Statistical analyses primarily consisted of summary statistics, including means and standard errors, where appropriate; simple t-tests for mean differences; and one-way analysis of variance followed by mean comparisons using TukeyÕs procedure conducted at the 0.05 level (Montgomery, 1997). When appropriate, the wood hydrolyzate batch was included as a blocking factor. Most analyses were performed with SAS version 6.12 software (SAS Institute, 1989). The ดรgures do not include standard error bars since relative errors weredwarfed by graphical symbols.