Most types of behaviour, from muscle contraction to conscious thought, are mediated at the cellular level between thousands, if not millions, of cells within a single biological organ. Technological advances over the next decade will make it feasible to simulate these interactions on a computer, providing an invaluable tool for predicting how an organ behaves when presented with particular stimuli. Finite-element modelling techniques are particularly suited to this task, since, by dividing a system into a large number of small elements, they mimic the physical reality by which cell interactions, even over large distances, result from a large number of localized interactions between adjacent units. Finite-element techniques have been used in engineering for some time: and they are already being applied to a variety of biological organs. One example is the mammalian cochlea, where sound is transformed into electrical signals that are subsequently passed to the auditory nerve. The cochlea contains an amplifier of mechanical motion that operates on a microsecond time-scale at sub-nanometre displacements, and it enables the auditory system to respond over a dynamic range in excess of 120 dB. A simple finite-element model that represents the cochlea at a cellular level has already demonstrated the potential value of this approach by providing an explanation for contradictory experimental observations. Developing structurally realistic cell-level models of biological organs will improve our ability to properly characterize and quantify experimental observations, and dramatically reduce the need for animal experimentation. The finite-element approach could also provide a valuable tool in the design of new, simpler, cellular structures that would mimic the known operation of a biological organ. Given the impressive specifications of such organs, these new devices-manufactured in carbon or silicon-could have numerous research, clinical and industrial applications in the new millennium.