SUMMARYThree methods for testing additives for their ability to prevent moulding of hay are described. Storage of 25 g samples of hay in glass jars at 25oC was useful for screening many chemicals rapidly. Dewar flasks (4–5 1) allowed testing on 500 g hay samples with sufficient insulation to allow any effects of chemicals on spontaneous heating as well as on moulding to be measured. Large drums containing 20 kg hay were used for larger scale tests with partially dried fresh, or rewetted old hay. Proprietary additives were unsuccessful at their recommended application rates and some even failed to show antifungal activity with very much larger doses. Volatile fatty acids, particularly propionic acid, and ammonium propionate were the most effective chemicals tested. The amount of propionic acid needed to prevent moulding and spontaneous heating could be defined in terms of the amount of water in the hay. Where the ratio of propionic acid to water was greater than 1–25:100 heating was rare. Ammonium propionate was slightly less effective than propionic acid against moulding. However, it lacks a pungent odour, is less volatile, less corrosive and is more pleasant and safer to han