THE EFFECT OF STRATIGRAPHIC DATABASES ON PHYLOGENY-BASED TESTS OF MACROEVOLUTIONARY HYPOTHESES

An important nuisance parameter in many macroevolutionary studies is the amount of time over which a lineage might accumulate anatomical changes and/or daughter species. The ages of first and last occurrences play an important role, but these are never known with certainty. Even when fossiliferous localities are tied down to specific zones, those zones often span 2-3 million years: i.e., an appreciable proportion of individual species lifetimes. A simple approach is to assume to use a uniform probability distribution for first/last appearances in an interval. Using information about superposition and lateral correlations of rock units within zones can improve upon this. If formations A and B fall in the same zone but A underlies B, then distributions of appearances within A & B should not have the same uniform distributions; those for taxa appearing in B need be constrained to appear after A. Here, I use a database recording superposition within general stratigraphic sections as well as lateral correlations across stratigraphic section to provide first/last occurrence data for formations and members of Ordovician gastropods. I then test for continuous vs. punctuated modes of character change using Markov Chain Monte Carlo searches that vary evolutionary parameters (phylogeny, divergence times and rates of continuous and/or punctuated change), but also temporal boundaries between formations within zones and then first/last appearances of appropriate species within the bounds for those formations. For runs using only continuous change, this should favor using the shortest necessary temporal durations of static lineages and also maximize the temporal durations of divergence times, both of which would favor continuous change models. However, in the cases examined here, punctuated models still do far better than do continuous change models. Thus, even though our knowledge of absolute ages of first and last appearances are inexact, basic information provided by stratigraphic superposition still limits plausible dates sufficiently to test basic macroevolutionary hypotheses.