The hexagonal ferrites form an unusual group of complex, ferrimagnetic oxides embodying some 60 known crystal structures. These include phases for which the structural unit cell is larger than that in any known inorganic materials. The various hexagonal ferrite modifications fall into two distinct structural series, each formed by the ordered interlayering (stacking) of two discrete building blocks; these blocks stack along the c crystallographic axis in varying ratios and varying permutations to form strictly coherent, reproducible crystal structures. This mixed-layering aspect of the hexagonal ferrites permits direct, visual observation of the sequence of their subunit-cell stacking elements, after etching, by means of electron microscopy. The sequence of stacked blocks in such structures constitutes the only information lacking for a complete, three-dimensional structure determination. Direct access to this information provides an immediate, unique solution of the crystal structure problem in each case and thereby avoids the dilemmas of a classical diffraction approach to such large unit cells. Ferrite structures with hexagonal c dimensions of 1455 and 1577 angstroms have been uniquely solved by direct electron microscopic readout of surface etch features. One must exercise caution, however, in generalizing these findings to other materials. The method is successful in the case of the hexagonal ferrites because these are mixed-layer structures, wherein the building blocks react at different rates to a specific etchant. Mixed-layer systems are not uncommon in crystallography, and it is likely that similar techniques can be developed for other such materials. Regardless of the validity of this prognosis, however, it is quite evident that high-resolution replica electron microscopy is a most promising tool for the direct observation of surface structure on an ultramicro scale. During the studies reported here replica resolution capability was improved to about 10 angstroms; final resolution is limited by the particle size of the platinum shadowing material. Careful control of experimental conditions during replica preparation or an alternate choice of shadowing material, or both, might reasonably improve the resolution by a factor of 2. This resolution is within the range of most unit cell dimensions and approaches interatomic distances in solid-state materials. The potential of such an experimental capability needs no elaboration.