increase in terrestriality and independence of standing water. Wise (n.d.b.) suggested this may represent an ecological, rather than an evolutionary, gradient, and that the advance of the Flood waters from sea to land may be all that is needed to explain the small handful of examples where the stratigraphic order agrees with evolutionary phylogeny. The strong stratigraphic-clade congruence for vascular plant divisions and classes (SRC=0.994, p << 0.001) was one of the factors that led Wise (2003) to develop the floating forest model, in which he reconstructed the plants of the Paleozoic as a supercontinent-sized pre-Flood biome growing over the deep ocean and broken apart and buried from the outside in during the Flood. The importance of stratigraphic congruence is not merely as a test for evolutionary claims. While the first order interpretation is the correspondence between cladistic order and stratigraphic order, many higher order questions can also be explored. For example, as seen in Wise’s (n.d.b.) work, stratigraphic congruence can suggest alternative interpretations of the trends in the fossil record. Wise’s work suggests that there might be a difference between ecologies, possibly between marine and terrestrial vertebrates. Additionally, we might expect to see a difference between stratigraphic congruence for fossils preserved in the Flood vs. those from before or after the Flood. Questions around stratigraphic-clade congruence have also been explored in the conventional literature. The reported results have been mixed. For example: Norell and Novacek (1992a, b) examined 38 vertebrate groups and concluded that in most cases there was a general correspondence between superpositional order and the sequence of branching in the cladograms. However, the degree of fit varied widely. Benton and Hitchin (1997) examined 384 cladograms of echinoderms, fishes, and tetrapods with fewer than 40% showing statistically significant congruence between cladistic and stratigraphic data. They reported that this result contradicted findings on smaller samples of cladograms. Wills (2001) examined 179 cladograms of arthropods, along with 510 tetrapod and 157 fish cladograms for comparison. He found that arthropod cladograms showed significantly worse congruence than tetrapods. O’Connor and Wills (2016) analyzed 647 animal and plant cladograms and found significant variations in the degree of congruence across the Phanerozoic, with parts of the fossil record with a higher proportion of arthropods showing poorer overall congruence and parts of the record with a higher proportion of tetrapods showing higher overall congruence. Benton (2001) assessed 1,000 published phylogenies and found that there was little change in congruence throughout the twentieth century despite revolutionary change in methods and data sources. Here, we present a new database of 2,721 published phylogenies, all of which have been evaluated for their stratigraphic correlations based on first appearance dates recorded in the Paleobiology Database (PBDB; https://paleobiodb.org). Our taxon sample is heavily biased towards vertebrates, and among the vertebrates towards dinosaurs, but the database and our procedure provide the means for new creationist evaluations of the fossil record, by analysis of our current data and by application of our ongoing methodological work to nonvertebrate groups. Here we show the utility of this dataset by examining two important questions: 1) Is there an evolutionary order to the fossil record? 2) Are there differences between post-Flood and Flood strata? II. MATERIALS AND METHODS We assembled a dataset of 2,721 phylogenies, including most phylogenies compiled from published papers on Graeme T. Lloyd’s website (http://www.graemetlloyd.com/matr.html) as of June 2022. We chose to use these phylogenies for several reasons: (1) Lloyd is a well-known expert on this subject, (2) we needed phylogenies mostly based on fossil taxa, not extant species, and (3) the collection includes many phylogenies in digital formats, enabling us to process more trees. We also selected some additional published phylogenies reflecting the interests of our co-authors. We then assembled first appearance data for all the taxa in our phylogenies from the PBDB. In order to ensure reliable statistics, any phylogenies with fewer than 10 taxa were excluded from our dataset, as well as any taxa in the remaining phylogenies for which first appearance data were not available in the PBDB. This resulted in the exclusion of all phylogenies for plants and unicellular organisms. For phylogenies with exactly the same taxon sample, we calculated clade and stratigraphic rank correlation first, then selected the phylogeny with the best correlation. As a result, there should be no trees in our dataset that sample exactly the same set of taxa, but the overlap in taxic coverage could be considerable. A. Assigning Ranks To rank the phylogenies, we followed this procedure: Firstly, the root was identified and followed to the subsequent taxa(on) or node(s). Both the node(s) and/or taxa(on) were given the rank of 1 (see Fig. 1). When the node leads to a taxon, the taxon was assigned a fixed Figure 1. A sample tree illustrating the clade ranking strategy. Modified from Gahn and Kammer (2002, Fig. 2(1)). MCGUIRE et al. Testing the order of the fossil record 2023 ICC 479
RkJQdWJsaXNoZXIy MTM4ODY=