Abstract
3 m of annually laminated sediments were recovered from a small steep-sided lake, Diss Mere, Norfolk, UK. In an 8 cm section, known to include the mid-Holocene 'elm decline', the laminae were sampled separately, as far as possible, resulting in 181 samples covering 252 years. Relative pollen percentages, pollen influx, and charcoal were analysed. They show evidence of two phases of human disturbance within the woodland around the Mere, the first starting at least 160 years before the 'elm decline', and involving cereal growth. At the 'elm decline', Ulmus pollen values fall by 73% over six years. This is comparable in rate and magnitude with the fall, as recorded palynologically, in Castanea dentata from North America known to have been caused by a pathogen earlier this century, with the mid-Holocene Tsuga canadensis decline also attributed to a pathogen, and with the present-day Ulmus decline caused by Dutch elm disease at Scords Wood, England. Exponential and logistic models of population change are applied to the Diss Mere Ulmus decline data. The form, magnitude, and rate of decrease of the Ulmus decline at Diss Mere are all consistent with the hypothesis of a pathogenic attack. All the available evidence from Diss supports the hypothesis that the mid-Holocene 'elm decline' was caused by a combination of disease and human activity.