1. Diameter measurements of Rhizosolenia styliformis from the Antarctic, the subtropical Atlantic and Pacific Oceans and from the North Sea and neighbouring waters have made it appear necessary to set up two varieties, oceanica and semispina, in addition to the type of the species R. styliformis. The type as I describe it has been called var. longispina by Hustedt, but elsewhere it has often been figured as the var. oceanica of this paper. Var. semispina is synonymous with the form represented by Karsten as R. semispina Hensen. It differs from R. semispina as drawn by Hensen and its synonym R. hebetata forma semispina Gran, but is thought likely to be linked by intermediates. If this is so R. hebetata may have to be extended to include and suppress R. styliformis, as var. semispina is linked to the type by intermediates. Var. oceanica has no intermediate forms and, if R. hebetata is to be extended, this variety should be established as a separate species.

2. Var. oceanica is absent from the southern North Sea and appears to be an indicator species related to oceanic inflow.

3. Auxospore formation was observed for the type in the southern North Sea in 1935 and biometric observations suggest that a period of 3-4 years elapsed between the production of auxospore generations in that area. Outside the southern North Sea for the type, measurements give no indication of auxospore generations occurring at intervals exceeding a year. While auxospore formation has been seen in var. oceanica from the Shetlands area samples of June 1935 and July 1938, this phenomenon has not been observed for var. semispina.