A new method for the estimation of absolute microfossil numbers, with reference especially to diatoms

A new method for obtaining permanent quantitative slides of microfossils involves the use of an evaporation tray in which microfossil suspensions can be sedimented randomly onto cover slips on the tray floor. Slides are made after evaporation. The technique is statistically reliable. The results of diatom analyses from lake sediments are usually presented in the form of percentage diagrams in which the variation in the percentage frequency of individual taxa or groups of taxa are shown for successive levels in the sediment (Nygaard 1956; Behre 1962). These percentage analyses are easy to make, and the diagrams often give a good picture of the changing structure of fossil communities. However, they allow no estimation of the absolute abundance of diatoms in the sediment, and their interpretation must take into account the constraints of using a constant sum (Chayes 1960). In response to the latter problem, pollen analysts have devised methods for estimating total numbers of pollen grains per sample (Davis 1965; Benninghoff 1962; Jergensen 1967) and have expressed the results after correction for the rate of sediment deposition in terms of pollen grains deposited per square centimeter per annum ( Davis 1967). In paleolimnological studies it is also useful to express results in this way, to obtain information on the paleoproductivity of lakes and to make comparisons with the quantitative results of limnological analysis. The history of lake eutrophication provides especially apposite conditions for this approach. I wish to thank Prof. F. Oldfield for reading the manuscript. The work was carried out while I held a research studentship at the New University of Ulster, Northern Ireland. PROBLEMS OF THE PIPETTE METHOD The difficulty of adapting existing absolute pollen techniques to the characteristics of diatoms has been one of the reasons why diatomists have rarely attempted absolute counting. The aliquot method of Davis ( 1965)) for example, is unsuitable for diatoms because it is necessary to count the total number of individuals over the whole of a cover slip. Round ( 1957, 1961) used an aliquot method for diatom counts and attempted to surmount the counting problem by only counting along sample diameters of circular cover slips. However, both Eaton and Moss ( 1966) and Battarbee (in prep.) have shown that diatom samples pipetted onto a cover slip follow a distinctly nonrandom distribution after evaporation, due to the effects of surface tension. The pattern of diatoms across a circular cover slip prepared in this way is shown in Fig. 1, in which the number of valves counted for each 3-mm interval along diameter traverses of a series of cover slips is plotted in bar diagrams. Comparison of horizontal and vertical traverses for each cover slip shows that the shape of the bars mainly reflects the shape of the meniscus of the original sample before evaporation and, to some extent, the tendency for the water to evaporate concentrically, pulling the diatoms toward the center. If the total numbers for the whole cover slip are calculated from diameter traverses, the results will usually be a large and variable overestimate (Table 1) since the ratio of high to low density areas is higher for the sampled area than for the cover slip as a whole. The nonrandom nature of the distribution is also seen in Table 2a, where it is compared with a Poisson series. EVAPORATION TRAY TECHNIQUE The problem is solved if it is possible to prepare cover slips on which valves are LIMNOLOGY AND OCEANOGRAPHY 647 JULY 1973, V. 18(4)