An island is shown with three different habitat types: grass, bluebells and boulders/rocks. The first premise is that response rates to playback might vary across these habitat types (or some other known variable). This could be because sound permeates the habitat differently, or because birds differ in condition etc. The reason is, however, of secondary importance. Of primary importance is measuring response rates independently for areas where there are biological reasons to think they might vary, or randomly selecting areas to measure response rates for. The multi-visit calibration areas aim to do this, providing sites are chosen properly. At these sites, repeated visits to representative plots will allow a habitat specific/ sub-colony specific response rate to be calculated. Then survey plots (the single playback survey squares) can receive single-visit playback and using the response rate suitable for that habitat, or sub-colony, a number of AOS can be estimated for the single-visit survey plot. This can then be extrapolated to the area that is the same habitat to give a global estimate, and this can be summed across habitat types.