The replacement of fetal by adult hemoglobin after birth is not complete, since small amounts of fetal hemoglobin are consistently present in the blood of normal adults (6-9). It is, therefore, of interest to learn how continued synthesis of fetal hemoglobin in the adult is regulated and whether it can be further stimulated. We examined this in vitro, with a bone marrow culture system which permits selective proliferation of erythroid cells (10). We report here data which suggest appearance of increased and clonal synthesis of fetal hemoglobin in the adult erythroid cells, in vitro. METHODS Bone Marrow Cultures. Bone marrow aspirates from healthy volunteers without a hemoglobinopathy and from a patient homozygous for hemoglobin S were used; on a few occasions ribs surgically removed from patients were the source of marrow cells. Cells were washed in medium (Microbiological Associates, Bethesda, Md.) containing 2% (vol/vol) fetal calf serum, penicillin (50 units/ml) and streptomycin (50 ,gg/ml); the buffy coat layers were removed, and resuspended in fresh media. Antibodies and Immunofluorescent Labeling. Antibodies against hemoglobin F (anti-Hb F) or A (anti-Hb A) were raised in rabbits (anti-Hb F) or horses (anti-Hb A) and purified by affinity chromatography. The purification procedure, the conjugation with fluorescein isothiocyanate (FITC), and evidence of specificity of the anti-Hb F antibodies have been described (9). Peripheral blood smears were fixed and labeled with anti-Hb F antibodies conjugated with FITC (anti-Hb F-FITC), as previously described (9); smears prepared from bone marrow cell suspensions were treated similarly. Flattened plasma clots were fixed for 12 min in an acetone-methanol (9:1 vol/vol) solution, rinsed for 2 min in phosphate-buffered saline at pH 7.0 (9), rinsed in distilled water for another 2 min, and dried. They were labeled with anti-Hb F-FITC in the same fashion as the smears, except that the whole area of the clot was covered with the fluorescent antibody. After incubation for 1 hr in a humidified chamber at 370, the slides with clots were rinsed in phosphate-buffered saline, then in water, and then dried. They were mounted with clear serum, and viewed in the fluorescent microscope (Zeiss, Universal fluorescence microscope with reflected light excitation) under 150-times magnification. Freshly thawed antibody solutions were used for each experiment and conditions of fixation, staining, and viewing were kept constant. Evaluation of fluorescent labeling was best when plasma clots contained fewer erythroid colonies: cultures inoculated with 5 X 104 nucleated cells per clot were optimal. The reaction of the fixed plasma clots with the fluorescent antihemoglobin antibodies preserved the morphological characteristics of the cells and permitted identification and quantitative estimates of total erythroid colonies under fluorescent light. Erythroid cells and colonies were classified into categories according to intensity of fluorescent labeling. Fluorescence of moderate to bright intens...