The accuracy of the copper sulfate method of determining hemoglobin was investigated under usual blood bank conditions. In our hands this method was found to be grossly inaccurate. Using the cyanmethemoglobin technic, it was demonstrated that 83.9 per cent of females and 80.4 per cent of males rejected as having a low hemoglobin by the copper sulfate method, actually had hemoglobin values above minimum requirements for blood donors. Detailed instruction as to technics and pitfalls of the copper sulfate method was ineffectual in correcting errors as encountered by a large nursing staff carrying out the procedure. Donors found to have acceptable hemoglobin values by the copper sulfate method appeared to be safely accepted as donors; cyanmethemoglobin duplicate checks did not reveal any discrepancies in this group.