Short-term variability of plant populations within a regularly disturbed habitat

The changing populations of weeds during 13 years of the Broadbalk continuous wheat experiment were analysed to investigate the extent of differences in shortterm variability of cover between species. The data were from two sections of the experiment where winter wheat was grown continuously under herbicide treatment for 13 and 6 years respectively. Logistic regressions were fitted to the data. Equisetum arvense showed significant long-term increases on both sections; long-term trends were also detected in the longer data run for Agrostis stolonifera, Cirsium arvensa, Poa trivialis, Ranunculus arvensis and Vicia sativa, and for Medicago lupulina on the shorter data run. Variation around long-term trends was low in the case of Equisetum, and, in the longer data run, for Cirsium and Tussilago farfara, and high for Poa spp. and Vicia. Cover values on the two sections were positively correlated for Alopecurus myosuroides, Equisetum, Poa annua and Tripleurospermum inodorum. There was a weak correlation between C-S-R strategy and short-term variability; the more competitive species displayed less variability than the ruderal species. Furthermore, species regenerating from persistent seed banks were more variable in the short term than those regenerating from short-lived seed or bud banks. This can be explained by differences in response to year-to-year variation in environmental conditions, those species with persistent seed banks being typically more sensitive to annual fluctuations in the environment than those without.