MORE THOUGHTS ON HOW TO COUNT FOSSILS

Counting the fossils contained within a bulk sample or quadrat is complicated by the fact that most species have bivalved or multi-element skeletons, which confuses the relationship between number of fossil specimens and number of individual organisms. A variety of mathematical corrections have been proposed and applied to raw counts of fossils to derive numbers of individual organisms, but a survey of recently published studies shows that none appear to be used widely or consistently. When do we need to correct raw counts of fossils and what are the appropriate corrections to use? Many paleoecological studies are comparative in that successive fossil assemblages are sampled to assess temporal and spatial stability and change in the biological community or depositional environment. Other studies are reconstructive, meaning that fossil data are used to estimate parameters of the living community (for example, predator-prey ratios). Comparative studies do not require that the raw data be corrected, however, species that contribute two or more skeletal elements to the fossil assemblage will have a statistically greater impact on the conclusions of the comparison than species that contribute a single element. This is undesirable because it creates an artificial relationship between skeletal complexity and ecological importance. Therefore, correcting the raw data to obtain the density of individual organisms in a sample is usually warranted. Only reconstructive studies require that the counts of individual organisms in the fossil assemblage be further corrected to reflect living standing crops by accounting for differences in rates of shell production and other taphonomic biases. Such corrections are discussed in the literature but not widely used because they rely on information about the life histories of the organisms that is usually unavailable. Some authors have also confused estimates of the number of unique individuals with the total density of individuals. Twenty pelecypod valves in a sample equals a density of ten organisms, even if taphonomic mixing has derived each valve from a genetically unique individual. With the increasing emphasis on establishing widely available fossil databases for research, generating data that is consistent and comparable should be a priority for paleontologists.