The results showed no effect on diversity of species across the isolation gradient. We even found that the site with highest diversity was an isolated area. However, the number of species found does not tell about the species relevance, some of the species found were forest species or arable field species and therefore not strictly tied to semi-natural grassland. Gaublomme et al. (2008) highlighted that specialists in general are sensitive to changes in habitat and to simply look at diversity is misleading because it favours widespread species that can inhabit multiple habitats. The analysis for the habitat preference showed that grassland species were significantly different between the areas. In the correspondence analysis (CA). The grassland species seems to favour the isolated and intermediate areas, while carabids that favours fields and arable land are more associated to the connected group. A study on butterflies show that the surrounding habitat (the matrix) influenced species composition, it showed that arable land had a stronger negative effect on species diversity than forest (Öckinger et al. 2012). The connected areas in this study was all found in agricultural dominated landscapes while the more isolated patches were found in more mixed areas. If then arable fields have a larger negative effect on species diversity, then that might have evened out the diversity between the areas despite isolation. 

Poecilius cupreus was highlighted as an indicator species for the connected areas, and while the indicator species value was relatively low (0.59). It can be difficult to estimate what a high and low abundance is, and with the high differences in abundance in the intermediate group (lowest: 1, highest: 38), deciding on the isolation will be very hard. Other species besides. P. cupreus is also more associated with agricultural fields than grasslands so it may be more telling about the surrounding than the grassland itself.