It is shown that the fluid friction and mass transfer rates measured as described in Part I do not exhibit the expected analogue relationship, changing with the flow variables in quite different ways. It is demonstrated that, if the transport processes are supposed to be edgy-diffusional, this difference implies that the eddy diffusivities of mass and momentum in the wall region are quite dissimilar and that their relationship alters with the flow conditions. This is equivalent to the conclusion that there is no mass-momentum transfer analogy in this system-in which, however, the ordinarily accepted similarity criteria are apparently satisfied, since the momentum transfer is solely by skin friction in a uniform channel and the Schmidt number is near unity. It is proposed that the dissimilarity between the mass and momentum transfer rates may be otherwise explained on the assumption that, in the system examined, the agents of transfer are large laminar vortices spanning the annular channel. The transfer properties of such a convective vortex system are described by elliptic partial differential equations, and are consequently influenced by the conditions at both bounding walls. On this basis it is suggested that the mass and momentum transfer characteristics measured as described in Part I are different because the boundary conditions are dissimilar; the flux at the inner (non-transfer-ring) cylindrical wall is zero in the mass transfer system and approximately equal to that the outer wall in the case of momentum transfer. Some experimental evidence is adduced in support of this suggestion, indicating that the conditions at one boundary may influence the transfer rate at the other in similar rotating coaxial cylinder systems. As a generalisation of the above, it is pointed out that natural rivers provide many notable instances in which differences appear between the (atmospheric re-oxgenation) transfer properties of apparently similar streams known to contain secondary flows. The possibility is suggested that the oxygen absorption characteristics of the beds of rivers ought to be included among the similarity criteria governing the rates of atmospheric oxygen uptake at their free surfaces. © 1996.
