Skip to main content

The extended ‘common cause’: causal links between punctuated evolution and sedimentary processes

Cite this dataset

Polly, P. David (2024). The extended ‘common cause’: causal links between punctuated evolution and sedimentary processes [Dataset]. Dryad.


The common-cause hypothesis suggests that the factors that control the availability of the Earth’s sedimentary record may also affect probabilities of speciation and extinction and thus exert macroevolutionary controls on standing biodiversity.  Here I show through computational modeling that common causes may also link sedimentary biases and microevolutionary processes of trait evolution.  Using Gould’s classic “evolutionary microcosm” of Bermuda and its diverse endemic clade of land snails, Poecilozonites, I show that the glacial-interglacial sea level cycles that toggle local sedimentation between rapid eolian accumulation and slow pedogenesis could easily toggle trait evolution between rapid bursts of morphological change driven small effective population size, disruptions in gene flow, and “genetic surfing” expansion events punctuated with long periods of slow morphological evolution associated with geographic range coalescence, large effective population size, and panmixia.  The pattern produced by this interaction is expected to be similar to that produced by punctuated equilibria, even without accompanying speciation events.  The spatial dynamics of this system are expected to produce patterns of random trait evolution that are more likely multi-rate evolutionary models than like the standard single-rate Brownian motion models that are currently used as the null model in many phylogenetic comparative methods.  The Bermudian example of links between sedimentation and evolution is arguably extreme, but the principles of the extended common cause are likely to extend to many other paleontological systems.

README: The extended ‘common cause’: causal links between punctuated evolution and sedimentary processes

P. David Polly,

The files in these archives contain the method descriptions, code, morphometric semilandmarks, raw output data, and visualized raw output data of the computational modeling of the evolution of Bermudian snails by genetic drift (Brownian motion) in response to change in spatial population structure driven by rising and falling sea level over the last 500,000 years of Earth history. These data underpin the interpretations and conclusions in the accompanying paper and the code and methods allow users to replicate them.

Description of the data and file structure

S1 - Extended methods. This PDF file describes the computational model in more detail, including details about the DEM grid pattern and scale, describes the geometric morphometric analysis of snail aperture shape that was used to derive the shape parameters for the Raup shell coiling equations, shows the phases of the sea-level (eustatic) cycle for Bermuda over the last 0.5 million years, and presents an analysis of evolutionary parameters (rate and directionality) derived from the modeling experiments.

S2 - Computational model code. This ZIP archive contains files needed to run the computational model in Mathematica(c). It contains the following files:

  • Bermuda Computational Model V3.0.nb - this file is a Mathematica notebook that can be executed using the desktop interface. It requires the CSV data files below.
  • Bermuda Computational Model V3.0.pdf - this is a PDF of the Mathematica notebook that can be read by those who do not have access to the Mathematica package.
  • Bermuda Computational Model V3.0.txt - this is an ASCII text file of the same notebook that can be read by those who do not have access to the Mathematica package or which can be executed by command-line versions of Mathematica.
  • IslesThruTime.csv - this is an ASCII comma delimited (CSV) format data file with pre-calculated sea levels that is called by the code in the computational model. The table contains 200 lines. The first column is an ID, the second is sea level in meters relative to modern sea level datum point, and the third is time in millions of years ago. The data are from Miller et al's (2005) reconstruction of sea level referenced in the Supplement 1 file. These data were used to derive an interpolation function for modeling sea level as described in the Supplement 1 file.
  • Miller2005OxyIso.csv - this is an ASCII comma delimited (CSV) format data file that is not used in the computational code itself, but records the sea levels interpolated for each rendering frame of the animated maps in Supplement 3. It has the same format as the IsleThruTime file but the first column is ID, the second column is the interpolated sea level for that time, and the third column is the age in millions of years ago.
  • MISDates.csv - this is an ASCII comma delimited file (CSV) that reports the start and end dates of Marine Isotope Stages (MIS) from Lisiecki and Raymo (2005). These boundaries are not directly used in the computational model, but they do form the basis for comparing modeling results to Bermudian geology in the paper.
  • ModelResBermuda.csv - this is an ASCII comma delimited file (CSV) that describes the grid used in the computational model. As described in the text and Supplement 1, this grid was derived by resampling the digital elevation model (DEM) of Sutherland et al. (2013) into cells approximately 0.5 km per side. This file contains the grid ID (column 1), real world latitude and longitude of the cell (columns 2 and 3), elevation of the cell relative to modern sea level (column 5), graphical x and y coordinates of the cell (columns 6 and 7), and a list of the grid cells that are immediately adjacent into which a snail population can disperse during a model run. Most cells have 8 adjacent cells, but those at the edge of the sea mount can have fewer.
  • Results processing.nb - this is a Mathematica with code that was used to obtain the results in the publication, including graphical processing of model output and fitting evolutionary models.
  • Results processing.pdf - this is a PDF of the Mathematica notebook of the same name.
  • Results processing.txt - this is an ASCII text file of the Mathematica code used in result processing, which is human readable and can be run from the Mathematica command line.

S3 - Model output. This ZIP archive contains the complete output of the 10 model runs. Each model run is in a folder with a unique name that includes the date and time when the run started and a random five-character hash code. For each run, the following files are available:

  • Parameters.csv - an ASCII comma-delimited (CSV) file that records the parameters used for that model run. Note that some parameters like StartSequence were not used in this paper.
  • RunProgress.csv - an ASCII comma-delimited (CSV) file that contains output used to monitor the run. Each line contains the model step (Iteration), the model time in millions of years, the number of populations that were extant at that step, the number of DNA haplotypes at the step (unused in this paper), and the timestamp at which the model step was processed.
  • Summary.csv - an ASCII comma-delimited (CSV) file that reports a summary of trait phenotypes that were extant at the end of each model step. The first five columns contain the same data as RunProgress.csv and the remaining ones report the mean, minimum, maximum, and variance for each of the five traits (W, T, D=Di, S1=PC1, S2=PC2) for all the populations extant at that time. These data form the time series used to estimate rates and modes of evolution during different eustatic phases.
  • TraitChange.pdf - a PDF file that shows graphically the change in the overall mean of each trait (columns 6-10 of Summary.csv) and the overall disparity (sum of columns 21-25 of Summary.csv) for each step in the model run.
  • AICresults.xlsx - a Microsoft Excel notebook (XLSX) with the AIC results for evolutionary model selection. Each tab in the notebook reports results for one of the five traits (W,D,T,PC1,PC2). Notation is the same in the text (which follows Hunt, 2006), where URW is unbiased random walk, or pure Brownian motion, GRW is generalized random walk with a directional parameter, and GRW/stasis is a generalized random walk with an Ornstein-Uhlenbeck parameter. The x-bar and sigma columns describe the expectations of the model for the directional parameter (x-bar) and evolutionary rate (sigma) for the 15 scenarios compared in this paper. "all diff" indicates that transgressive, regressive, and low stand phases have different x-bar and sigma values, "high diff" indicates that transgressive and regressive phases collectively have one value that is different from low stand phases, "same" indicates that all phases of the sea level cycle share the same x-bar and sigma. The maximum likelihood for each of the 15 scenarios is reported, along with the AICc and AIC weight values. See text for more information. Note that the XLSX file can be opened in LibreOffice as well as Microsoft Excel.
  • .mov - five animations (MOV) files, each showing graphically the geographic patterning of trait values through the entire model run. Trait values were color coded using a rainbow map with red showing the highest value in the model run and blue the lowest. There is one animation for each of the five traits (W, T, Di, PC1, PC2). These files can be viewed with QuickTime and many other animation viewers, including some web browsers.

S4 - Poecilozonites aperture landmarks.txt - an ASCII plain text file in thin-plate spline (TPS) format containing the raw 3D semilandmark data for the aperture margin of each snail which are required to replicate the morphometric analysis and data visualizations.


The code used to generate the data in this package is included in this repository. It was written for Mathematica, which is normally run through a "notebook" (nb) interface that doubles as execution control and an archival document similar to a word processor file. The .nb file can be opened in a text editor and parsed, but the text is filled with Mathematica formatting tags like "Cell" and "RowBox" as well as image data that is in machine-readable format that are unrelated to the computational code itself. Users who do not have access to Mathematica can refer to the PDF copy of the notebook to see the code in a more readable format, as well as the graphics and comment lines that explain its implementation. An ASCII plain text version of the notebook is also included, which can be read in a text editor or executed from the comman-line interface of Mathematica.


The data in this archive were generated using computational modeling.


Yale University, Yale Institute for Biospheric Studies

Indiana University, Shrock Professorship in Sedimentary Geology

Lilly Endowment, Indiana University Pervasive Technology Institute

National Science Foundation, Award: EAR-1338298, Earth