Handout to the R package divDyn v0.8.1 for diversity dynamics using fossil sampling data Adam T. Kocsis, Carl J. Reddin, Wolfgang Kiessling 2021-03-01 1. Introduction The purpose of this vignette is to guide users through the basic capabilities of the ‘divDyn’ package. Fossil occurrence databases, such as the Paleobiology Database (PaleoDB, http://www.paleobiodb.org/, http://fossilworks.org) are readily available to be used in analyses of diversity, extinction and origination patterns (the dynamics of biodiversity), with a certain toolkit that has become standard since the creation of the database. Until now, the implementation of most of these tools have been the responsibilities of individual researchers, with no software package to rely on. This R package intends to fill this gap. 1.1. Installation To install this beta version of the package, you must download it either from the CRAN servers or its dedicated GitHub repository (http://www.github.com/divDyn/r_package/). All minor updates will be posted on GitHub as soon as they are finished, so please check this regularly. The version on CRAN will be lagging for some time, as it takes the servers many days to process everything and updates are expected to be frequent. All questions should be addressed to Adam Kocsis, the creator and maintainer of the package (adam.kocsis@fau.de). Instead of spending it on actual research, a tremendous amount of time was invested in making this piece of software useable and user-friendly. If you use a method implemented in the package in a publication, please cite both its reference(s) and the ‘divDyn’ package itself (Kocsis et al. 2019). 2. Necessary Data Most functionality in the ‘divDyn’ package assumes that the time dimension is broken down to discrete intervals. Accordingly, most functions are built on two fundamental data structures: a time scale table and an occurrence dataset. 2.1. Time scales The workflow presented here is based on the discretization of geological time, which is constrained by stratigraphy. These intervals of time (bins) represent the basic units of the analysis, and their sequence is coded in the time scale table. Even if we develop a geological model that outputs robust estimates in a continuous time axis, the calculation of metrics presented in the package will require discretization. We added implementations of the basic functionalities for continuous time (chapter ‘4.3. Slicing’) as well, but we do not deem it as reliable as using stratigraphic bins for million-year-scale, deep-time analyses. As age estimates are dependent on the different geological ‘time scales’, binning the data can change more than necessary, which can have random effects on the resulting series. In order to demonstrate the workflow of binned analyses, we added an example table to the package. This table represents a somewhat altered form (see below) of the stage-level geological time scale of Ogg et al. (2016). You can attach this table using the data() function. 1