A number of optimization models have been developed for natural reserve design and reserve site selection. The most common approach seeks to maximize the number of individual species that occur among chosen sites. A number of heuristics and mathematical programming algorithms have been applied to solve this problem. While attaining maximum overall species representation is important, the relative quality of representation (which could be affected by site attributes such as habitat value, adequate population size, presence of critical resources, existence (or lack thereof) of exotic competitors, etc.) has been absent from most representation models. Yet issues of site quality should be considered in order to have any assurance of long-term species persistence in a reserve system. Here we present a multiobjective optimization model that addresses the issue of balancing species presence with habitat quality. One type of interesting alternative yields more high quality representation at the price of some reduction in overall representation. We present an application using a large dataset from California Gap Analysis to demonstrate this and other tradeoffs. Optimal solutions are attained using commercial integer programming software with reasonable computational effort.