Ecosystems perform numerous functions and provide a wide variety of services.  Effective restoration often requires that we restore as many of these functions and services as possible.  Maximizing the range and magnitude of functions and services requires that restored communities fill all available niches.  We combine ideas about functional diversity with those of habitat filtering to quantify the proportion of trait space occupied by a given set of species.  Advancing prior work using convex hulls, we employ a hulls-within-hulls approach that sums intraspecific hulls, and thus trait space, as a proportion of the total available trait space (the total hull volume of intact communities or of the candidate species pool).  We then demonstrate how this approach can be used to select the minimum number of species required to fill the trait space of a given community and restore a complex set of ecosystem functions and services, using a model system based on restoration of abandoned agricultural land in a landscape formerly dominated by tall-grass prairie.  We find that while a reduced set of species may fill the available functional trait space, this may come at the price of both reduced redundancy and limited ability to respond to environmental change.