On allometric equations for predicting body mass of dinosaurs

Cawley, G. C. ORCID: https://orcid.org/0000-0002-4118-9095 and Janacek, G. J. (2010) On allometric equations for predicting body mass of dinosaurs. Journal of Zoology, 280 (4). pp. 355-361. ISSN 0952-8369

Full text not available from this repository. (Request a copy)


Packard and colleagues investigate the prediction of the body mass of dinosaurs, using allometric models, advocating parameter estimation via direct optimization of a least-squares criterion on arithmetic axes rather than the conventional approach based on linear least-squares regression on logarithmic axes. We examine the statistical assumptions underpinning each approach, and find the method of Packard to be conceptually unsatisfactory as it assumes absolute rather than relative variability in body mass for a given long-bone circumference, which is biologically implausible. Their proposed approach is thus unduly sensitive to small relative errors for large mammals; as the largest (the elephant) is comparatively light for its large-bone circumference, the resulting model grossly overestimates the body mass of small mammals and is likely to substantially underestimate the body mass of dinosaurs. It is also important to note, however, that the error bars for the conventional model already indicate substantial uncertainty in body mass, such that for example, the body mass of Apatosaurus louisae may be as high as 63 metric tonnes, or as low as 23 metric tonnes, with a modal value of 38 metric tonnes.

Item Type: Article
Faculty \ School: Faculty of Science > School of Computing Sciences

UEA Research Groups: Faculty of Science > Research Groups > Data Science and Statistics
Faculty of Science > Research Groups > Computational Biology
Faculty of Science > Research Groups > Centre for Ocean and Atmospheric Sciences
Depositing User: Vishal Gautam
Date Deposited: 11 Mar 2011 16:42
Last Modified: 05 Aug 2023 00:15
URI: https://ueaeprints.uea.ac.uk/id/eprint/22037
DOI: 10.1111/j.1469-7998.2010.00777.x

Actions (login required)

View Item View Item