Successfully modeling the surface charge of goethite and anion adsorption on goethite using a surface complexation model (SCM) alone cannot verify the assumptions of this model. In this study, the assumptions of 2-pK triple layer model (TLM) and two-site 1-pK basic stern model (BSM) were assessed with respect to their ability to interpret both the proton-anion adsorption ratios of dimethylarsinate (DMA), monomethylarsonate (MMA), and arsenate and their effect on the zeta-potential. The proton-DMA adsorption ratio is around 0.9 at PH 4.25 and 1. 1 at pH 6.75 at DMA surface coverage ranging from 0 to 2 mu mol m(-2), and the zeta-potential is independent of DMA adsorption at these two pH values. The proton-MMA adsorption ratio increases to 1.5 at pH 4 and 2.1 at pH 6.75 as the MMA surface coverage decreases to 0.5 mu mol m(-2). The zeta-potential is less dependent on MMA adsorption at a surface coverage range of 0 to 1.8 mu mol m(-2), and it then decreases with a further increase in the MMA surface coverage at pH 4 and 6.75. The proton-arsenate adsorption ratio decreases to 2 as the arsenate surface coverage approaches zero, and the zeta-potential decreases linearly with the increasing arsenate surface coverage at pH 4 and 6.75. Neither the 2-pK TLM nor the 1-pK BSM give a consistent interpretation of both the proton-arsenic adsorption ratio and the effect of arsenic on the zeta-potential. The results suggest that the I -pK MUSIC model in which each type of surface hydroxyls has its own intrinsic proton-affinity constant and only one type of surface hydroxyls is involved in DMA, MMA, and arsenate adsorption is preferably pursued. The protonation degree of reactive hydroxyls estimated from proton-arsenic adsorption ratios is 0.2 at pH 4 and 0 at pH 6.75 in 0.001 M NaNO3 (c) 2007 Elsevier Inc. All rights reserved.