Résumé
The search to improve drought survival in temperate perennial grasses has led to a renewed interest in summer dormancy andhowto quantify it. This endogenously controlled trait, found in some temperate perennial grasses, is associated with drought that normally occurs in summer. While cessation of leaf growth and senescence of herbage occurs in all grasses in response to drought, it is under summer irrigation that these same responses are observed only in summer-dormant germplasm and hence the trait can be identi.ed in germplasm. Across the spectrum from completely summer-dormant to non-dormant, there is a range of expression. Our objective here is to highlight differences in characteristics of indices which measure summer dormancy and to identify aspects for incorporation into a superior index for use in measuring this trait. The experimental program comprised three .eld trials that compared 6 cultivars and a fourth that assessed a larger group of 12 cultivars of the same three species, cocksfoot (Dactylis glomerata L.), tall fescue (Festuca arundinacea Schreb.), and phalaris (Phalaris aquatica L.). Seasonal herbage yield and foliage senescence were measured under three summer watering regimes: complete drought, mid-summer storm, and full irrigation at Mauguio, France. Different indices were calculated to compare against the approach which evaluates senescence under drought. The key outcomes are as follows. (1) The assessment of summer dormancy needs to be viewed as the plant response to a period of non-limiting water supply over summer. It makes little difference whether this is produced by full summer irrigation or a mid-summer simulated storm after a drought. Assessment of this trait under conditions of unbroken drought is discouraged because it can result in false scores. (2) The determination of summer dormancy intensity under full summer irrigation is most appropriate for the intensive study of the dynamics of dormancy expression over the entire summer.Asimulated mid-summer storm within a drought gives an instantaneous view of dormancy intensity at a speci.c observation date and may be well adapted to the requirements of plant breeding. These methods are complementary. (3) Summer dormancy intensity can be assessed either by measuring herbage production or by a visual assessment of the level of herbage senescence. (4)An index of summer dormancy based on comparing irrigated summer herbage yield of any cultivar with that of a high, summer-yielding, non-dormant control cultivar was able to provide a reliable score of dormancy intensity. This index functions across a range of cultivars and species of perennial grasses. Further re.nement of the index is needed to identify ‘standard’ high and low summer-dormant populations