In order for either lipid bilayer membranes or biological membranes to undergo fusion, stresses must somehow be generated in the region of membrane contact. In the fusion of phospholipid vesicles with planar bilayer membranes, the stress can be produced by osmotic swelling of vesicles contacting the planar membrane. On the other hand, fully swollen vesicles may be sufficiently stretched that the additional stress experienced from their adhesion to the planar membrane may in itself suffice to produce fusion (see also Rand & Parsegian, this volume). There is considerable circumstantial evidence that osmotic swelling of vesicles may also be a driving force in exocytosis. This evidence centers both on experiments demonstrating inhibition of exocytosis when vesicles are in a hyperosmotic medium, and on observations of vesicle swelling during the secretory process. This article has not reviewed all of the examples in the literature supporting an osmotic mechanism for fusion, but has attempted to suggest the diversity of cell types from which the examples are drawn and to indicate that the evidence is not conclusive. The unambiguous establishment of vesicle swelling prior to fusion would go far in establishing an osmotic mechanism of exocytosis. We must also be prepared to find that osmotic swelling may not be the only biological mechanism of stressing vesicle membranes contacting plasma membranes. The viral membrane fusion proteins provide the precedent for agents that can apparently sufficiently perturb membranes to cause fusion without any additionally imposed stresses, and even direct membrane mechanical stretching may act biologically as a fusogenic stress.