The four closely related taxa considered in this paper have all been referred to Metrosideros at some time. Nani Adanson is still generally treated as a section of Metrosideros Banks ex Gaertn. (nom. cons.) despite the objections of several earlier authors (Merrill, 1952; Rock, 1917; Valeton, 1901; Miquel, 1885). A comparison of the description of Nani in the present paper with that of typical Metrosideros (Dawson, 1970) shows that the two genera differ widely in many features. I agree with Merrill (1952) that the affinities of Nani he with Xanthostemon and further suggest that a detailed revision at the specific level may lead to their being merged. If such a merger takes place, then Nani would have priority over Xanthostemon, although a strong case for conservation of the latter could be made.