This study was conducted to characterize better the response of rats to blood loss and hemolysis and to incorporate automated methods into the routine evaluations of those responses. Serial phlebotomies of 1.5–2.0 ml of blood per day for 5 days, or intraperitoneal injection of 50 mg kg⁻¹ phenylhydrazine (PHZ) for 3 days, were used to cause anemia associated with blood loss or hemolysis, respectively. Maximum decreases in red blood counts were observed on Day 3 in PHZ‐treated animals (68%) and Day 4 in blood‐loss animals (35%). In the routine complete blood count (CBC), hemoglobin, hematocrit/hemoglobin ratio and erythrocyte indices could be used to discriminate between the two treatments. Free plasma hemoglobin in PHZ‐treated animals resulted in marked elevations of mean corpuscular hemoglobin (MCH) and mean corpuscular hemoglobin concentration (MCHC) with a 2:1 hematocrit/hemoglobin ratio rather than the anticipated 3:1 ratio. Although both groups of animals had elevated white blood cell counts, PHZ‐treated animals also had monocytosis and basophilia. Reticulocyte counts were more sensitive than erythropoietin (EPO) concentrations in predicting erythroid changes. Maximum mean reticulocyte values were ca. 24% in serially phlebotomized animals and >99% in PHZ‐treated rats. Plasma EPO levels were 4–10‐fold higher than EPO levels in urine, kidney or liver. Flow cytometric differentials of rat bone marrow using 2,7‐dichlorofluorescin successfully predicted erythroid hyperplasia in both experimental groups. Erythrocyte indices returned to normal within 14 days and the remaining CBC parameters were normal within 28 days for both treatment groups. Reticulocyte counts remained slightly elevated on Day 28, but were normal when assessed at Day 56 in blood‐loss and PHZ‐treated animals. Copyright © 2000 John Wiley & Sons, Ltd