Respiration rates and chemical characteristics of soil organic layers were measured at 40 experimental plots, 5 sampling sites per plot, in a moderately polluted Niepolomice Forest, S. Poland. The respiration rate was positively related to pH, water content and concentrations of Ca and K, and negatively to N (tot), Zn and Pb (p < 0.001 for all variables). No significant correlation was found between the respiration rate and Na, Cu or Cd (p > 0.25 in all cases). The regression model explained 73% of the total variance. Analysis of variance components revealed that ca. 35% of the total variance in the respiration rate can be explained by the vegetation types covering the area: oak-hornbeam vs. pine-oak forests. The next 40% was explained by the variability between sampling plots and the remaining 25% by within-plot variability among sampling sites. Similar results were obtained for water content. The variance in pH was split 30%:39%:31% between vegetation types: plots: sampling sites. No variance in Ca and Na was explained by the forest type, and approximately half of the variance was due to between-plot and half to within-plot variability. In contrast, potassium concentration differed between forest types (58% variance explained), more than 25% of total variance was due to between-plot variability and only 15% due to within-plot variability. For Zn the results were 1%, 66% and 33%, for Cu 0%, 38% and 62%, for Pb 48%, 18% and 34%, and for Cd 0%, 33% and 67%, respectively. The study clearly shows (1) substantial variance in some soil characteristics between sampling sites and (2) a different split of variance among spatial scales for different soil characteristics.