In the past, macroscopic models of microcirculation have used either gas or liquid droplets or rigid particles to simulate red cells. These models cannot simulate the boundary conditions correctly at the cell wall. In the present paper we used flexible, thin-walled rubber models which were liquid-filled and geometrically similar to the human red blood cells. These cells were made to flow in a silicone fluid into circular cylindrical tubes with diameters comparable to that of the cell model. The Reynolds number based on the mean flow velocity and tube diameter was in the range of 4 × 10 −4−4 × 10 −2 similar to the range encountered in microcirculation. Measurements were made of the entry condition, the entry length, the deformation pattern of the cells, the cell velocity and the mean flow velocity, the local pressure difference on the tube wall across the red cell, and the loss of pressure head due to a single red cell. All these quantities are functions of the ratio of the cell diameter D c to the tube diameter D T . A dimensional analysis yields functional forms with which experimental data are organized. The entry length is less than a cell diameter in all cases. The cells are severely deformed if D c D T > 0.98 . For larger D c D T , large buckling occurs at the trailing end of the cell at higher velocities of flow; the buckled cell shape resembles the familiar pattern seen in vivo. The cell velocity V c is related to the mean flow velocity V M by the relation V c = k( V M − α); the value of k increases as D c D T decreases. The disturbance to the pressure field due to a red cell is rather localized. Data on the resistance (pressure drop) to cell motion when D c D T = 1.36, 1.13, and 0.98 are presented as a function of the cell velocity.