Amphora angusta Gregory, Amphora ventricosa Gregory and Amphora macilenta Gregory have been misunderstood for the last 140 years. Gregory described these species from sublittoral sediments off the west coast of Scotland. Although the illustrations were excellent for their time (1857), it is impossible to be confident of identification using Gregory's paper, and subsequent authors have created such confusion that records of these taxa must be regarded as untrustworthy unless backed by photographic evidence. In this paper, the first to describe in detail the recently established genus Seminavis, we show that A. angusta is in fact a species of Amphora Ehrenberg, whereas A. ventricosa sensu Gregory represents two independent species of the genus Seminavis D. G. Mann, namely Seminavis ventricosa (Gregory) Garcia-Baptista (non S. ventricosa sensu Garcia-Baptista 1993) and Seminavis arranensis Danielidis & D. G. Mann, sp. nov. The form usually known as A. ventricosa is neither of these and requires a new name, Seminavis robusta Danielidis & D. G. Mann, sp. nov. The long forgotten and misclassified species A. macilenta Gregory is shown to be yet another Seminavis species, requiring a new combination as S. macilenta (Gregory) Danielidis & D. G. Mann, comb. nov.; contrary to previous claims, it is separate from A. ergadensis Gregory, which is a true Amphora. Valve and girdle ultrastructure, plastid arrangement and auxosporulation are described for Seminavis and resemble those in the Naviculaceae sensu stricto.