The representative quality of fossil diatom assemblages in the recent sediment of a lake is compared with its contemporary diatom flora. In April 1986 experimental liming of the catchment of a small acidified lake, Loch Fleet (Galloway, U.K.), produced immediate changes in water quality. Lakewater pH rose from a mean of approximately 4.5 to 6.5, and in the two year period following liming a consistently higher pH was maintained. The marked response of diatom species to changing water quality provided a means of tracing events from living communities to the fossil assemblages. Diatom periphyton and plankton were sampled during a 20 month period and archived material was used to characterise earlier diatom communities. A comparison is made between living diatom communities and diatom assemblages collected by sediment traps and from sediment cores taken during the same period. Following liming, the diatom communities were found to respond within days or weeks to the changes in water quality. There is an initial change from acidobiontic communities, dominated by Tabellaria quadriseptata, to dominance by the acidophilous species Eunotia incisa and Peronia fibula. However, in the epipsammic community the acidobiontic species Tabellaria binalis fo, elliptica remains abundant after liming. Approximately one year after liming the abundances of species such as Achnanthes minutissima and Brachysira vitrea increase in the epilithon, epiphyton and epibryon, whilst in the epipsammon T. binalis fo. elliptica is replaced by small Eunotia spp. and Achnanthes altaica. During the latter part of 1987 and in 1988, despite a stable pH, fluctuating patterns of species abundances are seen in the epilithon, epiphyton and epibryon whilst the species composition of the epipsammon remains relatively stable. Spring blooms of the planktonic species Synedra acus and Asterionella formosa occur during 1988 and 1989 respectively. Sediment trapping, which began in April 1987, records shifts in species composition corresponding with those seen in the epilithon, epiphyton and epibryon and with the blooms of planktonic species. The signal from the smaller, and probably less easily transportable, epipsammic community is not so clearly discernible. Although the fundamental record of the sediment traps is one from living diatom communities, the appearance of taxa 'extinct' during the post-liming period reflects a low, but significant level of sediment resuspension. In contrast to the rapid response of living communities and their record in sediment traps, sediment cores do not begin to reflect changes in diatom composition until about 14 months after the initial liming. The first appearance of circumneutral taxa in significant abundance occurs only approximately 17 months after liming. The delayed reaction of sediment assemblages cannot be attributed principally to a slow rate of transport from the littoral to the profundal zone. Time-averaging processes within the sediment appear to be the main cause of the lag in core response. In contrast, blooms of planktonic species are quickly reflected in the stratigraphy of cores, but indicate that a considerable degree of downward mixing occurs. Comparison of the time trajectories of whole species assemblages in living communities, sediment traps and core surface sediments shows that the direction of change is similar in all three, but that the magnitude of change is attenuated in sediment assemblages.
