4265275608

The highest value of Shannon index has the site 2 m 2009 (H'=2.83; H_max=3.26 and e=0.87), whereas the lowest has the site 1 in 2007 (H'=1.82; H_max= 2.83 and e=0.64). From aspect of occurrence of rare and endangered species, relatively high value of diversity (H'= 2.14 to 2.83) and equitability (e=0.73 to 0.86) we can consider the site 2 as the most valuable ecologically.

The matrix datę of species abundance was subjected to factor analyses of main components for determination of correlation value of explained (response) variables (species) depended on environmental factors. The normalize varimax rotation was used for reduction of factors. On the base of factor analyses we identified two factors with higher value than 1, from which the factor 1 explains 66.34% and second factor 18.68% total variance. The data of hoverflies were strongly correlated with the environmental factors. The significant factor loads (>0.7) in factor 1 revealed in 2007 in the site 1-3; in 2009 in the site 1 and 3. The factor 2 was significant in 2008 in the site 2 and 3; and in 2009 in site 2. But from our data matrix the factors are unable to identify. It is important that we confirmed the existence of factors. The next research will have to be focused on determination of important environmental variables that influence hoverfly coenoses.

The species structure of hoverfly coenoses, from particular sites, was analysed using the data matrix of abundance of adults. Using detrended correspondence analysis (DCA) and log transformation of species data we ascertained the highest value of lengths of gradient = 1.72. The model explains only 1.22% of variability of species data. The cumulative percentage of variability of species data explained by l^sl ordinary axes was 25.4% and by the 4^th ordinary axes up to 74.9% which is with concordance with factor analyses. The ordinary graph - biplot (Fig. 3) indicates ecological species structure and their sites in accordance with l^sl ordinary axes (the most significant environmental gradient) as well as with the 2^nd ordinary axes both explaining 44.9% variability of species data. Among 59 species only 44 are documented in ordinary graph (species fit rangę = 10%). Because an indirect gradient analysis was used, ecological gradients expressed by l^st and 2^nd ordinary axes are determined merely according ecological characteristic of hoverflies. Assuming that the l^sl ordinary axes expresses humidity gradient. Ecological characteristic of hoverflies according SPEIGHT (2010) for example E. arbustorum, A lunulata, L. tarsata indicate binding to wetlands.

Species in negative part of ordinary axis, for example E. balteatus, E. tenax etc., prefer sunny places. The second ordinary axis can express in positive part tree and bush bank vegetation (Pipiza fenestrata Meigen, Pipizella viduata (LlNNAEUS), Cheilosia urbana (MEIGEN), P. perpallidus, Scaeva pyrastri (LlNNAEUS) etc.) whereas P. ftilviventris, N. interrupta, E. sepulchralis, Platycheirus occultus GOELDLIN, MAIBACH & SPEIGHT etc. prefer macrophyte vegetation.

-18-