Investigates mathematical, computer-based methods of simulating the blood composition of individual clinical patients. Earlier studies developed procedures for constructing models of the respiratory function and acid-base biochemistry of statistically normal human blood. This Memorandum extends these procedures and shows that, with sufficient laboratory data, such a model can be derived from individual patients, and conjectures that the blood's chemical pattern may indicate the patient's physical condition. Such models may be useful for clinical and experimental fluid therapy. The models described simulate the steady-state distribution of chemical species (particularly proteins, electrolytes, and water) between the plasma and red blood cells at one instant, but not the kinetics or time-dependent states of such systems. In 30 separate laboratory experiments, a model of individual blood was tested against real blood in vitro, under various chemical stresses. Results indicate satisfactory agreement.