Prediction of yield in the Rothamsted Park Grass Experiment by Ellenberg indicator values

Abstract

Abstract. The Rothamsted Park Grass Experiment was established in 1856, with experimental plots subjected to annual applications of fertilizer and twice‐yearly cutting of hay. There were two major responses to fertilizer, one reflecting high ammonium‐nitrogen and increased acidity and the other reflecting high herbage yield without increased acidity. We calculated mean Ellenberg indicator values for N (nitrogen) and R (soil reaction) for the hay harvested between 1948 and 1975, using both unweighted and abundance‐weighted means. Plot Ellenberg values were compared with herbage yield and with fertilizer application rates and published soil data. Annual yield of hay varied from 1.5 to 7.4 t/ha and was well predicted by the unweighted mean Ellenberg N‐values ( r = 0.91). Relatively large negative residuals from the relationship were found in plots whose soil combined low K and low pH. Soil pH was poorly predicted by the unweighted mean R‐value, but showed a moderately good relation with weighted mean R ( r = 0.73). The fact that Ellenberg N‐values correlated better with yield than with applied nitrogen suggests that they might rather be called productivity values.