Reticulocyte counting by flow cytometry with thiazole orange was compared to manual or automated counting of new methylene blue stained blood smears. Forty-nine samples were compared for manual counting from randomly chosen clinical samples. Two hundred and eighty-nine samples from bone marrow transplant patients were compared during the period before and through chemo-irradiation and engraftment. The slopes of correlation plots were less than 1 when flow cytometric data were the dependent variable, suggesting that thiazole orange is less sensitive than new methylene blue. In a third study, 407 samples from bone marrow transplant patients were compared after increasing the thiazole orange concentration. The reticulocyte fluorescence distribution was divided into four groups of the brightest (youngest) 40, 60, 80, and 100% of reticulocytes. The slopes from regression analysis were 0.25, 0.49, 0.78, and 1.14, respectively. This demonstrates that thiazole orange is more sensitive than new methylene blue because the window of analysis includes an increased fraction of mature reticulocytes. In addition, the precision of each assay as measured. The rank order of precision from high to low was flow cytometry > image analysis >> manual counting.