New and published data on the distribution and speciation of manganese and iron in seawater are analyzed to identify and parameterize major biogeochemical processes of their cycling within the suboxic (similar to15.6less than or equal tosigma(t)less than or equal tosimilar to16.2) and anoxic layers (sigma(t)less than or equal tosimilar to16.2) of the Black Sea. A steady-state transport-reaction model is applied to reveal layering and parameterize kinetics of redox and dissolution/precipitation processes. Previously published data on speciation of these elements in seawater are used to specify the nature of the transformations. Two particulate species of iron (Fe(III) hydroxide and Fe(II) sulfide) are necessary to adequately parameterize the vertical profile of suspended iron, while three particulate species (hydrous Mn(IV) oxide, Mn(II) sulfide, and Mn(II) carbonate) are necessary to describe the profile of suspended manganese. In addition to such processes as mixing and advection, precipitation, sinking, and dissolution of manganese carbonate are found to be essential in maintaining the observed vertical distribution of dissolved Mn(II). These results are used to interpret the observed difference in the form of vertical distribution for dissolved Mn(II) and Fe(II). Redox transformations of iron and manganese are coupled via oxidation of dissolved iron by sinking suspended manganese at sigma(t)similar to16.2+/-0.2kg m(-3). The particulate manganese, necessary for this reaction, is supplied through oxidation of dissolved Mn(II). The best agreement with observations is achieved when nitrate, rather than oxygen, is set to oxidize dissolved Mn(II) in the lower part of the suboxic layer (similar to15.90less than or equal tosigma(t)less than or equal tosimilar to16.2). The results support the idea that, after sulfides of these metals are formed, they sink with particulate organic matter. The sinking rates of the particles and specific rates of individual redox and dissolved-particulate transformations have been estimated by fitting the vertical profile of the net rate. (C) 2004 Elsevier Ltd. All rights reserved.