Methods of assessing the freezing tolerance of winter cereals must be improved in order to distinguish small differences due to genotype or environment. Seed of eight winter wheat (Triticum aestivum L. em Thell) cultivars, ranging in winter hardiness, were sown either in mid-August, the first week of September or mid-September. Individual plants of each were collected in late October and stored at either −4° or −8 °C. In December controlled freeze tests, employing a cooling rate of 2 °C h−1, could not distinguish the less freezing-tolerant cultivars stored at −4 °C. However, by March the less winter-hardy cultivars from the third seeding date stored at −4 °C could be distinguished. Seedlings stored at −8 °C declined in freezing tolerance more rapidly than seedlings stored at −4 °C. In December, the less hardy winter wheat cultivars, Rose, Rita and Siouxland, were less freezing tolerant than the hardy cultivars (e.g. Norstar). Seedlings of Rita and Siouxland from the second and third seeding date died by February when stored at −8 °C. Seedlings of all winter wheat cultivars were dead by March, except Norstar and Alabaskaja, the most winter-hardy cultivars. Storage of seedlings of wheat, triticale (× Triticosecale Rimpani Wit.) and rye (Secale cereale L.) at either −12° or −15 °C readily identified the superior winter-hardy genotypes. For example, in mid-winter both Siouxland and Norstar winter wheat had a similar LT50 (temperature at which 50% of the plants are killed). However, Siouxland could not tolerate storage at either −12° or −15 °C for the same length of time as Norstar. These results support the theory that winter kill in nature is more a function of duration of exposure to sub-lethal temperatures rather than exposure to a minimum low temperature for a short duration as programmed in a conventional freeze test. A more realistic and precise freeze test would be to determine the ability of genotypes to survive lengthy exposure to sub-lethal low temperatures. Key words: Winter cereals, freezing tolerance, winter injury, duration effects, screening method