ALBERT

Description: SummaryField plot experiments were carried out for 3 years at four sites to study the effect of date of application of ammonium nitrate/calcium carbonate (CAN) and urea (U) on perennial ryegrass production in spring. Fertilizer (70 kg N/ha) was applied at weekly intervals for 10 weeks from 1 February. Herbage was cut on the same day at all sites, 3–4 weeks after the last N application. CAN and U at 50 kg N/ha were immediately re-applied and a second cut of herbage was taken after 4–5 weeks. From meteorological data, the dates after 1 February when soil temperature at 100 mm depth increased to 5·5 °C and the dates when cumulative average daily air temperatures 〈 0 °C from 1 January reached 200 °C were calculated for each site and year.The date of application for maximum dry matter (D.M.) yield at the first cut differed with site and year, but for 11 of the 12 site/years was in February. The number of dates of application resulting in optimum yield (i.e. at least 90% of the average maximum yield response to CAN or U) varied also with site and year from one to six. Soil and air temperature predictive systems implied a precision in choice of application date that was unjustified and were no more successful at predicting the optimum application date than a simple date range. The first fertilizer application had a pronounced positive residual effect on D.M. yields at the second cut. Less precision on date of the first application was required to obtain optimum cumulative yields over both cuts than optimum yield at the first cut only.Differences in performance between CAN and U were only significant for three of the 120 fertilizer applications at the first cut. On these occasions, all in one year at two sites, U gave higher yields than CAN. Correlations were sought between D.M. yield response and growth period, air temperature, long-term rainfall and short-term rainfall for CAN and U separately. Factors relating to rainfall had no significant effect on response to U but response to CAN showed a significant negative correlation with short-term rainfall. The short-term weather forecast may therefore be another criterion to be considered in deciding when to apply N in early spring.