Data, scripts and logs for: Additive coding of morphometric characters as a source of artefactual paraphyly: a re-analysis of the osteology of skuas (Aves: Stercorariidae)

Abstract

Data, scripts and logs accompanying: Carlos, C. J. Additive coding of morphometric characters as a source of artefactual paraphyly: a re-analysis of the osteology of skuas (Aves: Stercorariidae). Systematics and Biodiversity. This deposit holds everything needed to repeat every analysis reported in the paper: the character matrices, the command script of each analysis, the output log each script returned, and an implementation of equal-weights parsimony written outside TNT to verify the reported lengths. The file README.md is the index. It names the table that each script feeds, and it states which scripts are deliberately absent and why. Contents: Matrices. skua_chu_S2.tnt, the faithful matrix, being the 141 characters and 19 terminals of Chu et al. (2009, appendix S2) with their coding and a priori weights reconstructed in full; skua_central.tnt, the discrete matrix, in which the same qualitative variation is recoded as direct discrete polymorphism; skua_TE_nonmolecular.tnt, the combined non-molecular matrix of 163 characters and the same 19 terminals; and skua_qual_mb.nex, the reduced matrix of 132 qualitative characters read by MrBayes. Scripts. Twenty-nine command scripts, twenty-seven for TNT and two for MrBayes, each under a header naming what it performs and the cells of the design it produces. Logs. The directory logs/ holds the thirty-three output files those scripts returned, so that any value reported in the paper can be traced to the commands that generated it and to the output they returned. Independent verification. parsimonia_independente.py implements equal-weights parsimony outside TNT, with Fitch optimization for the non-additive characters and Farris interval optimization for the additive ones, and reproduces the unconstrained optimum and every constraint cost reported for cells B1 and B2 of Table 1, for cells TE0a to TE6 of Table 3, and for both cells of Table S4. Software. TNT v1.6 (Goloboff & Morales, 2023) and MrBayes 3.2.7a (Ronquist et al., 2012). The Python file requires Python 3 and the standard library alone.