The classic gap hypothesis suggests that canopy gaps promote forest diversity by favoring the regeneration of tree species with high light requirements (i.e., shade intolerant species).  However, whether gaps act to maintain forest diversity remains an unanswered question.  We argue that canopy gaps can promote coexistence between tree species through the ‘storage effect.’  Certain combinations of environmental conditions and species’ traits can advantage species when they are rare and promote their long-term persistence in the community through the ‘storage effect.’  These combinations are likely to be found in canopy gaps.  For example, gaps (and subsequently light environments) vary in both space and time and tree species appear to be differentiated in terms of their ability to tolerate shade and respond to increases in light.  Furthermore, when gap environments are favorable to abundant species their ability to increase in dominance is likely limited (relative to rare species) because of increased within-species competition (i.e., covariance of environmental favorability and competition).  Finally, rare species will likely persist even when favorable gap conditions are infrequent because adult trees are long-lived.  We evaluate the ‘gap storage hypothesis’ with quantitative analyses that incorporate realistic patterns of variation in light environments (i.e., gaps) and trade-offs between species ability to tolerate shade and grow in high light.  We show that species differentiated along this trade-off may coexist via the ‘storage effect.’  We conclude that the ‘storage effect’ may act to maintain forest diversity because of canopy gaps.