Shifts in stability and control effectiveness during evolution of Paraves support aerial maneuvering hypotheses for flight origins

Abstract

The capacity for aerial maneuvering was likely a major influence on the evolution of flying animals. Here we evaluate consequences of paravian morphology for aerial performance by quantifying static stability and control effectiveness of physical models for numerous taxa sampled from within the lineage leading to birds (Paraves). Results of aerodynamic testing are mapped phylogenetically to examine how maneuvering characteristics correspond to tail shortening, forewing elaboration, and other morphological features. In the evolution of Paraves we observe shifts from static stability to inherently unstable aerial planforms; control effectiveness also migrated from tails to the forewings. These shifts suggest that a some degree of aerodynamic control and capacity for maneuvering preceded the evolution of a strong power stroke. The timing of shifts also suggests features normally considered in light of development of a power stroke may play important roles in control.

Keywords: Biomechanics; Control effectiveness; Directed aerial descent; Evolution; Flight; Maneuvering; Paraves; Stability.

Figure 1. Model construction.

Models were developed in Blender (A) from fossils (Archaeopteryx shown) and constructed using previous methods (McCay, 2001; Koehl, Evangelista & Yang, 2011; Munk, 2011; Evangelista et al., 2014b). Models for fossil paravians studied are shown in (B)–(H), scale bars indicate 8 cm snout-vent length. Anchiornis (B) (LPM B00169, Hu et al., 2009), hind limbs rotated out of test position to show plumage for illustration only. Microraptor (C) (IVPP V13352, Xu et al., 2003; tape covering proximal wing to body not shown). Archaeopteryx (D) (Berlin specimen as reconstructed in Longrich, 2006). Jeholornis (E) (IVPP V13274, 13553, Zhou, Zhang & Science, 2002; Zhou & Zhang, 2003b). Sapeornis (F) (IVPP V13275, Zhou & Zhang, 2003a). Zhongjianornis (G), (IVPP V15900, Zhou & Li, 2010). Confuciusornis (H) (Hou et al., 1995; Chiappe et al., 1999; Chiappe et al., 2008).

Figure 2. Testing and measurement of moments.

Models were tested (A) using previous methods (McCay, 2001; Koehl, Evangelista & Yang, 2011; Evangelista et al., 2014b). For simple cases such as a sphere or a weather vane, the relationship between slope and stability (B) is observed by plotting pitching moments versus angle of attack; negative slopes indicate restoring moments and stability while positive slopes indicate instability. Moments for sphere are not statistically different than zero, indicating marginal stability as expected, further validating the methods.

Figure 3. Appendage movements tested to determine control effectiveness.

Light gray indicates baseline posture, dark gray indicates appendage deflection. Appendage movements were selected based on those observed to be effective in previous work (Evangelista et al., 2014b), including (A) symmetric wing protraction (e.g., wing sweep to ±45°); (B) tail dorsiflexion to ±15°; (C) tucking of one wing; (D) tail lateral flexion to 30°; and (E) asymmetric wing pronation/supination to 15° (e.g., left wing pitched down, right wing pitched up).

Figure 4. Phylogenies.

(A) Phylogeny used in analyses, assembled from strict consensus of Zhou & Li (2010), Li et al. (2010), O’Connor, Chiappe & Bell (2011) for paravians and family relationships in (Cracraft et al., 2004) for extant birds. (B) Updated phylogeny from Turner, Mackovicky & Norell (2012) with revised position of Sapeornis, changes shown in blue. Additional proposed phylogenies (Godefroit et al., 2013; Xu et al., 2011), which alter the position of Archaeopteryx, are available in the .nex file at bitbucket.org/devangel77b/comparative-peerj-supplemental. Nodes 1–4 are discussed further in the text and are provided for reference between the trees.

Figure 5. Representative aerodynamic measurements for pitching stability and control effectiveness.

All plots show nondimensional pitching moment coefficient as a function of angle of attack. Long-tailed taxa (A) have a stable equilibrium point around 10–25° (yellow line) and the tail is effective in generating pitching moments at low angles of attack (pale yellow box indicates measurable moments for given tail deflections). In short-tailed taxa (B), including extant Larus, the equilibrium point at 0–5° is unstable (red line) and the tail control effectiveness is reduced (no measurable moments for the given tail deflections). Examples drawn from pterosaurs (Rhamphorhynchus and Pteranodon) illustrate similar patterns in phylogenetically distant taxa with contrasting tail lengths.

Figure 6. Evolution of pitch stability and control effectiveness.

Trees show (A) stability at equilibrium; (B) control effectiveness of the tail in pitch; (C) control effectiveness of symmetric wing protraction/retraction. Ancestrally, taxa are stable in pitch (A) and possess large, highly effective tails (B) but only moderately effective wings (C). Stability and tail control effectiveness are lost as tails shorten (AB, node 1), but more effective wings (C, node 1) are able to provide control. Control migrates from the reduced tail to the wings, which become larger and are associated with skeletal features that would enhance control and the production of left–right and fore-aft asymmetries.

Figure 7. Evolution of roll stability and control effectiveness.

Trees show (A) roll stability at low angle of attack; (B) roll stability at high angle of attack; (C) control effectiveness of asymmetric wing tucking in roll; (D) all of these together. Taxa are stable at high angle of attack (B), but mostly unstable at low angle of attack due to symmetry (A; Sapeornis and Confuciusornis marginal). Asymmetric wing tucking is always effective in roll (C). Thus, as animals developed the ability to fly at reduced body angles of attack, perhaps in shifting from steep-angle directed aerial descent (B) to shallower angles (A), more active control of roll would have been necessary. Ancestrally, inertial modes of the tail (Jusufi et al., 2008; Jusufi et al., 2011) would also have been available to assist the forewings, with function taken on solely by the forewings as tail inertia is reduced in derived taxa (after node 2).

Figure 8. Evolution of yaw stability and control effectiveness at high angle of attack (A)–(C) and at low angle of attack (D)–(G).

Trees show (A) yaw stability at high angle of attack; (B) tail and asymmetric wing pronation/supination control effectiveness; (C) yaw characters at high angle of attack together; (D) yaw stability at low angle of attack; (E) tail control effectiveness; (F) asymmetric wing control effectiveness; and (G) yaw characters at low angle of attack together. At high angles of attack (A)–(C), taxa are mostly marginally stable as might be expected for high angles (e.g., at 90° angle of attack all forms are marginal). Asymmetric pronation/supination of the wings are always effective in generating yaw at high angles of attack. At low angles of attack (D)–(G), by contrast, long-tailed taxa are stable and can control yaw with the tail. As tails reduce in size (nodes 1–2), taxa become unstable in yaw at low angles of attack and lose the ability to control yaw with the tail as well as any assistance from inertial modes of the tail. However, asymmetric movements of the wings are effective in producing yaw throughout the evolution of this clade, and control would thus have shifted from the tail to the forewings, paralleling the shifts seen in pitch.