Restoration ecology would be easier in a world of linear, deterministic, ordered, predictable change tending toward stable equilibria. In such a world, many restoration projects would require only that the restorationist give a degraded or damaged ecosystem an initial push, and then stand back and watch the system heal itself. But this is not the world that most ecologists believe we inhabit (Botkin 1990; Wu and Loucks 1995). Contemporary ecology describes a world characterized largely by nonlinear, stochastic, imperfectly predictable processes where historical contingencies, spatial context, and initial conditions are strong determinants of change following perturbation, and in which equilibria, if they exist at all, are likely to be unstable (Maurer, Menninger and Palmer, Suding and Gross, this volume). Contemporary ecology sees constant interactions between intrinsic or endogenous dynamics (for example, population cycles) and a nonstationary physical environment with multiple frequencies and amplitudes of change. What we now understand about climate variability suggests that the physical environment is nowhere near as stable—even on “ecological” time scales—as was once supposed (Cayan et al. 1998; McCabe et al. 2004; Millar and Brubaker, this volume). Indeed, ecological and evolutionary adaptation to spatial and temporal variability is a powerful new line of ecological inquiry (Chesson 2000; Clauss and Venable 2000; Reed et al. 2003). These emerging views of how the world works pose a fundamental challenge for restoration ecology (Pickett and Parker 1994; Hobbs and Norton 1996; Anand and Desrosiers 2004): Given that ecosystems are in a constant state of dynamic flux, what state should be restored? The contributors to this volume offer some novel and important answers, if only as working hypotheses. On the whole they emphasize ecological processes that underlie the visible composition and structure of ecological communities. Although “saving the parts”(Leopold 1953) is often used as shorthand for restoration, restoration ecology shows that how the pieces are assembled, and how they work together, are at least as critical (Naeem, this volume). Retaining all the individual components (species) of communities and ecosystems remains important, however. Restoration is becoming more attuned to underappreciated keystone functional groups, such as soil microflora and microfauna, cryptobiotic crusts, and dispersal agents. Uncommon and rare species may also play unknown ecological roles at small spatial scales. Nonetheless, there is a world of difference between having all the parts of an automobile laid out neatly on the garage floor and an assembled machine that can take you