Morphometric data for Mesodma species discrimination
Data files
Dec 08, 2023 version files 267.95 KB
-
curveslide.csv
169 B
-
Linear_measurements.txt
6.16 KB
-
M_Wilson_data.TPS
257.75 KB
-
README.md
3.86 KB
Abstract
Multituberculates remain one of the more poorly understood mammalian clades. The North American multituberculate record is comprised mostly of isolated teeth and incomplete jaws leading to interpretations of relationships based on limited anatomy. Despite the fragmentary record, the p4 of cimolodontan multituberculates is both common and a source of diagnostic characters in systematic studies. The results of a recent morphometric study on the neoplagiaulacid Mesodma suggest that p4 size may be more useful than shape in diagnosing the various species referred to this genus. We tested this hypothesis by applying two different morphometric methods (2D geometric morphometrics and linear measurements) to two samples – (1) one including the p4s of four known species (M. ambigua, M. thompsoni, M. formosa, and M. pygmaea), and (2) a sample of unidentified p4s of Mesodma from the Bug Creek Anthills locality of Northeastern Montana. Our results indicate that while form explains most of the morphological variation in p4s of the various species of Mesodma, linear measurement data support differences in p4 morphology that are not recovered by form data alone. Depending on the methods used, we found evidence for the presence of one or more species of Mesodmain the Bug Creek Anthills fauna. Although shape and size both contribute to morphological variation in the p4 of Mesodma, our results suggest that the diagnostic power of each in isolation, or in combination, varies significantly with the type of methodology employed.
https://doi.org/10.5061/dryad.fxpnvx0zd
We use the following data sets, code, and supplemental information to run all of our morphometric analyses in our study including the following:
- Procrustes analysis
- Principal component analysis
- Linear discriminant analysis
- Canonical Variate analysis
- MANOVA + Pairwise comparisons
- Figure generation using GGplot2
- Both Appendix 1 and 2
Methods for the files used
2D GMM: To capture the relevant morphology of the p4, digital landmarks were placed at the anterior-most and posterior-most projections of the serrate crest (TL1 and TL2 respectively). TL1 was positioned at the point of maximum curvature at the anterior margin of the crown. TL2 was placed at the posteriormost part of the crown-root junction at the posterior root. After TL1 and TL2 were located, a curve was traced in TPSDig (version 2.31; Rohlf 2013a) along the apical portion of the crown to join the two landmarks. Eighteen semilandmarks were then resampled from the curve with 19 helper points (HP) placed between semilandmarks to assist in later standardization. We resampled all landmarks to be equally spaced using TPSUtil (version 1.78; Rohlf 2013b). Landmark configurations were then prepared for 2D GMM analyses.
Linear Measurements: Standard linear measurements taken in millimeters were collected for specimens in both the known species data set (KDS) and the Bug Creek Anthill data set (BCADS). Three standard linear measurements have traditionally been employed in describing the shape of multituberculate p4s: crown length (L), crown height (H), and the length from the anterior margin of the crown to the intersection with the crown apogee (L1; Clemens 1964; Krause 1977; Novacek and Clemens 1977; Sloan 1981; Archibald 1982a; Sloan et al. 1987). L refers to the length of the crown, as measured along a baseline from the anteriormost point on the crown to the point where the posterolabial shelf intersects the posterior margin. H refers to height, the distance from the baseline to crown apogee. Lastly, L1 is the distance between the anteriormost point on the crown and the point where the lines used to measure L and H intersect. Measurements were taken in Fiji version 1.53c and recorded for further analysis (Schindelin et al. 2012). Linear measurements for all specimens are compiled in Appendix 1.
Description of the data and file structure
We include four files here associated with our analyses and :
1. Mesodma_2023_AJA_DataDryad.Rmd = Annotated R markdown file with all the code necessary to conduct our analyses. [in Zenodo]
2. curve slide.csv = CSV file indicating which landmarks are fixed and which are sliding. Should also be automatically generated after running the define.sliders line of code.
3. M_Wilson_data.TPS = TPS file exported from TPSDig suite of all KDS and BCADS specimens.
4. Linear_measurements.txt = All linear measurements taken in mm (L, H, and L1) for all KDS and BCADS specimens.
5. Ashbaugh_2023_Appendix1_FINAL – Additional tables; Specimen metadata, PC loadings, CVA classifications [in Zenodo]
6. Ashbaugh_2023_Appendix2_FINAL – Additional figures; PC2 – PC3 plots for KDS and BCADS [in Zenodo]
Sharing/Access information
If using the code or modifying the code and using it please contact austin.ashbaugh@ucalgary.ca to collaborate!
Code/Software
We have included the R markdown with detailed comments on how each of the morphometric steps were performed. Please feel free to contact the lead author of the manuscript to discuss the code as it is listed.
Note that we list the required library needed to run the code within the markdown file itself - You will need to modify directory information seeing as these will be filled with a template as it stands.
2D GMM: To capture the relevant morphology of the p4, digital landmarks were placed at the anterior-most and posterior-most projections of the serrate crest (TL1 and TL2 respectively). TL1 was positioned at the point of maximum curvature at the anterior margin of the crown. TL2 was placed at the posteriormost part of the crown-root junction at the posterior root. After TL1 and TL2 were located, a curve was traced in TPSDig (version 2.31; Rohlf 2013a) along the apical portion of the crown to join the two landmarks. Eighteen semilandmarks were then resampled from the curve with 19 helper points (HP) placed between semilandmarks to assist in later standardization. We resampled all landmarks to be equally spaced using TPSUtil (version 1.78; Rohlf 2013b). Landmark configurations were then prepared for 2D GMM analyses.
Linear Measurements: Standard linear measurements taken in millimeters were collected for specimens in both the known species data set (KDS) and the Bug Creek Anthill data set (BCADS). Three standard linear measurements have traditionally been employed in describing the shape of multituberculate p4s: crown length (L), crown height (H), and the length from the anterior margin of the crown to the intersection with the crown apogee (L1; Clemens 1964; Krause 1977; Novacek and Clemens 1977; Sloan 1981; Archibald 1982a; Sloan et al. 1987). L refers to the length of the crown, as measured along a baseline from the anteriormost point on the crown to the point where the posterolabial shelf intersects the posterior margin. H refers to height, the distance from the baseline to crown apogee. Lastly, L1 is the distance between the anteriormost point on the crown and the point where the lines used to measure L and H intersect. Measurements were taken in Fiji version 1.53c and recorded for further analysis (Schindelin et al. 2012). Linear measurements for all specimens are compiled in Appendix 1.