The line widths of 35Cl- nuclear magnetic resonances were used to measure chloride binding by Band 3. Since this procedure relates directly to binding, the data obtained may be interpreted more unequivocally than affinities derived from kinetic data which could be related to either translocation or binding. Chloride binding to the active sites in Band 3 was assessed from that portion of the total line width which was sensitive to 4,4'-dinitrostilbene-2,2'-disulfonic acid. These sites appeared to be completely inhibited by treatment of erythrocyte membranes with diethylpyrocarbonate. This result is consistent with our previous observation that this reagent inhibits anion transport in resealed erythrocyte ghosts (Izuhara, Okubo & Hamasaki, 1989, Biochemistry 28:4725-4728). Hydroxylamine could not reverse the diethylpyrocarbonate inhibition of chloride binding to Band 3. The pH-dependence of diethylpyrocarbonate reactivity suggests that the modified residues may be those of histidine.