In the guinea pig isolated perfused trachea contracted with serosal methacholine (MCh), increasing the osmolarity of the mucosal bathing solution elicits relaxation of smooth muscle mediated by epithelium-derived relaxing factor (EpDRF). The present study was undertaken to determine whether a specific modality of the hyperosmolar stimulus induced the relaxation response. Mucosal hyperosmolar challenge with D-mannitol, N-methyl-D-glucamine (NMDG)-chloride, NMDG-gluconate (NMDG-Glu), or urea elicited relaxation with equal potency. In contrast, hyperosmolar solutions at the serosal surface induced diverse, osmolyte-specific responses. In tracheae contracted with MCh, abrupt replacement of the mucosal modified KrebsHenseleit solution (MKHS) with isosmolar osmolyte solutions to stimulate cell shrinkage elicited five discrete response patterns related to the membrane permeance of the solute, but increasing the osmolarity of the isosmolar solution via the further addition of the same solute always induced relaxation. Similarly, perfusion of the lumen with water induced a transient contraction, but subsequent addition of MKHS, or isosmolar D-mannitol, urea, NMDG-Glu, NaCl, or KCl induced relaxation. Subsequent hyperosmolar addition of the same osmolyteevoked relaxation. Compatible osmolytes had no effect on smooth muscle tone and did not affect responses to hyperosmolar challenge. The results suggest that the airway epithelium acts as an osmolarity sensor, which communicates with airway smooth muscle through EpDRF. The mechanical responses of the smooth muscle resulting from changes in the osmotic environment are associated with discrete modalities of the osmolar stimulus, including membrane reflection of the particles, incremental change in osmolarity and directionality, but not cell shrinkage.