Experiments with spherical iron particles 28 A to 265 A in diameter demonstrate that the proposed nonferromagnetic surface layer on an iron particle must be less than 1 A thick, i.e., about one atomic radius. This conclusion is based upon a comparison of the ferromagnetic iron calculated from the magnetic saturation, and the total amount of iron as determined by chemical analysis. The 1/Hextrapolation of the magnetic induction appears adequate to obtain the saturation induction at room temperature of particles greater than about 100 A in diameter. For particles less than this size, in the superparamagnetic size range, the Langevin function must be used taking into account the size distribution of the particles in the sample. Thus there will be a negligible decrease in saturation induction due to the loss of exchange coupling at the surface of iron particles in the size range above 100 A of interest for fine‐particle permanent magnets.