It was previously shown, in theory and in practice, that a blood component program utilizing 80--85% of all donations as red cell concentrates with a hematocrit of 70% contributes substantially to the coverage of national albumin demands without increasing the need for that protein. The current trend towards the production of red cell concentrates with hematocrits around 85% by the removal of more plasma prompted a computer simulation study of the effects of this modification on the protein and monetary economy of a component program. The results suggest that a hematocrit around 70% constitutes an optimum in these respects. In a general way, they point to the desirability of examining the repercussions of isolated modifications of the processing of blood donations on the system as a whole prior to their large-scale application.