A new immunomagnetic separation process that uses protein A-coated magnetic nanoclusters (PACMAN) as the separation vehicles has been developed. The nanoclusters are produced by sonicating egg yolk phosphatidylcholine and the transmembrane F-c receptor protein-A in a buffered aqueous ferrofluid suspension. The phospholipids appear to form a coating around clusters of 5-50 magnetic nanoparticles, while some of the protein-A associate with the coated layer. The PACMAN suspension shows no flocculation for several months. When this suspension is added to a mixture of chicken (CRBC) and antibody-bound sheep red blood cells (SRBC), the nanoclusters bind to the exposed F, domains of the antibodies attached to the target SRBCs. Exposure of this mixture to a magnetic field gradient oriented against gravity causes selective migration of the "magnetized" SRBCs to the high magnetic field region, thus affecting a separation. Because of the overwhelming concentration of the nanoclusters in the high field region in the proximity of the magnet poles and their virtual absence in the rest of the sample away from the poles, these regions appear visually distinct. Considering these regions as different phases, the efficiency of this technique is quantified by defining a separation factor as the concentration ratio of sheep to chicken red blood cells in the high magnetic field region divided by this ratio in the low magnetic field region. Single-stage separation factors of similar to 5.0 have been obtained consistently on this system. In contrast, three control experiments where the linkage between the nanoclusters and the target cells were severed at three different locations produced negligible separation, indicating minimal non-specific binding under our experimental conditions. Using a negative selection procedure, an initial mixture containing an equal number of chicken and sheep red blood cells was concentrated to 80% CRBCs in a single stage, while in three stages the concentration of CRBCs was increased to over 99%. Potential strengths of this method are the ease of PACMAN formation, the ability to have an indefinitely stable suspension of possibly biocompatible magnetic clusters before addition to the target sample, and provision of a natural lipid environment for the protein-A, thus maximizing its functionality.