INTEGRATING PROBABILISTIC PHYLOGENETICS WITH THE THEORY OF PUNCTUATED EQUILIBRIA TO TEST THE TEMPO AND MODE OF SPECIATION IN THE FOSSIL RECORD

The application of phylogenetic models to Paleozoic invertebrate taxa formed a crucial component of Eldredge’s (1971) insights regarding how patterns of morphological change and speciation are distributed in the fossil record, which ultimately led to the development of punctuated equilibria (PE) as a key theory in macroevolutionary biology. Since Eldredge and Gould’s paper on PE in 1972, the disciplines of paleobiology and phylogenetics have each experienced “revolutions” related to theoretical and/or methodological advancements, but have largely developed in isolation of one another. For example, many paleobiological tests of PE analyze time-series data for ancestor-descendant taxa determined a priori (often by taxonomic opinion/expertise), which could bias which taxa and trait(s) are selected for analysis. Similarly, modern phylogenetic methods rarely incorporate aspects of PE theory directly (Bokma, 2008) and even most paleontological phylogenies in the literature were inferred using methods that do not consider the temporal distribution of species as data, let alone stasis.

In this talk, I develop and apply a novel step-wise approach to empirically testing PE’s predictions using probabilistic phylogenetic methods, and apply this approach to multiple taxonomically and stratigraphically vetted species-level datasets of Paleozoic echinoderms. The step-wise approach involves first estimating a posterior distribution of evolutionary histories using Bayesian tip-dating methods, which combine models of character evolution, diversification, and fossil sampling. Notably, ancestor-descendant relationships are plausible to infer using this model without conditioning on any particular species-pair a priori. Using the distribution of phylogenies for these datasets, I apply a series of probabilistic methods to document evidence for several key predications of PE theory: including the frequency of “budding” speciation vs. anagenesis, the proportion of a species’ history in stasis vs. change, and whether character models featuring punctuated change at speciation events are favored over standard Mk model assumptions. Notably, methods in the first step are agnostic to PE and therefore evidence supporting PE’s predictions in the second step represent unambiguous cases of support.