The RAMSAR Convention is the intergovernmental treaty that offers the framework for the conservation and rational use of wetlands and their resources. The term wetland is recent in the world of science and conservation (Canevari et al., 1999). In general, wetlands are intermediate systems between permanently flooded environments and normally dry environments (Finlayson & Moser, 1991). Currently, the RAMSAR List is the most extensive network of the world's protected areas, with 2341 sites of international importance, 23 of these are in Argentina (RAMSAR, 2019).

The RAMSAR Chaco wetlands site encompasses the eastern strip of the province of Chaco (Fig. 1). It is one of the three biomes with the highest biological diversity in Argentina due to its very heterogeneous environment with various units, such as the forest in gallery, lowland flood forest, palm savanna, grasslands, flooded areas with cattle raising and areas with agriculture (Alberto, 2006). In fact, despite being a priority area, cattle raising and agriculture activities are carried out. These activities affect large natural areas such as savannas, grasslands, ravines, marshes and “albardones” (Ginzburg & Adámoli, 2006) and consequently, they affect the flora and fauna diversity of the region. In fact, this is a global problem, the expansion of agriculture in the world has led to the loss and fragmentation of forests, which has resulted in an impoverishment of biodiversity (Barrera et al., 2015).

This problem is further compounded by the Territorial Ordering law of native forests presented by the Executive Branch of the province of Chaco in 2008 (Law: 6409) which awards this site category II allowing modifications to the forests. In this context, studies of spider diversity in native environments or in natural areas become important for conservation actions, as well as in fragmented or remnant forest sites due to human activities such as agriculture, cattle raising and urbanization processes (Tigas et al., 2002). The fragmentation of natural areas leads to a homogenization of the environment; this is associated with a decrease in the number of spider species that can coexist and in the diversity of trophic guilds (Avalos et al., 2009). In addition, as Pinkus-Rendón et al. (2006) point out, knowing how the landscape composition of a region affects the distribution and diversity of spiders can be used to characterize this ecosystem.

In this regard, studies on the diversity of spiders in the RAMSAR Chaco wetlands site are still incipient (Escobar et al., 2012; Achitte-Schmutzler et al., 2016), and none of these focused on Thomisidae Sundevall, 1833.

The structure of the vegetation is an important factor that influences the diversity of spiders (Jiménez-Valverde & Lobo, 2006). Additionally, it is postulated that the more diverse or complex the plant organization of an area, the more species of spiders can be found and in greater abundance, due to foliage biomass and prey availability (Hore & Uniyal, 2008). In addition, the habitat structure, mainly due to the complexity of the vegetation, affects the specific interactions of spiders. This influences the presence of species, richness and the composition of communities and reduces the mortality of these communities by providing refuges or influencing interactions with other guilds (Jiménez-Valverde & Lobo, 2007; Malumbres-Olarte et al., 2013). In this way, the variation in vegetation architecture between habitats could lead to different sets of spiders throughout the landscape (Gómez et al., 2016). Thus, forests with greater complexity tend to house greater diversity of spiders as well as many unique species (Simó et al., 2011).

The spiders of the families Araneidae Clerck, 1757 and Thomisidae have species with broad environmental tolerance but with a great dependence on the physical structure of the environment (Jiménez-Valverde & Lobo, 2007). Indeed, several species of Thomisidae preferably inhabit foliage and they are commonly abundant in natural areas (Podgaiski et al., 2007; Ricetti & Bonaldo, 2008; Rubio et al., 2008). Furthermore, these spiders build refuges of silk threads where they move, oviposit, reproduce and remain for long periods of inactivity (Foelix, 1982) which could imply some dependence on the substratum where they live. Tomisids are wandering spiders, that do not build webs to trap prey, and live almost exclusively on branches, leaves or flowers in which they hide thanks to their mimetic colors (Mello-Leitão, 1929). They are popularly called “crab spiders” because they walk sideways like the crabs, this similarity is accentuated by the stalking posture that they adopt when waiting for their preys (Grismado, 2007). For these characters, the family is placed within the trophic guild of ambush hunters (Cardoso et al., 2011).

Currently there are more than two thousand species of tomisids distributed worldwide and only 40 of them were reported for Argentina (WSC, 2019). However, spider diversity studies in Argentina report low tomisids richness (Avalos et al., 2007, 2009, 2013; Rubio et al., 2008; Achitte-Schmutzler et al., 2016; Nadal et al., 2018). This could be due the fact that these works focus on spiders in general and not on a particular family or that actually these are not very diverse in a certain area. As Grismado et al. (2014) point out, only Araneidae and Salticidae Blackwall, 1841 are well studied, while Thomisidae, the seventh spider family in terms of species richness is poorly researched. In this study we characterized and compared the assemblages of tomisids, in terms of structure, richness and abundance, in the gallery forest environments, low open forest, grassland and palm grove of the RAMSAR Chaco wetlands site. Besides, we evaluated the change of species between these environments with different degrees of structural complexity.

We collected 288 adult individuals from the total of specimens collected (n = 464), distributed in 34 species / morphospecies of Thomisidae spiders. In this work, Titidius albifrons (Mello-Leitão, 1929) and Epicaudus camelinus (Pickard-Cambridge, 1869) are registered for the first time for Argentina (WSC, 2019).

The species exclusive of gallery forest (GF) were: Epicadus trituberculatus (Taczanowski, 1872), E. camelinus, Tmarus sp 6, T. sp 9, T. sp 11, T. sp 13, T. sp 16, T. sp 18, Synaemops pugilator Mello-Leitäo, 1941, Runcinioides aff sp Mello-Leitão, 1929; the species exclusive of open forest (LOP) were: Thomisus aff. sp (Walckenaer, 1805), Tmarus sp 7 and T. sp 14; and the species exclusive of palmar groves (PalmGrv) were: Tmarus sp 17, Misumenoides sp (Pickard-Cambridge, 1900) and M. sp2. Grassland (Grassl) did not have any exclusive species.

The dominant species of both forested environments was Tmarus pugnax (Mello-Leitäo, 1929); the dominant species of Grassland was T. aff. humphreyi (Chickering, 1965) and dominant species of the Palm groves were Misumenops maculissparsus (Keyserling, 1891) and Uraarachne sp (Keyserling, 1880) with equal abundance. In general, in the GF, Grassl and Palm Grv the most abundant species correspond to different genera, while in the LOF the Tmarus species dominate (see Appendix).

In general, the completeness of the sampling in the environments exceeded 90% and more than 60% of the expected species were obtained according to the Chao 1 estimator (Table 2, data in brackets). The greatest richness and abundance was obtained in the gallery forest followed by the low open forest (Table 2). However Kruskal-Wallis test results showed that there was no significant difference between the GF-LOF (H = 1.85, p > 0.05) and Grassl-Palm Grv (H = 0.33, p > 0.05). The only significant difference was between the forested and non-forested environments (p < 0.05).

The rarefaction/extrapolation curves for the three effective numbers of species showed that the specific richness in GF is significantly higher compared to the other environments. However, with extrapolation, the differences with LOF are not significant. On the other hand, the richness of tomisids in the Grassl and Palm Grv were significantly different of both forests (Fig. 2). The GF was also the most diverse (q = 1) and equitable (q = 2) compared to the other environments, but without statistical significance. On the contrary, grassl had the lower richness, diversity and (in part) equitativity compared with other environments, and this was significant (Fig. 2).

The ordination analysis NMDS showed an evident separation between forests (GF and LOF) and Grassl + Palm Grv, with a stress of 0.28 (Fig. 3); and the ANOSIM indicated significant differences between groups (R = 0.38, p = 0.02).

Likewise, the Chao-Jaccard index indicated a clear separation between the composition of tomisids with greater similarity between forest environments (J-C = 0.85) and between the Grassl and Palm Grv (J-C = 0.70) (Table 3).

This is the first study of the diversity of Thomisidae in the northeast of Argentina and it reveals a high specific richness, since the number of species/morphospecies reported represents almost 70% of the species registered for Argentina (49 spp.) according to the Catálogo de Arañas de Argentina (CAA, 2019).

In most of the articles about spider diversity in natural environments and crops in the northeast of Argentina, the richness of Thomisidae is generally represented by a dozen or less species, although with a remarkable abundance (Avalos et al. 2007, 2009, 2013; Rubio et al., 2008; Achitte-Schmutzler et al., 2016; Nadal et al., 2018). Although the same capture techniques were used in these works and most of the samples were seasonal, the tomisids were poorly represented compared to this research.

It should be noted that only seven species of the total collected in this work are mentioned in the CAA (2019), although several Misumenops and Tmarus have not been identified at a specific level. Tmarus is represented by numerous species globally distributed (WSC, 2019); they are cryptic foliage spiders and are found in certain tree species, many of them are Mirtáceas or trees with dry branches (personal observation). Therefore, foliage beating was the most effective technique to collect them, as reported by Jiménez-Valverde & Lobo (2006). Only two species were found in non-forested environments: T. aff. humphreyi, which was dominant in grassland, and Tmarus sp17 found only in the palm grove. T. aff. humphreyi was also reported in grasslands of the Espinal Region by Nadal et al., (2018), although predominantly in cold seasons of the year.

Tmarus pugnax, which was dominant in both forests, was also dominant in the forests of a reserve in the Espinal region (Nadal et al., 2018). Likewise, species of Tmarus were dominant in forests of Chaco Húmedo (Avalos et al., 2007; Achitte-Schmutzler et al., 2016) and in the Pampeana region (Grismado et al., 2011). However, comparisons at a specific level are impossible since in most spider diversity studies tomisids are reported as morphospecies (Avalos et al,. 2007, 2009; Schwerdt et al., 2014; Rubio, 2015). The lack of reviews at the family and genus level is associated with taxonomic problems (Grismado, 2007), and makes it difficult to compare tomisid species assemblages in different regions. This problem extends to several countries in South America (Bizuet-Flores et al., 2015; Castanheira et al., 2016).

The species of Misumenops that dominated in the palm grove and in the grassland seem to prefer these types of environments since their abundance decreased in the forests. This was also verified in other studies in which they were also not very abundant or were absent in forested areas (Grismado, 2007; Bonaldo & Dias, 2010; Ibarra-Nuñez et al., 2011). In contrast, they were abundant in open environments and/or with herbaceous strata, as in grasslands (Nadal et al., 2018; Schwerdt et al., 2014), in various crops (Armendano & González, 2010; Almada et al., 2012) and in urbanized areas (Argañaraz & Gleiser, 2017). Moreover, Rubio (2015) also reported a greater abundance of Misumenops species in the Yungas in transition, an area characterized by a simpler and less diverse plant structure (Brown et al., 2002).

The richness and diversity of tomisids in forested environments were notorious, as well as the presence of numerous rare species in such environments when compared with grasslands and palm groves. Halffter & Moreno (2005) affirm that the variation in the alpha diversity of the communities is related to local factors and to the interactions between the populations. Nevertheless, the structural complexity of the habitat is also an important factor that influences spider diversity (Avalos et al., 2009), probably because complex habitats increase the availability of niches (Jiménez-Valverde et al., 2010). So, according to the results of this work, forested environments seem to have favorable ecological conditions for tomisid assemblages.

The specific richness of a site is influenced not only by high environmental heterogeneity, but also by other elements such as the presence of rare species, tourist species, demographic phenomena and the area occupied by the study community (Halffter & Moreno, 2005). In this work, the high proportion of tomisids that appear to be missing in both forests could partly be explained by the numerous species that were represented by few individuals, which are generally difficult to collect in environments of greater structural complexity (Almada & Sarquis, 2017).

However, the high proportion of rare species can also be attributed to edge effects, to species that are collected outside their mating season, to species that are difficult to collect with the methods used or due to biases in the procedures during collection (Cardoso et al., 2008; Rubio et al., 2008). Rare species are very important in terms of conservation (Elphick, 1997) because they are more prone to extinction due to the loss and/or fragmentation of their habitat by mismanagement and various other human activities (Isik, 2010; Mouillot et al., 2013). In this regard, in this work several species of Tmarus were found only once in the entire sampling, possible because they are more sensitive to change in the environment.

In this study, the gallery forest presented the most richness of tomisids with species collected only in this environment, such as several of the genus Tmarus and Epicadus. This type of environment showed greater spider diversity in the studies conducted by Simó et al. (2011) and those carried out with other groups of organisms by Sánchez et al. (2004). These authors emphasize the importance of these types of environments for biodiversity conservation by providing food and refuge to the species.

According to Pearce et al. (2004), the composition of spiders tends to change as long as a plant succession exists which provides microhabitats and availability of specific prey, thus certain species are distributed in particular areas of vegetation (Jiménez, 1996). In this regard, exclusive species were found in each environmental unit except in the grassland. It is interesting to note that in the open forests the species that are at the top with the highest range of abundance are those of the genus Tmarus. This genus was also representative in several forests in the northeast of Argentina (Bar et al., 2008; Rubio et al., 2008; Rubio, 2015).

The lower diversity in the grassl compared with other environments suggests that environments with less complexity are only favorable for a few species which significantly increase their abundance in detriment of others. Moreover, most of this grassland is used for cattle raising and this could be affecting the diversity of Thomisidae; according to Almada (2014), cattle raising could have negative effects on spider fauna.

Probably, the presence of shrubs and low trees in the palm grove has some influence on the composition of the tomisids assemblage. In this sense, Weeks & Holtzer (2000) reported significant differences between the grassland and the mixed grassland (with bushes), which presented the greatest abundance of Thomisidae. Additionally, the abundance of tomisids reported in the understory of low tropical forests, in edge habitats and in more open areas (Halaj et al., 2000; Álvares et al., 2004; Sørensen, 2004), suggests that shrubs along with other low-bearing plant species, such as those present in sampled palm groves, are conducive to the establishment of several species of tomisids. Although palm trees were not sampled in this study, in another study conducted in Brazil tomisids were reported to be rare on them (Battirola et al., 2004).

Taking into account that the composition of spiders is strongly influenced by the type of habitat (Weeks & Holtzer, 2000; Jiménez-Valverde & Lobo, 2007) and that habitat heterogeneity could involve strong differences in the composition of spider assemblages (Cabra-García et al,. 2010), this work shows how tomisids assemblages differ according to the different environmental units. The results show that the change in plant structure and composition in each environment reflect a change in the structure and diversity of the Thomisidae, with unique species and others that vary in terms of abundance in each environment. Furthermore, forested and non-forested environments showed low similarity and in consequence a high species turnover.

Considering that the western strip of the Province of Chaco was declared of interest for conservation and that the increase in the degradation of natural habitats has strengthened the need to know and evaluate biodiversity patterns, this work becomes important in assessing the tomisids assemblages in this site.