Frozen saliva samples demonstrate a variable amount of precipitate on thawing depending on the type of secretion [submandibular-sublingual (SML) greater than parotid]. This precipitate has been resuspended using EDTA or removed by centrifugation by some workers and others do not mention it. Yet others collect the salivas into EDTA or centrifuge them before freezing. To determine the adsorption of proteins to hydroxyapatite, prior treatment with EDTA would be disadvantageous. The aim here was to determine if the protein pattern in parotid and SML saliva as demonstrated by sodium dodecyl sulphate gel electrophoresis is affected by the formation of precipitates. Portions of parotid and SML saliva were thawed and treated in the following ways: (a) mixed vigorously with a vortex mixer; (b) centrifuged to remove the precipitate; (c) mixed with EDTA (1 and 5 mmol final concentration for parotid and SML samples, respectively) to resuspend the precipitate. The samples were loaded on to gradient (5-20%) SDS gels and, following electrophoresis, the gels were stained with Coomassie brilliant blue R-250. The protein patterns obtained for (a) and (c) were the same. The centrifuged samples demonstrated loss of a specific band of less than 14 kDa, although this was less obvious in the parotid samples. The SML samples also showed a reduction in other lower molecular-weight proteins. This study demonstrates that precipitates in thawed frozen salivas contain specific proteins and that these samples require careful handling to avoid any alteration in the overall protein composition. (C) 2000 Elsevier Science Ltd.