Species selection, differential rates of speciation or extinction resulting from species level characters, is often invoked as the main mechanism of macroevolution that is not simply an extension of microevolutionary processes. So long as we are careful in defining "species", the logic of species selection is sound. This does not mean, however, that this process can influence evolutionary dynamics under realistic conditions. The principal challenge to the efficacy of species selection as an evolutionary mechanism is the idea that selection between individuals within species will be so much more efficient as to swamp out any effects of selection between species. To assess this, a genetic model is constructed that includes simultaneous selection within and between species, and this is used to ask: under what conditions could species selection influence evolutionary dynamics, even in the face of opposing selection between individuals? The model shows that the efficacy of species selection is strongly determined by the time between speciation events (measured in individual generations), the mutation rate of the character under consideration, and the initial size of a newly formed reproductively isolated population. Data indicate that a few studied lineages have shown sufficiently high speciation rates to make species selection an important mechanism in the evolution of characters with mutation rates on the order of 10(-6) per generation. Quantitative characters, such as body size, generally change too readily for species selection to be relevant to their evolution. Complex characters, however, may be good candidates to be influenced by species selection. The interaction of selection within and between species can be subtle, with individual selection looking, from the standpoint of a species, very much like development of an individual. Furthermore, selection between individuals may be the main process assembling complex adaptations, while species selection allows them to persist over long periods of time.