Extinct sloths (Xenarthra, Folivora) are morphologically diverse, despite some similarities among some Pleistocene genera. Cranial and diagnostic postcranial elements (especially limb bones) are taxonomically informative but the axial postcranial skeleton can prove difficult to classify, as in cases with only vertebral remains or when closely related taxa are found together. Here, 24 linear and angular measurements of presacral axial skeletons from eight genera of Pleistocene ground sloths were analysed through multivariate methods for assigning their vertebrae to a genus and to a position along the vertebral column. Both isolated vertebrae and vertebrae associated with partial and complete skeletons of each genus were included. Principal components and linear discriminant analyses show a high percentage (over 90%) of correct taxonomic reclassification. For the position of vertebrae, the accuracy increased significantly when cervical, thoracic and lumbar regions were separately considered instead of the whole column, misclassifying the position of the vertebrae only by a few positions in most cases. Our database is useful for confidently assigning isolated vertebrae of extinct sloths to a specific genus and position, showing that beyond the apparent homogeneity in the axial postcranial skeleton of ground sloths, quantitative studies allow the evaluation of diversity and variation between groups.

Extinct sloths (Xenarthra, Folivora) are morphologically diverse, despite some similarities among some Pleistocene genera. Cranial and diagnostic postcranial elements (especially limb bones) are taxonomically informative but the axial postcranial skeleton can prove difficult to classify, as in cases with only vertebral remains or when closely related taxa are found together. Here, 24 linear and angular measurements of presacral axial skeletons from eight genera of Pleistocene ground sloths were analysed through multivariate methods for assigning their vertebrae to a genus and to a position along the vertebral column. Both isolated vertebrae and vertebrae associated with partial and complete skeletons of each genus were included. Principal components and linear discriminant analyses show a high percentage (over 90%) of correct taxonomic reclassification. For the position of vertebrae, the accuracy increased significantly when cervical, thoracic and lumbar regions were separately considered instead of the whole column, misclassifying the position of the vertebrae only by a few positions in most cases. Our database is useful for confidently assigning isolated vertebrae of extinct sloths to a specific genus and position, showing that beyond the apparent homogeneity in the axial postcranial skeleton of ground sloths, quantitative studies allow the evaluation of diversity and variation between groups.