Comment on: Bi nanolines on Si(001) : registry with substrate. Authors' reply

A recent article (Miwa et al 2005 Nanotechnology 16 2427) casts doubt on the four-dimer-wide Haiku model for the Bi nanoline on Si(001), suggesting instead that the three-dimer-wide Miki model (which had been ruled out) is a better fit in particular to x-ray data. The reasons why the Haiku model provides the best fit to all published data among currently proposed structures are discussed, concentrating on the width and registry of the Bi nanoline, and mentioning new data which shows that under appropriate conditions the two structures coexist in the same surface.