RELATIONSHIP BETWEEN PASTURE DRY-MATTER YIELD AND SOIL OLSEN-P FROM ASERIES OF LONG-TERM FIELD TRIALS

Forty-six data-sets from a series of 17 long-term field trials measuri ng the effects of rates and forms of phosphate (P) fertiliser on pastu re production and soil fertility in New Zealand were used to character ise the relationship between pasture dry matter (DM) yield in any year (expressed as relative yield, RY) and soil Olsen P test at the end of the annual DM measurement period. Average coefficients of variation i n the measurement of these parameters were 7% and 17% for DM yield and Olsen P, respectively. The results failed to demonstrate that the RY at any site in any year could be reliably assessed from a common relat ionship between RY and the Olsen P test. However, the mean relationshi p did take the general ''diminishing returns'' form as expressed by th e Mitscherlich equation, and as Olsen P levels increased the probabili ty of low RY decreased. There was no evidence that RY increased with i ncrease in Olsen P test values above 20, with the exception of the one yellow-brown pumice soil in the series. The largest component of vari ability in the RY v. Olsen P relationship occurred within years within individual sites, reflecting inherent within site variability and tre atment and measurement errors. Variability was also large between year s within sites; this appeared to arise from erratic patterns of respon se within sites over time, possibly because of climatic, botanical, an d edaphic fluctuations. The least variability in the RY v. Olsen P rel ationship arose from differences between sites. Soil pH, P retention, and rainfall did not significantly affect the relationship, but the ye llow-brown pumice site and sites with high DM production gave lower RY for any given Olsen P than other sites.