Many of the interesting functions of human lymphocytes in vitro are known or suspected to be dependent on the presence of monocytes. A practical limiting factor in cell interaction experiments is the ability to separate and recover functional monocytes and lymphocytes rapidly from the same blood specimens. Here we describe a simple, rapid and inexpensive solution to this problem. Incubation at 37 degrees C on Sephadex G-10 columns consistently depleted monocytes to low levels (less than 0.1%) and gave maximum lymphocyte yields (greater than 60%) in a short time. Half the esterase-positive monocytes entering the column may be easily recovered from the beads by a 4 degrees C chilling/shaking technique; by using 0.5% lidocaine, an additional 15% can be recovered. Simultaneously, the G-10-nonadherent lymphocytes are being further separated into T and B cells by E-rosetting. These recovered G-10-adherent cells were above 90% esterase positive, and showed a cell size distribution on the Coulter channelyzer distinctly larger than nonadherent lymphocytes. To illustrate the use of this protocol we present a monocyte titration experiment which demonstrated a plateau of optimal pokeweed mitogen-stimulated Ig production at 5-10% monocytes.