The Rengen Grassland Experiment: Plant species composition after 64 years of fertilizer application

A long-term fertilizer experiment (the Rengen Grassland Experiment, RGE) was established in 1941 in the Eifel Mountains of Germany on low productive grassland naturally dominated by Calluna vulgaris and Nardus stricta. Six treatments combinations of Ca, N, P, and K fertilizer were applied annually: an unfertilized control, Ca, CaN, CaNP, CaNP–KCl, and CaNP–K2SO4. In mid-June 2005, plant cover was visually estimated and sward height was measured aiming to detect changes in floristic composition caused by long-term fertilization.Calculated by redundancy analysis (RDA), the effect of treatment was found to be a significant predictor of sward structure in the experimental area and explained 62% of cover data variability. The largest difference in vegetation structure and composition was between the treatments without and with P application. Briza media was the dominant short grass in the control, Ca, and CaN treatments. Lathyrus linifolius was the dominant legume in the control and Carex panicea was dominant in the CaN treatment. Among treatments with P application, plant species composition was similar with tall grasses such as Alopecurus pratensis, Arrhenatherum elatius, and Trisetum flavescens dominating in the sward. Sward heights were lowest in treatments without P addition in contrast to tall canopies in the P fertilized treatments. Cover of the moss layer was continuously developed in treatments without P application only and was negatively correlated with sward height as was species richness of vascular plants.Long-term fertilization caused significant diversification of plant communities. Species indicative of low productivity grasslands (short grasses, orchids, and sedges) survived in the CaN treatment but not in plots with P or K fertilizer. It was often thought that N enrichment is detrimental to vascular plant richness. However, the RGE shows that this is not necessarily so if N application is not accompanied by another limiting nutrient like P.