Diversity of harvestmen ( Arachnida , Opiliones ) in Parque da Onça Parda , southeastern Brazil

The environment most diverse in harvestmen species is the Atlantic Forest of São Paulo. However, there remains a lack of studies regarding their communities in certain regions. Among these regions is one south of the Paranapiacaba mountain range in the state of São Paulo, the Parque da Onça Parda (POP). Through nocturnal collections and pitfall traps, the region’s harvestmen community has been studied. The observed richness of this site included 27 species, with dominance of three species: Holcobunus nigripalpis Roewer, 1910, Neosadocus maximus (Giltay, 1928) and Munequita sp., accounting for 68.4% of harvestmen abundance. This makes the diversity of POP more similar to the semideciduous Atlantic Forest communities of the interior than to those of the Coastal Atlantic Forest that contains the park. Its geographic location places it within the Southern São Paulo State (SSP) area of endemism, along with the Parque Turístico do Alto Ribeira (PETAR), with which it shares up to 12% similarity regarding harvestmen fauna. Richness and abundance of harvestmen were positively related to temperature and humidity. The period of animal activity (as measured by abundance and richness) varied throughout the night, being highest in the early hours during both studied seasons (summer and winter).

Harvestmen comprise the third most diverse order among the class Arachnida (over 6,000 species), being second only to mites and spiders (Shultz & Pinto-da-Rocha, 2007).The group is divided into four suborders: Cyphophthalmi, Dyspnoi, Eupnoi and Laniatores.Having a mainly pantropical distribution, Laniatores comprises the most diverse group with 3,500 species described (Machado et al., 2007).
Harvestmen are omnivores and can feed on other invertebrates, decaying plants and animals, as well as fungi.Yet there is a predominance of predatory habits (BRagagnolo & Pinto-da-Rocha, 2009).They can be found in most terrestrial ecosystems, but their greatest diversity is reported for tropical forests (Pinto-da-Rocha et al., 2005).The environment with greatest diversity of this group is the Atlantic Forest in São Paulo, where 232 species have been recorded (Pinto-da-Rocha, 1999).Apparently this occurs due to their being highly sensitive to environmental conditions, especially humidity (SantoS, 2007).Species richness declines at higher elevations in the tropics due to restrictions imposed by temperature and humidity (alMeida-neto et al., 2006).Moreover, not only the diversity, but also the level of endemism is especially high in tropical forests, where several species present restricted distributions.For example, the Brazilian Atlantic forest harbors 600 of the nearly 6,000 described species of harvestmen in the world, of which around 97.5% are found exclusively in this biome (Pinto-da-Rocha et al., 2005).
Studies of Atlantic harvestmen diversity began in the middle of the last century, when only lists of species were published, in which sampling protocols were not followed while collecting (for example SoaReS 1944a(for example SoaReS ,b,c,1945;;SoaReS & SoaReS, 1970).The standardization of sampling in harvestmen inventories began with Pinto-da-Rocha (1999).Sampling methodologies were best developed and used in subsequent work (BRagagnolo & Pintoda-Rocha, 2003;BReScovit et al., 2004;BRagagnolo et al., 2007).Finally, the work of Pinto-da-Rocha et al.
(2005) provides a detailed overview of the biological and geomorphological processes that formed the Atlantic rainforest itself, leading to the current spatial distribution of harvestmen within this biome.This work further determined areas of endemism of harvestmen and highlighed their high diversity.
The aim of this work is to analyze the harvestmen community in an Atlantic forest remnant (Parque da Onça Parda) and verify its temporal distribution throughout the night.A further aim is to investigate how the richness and abundance of these arachnids are affected by season (summer and winter) and environmental parameters (temperature and relative humidity).

MATERIAL AND METHODS
Studied Area.Parque da Onça Parda (POP) is a Private Reserve of Natural Heritage ["Reserva Particular do Patrimônio Natural" (RPPN)], located in the municipality of São Miguel Arcanjo (24°05'01.89"S,48°00'57.58"W),state of São Paulo.The park area is made up of 50 ha covered with vegetation typical of the Coastal Atlantic Forest.It is part of the Paranapiacaba System, a prominent remnant of this biome with 150,000 ha located in the southern region of the state.However, the area had been subjected to logging during the 1940s and 1950s, and secondary vegetation predominates.POP is drained by the Monjolinho and Ribeirão Bonito Rivers, which belong to the Ribeira River basin.Average annual rainfall of the region is from 1,700 to 2,400 mm, and average temperature is around 17-22°C.According to MonteiRo (1973) the Paulista Coastal Zone covering the park is inserted into a large South Coast climactic unit within the subtropical area, and is subjected to masses of tropical climates and polar fronts.
Sampling.Two nights of sampling were carried out in the summer (January 15-17, 2011), and two in the winter (July 22-24, 2010).During each night of sampling one area of the park was sampled (area 1, shore of Monjolinho stream, 24°04'45"S -48°01'03"W; area 2 trail Onça-Mata Seca: 24°05'07"S -48°00'58"W).Sites were chosen with no physiognomic differences between them.In each area, eight sampling points were selected independent of each other, and with a distance of at least 100 meters.Therefore, each point corresponded to a sampling unit.At each of these points, four transects were established (10 m x 30 m, 10 m apart from each other) and sampled by four collectors.At each transect, harvestmen that were found in litter, leaves and stems to a height of one meter, were collected and fixed in 70% ethanol.Thus, the total number of samples was 16 during each season, each sample being composed of four pseudoreplicates (standardized transects sampled by the same four collectors at all units).All sampling began at 6:00 pm and finished about 3:00 am, so that a temporal distribution analysis of harvestmen could be performed.Hourly temperature and relative humidity were measured with a digital thermohygrometer.In addition to nocturnal manual samplings, pitfall traps were assembled within the sampled areas.In total, 48 pitfall traps (made from PET bottles) were installed randomly in the spaces between transects.They were filled with 70% ethanol and covered by plastic plates on bamboo sticks to prevent entry of leaves and branches.These traps were installed nearly 5 m apart from each other, mounted on the first day of collection, and retrieved after 10 days.
Statistical analysis.Alpha diversity (mean diversity within the community) of POP harvestmen was determined by the Shannon -Wiener index (H'= Σ pi ln pi); whereby pi is the proportion of species in relation to the total number of individuals in the community, and ln is the natural logarithm or neperiane log (≈ 2.72).Conversion of Shannon index values for an effective number of species (value that represents the number of species if the community submits all species with equal abundance) was done using the formula: exp (Σ pi ln pi).This is the same as exp (H'), whereby exp is the natural logarithm raised to ln H' power, and H' is the value of Shannon index (JoSt, 2006).To calculate beta diversity (the change in species composition between places) we used the Jaccard index (Sj = a / (a + b + c)), which produced the dendrogram of similarity among localities based on data of present / absent species (MaguRRan, 1988).To assess the relationship between data (environmental parameters, temporal distribution of abundance, and richness of harvestmen), Multiple Linear Regression was used (relations between temperature, relative humidity, and richness and abundance), as well as linear correlation (temporal distribution of harvestmen -richness and abundance correlated with the nighttime hour).To evaluate the differences in abundance, richness and environmental parameters between seasons, ANOVA (Analysis of Variance) was used.These tests were developed in BioEstat 5.0 (ayReS et al., 2007).

RESULTS AND DISCUSSION
Environmental parameters.In winter the average temperature was 15°C (SD: 1.041), and in summer 22°C (SD: 0.618).Relative humidity in winter showed a mean value of 85.5% (SD: 3.22).In summer, average relative humidity was 82.5% (SD: 2.75).Parameters varied little within the collection periods.However, between winter and summer there was a difference of 7°C in average temperature (F = 29.280,p = 0.005), and a marked variation in relative humidity (F = 2.867 p = 0.051).
Richness.In total, 27 species of harvestmen were identified at POP belonging to two families and nine subfamilies (Tab.I).In total, five species were singletons and two were doubletons.Among Laniatores, only the Gonyleptidae species were collected.This is expected for a locality within the Atlantic Forest in São Paulo, a location in which this group represents 96% of species in the suborder (KuRy, 2003).Within this family, the subfamily Pachylinae was the most important, representing 48% of species.This result represented a dominance also expected for this group in the state of São Paulo, as Pachylinae accounts for 44% of the Gonyleptidae species (KuRy, 2003).Eupnoi Sclerosomatidae (subfamily Gagrellinae) was represented by two species (Holcobunus nigripalpis and Munequita sp.).The manual collecting method was most efficient, obtaining all the observed richness, as well as a higher number of individuals.Pitfall traps did not add any distinct species to those collected manually, and the number of individuals obtained was very low (Tab.I).In summer, Diversity of harvestmen (Arachnida, Opiliones) in Parque da Onça Parda... ReSende et al.
14 individuals representing six species were captured, while in winter only two individuals of two species were captured.Of the 48 traps set in winter, only two collected harvestmen, an efficiency of 4.16%.Of the 48 traps set in summer, 12 collected harvestmen, representing an efficiency of 25%.
Rarefaction curves showed different behavior between the two seasons (Fig. 1).Although both curves did not approach an asymptote, in winter the curve was steeper with no asymptotic trend (with half the number of individuals, 75 percent of richness had already been collected).In summer, the curve was less steep and an asymptotic trend could be observed (with half the number of individuals, more than 90 percent of richness had already been collected).More than twice the number of harvestmen was collected in the summer when compared to winter.Therefore, it is concluded that despite the large increase in abundance, richness of these animals did not increase in similar proportions between the two seasons of the year, and remained close to 20 species.Despite the small difference between total richness collected in the summer and winter, average richness in the two seasons was significantly different (t = 16.85,p = 0.0005) (Fig. 2).
Abundance.During the winter, 536 individuals were collected (including 173 juveniles) belonging to 20 species; in summer 1,231 individuals were collected (including 45 juveniles), representing 23 species.There was a significant difference (t = 9.303, p = 0.005) (Fig. 3) between the means of harvestmen collected in winter and summer.Distribution of harvestmen species abundances showed a notable difference between winter (Fig. 4) and summer (Fig. 5).In winter, the dominant species were Neosadocus maximus and Munequita sp., as both represented 44% of the total individuals collected.In summer, Holcobunus nigripalpis appeared as the dominant species, representing 51.5% of the total harvestmen collected.These three species together constituted 68.4% of all individuals collected during two seasons.According BRagagnolo et al. (2007), the abundance of Munequita sp. and Neosadocus maximus is not affected by quality of habitat nor fragment size of the forest in which they live.There are no published data on this subject for Holcobunus nigripalpis.However,   With a richness of 46 species and a diversity of 19.86, its richness is almost double the richness of POP (27 species), yet diversity is more than three times higher.The diversity of POP is closer to that of semi deciduous Atlantic Forest areas, like Serra do Japi (6.11).According to BRagagnolo et al. (2007), communities of harvestmen are strongly affected by habitat fragmentation.In POP, the exploitation of wood for charcoal production within 50 years (the current vegetation is secondary) is probably the cause of low diversity.The beta diversity, represented by a dendrogram (Fig. 6), puts POP in a group with Cubatão Miracatu, Parque Estadual do Morro Grande, Estação Ecológica Juréia Itatins, Estação Biológica de Boracéia, Estação Biológica Serra de Paranapiacaba, and Parque Estadual Turístico do Alto Ribeira (PETAR).This is based on a similarity of approximately 12%.Similarities between localities concerned mainly the distance between them, given the low dispersal ability of this group.Natural barriers, such as a chain of hills and mountains in the Serra do Mar, also greatly influence the isolation of the areas (Pinto-da-Rocha et al., 2005).ReSende et al.Relation between abundance and richness with environmental parameters.Temperature and humidity are environmental factors that most influence the occurrence and habitat use by harvestmen (cuRtiS & Machado, 2007).In summer, the abundance of harvestmen at POP correlated significantly with temperature (p = 0.009) and relative humidity (p = 0.04) (r 2 = 0.368, p = 0.019), increasing in warmer, moister environments.Richness was not associated with any of the parameters analyzed (r 2 = 0.070, p = 0.230).In winter the abundance of harvestmen was positively related to environmental parameters (r 2 = 0.530 p = 0.003), temperature (p = 0.001), and relative humidity (p = 0.035), as more animals were collected during the greatest heat and humidity.Richness was also related to temperature (p = 0.005) and relative humidity (0.007) (r 2 = 0.490, p = 0.004).Abundance decreased significantly over the period of activity (distribution of Diversity of harvestmen (Arachnida, Opiliones) in Parque da Onça Parda... 1 Recebido em 21 de outubro de 2011.Aceito em 23 de março de 2012.ISSN 0073-4721 Artigo disponível em: www.scielo.br/iszabundance over time) in summer.More animals were collected in the early hours of the evening than late at night (r = 0.649, p = 0.006).Richness in summer was not associated with period of activity (r = 0.304, p = 0.251).In winter abundance and richness was correlated to the period of activity (r = 0.800, p = 0.0001 and r = 0.727, p = 0.001, respectively), wherein fewer individuals and species were captured as nighttime advanced.
Harvestmen respond better to changes in temperature and humidity during the winter.Another important factor is that the richness of these animals is less susceptible to environmental changes than is their abundance.Finally, the activity of these animals (as measured by abundance and distribution of richness over time) seems to be very different throughout the night, both being higher in early nighttime hours during both seasons (except richness during summer).
Tab. I. Summary of harvestmen collected with two methods in two seasons at the Parque da Onça Parda São Miguel Arcanjo, state of São Paulo, Southeastern Brazil.
Tab. II.Comparison among alfa diversity in several quantitatively inventoried localities of Atlantic Rain Forest and one Amazonian (E.N.S*, Effective Number of Species; sources: BRagagnolo et al., 2007 and Pinto-da-Rocha & Bonaldo, 2006) (RJ, state of Rio de Janeiro; SP, state of São Paulo; PA, state of Pará).