Columns of DEAE-agarose beads of high maximum capacity (0.18 M inside the gel) were examined for their suitability for protein separation in the system of chromatofocusing. It was concluded that: (a) large sudden changes of ionic strength should be avoided; (b) when an anion exchanger is used, the component of the pre-equilibration buffer and the eluction buffer should be of the cationic type, whereas the counter ions should not be subject to association—dissociation equilibria in the pH range of operation. Furthermore slow equilibration between ion exchanger and surrounding medium caused small maxima in the pH gradient. These complications were mainly circumvented while maintaining the advantages of chromatofocusing by using the following buffer system: pre-equilibration of the column with 10 mM ammonia, pH 9.2, and elution with a linear gradient from 10 mM ammonia, pH 7.5, to 10 mM triethanolamine, pH 7.5, produced in a gradient mixer. In the approximately linear pH gradient which results a commercial preparation of myoglobin could be separated into several components. The resolution was twice as good as in previous experiments, corresponding to theoretical calcultions. However, the sluggishness of the exchanger has to be taken account when designing new ion exchangers.