Evidence is reviewed that freezing and dehydration are fundamentally different stress vectors: (a) Proteins, membranes, phospholipids, and living cells and organisms all contain about 0.25 g nonfreezable H2O/g dry weight. By definition, this H2O is not removed by freezing. (b) Dehydration, by contrast with freezing, can remove the nonfreezable H2O. Removing this H2O results in profound changes in the physical properties of biomolecules, particularly phospholipids and proteins, (c) The mechanisms of preservation of proteins during freezing and drying are completely different. The specificity for solute requirements for stabilization of proteins during freezing is low; any solute that is preferentially excluded from the hydration shell of a protein is also a cryoprotectant. (d) By contrast, stabilization of proteins during drying requires direct interaction between the stabilizing molecule and the protein, probably involving hydrogen bonding between the stabilizer and polar residues in the protein. The specificity is very high in this case; only carbohydrates are effective, and of those that have been tested trehalose is the most effective, (e) Less is understood about the mechanism of stabilization of phospholipid bilayers during freezing, but it is clear that while many solutes will preserve liposomes during freezing, only a few (of which trehalose is the most effective) will preserve them during drying. Stabilization of bilayers during drying requires direct interaction between the sugar and polar head groups of the phospholipids. © 1990.