On 20 consecutive work days during four weeks, one technologist performed 24 microchromatographic determinations of hemoglobin A2 (Hb A2) by each of four methods: the Efremov procedure requiring Tris/HCl buffer, the original Huisman technique with use of glycine developer, and two commercial test kits in which a modified glycine developer is used. The bloood samples tested were obtained from 12 adults with no hematological abnormality and from 12 beta-thalassemia carriers previously diagnosed by familial and hematologic studies. Results by the first method and the two commercial kits (one from Helena Laboratories and one from Isolab, Inc.) did not differ significantly in precision for either the normal or beta-thalassemia trait samples. For both sample types, the second method yielded larger coefficients of variation than those obtained with the other methods. Moreover, the second method was the only one with which values overlapped for normal samples and samples with above-normal Hb A2 concentrations.