We have measured separation distances between live human red blood cells and simple or modified glass surfaces, using the finite aperture technique of microscope interferometry. In general, separation increases as the ionic strength falls, in isotonic solutions. Restriction on movement parallel to the glass in all except the most dilute salt solutions, coupled with the absence of Brownian motion, indicates direct molecular contact with the substratum. Thus increased separation must be due to swelling of the glycocalyx under electrostatic forces. However, at approximately less than to 2mM adherent cells show a separation greater than 100 nm, execute Brownian motion and the restriction on lateral motion is less evident. This suggests that secondary minimum adhesion by long-range forces with little or no direct molecular connection occurs at extreme dilution only. Treatment of cells with trypsin reduces separation by up to 40 nm, but the extent to which this reflects reduced double-layer repulsion due to loss of surface charge, as opposed to the reduced opportunity for swelling in a trimmed-down glycocalyx, is unclear. Adhesion at a separation approximately 100 nm in 1 mM buffer after trypsinization supports the view that adhesion can occur without very long glycoprotein connections, but does not prove it. Adhesion to unwettable methylated glass and completely wettable unmethylated glass, with an identical ionic strength dependence of the separation, shows that hydrophilicity is not an absolute requirement. Red cells interact closely at all ionic strengths with glass made polycationic with poly-L-lysine, owing to electrostatic attraction. The interference technique also shows that adherent cells can be spaced from the glass by an intervening layer of previously absorbed serum albumin.