Background and Aims Dormancy has been extensively studied in plants which experience severe winter conditions but much less so in perennial herbaceous plants that must survive summer drought. This paper reviews the current knowledge on summer dormancy in both native and cultivated perennial temperate grasses originating from the Mediterranean Basin, and presents a unified terminology to describe this trait. Scope Under severe drought, it is difficult to separate the responses by which plants avoid and tolerate dehydration from those associated with the expression of summer dormancy. Consequently, this type of endogenous (endo-) dormancy can be tested only in plants that are not subjected to moisture deficit. Summer dormancy can be defined by four criteria, one of which is considered optional: (1) reduction or cessation of leaf production and expansion; (2) senescence of mature foliage; (3) dehydration of surviving organs; and (4, optional) formation of resting organs. The proposed terminology recognizes two levels of summer dormancy: (a) complete dormancy, when cessation of growth is associated with full senescence of foliage and induced dehydration of leaf bases; and (b) incomplete dormancy, when leaf growth is partially inhibited and is associated with moderate levels of foliage senescence. Summer dormancy is expressed under increasing photoperiod and temperature. It is under hormonal control and usually associated with flowering and a reduction in metabolic activity in meristematic tissues. Dehydration tolerance and dormancy are independent phenomena and differ from the adaptations of resurrection plants. Conclusions Summer dormancy has been correlated with superior survival after severe and repeated summer drought in a large range of perennial grasses. In the face of increasing aridity, this trait could be used in the development of cultivars that are able to meet agronomic and environmental goals. It is therefore important to have a better understanding of the genetic and environmental control of summer dormancy.