Frost hardiness in Calluna vulgaris (L.) Hull, which had received ammonium nitrate applications in the field for 30 months, was assessed using scores of visible injury and measurements of the rate of total electrolyte leakage from excised shoots following controlled freezing treatments in the laboratory. There was good overall correlation between the two methods (Spearman correlation coefficient 0.77), but leakage measurements were more sensitive than injury scores to the effects of nitrogen. Visible injury was not significantly altered by nitrogen supply (Kruskal-Wallis non-parametric test). Ion leakage was analyzed in different ways, using either calculations of the first-order rate coefficients or expressions of relative conductivity. These analyses produced similar results with respect to the effect of frost and nitrogen. Shoots of nitrogen-fertilized (40, 80 and 120 kg ha−1 yr−1) C. vulgaris sampled in October 1991 showed significantly (P < 0.05) less leakage after overnight frosts of −15 and −20 °C than did the water-treated controls. In October the temperature which killed 50% of the shoots (LT50), derived from the leakage data, was raised by at least 4 °C by the highest nitrogen treatments compared with the control plants. Frost treatments to pot-grown C. vulgaris in November produced similar visible injury to attached and excised shoots from the same plants, both being significantly less damaged by a −18 °C frost after a 7-month exposure to an NaNO3 mist solution (1.0 mM, pH 4.5) than were water-misted controls. Ammonium-mist treatments also marginally reduced frost injury, but the effects were not statistically significant. These results demonstrate that frost hardiness of C. vulgaris in the field can be assessed rapidly and accurately in the laboratory by analysis of electrolyte leakage from excised shoots.