It has long been assumed that unstimulated tears are more thoroughly equilibrated with epithelial secretions than stimulated tears, since they are in contact with tarsal, bulbar and corneal surfaces for longer. It was also believed from results with model solutions that soluble mucin is responsible for the observed surface tension and viscosity of tears. If longer contact means more mucin is dissolved in the aqueous tears, then the surface activity (surface tension lowered by mucin) and viscosity (raised by mucin) of tears should therefore be enhanced in unstimulated over stimulated tears. Pools of stimulated and minimally-stimulated tears were collected from a group of healthy adult volunteers by glass capillary. Viscosities were measured in the Contraves Low Shear 30 rheometer over the range of shear rates 0-130 sec-1. Surface tension was measured in the collection capillaries by a micro-technique, before and after refrigerated storage. Both surface tension and viscosity were determined for a variety of tear proteins and mucins. No significant difference was found between the viscosity/shear rate plots of stimulated and unstimulated tear samples. The viscosities of solutions of individual tear proteins were low, except for the combination of lysozyme and secretory IgA. Surface tensions were also similar in both cases, and unchanged by storage at room temperature or refrigeration, indicating no significant loss of surface-active material by adsorption on the capillary walls. Results with model mucin solutions gave a variety of results indicating either little surface activity or losses due to wall adsorption. Tear proteins, individually or in combination, did not lower surface tension to the level of tears. Tear viscosity seems not to depend on the level of dissolved mucins. This suggests either that a constant level of these is picked up even by short-term contact with ocular surfaces, or that viscosity arises from currently unknown materials which vary little with tear flow rate. This type of shear-dependent viscosity is most easily simulated in model solutions with polyionic linear macromolecules, including mucins. The contribution of individual proteins to overall viscosity is small, but combinations including lysozyme show tear-like characteristics, and may indicate that proteins whose concentration is relatively independent of tear flow rate combine with other tear components (possibly including mucins or lipids) to produce their full effect on tear viscosity. The surface tension results suggest that mucins are not of primary importance. Theories of tear film structure and performance need revision.