The Sea-Ducks (tribes Mergini and Somateriini), stiff-tailed ducks (tribe Oxyurini) and the pochards and scaups (tribe Aythyini), all of which are diving ducks with a lobed hallux, are considered to be polyphyletic, but their systematic affinities are extremely uncertain. Until 1958, the Sea-Ducks were considered as one tribe but have now been split into two unrelated groups, on the basis of differences in tracheal and plumage characters of the eiders which closely resemble those of the surface-feeding ducks (Anatini). This study, made in British Columbia and Alaska, was begun in 1955 and continued through 1958. It was planned to describe the display characters of Sea-Ducks and to use these to suggest systematic relationships among the eight genera in which the twenty species are grouped. The behaviour of six species (Melanitta and Bucephala) was studied in some detail, and observations were made also on three other species (the monotypic Clangula and Hlstrionlcus, and on Somaterla spectabilis). Published information was also utilised (particularly that on Mergus albellus, M. serrator and Somateria mollissima), so as to make the study comparative for the whole group. In ducks the pre-copulatory displays of male and female, and the general displays peculiar to the female, are highly conservative. They may be used as taxonomic characters at the generic and tribal levels of classification. The general "courtship" characters of the male are of little value at any taxonomic level, except within a large genus (e.g. Anas). This is because signal characters may evolve from basic non-social or agonistic behaviour independently in related lines, and because of convergence (at the generic or tribal level) in isolating mechanisms such as male plumage and display characters. In female and pre-copulatory behaviour characteristics, both Somateriini and Mergini are shown to be quite distinct from the Anatini. In female courtship behaviour Somateria resembles some of the Aythyini, and Melanitta. In female behaviour Clangula resembles Melanitta and both are distinct from Bucephala. Mergus has been considered closely related to Bucephala for various reasons. There are striking (? convergent) similarities in some of the displays of the male of M.merganser and M.serrator with displays of B. clangula, but further study is required of the behaviour of Mergus before ethological evidence linking the two genera is acceptable. B. albeola and M.albellus have been supposed to be links between the two genera, but they have little similarity, at least in the behaviour of the male. The pre-copulatory behaviour of B. albeola is more like that of Melanitta than the goldeneyes (B. islandica and B.clangula) in which it is highly developed. The behaviour of the female of B. albeola is unlike either the goldeneyes or Melanitta. B. albeola probably deserves generic distinction from the goldeneyes, although in nesting behaviour, territoriality, diving and agonistic behaviour it resembles them. On the behavioural evidence the Mergini appear to consist of at least two unrelated lines (Melanitta—Clangula and Bucephala—Mergus). The position of Histrionicus is not clarified. On the pattern of its downy young it seems close to Clangula, but Humphrey noted that it had an eider-type tracheal bulla. The male of both Histrionicus and Clangula has a paucity of visual displays which is accompanied by rather greater vocalisation than in other Sea-Ducks. It is suggested that paucity of visual displays is an indication of isolation from related groups, due to ancestral divergence, so that the generic rankings of Clangula and Histrionicus should perhaps stand. / Science, Faculty of / Zoology, Department of / Graduate