For the dependent variable of total soil carbon stock within each set of pairs (Broadleaf–Pasture and Broadleaf–Conifer), the independent class variable ‘‘vegetation’’ was the main treatment, and paired plots were considered the ‘‘block’’ class variable to take into account site variability and make the test of the treatment (vegetation type) more powerful in the ANCOVA. Comparisons for differences between vegetation treatments in ANCOVA were analyzed using pair-wise comparisons protected Least Significant Difference (LSD) with a significance level of P < 0.05. Clay content was included in the ANCOVA as a continuous independent covariate. Forest stand age could not be included as a covariate in the ANCOVA models of Broadleaf–Pasture because identical pasture age (54 years) would skew the model.