Impact of disturbed areas on Theraphosidae spiders diversity (Araneae) and first population data of Grammostola rosea (Walckenaer) in Panul Park

Abstract Soil fauna constitutes one of the most abundant and richest environments on earth (Coleman et al. 2004, Fundamentals of solil ecology, 2nd ed. Elsevier Academic Press, London, UK). Different degrees of soil disturbance can affect arthropod diversity, which allows a correlation of biodiversity to quality of habitat. The present study aimed to evaluate the impact of habitat on Theraphosidae spiders, with special focus on Grammostola rosea. Slight differences in the diversity of Theraphosidae between the disturbed area of Cerro Huechuraba and the undisturbed Panul Park were found. However, a high dominance of G. rosea was observed in both study areas. G. rosea density 1,350 ind/ha in Panul Park, and 750 ind/ha in Cerro Huechuraba. UPGMA cluster analysis did not show significant differences between established environments. A standard methodology to develop inventories of Theraphosidae was proposed. The distribution of G. rosea and its natural history were reported.

Among the Theraphosidae present in Chile, Grammostola rosea (Walckenaer, 1837) is the most frequently mentioned in informal reports, but no data about its population distribution are given.
Moreover, historical distributions may no longer be useful because humans in their imperative need for expansion have notoriously modified the environment. Consequently, land use has shifted to urban use along with alterations of vegetation and geography (Argañaraz & Gleiser, 2017), thereby producing local changes in arachnid diversity. In addition, some historical distributional data may be wrong, as is the case with G. rosea, which was cited for Mexico, a labeling error by Pickard-Cambridge (1897) detected by Pocock (1903).
Another important factor is that wild specimens of this species are under great capture pressure to be sold abroad as pets, which could eventually cause a drastic decline in populations and possibly local extinctions (Aguilera, 2015;ODEPA, 2017).
The present study aim was to evaluate habitat impacts on communal and distributional parameters of Grammostola rosea, in two areas with different levels of anthropic disturbance and land use, and especially in an ecological preservation area. Furthermore, data related to habitat and natural history of the species, proposed methodologies to correctly evaluate Grammostola populations in environmental baseline are given.

| Study area
Forest of Panul Park, with an approximate sampling area of 13.5 ha (Figure 1a) in the eastern sector of La Florida (357,928 E, 6,288,800 S). Located in the foothills of one of the closest mountain ranges of Santiago, known as the Sierra de Ramón, where the highest elevations exceed 3,000 m above sea level. Recently, this sector has had an economic interest for real estate development, and therefore, some environmental studies have been performed in the area (Gesterra, 2011), but none about arachnids. Due to the importance of the area and the imminent urbanization, the Corporación Municipal de Fomento al Desarrollo Comunal y Productivo of La Florida (COFODEP, 2015) suggests Panul Park as an "environmental ecological preservation area of the foothills, municipality of La Florida." Panul Park is known as the last native sclerophyllous forest in Santiago, alternating secondary sclerophyllous scrubs (Mann, 1964; The second study area was located in Cerro Huechuraba (343,367 E, 6,308,994 S)

| Sampling
Collection was performed during the spring of November 2016, using stratified random sampling with eleven transects of 200 m length by 2 m width, with a pedestrian path at low speed for two specialists in each area for a minimum of 30 min by transect. To confirm inhabitation of burrows, a flexible endoscope with a video microcamera with a lighting source connected by fiber optic to a screen, which allowed insertion of the fiber optic cable inside the burrow for inspection, was used. A manual collection was performed (Sorensen, Coddington, & Schaft, 2002;Upton, 1991), and the individuals were stored in 95% ethanol prior to labeling, and later de-
Artificial burrows attached to glass were built for observation.
All containers had a substrate of soil, Sphagnum sp. and water provision. They were fed once every 2 weeks mainly with cockroaches (Blatta sp.), crickets (Gryllidae), Tenebrio sp., and Zophobas sp. Temperature in laboratory varied between 16°C and 34°C.
Behaviors were studied by direct observations and registered by notes and photographs.

| Taxonomy
For identification of theraphosids, the morphology of the tibial processes by Montes de Oca, D'Elía, and Pérez-Miles (2016) and Montenegro, Aguilera, and Casanueva (2018), the palpal bulb structure by Bertani (2000), the spermatheca morphology by Schiapelli and Gerschman (1961), and the urticating setae by Cooke, Roth, and Miller (1972) were used. Identifying images were obtained with a stereoscopic magnifier Motic SMZ-140 and ZeissStemi SR with its supplement for a Nikon Coolpix P600 camera.

| Statistics
In order to quantify Grammostola rosea and the other Theraphosidae in the study sites, a diversity analysis was performed using the computer program PAST 3.16 (Hammer, Harper, & Ryan, 2001) to determine S richness, H′ diversity according to Shannon-Weaver (Shannon & Weaver, 1949), and J′ uniformity according to Simpson (1949). The Shannon-Weaver and Simpson indices were used because they allow comparisons of Araneofauna in terms of richness proportional to abundance, and evenness of abundance among the species in the two study areas. Moreover, these indices are frequently used in the literature (Jaksic, 2001) and therefore facilitate comparisons with similar studies (Aguilera, Casanueva, & Hernández, 2006). In order to compare alpha biodiversity estimators between both study sites, a comparison test was performed with 1,000 iterations in the PAST program, based on the bootstrapping technique (Manly, 1997). To evaluate the degree of similarity of tarantulas between the two study areas, the Jaccard similarity coefficient was used based on presence/absence taxa records. In order to evaluate the existence of groupings according to localities and/or vegetation formation, a cluster analysis was performed using Jaccard similarity values. For this conglomerate, the abundance data of each transect were analyzed and the unweighted pair-group method using arithmetic averages (UPGMA) was used (Sokal & Rohlf, 1995). Species accumulation curve was calculated (Collwel & Coddington, 1994;Jimenez-Valverde & Hortal, 2003) to evaluate the information collected in the field and also to determine the efficiency of the characterization of the influence area. Additionally, density of species per m 2 was also estimated for each study area. In relation to the communal parameters (Table 2), it can be observed that Panul Park had a greater H′ diversity (H = 0.178) compared to Huechuraba (H = 0.146), but without significant statistical differences, while the J' index indicates that the abundances are similarly distributed in both study areas, but since the J′ values are considerably low indicate a dominance given by G. rosea. On the other hand, G. rosea density values are greater in Panul Park than in Huechuraba. The cluster analysis (UPGMA + Jaccard) evaluated Theraphosidae similarity between the two study sites and did not show significant groupings according to environments, or they would be expected to occur by chance (Figure 2a).

| RE SULTS
In the species accumulation analysis, it was observed that the asymptotic number of species was not reached in any of the two study areas, so objective criteria cannot be applied to determine whether the species inventory was sufficiently complete (Figure 2b).
It has been observed that G. rosea has various color patterns, frequently with reddish to dark brown tones in the setae, cephalothorax, and abdomen (Figure 1a,b). Additionally, the characters of the spermatheca of the female, bulb, and tibial apophysis of the male for G. rosea are given in Figure 1c- In particular in the areas of study, high burrow density was observed for some of the sectors sampled, with distances of no more than 1 m between burrows. In Panul Park, burrows were found among different vegetational formations, both in sclerophyllous forests and in environments devoid of vegetation or with annual isolated individual poaceae and A. caven, while in Cerro Huechuraba burrows were mainly found in vegetational formations of forest and scrub. This showed high similarity between the environments; therefore, these would behave as habitat units structurally homogeneous for the spiders, and G. rosea would not be restricted by the types of environments present (forest/scrub and cover of poaceas).

| D ISCUSS I ON
Despite this, in Cerro Huechuraba, the relative abundances were considerably lower in wild allochthonous environment than in the forest/scrub and cover of poaceas environments. Furthermore, density followed the same pattern indicated for relative abundance and was estimated at 1,350 ind/ha for Panul Park, and 750 ind/ha for Cerro Huechuraba.
As for the obtained species accumulation curve, it was not possible to reach asymptotic numbers of species. In this context, there are no objective criteria that determine when the Theraphosidae inventory in both study areas is sufficiently complete. As reported by Jimenez-Valverde and Hortal (2003), the fact that it was not possible to reach asymptotic numbers could be due to temporal or spatial biases in the sampling effort distribution that could affect the curve shape. In the present study, collection was performed during spring; therefore, temporal bias was considered as spring and to a lesser extent summer is the ideal sampling times for Theraphosidae. This is on account of its seasonal life cycles, and these times correspond with a higher activity cycle, including reproductive and growing seasons (Aguilera, 2015). During these seasons, adult males and females are more easily found, along with younger populations (personal observations), therefore facilitating proper estimation of Theraphosidae population densities. Additionally, in order to avoid generating spatial bias of the sampling effort distribution a stratified random technique was employed, with a greater sampling effort in those strata that presented conditions of specific environments or microhabitats suitable for the presence of Theraphosidae and also to maximize the probability to inventory all species. It is important to mention that it has been observed that local distribution of Theraphosidae population can be fragmented in other areas of the country and that even though the sampling sites are adequated, these spiders may not be registered. For this reason, it is recommended to perform samplings in more extensive areas and thus increase the possibility of registering these fragment distributions.
Although possible errors associated with the species accumulation curve were considered, an initial exponential growth was not possible, neither an asymptotic number of species. This fact has been observed in other studies of spider diversity, in which the quality of spider inventory is evaluated and subsequently it is observed that it is not possible to register all species, causing accumulation curves to vary greatly from the asymptote (Brennan, Majer, & Reygaert, 1999;Coddington, Young, & Coyle, 1996;Edwards, 1993;Jimenez-Valverde & Hortal, 2003;Samu & Lövei, 1995;Sorensen et al., 2002;Toti, Coyle, & Miller, 2000).  (Aguilera, 2015;Legendre & Calderón, 1984;Pocock, 1903). This work establishes the northern limit of the 28.5°S Atacama Region (Vallenar) and the southern limit of 37.8°S of the Biobio Region (Angol) for the studied species. This wide distribution of G. rosea facilitates evaluating the state of the different populations throughout the country, with studies of the genetic structure of each population, to analyze the distribution of interand intranatural population variability. Additionally, future studies could examine the processes involved, to later infer regarding relationships among different entities, as well as deterministic evolutionary mechanisms (natural selection, migration and mutation) and stochastic (genetic drift), as well as demographic processes (changes in population size) involved in the maintenance of variability in a certain group of organisms. This context could facilitate interpretation of the adaptive strategies of these spiders or be used to implement management units in biodiversity conservation plans (Remis, 2011), and also analyze phylogenetic aspects to better understand the presence of this genus in Chile.
The methodology proposed here could be used in future records of Theraphosidae and especially for environmental impact studies and environmental baselines.

ACK N OWLED G M ENTS
Publication funded by VRID No. 214.113.088-1.0 University of Concepción and Aracno Inc. SpA.

CO N FLI C T O F I NTE R E S T
None declared.

AUTH O R CO NTR I B UTI O N S
MAA and RM conceived the ideas and did the field trip and collect data; MAA and MEC designed methodology and analyzed the data.
All authors wrote the manuscript and contributed critically to the drafts and gave final approval for publication.