THE INFLUENCE OF PRESENCE-ABSENCE VS. RELATIVE ABUNDANCE DATA ON INTERPRETATIONS OF FAUNAL STABILITY

Since its introduction by Brett and Baird (1992), coordinated stasis has intrigued paleoecologists as a pattern with potential implications for ecological and evolutionary theory. Over the last twelve years, many investigators have tried to identify the pattern in a variety of paleocommunities, only to find the results confusing and controversial as to whether coordinated stasis truly exists. One of the major obstacles in evaluating this pattern is the inconsistency of selected data format (presence-absence vs. relative abundance) and statistical methodology used by researchers, resulting in a broad spectrum of interpretations ranging from stable to non-stable. The goal of this research effort is to examine the effects of using presence-absence vs. relative abundance data on the outcome of hypothesis-testing for faunal stability.

An exposure of the early Oligocene Byram Formation near Vicksburg, MS, provides an ideal setting in which to test for faunal persistence through time. The outcrop exhibits an alternating pattern in lithology such that very similar fossiliferous facies reappear three times within the stratigraphic section. Ten replicate samples were collected along strike from each recurring horizon, allowing for multiple comparisons in space and time. Absolute abundance counts were converted to presence-absence and relative abundance data for calculation of pair-wise similarity coefficients, specifically the Jaccard and Bray-Curtis. When using presence-absence data, cluster dendrograms, multi-dimensional scaling plots, and ANOSIM (analysis of similarities) test results all indicate faunal persistence through time. However, when proportional abundance data are employed in each of these statistical analyses instead, all conclusions point towards non-stability through time. Our results support the presence of coordinated stasis in both guild and taxonomic composition, but not in abundance distribution. Since these paleocommunities can be regarded as both stable and non-stable based on the choice of presence-absence or relative abundance data, perhaps the next issue to be addressed is the type(s) of information provided by proportional abundance data and how they contribute to discussions of faunal composition in both space and time.