We show how the spatial character of unconstrained motion in a network of bonds can be directly inferred from the topological arrangement of constraints. Relaxation time scales of these soft modes are determined, and from this information we generate spatial maps of the heterogeneous distribution of relaxation times in the disordered network. We show that the nature of the dynamic heterogeneity and its sensitivity to changes in bond configuration depends dramatically on the proximity of the system to the rigidity percolation point.dynamic heterogeneity ͉ glass transition ͉ inherent structure ͉ local mode ͉ supercooled liquid T he kinetic behavior of supercooled liquids is dominated by those configurations, known as inherent structures (1), which correspond to the local minima of the potential energy. This perspective, articulated by Goldstein (2) in 1969, is now generally accepted (3-6). The inherent structures can be regarded as zero-temperature disordered solids. The continuous transition from fluidity to rigidity that characterizes the glass transition is accomplished by the supercooled liquid sampling these disordered solids at nonzero temperatures for durations that increase as the temperature drops. Eventually, a low enough temperature is reached such that 1 local minimum persists indefinitely. To understand the kinetic and structural features of the transitions between different inherent structures we must understand the types of motion available to disordered solids. In this article we start from the proposal that constraint networks provide a general description of such solids and then develop a method for identifying the spatial distribution of the motions that remain unconstrained in such materials (i.e., the soft modes) and the spectrum of relaxation times associated with these modes.In the absence of explicit structural correlates for dynamics, the descriptive approach of dynamic heterogeneities has proved a powerful way forward in developing microscopic accounts of glassy relaxation and in articulating questions about determining dynamics from structure (7). Recently, it was shown that the spatial distribution of low-frequency quasi-localized normal modes is strongly correlated with the spatial distribution of irreversible reorganization (8). This study and related work (9-13) underline the importance of understanding the relation between local soft modes and structure in disordered solids. Understanding how features of an inherent structure determine the spatial pattern of motions is of central importance in developing a useful microscopic treatment of glass transitions and is the goal of this article.Faced with the daunting variety of particle arrangements that typically comprise a disordered solid, geometrical characterizations of such solids are difficult to both perform and interpret. An alternative treatment of amorphous rigidity is based on networks of constraints. The constraint approach to rigidity focuses on topology, instead of geometry, and the global balance between constraints and deg...