This paper is a discussion of Rhoads and Kenworthy ( 1998)
`Time-averaged flow structure in the central region of a stream
confluence' Earth Surface Processes and Landforms, 23, 171-191, that
focuses upon the methods used to identify secondary circulation in
river channel confluences, It argues that the Rozovskii method that
Rhoads and Kenworthy use to rotate their field data to allow
identification of secondary circulation cells is flawed, and can result
in misleading conclusions about the nature of flow processes in
confluences. It recommends that there is a re-emphasis upon helical as
opposed to secondary circulation, and that recent developments in both
field monitoring and numerical modelling may help significantly in this
respect.