Figure 1.
Log10(femur length)-log10(body mass) relationships for three major dinosaur clades: (a) Ornithischia (y = 3.0587x−2.7042; r2 = 0.93), (b) Sauropodomorpha (y = 2.3459x−0.2935; r2 = 0.73) and (c) Theropoda (y = 3.1854x−3.1840; r2 = 0.97).
The data sources for these relationships are shown in Table 1.
Table 1.
Dinosaur taxa constituting the femur length-body mass relationships shown in Fig. 1, along with specimen numbers (where available), mass (in kg), femur length (FL in mm), source of mass measurement and reference to the paper containing the FL and mass estimate.
Table 2.
Exploration of body size distributions for major vertebrate groups, dinosaur clades, time periods and formations.
Figure 2.
Frequency distribution of species body size for eight different animal groups: (a) extinct dinosaurs; (b) extant birds; (c) extant reptiles; (d) extant amphibians; (e) extant fish; (f) extant mammals; (g) extinct pterosaurs; and (h) Cenozoic mammals.
Note that all distributions are positively-skewed except for dinosaurs, which are markedly negatively-skewed (see Table 2). A combination of kernel density estimation and smoothed bootstrap resampling was used to test for optimum modality of the body size distributions. Silhouettes of the largest and smallest animal in each group are also shown (provided by Matt van Rooijen).
Figure 3.
Frequency distribution of maximum species body size for Cenozoic mammals (in grey) overlaid on the distribution for extant mammals (in white).
Curve fitting is based on a combination of kernel density estimation and smoothed bootstrap resampling. The figure clearly highlights the reduced frequency of small-bodied species in the Cenozoic mammal dataset, while the frequency of large-bodied species is comparable between both datasets.
Figure 4.
Frequency distribution of species body size for three major dinosaur clades: (a) Ornithischia; (b) Sauropodomorpha; and (c) Theropoda.
The Sauropodomorpha and Ornithischia are significantly negatively-skewed, while the Theropoda exhibit a bell-shaped distribution (see Table 2). All three clades are best fitted by unimodal distributions.
Figure 5.
Frequency distribution of dinosaur species body size for six major time periods: (a) Late Triassic; (b) Early Jurassic; (c) Middle Jurassic; (d) Late Jurassic; (e) Early Cretaceous; and (f) Late Cretaceous.
The Early and Middle Jurassic were best fitted by unimodal distributions; the Early Cretaceous by a bimodal distribution. The Late Triassic, Jurassic and Cretaceous were all best fitted by negatively-skewed multi-modal distributions (see Table 2).
Figure 6.
Frequency distribution of dinosaur species body size for two major formations: (a) the Morrison and (b) Dinosaur Park.
Both formations showed negatively-skewed distributions, with the Morrison formation approximately unimodal and the Dinosaur Park formation best fitted by a bimodal distribution (see Table 2). These patterns should be interpreted with caution, however, due to the small number of data points for each formation.