Figure 2.

The relationship between body size and forelimb length. When the phylogeny is split into three regimes (A and B), Aves (gray), Paraves (black), and other theropods (red), differences are seen between the main phylogeny in the slope for Aves and the phylogeny of Godefroit et al. (2013) in which a shallower slope for Aves is seen. These differences appear to be due to species being moved from the Aves and Paraves; Godefroit et al. (2013) place Xiaotingia and Anchiornis in Paraves, whereas in the main phylogeny they are placed in Aves. When the phylogenies are split into two portions, Aves and non-Aves, (C and D) differences are once again seen between the trees. The main phylogeny, following conclusions from Dececchi and Larrson (2013), shows a smaller Aves intercept, indicative of an allometric change in scaling between body size and forelimb length (see Table 1). Again, differences are seen in the alternative phylogeny of Godefroit (D) where this lower Aves intercept is not found. Models that split the phylogenies into three portions (A and B) are better-fitting models (according to AICc scores) and explain more of the variance in the data (according to higher R-squared values) than models that split the phylogenies into Aves and non-Aves (C and D).