Amino acids have been cited as potential precursors of the disinfection byproduct cyanogen chloride in chlorinated drinking water. Screening experiments with 17 amino acids were performed in this study to comprehensively identify important CNCl precursors. Among this set, only glycine was found to yield detectable CNCl (i.e., > 0.6% yields). Additional experiments were conducted to estimate the relative significance of glycine as a CNCl precursor in water samples collected from the Huron River, Michigan, by concurrently characterizing the amino acid content and monitoring CNCl yields after chlorination. Chlorine was added at slightly less than the sample breakpoint dose to optimize CNCl formation and stability in the samples. On the basis of previous determinations that glycine-nitrogen is stoichiometrically converted to CNCl-N at pH > 6, it was estimated that glycine may account for 42-45% of the CNCl formed in the river water samples (pH 8.2). The kinetic profile of CNCl formation in the sample, with a half-life of about 20 min, indicated that both rapid and slower formation pathways were important. Glycine formation of CNCl, with a half-life of 4 min, is likely to contribute significantly to the rapidly formed CNCl, while unidentified precursors must account for the slower pathway. Non-glycine-derived CNCl precursors in this water source were further examined to determine if they were largely proteinaceous in character using a technique known as immobilized metal ion affinity chromatography (IMAC). These experiments demonstrated that copper-loaded IMAC resins were much more effective in removing glycine than other CNCl precursor compounds in the sample matrix. The unidentified CNCl precursor components, therefore, are not likely to be proteinaceous and are more likely to be associated with the fulvic/humic fraction of organic matter.