Two conflicting hypotheses have been proposed to explain large-scale species diversity patterns and dynamics. The unbounded hypothesis proposes that regional diversity depends only on time and diversification rate and increases without limit. The bounded hypothesis proposes that ecological constraints place upper limits on regional diversity and that diversity is usually close to its limit. Recent evidence from the fossil record, phylogenetic analysis, biogeography, and phenotypic disparity during lineage diversification suggests that diversity is constrained by ecological processes but that it is rarely asymptotic. Niche space is often unfilled or can be more finely subdivided and still permit coexistence, and new niche space is often created before ecological limits are reached. Damped increases in diversity over time are the prevalent pattern, suggesting the need for a new ‘damped increase hypothesis'. The damped increase hypothesis predicts that diversity generally increases through time but that its rate of increase is often slowed by ecological constraints. However, slowing due to niche limitation must be distinguished from other possible mechanisms creating similar patterns. These include sampling artifacts, the inability to detect extinctions or declines in clade diversity with some methods, the distorting effects of correlated speciation-extinction dynamics, the likelihood that opportunities for allopatric speciation will vary in space and time, and the role of undetected natural enemies in reducing host ranges and thus slowing speciation rates. The taxonomic scope of regional diversity studies must be broadened to include all ecologically similar species so that ecological constraints may be accurately inferred. The damped increase hypothesis suggests that information on evolutionary processes such as time-for-speciation and intrinsic diversification rates as well as ecological factors will be required to explain why regional diversity varies among times, places and taxa.