Theoretical and experimental investigation of high flow microinfusion into brain tissues has been undertaken. Theoretical transport models have been developed, treating the brain either as a rigid or deformable elastic medium into which small molecular weight or macromolecular solutions are infused at flow rates of 0. 5 to 6. 0 microliters/min. Mechanical stress, diffusion/bulk flow, and continuity relationships are being modified to account for the volume-dependent hydraulic conductivity characteristic of unconstrained tissues in vivo. Theory predicted that infusion from a single cannula would yield a penetration distance on the order of a centimeter or more for antibody-sized macromolecules. Experimental microinfusion of dextran-dye, In-111 transferrin, and C-14 albumin into the normal white matter of a cat has confirmed that such penetration depths are actually achievable. Autoradiography has yielded concentration profiles surrounding the cannula tip for total infusion volumes of 75 to 300 microliters. In agreement with theoretical prediction, these profiles have relatively sharp edges and a moving front character. These results suggest that residual brain tumor may be treated by infusing radioactive antibodies or immunotoxins by high flow microinfusion into neighboring normal tissue.