The evidence for most benefits is correlational, and although there are several experimental studies, little as yet is known about the mechanisms that are important for delivering these benefits. We do not know which characteristics of natural settings are most important for triggering a beneficial interaction and how these characteristics vary in importance among cultures, geographic regions and socio-economic groups. These are key directions for future research if we are to design landscapes that promote high quality interactions between people and nature in a rapidly urbanising world.