The dynamics of a grassland community: a simultaneous investigation of spatial and temporal heterogeneity at various scales

(1) In a permanent pasture, fine and field-scale heterogeneity were monitored at different intervals of time. For a period of two years, rooted frequency was recorded in March, May, July and October in permanent field plots at scales of 1 cm and 1 m. This gave a measure of spatial heterogeneity at those scales and simultaneously a measure of seasonal, annual and biennial temporal heterogeneity. On the larger scale of the whole pasture, changes in the cover patterns of six of the most common species were assessed over a seven-year period. (2) On the 1-cm scale, significant and consistent deviations from randomness were found, both in the spatial arrangement of species and in the patterns by which they replaced each other. Certain species, primarily the stoloniferous ones, continually colonized and left gaps in the sward more often than predicted from their abundance in the field. (3) Most species which were spatially associated also showed significant interactions in time. It is suggested that the common grasses may be divided into two guilds, depending on whether or not they are associated with Trifolium repens. (4) The probability that the same species would still occupy the same 1-cm space after one year was highly species specific. Viewed on a scale of 1 m over a one or two-year period, some species had a static, but others a highly dynamic pattern. It is concluded that the concept of a static or dynamic community pattern has no meaning unless referred to a particular species and viewed from a particular scale. Evidence for cyclic regeneration as postulated by Watt was not found. (5) While seasonal patterns of species replacements were significantly non-random, deviations from randomness were not detected on the time-scale of one or two years. This indicates that the outcome of interactions between plants is decided within each growing season and not on a time-scale of years. (6) With one exception, the small-scale distribution of plants in March and May was random but, as the growing season progressed, increasing deviations from randomness were detected. The repeated creation of the same patterns excludes the possibility that the patterns were the results of historical chance events. It is concluded that the patterns were generated and maintained by the plants themselves through species-specific interactions.