Astringency perception has been associated with interactions between tannins present in some foods and salivary proteins. A variety of laboratory methods to measure tannin protein interactions have been designed. Most of them, however, do not differentiate clearly between tannins and the protein fraction. The aim of this work was to characterize a method to measure tannin protein interactions by following the behavior of the protein fraction. Experiments were performed with a representative hydrophilic globular protein, bovine serum albumin (BSA), and a gallotannin-rich commercial product, tannic acid (TA). Using cellulose membranes, concentration dependency of the inhibitory effect of TA upon the diffusion of BSA on the membrane and the ability of TA to precipitate BSA were assayed. Selective staining of the protein fraction was obtained using Coomassie Blue. TA inhibited BSA diffusion in a concentration-dependent manner in the range from a 0.1-0.2 up to a 1 mu g of TA/mu g of BSA ratio. Likewise, maximal precipitation of BSA (equivalence point) occurred at a 0.55 mu g of TA/mu g of BSA ratio, that is, when about 36 representative molecules of TA (average molecular weight = 1000) interact with every molecule of BSA (molecular weight = 66000). The procedure may be readily adapted to measure interactions between different types of tannins and proteins that may be of relevance for taking decisions during food manufacturing.