Skip to main content

CarniFOSS: A database of the body mass of fossil carnivores

Cite this dataset

Faurby, Søren (2021). CarniFOSS: A database of the body mass of fossil carnivores [Dataset]. Dryad.


Motivation: Body mass is one of the most important determinants of animal ecology. Unlike other important traits it is also readily inferable from fossils and it is therefore one of the only traits that can be directly analyzed and compared between fossil and contemporary communities. Despite this, no comprehensive database of the body mass of larger clades of extinct species exists. Analysis of fossils has therefore been restricted to small clades or to smaller, potentially biased, subsets of species. We here describe CarniFoss, an open-access database of body masses of all 1322 extinct species of non-pinniped Carnivoramorpha and two related extinct groups of carnivorous mammals, Hyaenodonta, and Oxyaenida.

Main types of variables contained: We gathered lengths of teeth of fossil and extant species and body mass for extant species and a few of the best-known fossil species. Following this we estimated body mass for all species through phylogenetic imputation.

Spatial location and grain: Global, terrestrial

Time period and grain: We collected data on all known species within the focal groups. The known species all lived in the Paleogene, Neogene or Quaternary (i.e. the last 66 Mya).

Major taxa and level of measurement: We searched for data on reported tooth size of all described species of Carnivoramorpha (excluding pinnipeds) and selected extinct related groups (Hyaenodonta and Oxyaenidae). We combined this with measured body mass for all extant species and inferred body mass based on long-bones for selected extinct species, as well as a species-level phylogeny including all extant and extinct species in the group, and inferred the body mass for all species using phylogenetic imputation.

Software format: Data is provided as a series of .csv files, with all metadata in a separate PDF file.


All holes in the data are filled by imputation. Imputation is clearly marked in the dataset but we urge all users to consider when this imputation influences their downstream analyses.

Usage notes

The dataset may be updated in the future and we urge all users to check the page on GitHub to see if there is any newer version available (


Swedish Research Council, Award: 2017‐03862

Swedish Research Council, Award: 2015-04587