Published online by Cambridge University Press: 08 April 2016
The degree to which the relative abundances of the species in a fossil assemblage represent those in the living community from which it came is an indication of the fidelity of the assemblage in preserving most of the aspects of community structure that we may hope to find in the fossil record. I have developed a model for the formation of attritional fossil mammal assemblages that shows that the logarithm of the number of individuals of each species in an ideal unbiased assemblage, when regressed on the logarithm of body mass, will exhibit a slope of approximately −1.05 ± 0.25 (approximate 95% confidence interval). This value is based upon ubiquitous scaling relationships of population density and turnover rate with body mass among extant mammals. These relationships are the results of widespread energetic and metabolic regularities in community structure and physiology and can be expected to have been essentially the same in the remote past. The slope of −1.05 provides for the first time a reference standard for community structure of sufficient generality for use in the analysis of this aspect of fossil assemblages. We can use this value as a sort of “null hypothesis” of no bias against which we can compare actual fossil assemblages.
Analysis of some promising assemblages reveals that they exhibit badly biased relative abundances. However, it is likely that in these cases we can explain this bias as almost entirely the result of pre-burial differential destruction of the carcasses of different-sized species.
The combination of this new, biological approach with the largely physical approaches of taphonomy and sedimentology will allow us to ask more specific questions about bias in fossil assemblages, avoiding circular arguments. We can now distinguish, in a large number of situations, assemblages that preserve reliable community-structure information from those that do not.