Computational protocol: Forelimb bone curvature in terrestrial and arboreal mammals

Similar protocols

Protocol publication

[…] The sample included the ulna and humerus from a wide range of primate (n = 28) and marsupial (n = 42) species (see and ). Only adult or semiadult specimens were examined. The sex of most of the specimens was unknown, and so was presumed to be mixed in all analyses.Specimens were determined to be arboreal, semiarboreal or terrestrial (; ). Arboreal species live in trees and use their forelimbs in flexion to raise themselves and cling to branches. Terrestrial species are quadrupeds where triceps likely dominate forelimb function; if they do climb they do so by walking on top of branches. Semiarboreal species may spend most of their time on the ground, but are adept at climbing (quolls), or conversely, may be mostly arboreal but also spend much time on the ground (most semiarboreal primates).The ulna and humerus of each specimen were photographed in a standardised orientation that captured the curvature of the posterior (caudal) surface of the bone. Each bone was placed with the lateral surface facing up and the axis of the elbow joint directed vertically. The camera was mounted pointing downwards and the height adjusted to 1.75 times bone length. A centimetre scale was placed beside the bone. The images were prepared in PowerPoint by overlaying a line representing the chord with 13 equally-spaced lines crossing it perpendicularly (). For the ulna, the first intersection was placed on the posterior margin of the shaft in line with the coronoid process, and the thirteenth intersection on the posterior margin at the distal point of minimal circumference. For the humerus, the first intersection was placed on the posterior margin just distal to the head, and the thirteenth intersection on the posterior margin at the top of the olecranon fossa. The images were then transferred to tpsDig using tpsUtil software () and the points (constructed landmarks) where the thirteen lines intersected the posterior margin were digitised ().The 2D coordinate data were translated, scaled and rotated so that the first constructed landmark was at the origin (0, 0) and the distal constructed landmark had the coordinate (1, 0) (; ). The largest y-value among the constructed landmarks represented the normalised curvature (= max subtense/chord length). ANOVA was used to test for curvature differences among the locomotor categories (1 = arboreal, 2 = semiarboreal, 3 = terrestrial). The species mean normalised curvature values were regressed against the species body mass estimates (). To determine if larger species have more or less curvature, this regression was performed for the absolute value of the normalised curvature as well as the positive and negative curvatures separately.The normalised curvature provides a single value that best describes the overall curvature. Negative values are cranial and positive values are caudal curves. However, normalised curvature does not fully describe the sagittal curvature of the bone. For example, some bones have a caudal curvature proximally and a cranial curvature distally, and a single normalised curvature value does not adequately describe the shape of the bone shaft. Geometric morphometric analysis of the shape represented by the 13 two-dimensional landmarks can provide a more detailed analysis of the variation in curvature.The coordinates of the 13 semilandmarks representing the curve for each specimen were submitted to geometric morphometric analysis (GMA) in morphologika (). The data were Procrustes registered, with the “enable reflections” option unticked, and then the data were submitted to principal components analysis. A multivariate regression of all PCs against locomotor category (1 = arboreal, 2 = semi-arboreal, 3 = terrestrial) was performed to determine if there were significant differences in shape between arboreal and terrestrial species. […]

Pipeline specifications