The Bellows Dune site was excavated more than three decades ago (Pearson et al. 1971), and has been generally considered one of the earliest settlement sites in the Hawaiian cultural sequence. More than ten years later, in the now-classic summary of Hawaiian archaeology, Kirch (1985) considered it to be one of only two sites firmly identified as belonging to the Colonization phase in Hawai'i. This status has remained largely intact. Working independently, the authors of the present article found problems with the interpretations of the dating of this site. Combining our efforts and reviewing the general debate over the timing of human colonization of the Hawaiian archipelago, we suggest that the oft-quoted early dates for the Bellows site are in error, and that a site-based argument for pre-A.D. 800 settlement of Hawai'i is approaching a case list of zero. The most supportable conclusion is that of the two main layers at O18, the lower one (L. III) pre-dates A.D. 1000, and the upper one (L.II) post-dates A.D. 1000. The Bellows Dune site dating is deconstructed, dates from Bellows that have not been published are presented, the Bellows dates are placed in the context of new information from other sources on the date of Hawaiian colonization, and a new hypothesis for the age of the Bellows Dune site is proposed.