The contribution of red cell morphology to the diagnosis of beta-thalassemia trait in a predominantly healthy population was analyzed. The smears from 59 beta-thalassemia trait subjects already identified by DEAE-cellulose microcolumn chromatography, 60 subjects with other forms of microcytosis, and 64 nonmicrocytic subjects were randomly selected and evaluated blind. The "experimental" morphologic criteria were the presence of both basophilic stippling and either microcytosis or target cells. The "control" morphologic criteria were the presence of both microcytosis and poikilocytosis. Although the sensitivity of both sets of criteria were only moderate (73%), the specificity and predictive value of the experimental criteria were very high (99% and 98% respectively).