Landscape ecology of small mammals, with special reference to the cotton mouse (Peromyscus gossypinus).

摘要

I sampled 60 bottomland forest patches in the six southwestern-most counties in Illinois to determine the current status of the cotton mouse ( Peromyscus gossypinus) and to assess the effects of landscape characteristics on small mammal communities. Landscape characteristics included patch size, patch shape, overstory tree genera richness, overstory tree distribution patterns, average diameter at breast height (DBH), and landuse (percentages of deciduous forest, conifer forest, cropland, grassland, and urban/other) within a 300-m buffer around each patch. Identification of Peromyscus was based on a modified allozyme electrophoretic technique using the diagnostic GPI-1* locus. Allozymes were isolated from toe-clips, as opposed to liver, and run on a cellulose acetate medium. One hybrid Peromyscus gossypinus-leucopus and one small Peromyscus, carrying a cotton mouse allele at the GPI-1* locus, were identified from 384 individuals screened using this genetic marker. White-footed mice comprised 93% of all small mammals captured. Spearman Rank Correlation Coefficients revealed significant correlations between P. leucopus abundance and percentage of grassland, deciduous forest, and urban/other habitat surrounding a patch. There were significant correlations between small mammal species richness and patch size, as well as small mammal heterogeneity and percentage of urban/other and grassland habitat surrounding a patch. Finally, small mammal species richness was significantly correlated with overstory tree genera richness, average DBH, and percentage of grassland and urban/other habitat surrounding a patch. Although simple correlations were useful in determining relationships between the dependent and independent variables, stepwise multiple regression was needed to assess the combined affects of multiple landscape characteristics on small mammal communities. Further research into landscape approaches combined with multivariate analyses of multiple landscape characteristics, especially in landscapes with high species diversity, will provide much needed information for small mammal conservation and management.