Water-ligand observed via gradient spectroscopy (waterLOGSY) is a popular ligand-observe NMR technique for studies of protein-ligand interactions.1 However, when waterLOGSY was applied to measure dissociation constants (KD) through ligand titration, the observed KD values were found to be strongly dependent on experimental parameters and sample conditions.2 Most significantly, there was a positive correlation between the observed KD values and the protein concentrations used. In this presentation, we suggest that this deviation from the 'true' KD value is due to the rebinding with the protein of partially saturated ligands during the waterLOGSY mixing time period, and show that accurate KD values can be obtained if one conducts the titrations using protein concentrations much lower than the KD of the protein-ligand system.3