The search for the mechanisms that drive the species–area relationship (SAR) has been complicated by a number of factors. First, the different proposed mechanisms have generally been considered to be mutually exclusive. Second, due to the misinterpretation of curve shapes after (log) transformation of one or both of the axes. Third, due to confusion about different types of SAR, such as isolate and sample area SARs. These two SAR types are underpinned, at least partly, by different factors (or drivers), which act in combination. The different patterns and processes that drive the SAR can be organized according to a hierarchy: from underlying environmental patterns at the bottom, through the processes driving species diversity patterns, up to the research and sampling design. Environmental patterns include both biological and geographical patterns, while processes may be stochastic, evolutionary or ecological. The resultant species diversity patterns include species range patterns, species densities, species abundances and the spatial distribution of individuals of a species. Although the individuals of a species are rarely randomly distributed in space, the random placement model provides a useful null model that can be used to partition the effects of species abundances and the spatial distributions of individuals on SARs.