Representations used in science may be separated into two classes: atomistic representations, which model discrete entities interacting through pair-wise forces, and continuous representations, which model derived properties that vary continuously through space. The choice of an atomistic or a continuous model is governed primarily by the complexity of the system: atomistic models are useful in systems with several thousand interacting entities, whereas continuous models are necessary in larger systems. In molecular biology, atomistic models are used at two levels: at the atomic level, where the atomic structure of molecules is studied, and at the molecular level, where the molecular structure of cells is studied. Continuous models are also useful at both levels, for simplifying the atomic details of large, complex molecules, and for simplifying the molecular details of entire cells.