Experimental chambers were used in a Virginia salt marsh to partition the tidal flux of dissolved nutrients occurring at the marsh surface and in the water column. On five dates from June to October 1989, six replicate chambers in the short Spartina alterniflora zone were monitored over complete tidal cycles. When reservoir water, used to simulate tidal flooding in the chambers, was initially low in dissolved nutrients, the marsh surface was a source of both ammonium and phosphate to the water column. Calculations of the physical processes of diffusion and advection could not account for total nutrient release from the marsh surface. We hypothesize the primary source of nutrients was organic matter mineralization in surface sediments, which released nutrients into the flooding water column. Assimilation (uptake) of phosphate measured in water-column incubation experiments was nearly equal to phosphate released from the marsh surface. Surface release of ammonium, however, was somewhat greater than water-column uptake. In this salt marsh, benthic production and release of ammonium and phosphate is comparable in magnitude to pelagic consumption, thereby yielding only a small “net” transfer of these nutrients to the estuary.