Elution time measurements of colloidal particles injected in a symmetrical flow field-flow fractionation (flow FFF) system when the inlet and outlet cross-flow connections are closed have been performed. This no-field method has been proposed earlier for void time (and void volume) determination in flow FFF Giddings et al. (1977). The elution times observed were much larger than expected on the basis of the channel geometrical volume and the flow rate. In order to explain these discrepancies, a flow model allowing the carrier liquid to flow through the porous walls toward the reservoirs located behind the porous elements and along these reservoirs was developed. The ratio between the observed elution time and expected one is found to depend only on a parameter which is a function of the effective permeability and thickness of the porous elements and of the channel thickness and length. The permeabilities of the frits used in the system were measured. Their values lead to predicted elution times in reasonable agreement with experimental ones, taking into account likely membrane protrusion inside the channel on system assembly. They comfort the basic feature of the flow model, in the no-field case. The carrier liquid mostly bypasses the channel to flow along the system mainly in the reservoir. It flows through the porous walls toward the reservoirs near channel inlet and again through the porous walls from the reservoirs to the channel near channel outlet before exiting the system. In order to estimate the extent of this bypassing process, it is desirable that the hydrodynamic characteristics of the permeable elements (permeability and thickness) are provided by flow FFF manufacturers. The model applies to symmetrical as well as asymmetrical flow FFF systems.