The authors established the specificity, reliability, and precision of human erythrocyte insulin radioreceptor assay. On the basis of insulin binding, cell viability, and degree of hemolysis, heparin sodium was found to be a more suitable anticoagulant than sodium fluoride, ethylenediaminetetraacetic acid, sodium oxalate, or sodium citrate. In two sets of experiments carried out at 4°C and 23°C, human erythrocytes were stored as whole blood or isolated erythrocytes suspended in Tris-{4-(2-hydroxyethyl)-1-piperazine-ethanesulfonic acid} buffer. The effect of storage under these conditions was evaluated by erythrocytespecific insulin binding. Human erythrocytes can be stored for 72 hours at 4°C without any change in insulin binding, insulin receptor sites per cell, or average affinity constant at the empty sites. Isolated erythrocytes can also be stored in plasma for 72 hours or in buffer G for 24 hours at 4°C without any change in insulin binding. It is not advisable to store human erythrocytes in plasma or as whole blood for more than 24 hours at 23°. These findings are useful in preserving insulin receptor activity when storage of erythrocytes is unavoidable.