Advantages of the continuous magnetic flow sorting for biomedical applications over current, batch-wise magnetic separations include high throughput and a potential for scale-up operations. A continuous magnetic sorting process has been developed based on the quadrupole magnetic field centered on an annular flow channel. The performance of the sorter has been described using the conceptual framework of split-flow thin (SPLITT) fractionation, a derivative of field-flow fractionation (FFF). To eliminate the variability inherent in working with a heterogenous cell population, we developed a set of monodisperse magnetic microspheres of a characteristic magnetization, and a magnetophoretic mobility, similar to those of the cells labeled with a magnetic colloid. The theory of the magnetic sorting process has been tested by injecting a suspension of the magnetic beads into the carrier fluid flowing through the sorter and by comparing the theoretical and experimental recovery versus total flow-rate profiles. The position of the recovery maxima along the total flow-rate axis was a function of the average bead magnetophoretic mobility and the magnetic field intensity. The theory has correctly predicted the position of the peak maxima on the total flow-rate axis and the dependence on the bead mobility and the field intensity, but has not correctly predicted the peak heights. The differences between the calculated and the measured peak heights were a function of the total flow-rate through the system, indicating a fluid-mechanical origin of the deviations from the theory (such as expected of the lift force effects in the system). The well-controlled elution studies using the monodisperse magnetic beads, and the SPLITT theory, provided us with a firm basis for the future sorter evaluation using cell mixtures.