We recently reported the remarkable spontaneous self-assembly and hierarchical organization of amelogenin 'microribbons' and their ability to facilitate oriented growth of apatite crystals in vitro. In a letter of correction we communicated the finding that the X-ray diffraction pattern reported in our original report was that of cellulose contaminant and not amelogenin microribbon. We have re-evaluated our data and confirmed the protein nature of the microribbons using Fourier transform infrared and Raman microspectroscopy. Some microribbons were remarkably similar in their morphology to that of cellulose fibers. The size distribution of amelogenin microribbons was wider, particularly in width and length, and generally smaller than those originally reported. Here we present additional detailed information on the formation of a series of intermediate hierarchical structures of amelogenin assemblies prior to the formation of microribbon. The most significant finding was that full-length amelogenin nanospheres had a tendency to assemble into collinear arrays whose function is assumed to be critical at the initial stage of enamel mineral deposition. The present data gives an insight into the step-by-step assembly process of amelogenin from nanometer scale molecules to micrometer scale organized structures that can be used as templates for controlled and oriented growth of apatite mineralization in vitro.