Two 1-year experiments are described in which a single dressing of 120 lb N/acre (134 kg/ha) was applied to a grass sward as anhydrous ammonia (82% N), ammonium nitrate (34·5% N) or sulphate of ammonia (21% N) at various dates from December to April. The pattern of primary growth from each of the treatments was recorded from early April until mid-June by frequent sampling of the herbage to determine the dry matter and crude protein yields. At every date of fertilizer application from late January to mid-April the growth rate of the herbage was markedly lower with anhydrous ammonia than with ammonium nitrate. With the December applications the spring growth rate of herbage from the anhydrous ammonia treatment was similar to that from the ammonium nitrate treatment. Applications of sulphate of ammonia gave results almost identical with those obtained from ammonium nitrate. The calculation of the dates on which a yield of 2000 lb dry matter/acre (2242 kg/ha) would have been available for dairy cow grazing showed clearly a delay of 12–15 days where anhydrous ammonia rather than ammonium nitrate was applied. The response to anhydrous ammonia was improved by applying it early in the year but solid nitrogen fertilizers, applied either early in the season or when growth was starting, gave higher yields of herbage than anhydrous ammonia.