AR T ICLES Phenotypic Evolution in the Fossil Record: Numerical Experiments
Bjarte Hannisdal · 2006
Stratophenetic data document phenotypic changes in a fossil lineage and play a vital role in reconciling contemporary microevolution with long-term paleontological patterns. However, stratophenetic series represent multiscale geological and biological interactions, defying simple analysis and interpretation. A numerical model is presented that simulates stratophenetic series in shallow marine siliciclastic deposits. The model is driven by predictions of water depth, substrate properties, and sedimentation rate from a high-resolution sedimentary basin fill model. Species abundance is modeled as a probability density peaked with respect to environmental preferences. The Price equation is used to model phenotypic evolution based on phenotype-fitness covariance and drift. Preservation is a Poisson process controlled by population size, preservation probability, and sedimentation. Numerical experiments are used to investigate (1) the effects of sampling and depositional architecture on observed patterns and (2) the performance of various statistical tests in identifying evolutionary mode. As sample sizes decrease, the inaccuracy of sample mean values causes a stratophenetic pattern referred to as analytical stasis. Depositional architecture can cause nonrandom patterns through temporally irregular preservation, with the relative size and distribution of gaps being more important than the absolute size of gaps and overall completeness. For short series, statistical tests based on a random-walk null hypothesis lose power and should be abandoned in favor of a multidimensional approach. A model of the data is needed that can account for confounding factors, and all available information on time and environment as well as phenotypic data should be incorporated and analyzed jointly, with a greater emphasis on quantifying uncertainty. Online enhancement: appendix.