Background: There are few reports in dogs that have evaluated the utility of semiquantitative scoring of bone marrow iron stores in conjunction with reticulocyte hemoglobin (CHr) to identify iron-restricted erythropoiesis due to absolute iron deficiency or iron sequestration.Objectives: An established system for scoring iron stores in human bone marrow samples was applied to dogs. The objectives were to evaluate interobserver agreement (Κ ω ), determine marrow iron scores in dogs without detectable hematologic abnormalities, and assess combined interpretation of iron scores and CHr to evaluate for iron-restricted erythropoiesis.Methods: Four blinded observers independently scored iron in 139 Prussian bluestained canine marrow samples from 0 (none) to 6 (very heavy), including healthy controls (n = 12), clinically ill dogs with (n = 100) and without (n = 16) detectable hematologic abnormalities, and dogs with experimental nutritional iron deficiency (n = 11).Additional medical record data were available for 118 dogs to evaluate for other evidence of iron deficiency (abnormal CHr, RBC indices, serum iron variables, external blood loss, or nutritional deficiencies).Results: Mean Κ ω was 0.69 (substantial agreement) for all samples but was 0.44 (moderate agreement) for samples with iron scores <3, indicating distinguishing scores 0-2 may not be reliable. Dogs without detectable hematologic abnormalities had scores from 3-5. Dogs with scores <3 and decreased CHr often had more indicators of iron deficiency vs dogs only having low iron scores or low CHr.Conclusions: Evaluation of dogs with marrow iron score <3 for external blood loss or nutritional deficiencies is likely clinically worthwhile, particularly if there is also decreased CHr.