano and Bichuette, 2010).
Arachnids and insects were the most representative (in richness and abundance), as observed in studies conducted in caves from different Brazilian regions (Trajano and Bichuette, 2010).The occurrence of six species of Opiliones in a single cave (Gonyleptidae family) is similar to ob- Terrestrial fauna of the largest granitic cave from Southern Hemisphere, southeastern Brazil: A neglected habitat served in other studies for limestone caves in Minas Gerais state (five species, four Gonyleptidae) (Resende and Bichuette, 2016) and in Goiás state (eight species, three Gonyleptidae) (Bichuette et al., 2015).It is noteworthy that in these studies it was necessary at least four sampling occasions to reach that richness of opilionids and other species records, which indicates that the number of samplings is directed related to the diversity.Gallão and Bichuette (2015) show that even after six sampling occasions, species were added in their samplings, which highlights the inefficiency of Brazilian laws that demand only two sampling occasions (Trajano, 2010;Gallão and Bichuette, 2015).Besides, the high richness of opilionids in Riacho Subterrâneo cave likely is related to vegetation surrounding, which is remaining of Atlantic Rainforest.Gnaspini and Trajano (1994) estimated a richness range of 16 to 28 and ca. of seven troglomorphic and possible troglobitic species in four granitic caves inserted in Atlantic Rainforest physiognomy.Souza-Silva et al. (2011) sampled 33 magmatic caves also inserted in Atlantic Rainforest, but in different latitudes; the authors recorded richness from 5 until 81 and ca.11 troglomorphic organisms.Bernardi et al. (2012) recorded a richness of 52 in a single granitic cave from state of Minas Gerais (Atlantic Rainforest), most of them troglophilic and no one troglomorphic.In our study 199 species were recorded (with 13 troglomorphic), increasing the richness for Brazilian caves in igneous rock and shows its importance as a refuge to not-obligatory cave fauna.
All these works did not make replicas in the samples and overlooks the hidden diversity in subterranean environments.It is mandatory a robust sampling effort with standardized replicas to describe the taxonomic diversity and understand the community functioning (Trajano, 2013;Gallão and Bichuette, 2015).This affirmation is corroborated by estimators values observed at Figure 7, allied to the intermediary values of singletons and doubletons, which represent the occurrence of only one (single) or two (double) individuals in all samplings.Gallão and Bichuette (2015) and Bichuette et al. (2015), studying sandstone and limestone caves, respectively, stressed this question and proposed the necessity of replicas on subterranean studies Maria Elina Bichuette, André R. Nascimento, Diego M. von Schimonsky, Jonas E. Gallão, Leonardo P.A. Resende, Tamires Zepon Volume 12 number 2  may -august 2017 to discuss any distribution and diversity pattern, since they are related to intrinsic temporal variations (Trajano, 2013).
The spatial and temporal distribution of cave fauna is influenced by several factors which presents variation in their occurrence, such as patches of vegetable organic matter, guano, concentration of prey (Trajano, 2013), allied to the desiccation of surface habitats in climatic marked regions (in this case, cave habitats operate as seasonal shelters).The seasonality must influences the communities, as observed for the cave fauna of Riacho Subterrâneo cave, where several groups showed higher richness and abundance in the beginning of wet seasons when floods carry organic matter, the trophic basis of detritivores.The lowest richness values observed in the dry season are probably related to the availability of trophic resources, less abundant in this season.Our results are opposite to those observed by Simões (2013) in a single cave from Goiás state (Angélica cave, six replicas): the highest richness and abundance were recorded in the dry seasons and the lowest in the end of rainy season, possibly due the floods influence, carrying and washing the fauna, which reinforce that the cave morphology represents another factor regulating the distribution of the fauna.
Unpredictable events must be discussed herein: the occurrence of fires in 2010 certainly influence the high richness values observed in the first sampling (cave acting as shelter), which again, reinforces the necessity of replicas to detect if the fauna distribution is a pattern (influenced by seasonality for example) or if could be a stochastic process related to an unpredictable event.
Specificity for microhabitats and singularity of fauna show the importance of habitat and the necessity of conservation.In this case, we observed a high specificity for the cave fauna of Riacho Subterrâneo cave (measured by richness and abundance), with an unexpected preference by those showing more rocky components and not those with higher concentration of organic matter.We proposed at least four possible explanations for that: (i) igneous rocks are extremely rough with several crevices, forming humid microhabitats (with mousses in places close to the entrance) for the fauna; (ii) we observed several roots of Philodendron sp.(Figures 2 and 3) close to the sampled rocky substrates, representing a resource for the fauna nearby; (iii) the fact that granitic caves possess several routes (and contacts to the surface), the floods carry food amounts, but also could wash the fauna, at least in rainy season; (iv) the cave morphology, with several rounded blocks hampers the samples in the lowest level of the cave, where the organic matter accumulates, causing a biases in the samples.Again, only the application of replicas in the sampling can reduce these biases.
Riacho Subterrâneo cave is the largest granitic cave in the Southern Hemisphere and shows a high richness (199) with an importance as refuge for epigean fauna.These particularities highlight the importance of this cave.There are few granitic caves recorded in Brazil (ca.216), and certainly most of them are under threat since this kind of rock has a large-scale exploitation.Therefore, this study can contribute to establish comparable parameters for other repeatable studies, mainly those for environmental purposes.Indeed, the replicas showed that some aspects can influence the fauna distribution, as seasonality, which provides parameters that must be applied for cave faunistic studies in general.Finally, besides its role as refuge, it is clear that cave morphology must also be considered as a strong influence for ecology (communities) of cave organisms.

Figure 1 .
Figure 1.Map location of Riacho Subterrâneo granitic cave with a detailed map of the passageways conduits, Itu municipality, southeastern Brazil.

Figure 2 .
Figure 2. Detail of rock blocks and roots in a large gallery of Riacho Subterrâneo granitic cave, Itu municipality, southeastern Brazil.Source: Adriano Gambarini.

Table 1 .
Sampling Occasions with detailed informations of Riacho Subterrâneo cave, Itu municipality, southeastern Brazil.