BEYOND THE VEIL: USING MUSEUM COLLECTIONS TO EXTEND FIELD-BASED ESTIMATES OF SPECIES RARITY

Species abundance affects ecological interactions and extinction vulnerability. Yet attempts to quantify abundance in modern and fossil communities are complicated by sampling biases. Rare taxa are, by definition, encountered infrequently and their observed occurrences strongly controlled by sampling effort. This phenomenon has been described as a veil behind which rare species are concealed from observation. The number of species concealed is controlled by sample size and the shape of the underlying abundance distribution. I propose a method for sampling behind the veil through the integration of fossil species occurrence data archived in museum collections and quantitative estimates of species abundance gathered from replicate bulk samples. The occurrences of rare species in museum collections, often assembled to maximize taxonomic coverage, can be used to augment the abundance distributions observed in bulk samples. In a given sample, the lowest proportional abundance value observed can be used to estimate the maximum abundance of rare species known to have occurred at the site at least once but not found in that collection of individuals. Assuming a lognormal abundance model, the number of taxa unveiled for a given increase in sampling effort can be estimated and compared with the species richness observed in museum collections. Combining these two sources of data can provide a more comprehensive estimate of abundance and taxonomic richness without substantial increase in current sampling effort, thereby expanding the scale of abundance and the sample size of species that can be included in paleoecological and evolutionary analyses. The Paleogene fossil record of the Gulf Coastal Plain is ideal for applying this method, as the molluscan fauna is well represented in museum collections, has been the subject of taxonomic standardization, and many historical sites remain accessible thus facilitating resampling using standardized, quantitative approaches.