Although it has been thoroughly recognized that a relationship exists between particle size distribution (PSD) and cake porosity and permeability, a lack of prediction capability usually limits the possibility to forecast the change of permeability that occurs when the PSD of particles varies. Results obtained with limestone, glass spheres, aluminium hydroxide and magnesite, being the experimental matrix for limestone broader in both mean particle sizes and polidispersity are discussed. "Taylor-made," particle mixtures were prepared varying mean sizes between 5 and 150 micrometers and geometrical standard deviations between almost 1 (narrow mixtures) and 2.5 in order to ensure a wide data spectrum. Particle size distribution was measured by laser diffraction, but it is shown that similar results can be obtained using other sizer devices, by the definition of a novel device dependent-particle shape factor that relates assuming particles as spheres the actual specific surface to the specific surface estimated from PSD measurements. Further to the use of the particle shape factor it was necessary to include additional parameters take into account the fact that the same particle collective can build different porous structures depending on the way the bulk is formed. Therefore, two parameters that include the cake building mode were lumped with the particle shape factor to give a "pore-particle shape factor", which is a system-specific parameter that must be determined experimentally. The use of this approach allows that the well-known Kozeny-Carman equation fits very well the data, provided that the relationship between PSD and porosity is known. This relationship is discussed in [1] but the approach suggested by Yu et al. [4] was followed. The results can be easily illustrated in a permeability-porosity plot that has the advantage of concentrating all cake parameters in the y-axis, while the particle collective characteristics (namely PSD) are in the x-axis via the D-3,D-2, being the shift from the ideal case of spheres forming a random bulk a measure of the pore-particle shape factor.