Statistical analyses

Separate one way fixed effects analysis of variance (ANOVA) for each stockpile size and extraction period (8 and 16 months) was used to determine whether stockpiling affected seed viability, seed germination, rhizome emergence and soil chemical properties by comparing response variables before and after stockpiling. Significant main effects were analyzed using least squares difference (LSD) post hoc test. Two way fixed effects ANOVA for each extraction period was used to determine effects of stockpile size and burial depth. Significant interaction effects were analyzed using one way ANOVAs to determine significant differences between depths for each stockpile size; if main effects were significant, differences were analyzed using LSD. Residuals from raw data were tested for normality with the Shapiro-Wilk test and heterogeneity of variances with Levene’s test. Data were rank transformed when variances of raw data were heterogeneous. Spearman’s rank correlation coefficient was used to compare strength of viability relationships of each species to soil gases and chemical properties known to affect seed germination, separately for large and small stockpiles. Analyses were conducted using SPSS 18.0; significance was at p = 0.05.