Multivariate patterns in biosorption were visualized using Principal Components Analysis (PCA) from a correlation matrix with some elements log-transformed to create a normal distribution. Univariate analysis took the form of three-way fullyfactorial Analysis of Variance (ANOVA), with the factors biosorbent (fixed), pH (fixed) and time (random). Data were examined for normality and homogeneity of variance using normal-probability plot of raw residuals and predicted-residual scatter plot, and were transformed as necessary to meet assumptions [26]. Both the PCA and ANOVA test were conducted in Statistica for Windows (Ver. 10, C. Statsoft Inc. 1984–2011).