Ábrahám Levente (szerk.): Válogatott tanulmányok IX. - Natura Somogyiensis 24. (Kaposvár, 2014)
Szűcs D. - Kitti Horváth K. - Horváth Gy. F.: Comparing small mammal faunas based on barn owl (Tyto alba) pellets collected in two different lowland landscapes
308 Natura Somogyiensis Methods of owl pellet analysis The collected material included whole pellets as well as pellet fragments/debris in many cases. This is important to note because prey lists were compiled based on whole pellets on the one hand and relying on whole pellets plus pellet debris on the other. Taxonomic small mammal identification was done on the basis of skull characteristics and dentition (Schmidt 1967, Ács 1985, Újhelyi 1994). Neomys species (Neomys fodi- ens Pennant 1771, and Neomys anomalus Cabrera 1907) were differentiated by measuring the height of the corona-process of the mandible; if this was unfeasible, only the genus was identified (Neomys sp.). The wood mouse (Apodemus sylvaticus Linnaeus 1758), the yellow-necked wood mouse {Apodemus flavicollis Melchior 1834) and the pygmy field mouse {Apodemus uralensis Kratochvil and Rosicky 1952) were categorised commonly as wood mice {Apodemus spp.) The house mouse {Mus musculus Linnaeus 1758) was differentiated from the gleaner mouse {Mus spicilegus Petényi 1882) on the basis of the length proportions of the upper and lower zygomatic arches (Macholán 1996, Cserkész et al. 2008). Consequently, the summarized list of small mammal taxa which was involved in our comparative statistical analysis consisted of 25 components. Statistical methods Mesoregional scale The number of small mammal individuals was the basic data of pellet analysis. Because sample sizes were different, we used relative frequencies of taxa as derived data in our landscape level statistical analyses. The distribution of relative frequencies for each species was performed by G-test between two mesoregions. Based on data of total species list we investigated the hypotheses with homogeneity-test Of2) that the two samples derive from the same statistical population or not. We used detailed data of settlements to test the difference of small mammal abundance with non parametric Mann- Whitney-U test (Zar 2010). We calculated percentage overlap (Renkonen-index) to compare small mammal composition of the two landscapes. This index measures the percentage similarity of two assemblages (Krebs 1989): Px = n y^(min./?j/,ptt) i=i x 100 where n is the number of prey categories, py and pik is the relative proportion of the Ith prey in two samples (j and k). To compare the rank order of species frequencies, we used Spearman’s rank correlation. This statistical method shows how much the ranking of frequencies are similar between two landscapes. Microregional scale By refining the spatial scale, it is possible to evaluate the small mammals’ data of the detailed landscape-level with respect to microregions within mesoregions. In case of the Drava floodplain, the collected samples affected 3 microregions (Dráva-sík (DS), Fekete-víz síkja (FVS), Nyárád-FIarkányi-sík (NHS)), while in the Győr basin only 2: Csornai-sík (CS), Kapuvári-sík (KS). We calculated the percentage overlap (Pjf) between each microregion (Renkonen-index) and also compared the rank order of small
