The aim of the present study was to compare fluoride uptake of dentin with and without simulating dentinal fluid flow. Eighty-five dentinal discs were prepared from retained human molars. Seventeen discs were used to assess baseline fluoride content (controls). Sixty-eight discs were evenly distributed among two experimental groups and were fixed for 24 h in a two-chamber device allowing for simulating dentinal fluid flow. Thirty-four samples were treated with dentifrice slurry containing a low fluoride concentration (Elmex, dilution 1:5); the remaining discs were treated with a solution of a high fluoride concentration (Elmex Fluid). For the perfused group, each of the 17 discs of the high and low concentration fluoride groups were constantly perfused during the experiment. The remaining fluoridated specimens were not perfused. All specimens were assayed for KOH-soluble fluoride and structurally bound fluoride. Structurally bound fluoride was determined in three successive layers of 20 microns each. Uptake of KOH-soluble fluoride was significantly higher in the specimens fluoridated with Elmex Fluid compared to the samples treated with the dentifrice slurry. The uptake of KOH-soluble fluoride was not influenced by perfusion of the dentinal discs. However, perfusion resulted in lower acquisition of structurally bound fluoride in the samples treated with the dentifrice slurry. After application of Elmex Fluid, no significant difference was found in the amount of structurally bound fluoride between the perfused specimens and non-perfused samples. It is suggested that further studies on fluoride uptake in dentin should use a model simulating outward dentinal fluid flow.