WHERE’S YOUR HEAD AT: EXPLORATION AND EVALUATION OF EURYPTERID MORPHOSPACE

Natural History collections contain primary data spanning the history of life on Earth. Much of these data remain understudied and therefore has not been integrated into our current understanding of paleontology. One such collection is the eurypterid collection at the University of Michigan Museum of Paleontology (UMMP). Last summer this material was digitized and preliminary morphometrics work was conducted. Here, we leverage the work we completed to include material from other previously published studies on eurypterid morphometrics. Specifically, we are interested in evaluating landmark placement between studies and comparing results of the two studies separately and combined. 130 specimens from the UMMP possess intact prosoma and were landmarked in the StereoMorph package in R. Four fixed landmarks and two sets of sliding landmarks along curves of the prosoma were employed. The data were analyzed using a Principal Components Analysis (PCA) and results were visualized in R using ggplot2. Previous work utilized more landmarks, which were unobtainable with the UMMP dataset. So for a viable comparison, approximately 115 specimens (those used in Bicknell and Amati 2021) from the New York State Museum and Yale Peabody were landmarked using our smaller landmark set. This allows us to examine the efficacy of different amounts and types of landmarks (fixed versus curves) and the resulting distribution in morphospace. The resulting morphospace shows a broad occupation of the genus Eurypterus, which supports previous studies. Additionally, there is a difference in distance between groups in the combined morphospace compared to previous work. This is likely due to the variation in landmarks used to capture specific aspects of the prosoma. This case study in landmark variation provides evidence that landmark selection, research question, and reproducibility should be carefully considered. Furthermore, targeted digitization of museum collections will increase mobilization of primary datasets.