Comparison of the species composition of Gamasina mite communities (Acari: Mesostigmata) in selected caves of the Kraków-Częstochowa Upland (southern Poland) and their immediate surroundings

This paper presents the results of a study performed in eight caves located in the area of the Kraków-Częstochowa Upland. In total, 586 samples (400 cm3 each) were taken from the caves and sites located up to one metre from the main entrances of the caves. Of 109 species belonging to 13 families that were identified, four were species new for Polish fauna. In this study, cluster analysis and correspondence analysis were used to compare the species composition of mite communities living in the caves and their immediate surroundings. We also examined the possibility of isolating troglophilic species based on IndVal analysis. Those results showed distinct differences between Gamasina mites living in two distinct habitats. We also established indicator species for the examined caves.


Introduction
Research on caves is an important element of wider research on the factors contributing to species diversity. Caves are unique yet weakly researched habitats, in particular concerning their fauna, which is dominated by arthropods, including mites. Acaralogical research in caves has been scarce as a result of taxonomic and methodological difficulties associated with sampling (Ducarme et al. 2003). Mite fauna in European caves has been researched in several studies: Leruth (1939), Lundqvist et al. (1999), Fend'a andKošel (2000), Ducarme et al. (2003), , Mašán and Madej (2011) and Skubała et al. (2013).
The aim of this research was to determine the species composition of Gamasina mites inhabiting selected caves and determine if the mites had formed distinct communities and how they differed from communities outside the caves. We also attempted to establish troglophilic species.
IndVal is the result of multiplying the relative abundance (specificity) and relative frequency (fidelity) by 100. The index has a maximum value of 100% when all the individuals of a given species can be found in one group of samples and where the species is present in all samples of this group. IndVal was calculated for the two groups of sampling sites (cave and cave surroundings).
In order to determine the main levels of divisions into groups, a hierarchical classification based on Ward (1963) was used. IndVal was calculated using INDVAL 2.0 software. Randomization was performed for 1000 replicates. Multivariate analysis was carried out to explore relationships between the different habitats and the mite communities inhabiting them. The ordination algorithm used was correspondence analysis (CA) using the CANOCO community ordination program (TerBraak and Šmilauer 2002). The analysis took into account species with more than 100 specimens. The abundance of Gamasina mites was transformed (log[x + 1]) for the ordination analyses.

Results
At the examined sites, 109 mite species were identified, from 13 families ( Table 2). In the taxonomic spectrum the most represented families were: Ascidae, Ameroseiidae, Veigaiidae and Ologamaside in the caves, and Zerconidae and Veigaiidae in sites located up to 1 m from the main entrances of the caves. Cluster analysis showed distinct differences in species composition between the communities of mites inhabiting caves and those living outside caves (Figure 1), which was confirmed by the results of CA. The first canonical axis greatly diversified the sites and explained 16.5% of the variability of mite assemblages, dividing them into two groups: those inhabiting caves and those outside. The second canonical axis showed internal differences within the groups. The distribution of species      in the ordination space ( Figure 2) showed on the right side of the axis species found only in caves, and on the left side those only in external sites. This distribution of species indicates a strong environmental gradient represented by the first canonical axis, which has a significant impact on the observed differences in the species composition between the communities in the examined habitats. The analysis of data made it possible to distinguish indicator species for the two groups of sites, established on the basis of Ward's method (Figure 1). In group 1, the highest IndVal was observed for two species: Proctolaelaps pygmaeus (96.28%) and Parasitus loricatus (85.71%), with a statistically confirmed significance (Table 3). These results confirm the results of previous analyses, indicating significant differences in species composition of mite communities found in the studied caves and their immediate surroundings.

Discussion
The spatial variability of the land area covered in this study is key for the understanding of the structure of local mite communities. The presence of caves has a significant effect on the biodiversity, but only detailed research may explain how they influence soil mesofauna and whether it was possible to determine species specific to these habitats. The species found in the examined spaces did not reveal morphological changes characteristic for arthropods of deep caves (Howarth 1983). Among the species recorded in the Polish underground habitats, we recorded species well known from caves in other countries, some considered as cavernicolous or even troglobite. The indicator species of the examined caves was Proctolaelaps pygmaeus. This species has also been discovered in caves in Slovakia (Fend'a and Košel 2000) and Australia (Halliday 2001), and in old mines (Solarz et al. 2002). A high IndVal (96.28%) indicates that in the studied area this species preferred cave  Table 1).
habitats. Despite this, it needs to be remembered that this is a ubiquitous species, present in various types of habitats and substrates.
Comparative analysis of mite communities in the caves and their immediate surroundings showed differences in the species composition between these sites. A greater diversity of mites occurred near the caves, similar to a study by Culver and Sket (2000). Classification analysis and canonical analysis (CA) indicated large differences in the species composition between the mite communities living in caves, even though some of the caves were connected or existed in the same rock mass. Cave communities showed greater specificity than communities outside caves.
The use of IndVal helped to distinguish indicator species for the examined sites. The IndVal method has been used before in the analysis of Mesostigmata mites communities, by  and Salmane and Brumelis (2008). Based on the obtained results one cannot unambiguously indicate troglophilic species. In fact, IndVal indicated those species with a high level of ecological plasticity and adaptability that were well tolerant to the specific abiotic factors in the caves. Species found in the studied caves have often been found in non-subterranean habitats. The exception is Parasitus loricatus, the only species which could be classified as troglophilic. That is why IndVal may be more helpful in the determination of ecological preferences of a given species but does not give ground for global conclusions.
Analysis of available literature and obtained results also showed difficulties with the classification of Gamasina mites as troglobionts or troglophiles. Many mite species originally described as troglobionts or troglophilic species were later found outside the cave environment. The reason may be the still insufficient body of acarological research in caves. Especially in Poland, mite fauna inhabiting these specific habitats has been weakly researched. However, these habitats are essential for understanding the ecology of mites.
Journal of Natural History 1685