Water has long been assumed to be essential for biological function. To understand the molecular basis of the role of water in protein function, several studies have established a correlation between enzyme activity and hydration level. While a threshold of hydration of 0.2 h (grams of water per gram of dried protein) is usually accepted for the onset of enzyme activity, recent works show that enzyme activity is possible at water contents as low as 0.03 h (Lind et al., 2004). Diffusion limitation in these experiments was avoided by monitoring enzyme-catalyzed hydrolysis of gas-phase esters. However, since water is also a substrate for the enzyme used in these experiments, they cannot be used to probe the possibility of activity at zero hydration. However, the pig liver esterase and C. rugosa lipase B are able to catalyse alcoholysis reactions in which an acyl group is transferred between an ester and an alcohol. Therefore, by following this reaction and using a gas phase catalytic system, we have been able to show that activity can occur at 0 g/g. These results led to the question of the accuracy of determinations of very low water concentrations; i.e., how dry is 0 g/g? Although gravimetric measurements of the hydration level do not allow us to define the anhydrous state of the protein with sufficient sensitivity, using 18O-labeled water, we have been able to quantify the small number of water molecules bound to the protein after drying, using a modification of the method of Dolman et al. (1997). Testing different drying methods, we have been able to determine a level of hydration as low as 2 moles of water per mole of protein (equivalent to 0.0006 h in the case of pig liver esterase) and have shown that in the case of the pig liver esterase, activity can occur at this hydration level. At the molecular level, if the hydration level affects activity, we can expect an effect on the protein dynamics. Neutron scattering spectra of hydrated powders, for instance, show that diffusive motions of the protein increase with the hydration (Kurkal et al., 2005) To address the question of the protein motions involved in the onset of enzyme activity at low hydration, we performed neutron scattering experiments on a pico-second time scale on dried powders. Preliminary results show a dynamical transition at hydration levels as low as 3 h. Molecular dynamic simulations have also been used in this study to access the dynamics of the active site. Overall, the results here show that pig liver esterase can function at zero hydration, or as close to zero hydration as current methods allow us to determine. Since the experimental methodology restricts this work to a small number of enzymes, it is unlikely that it will ever be possible to determine if all enzymes can function in the anhydrous state: however, the results here indicate that water is not an obligatory requirement for enzyme function.