Discussion

The total body volume and mass obtained for Plateosaurus engelhardti falls within the range of previous estimates. Seebacher (2001) calculated 1073 kg for AMNH 6810, an animal roughly the same size as GPIT1. Gunga et al. (2007) constructed two CAD models based on a high resolution laser scan of the mount of GPIT1. One of the models has significant volumetric errors (Mallison in press) and results in a weigh estimate of 912 kg. The other model follows the contours of the mount closely and results in a 630 kg estimate. Sander and Klein (2005) estimate the largest individuals of Plateosaurus to have reached 4000 kg at an overall length of 10 m, which is nearly identical to the results presented here. Henderson (2006) derived a total mass of only 279 kg for Plateosaurus by 3D mathematical slicing. However, it seems that the scale of the drawing cited by Henderson (2006: p. 919) as source data (Paul 1987) has been unintentionally altered. The animal is roughly two thirds as long as GPIT1, much smaller than the smallest known individual of Plateosaurus (4.8 m total length, Sander and Klein 2005). Scaling the model used here down to equal size as Henderson's 2006 model results in a mass of 418 kg (d = 0.6 kg/l neck, 0.9 kg/l remaining body parts as in Henderson 2006). The main reason for the discrepancy in total mass appears to be the more slender belly of the model used by Henderson 2006, which is unusual for a herbivore. Also, the drawing the model is based on has a relatively slender tail and hindlimbs.

Posing the virtual skeleton of GPIT1 digitally resulted in a bipedal pose in which the COM is well supported by the support area formed by either one adducted foot, or both feet with the limbs held vertically in anterior view (Figure 1.2, Figure 9.1-2). Bipedal locomotion appears easily possible at any speed, because the feet can be adducted at mid-stride sufficiently to support the COM. Due to the wide pubes, very rapid locomotion with long strides requires significant abduction of the femora. Therefore, at a run Plateosaurus had to sway from side to side, or rotate the pelvis around the vertical axis. However, the claim by Moser (2003) that Plateosaurus could only take tiny steps in a bipedal pose is not supported by the virtual mount. A stride length of 1.34 m is easily possible for GPIT1, without requiring any abduction of the femora (Mallison in press). Rather, the stride length limitation is true for the extremely upright pose suggested by Huene (1926, see Figure 9.3). While rapid running gaits probably require massive abduction of the femora, rapid walking can easily be achieved bipedally with a subhorizontal back, as the femur can easily cover a 65° angle without significant abduction (Mallison in press). Additionally, a bipedal, digitigrade pose appears to create bending moments in the limb bones that conform to the observed shapes; however, a detailed biomechanical analysis is required. It also frees the manus for other uses than locomotion, e.g., inter- and intraspecific combat or manipulation of food.

None of the tested quadrupedal poses offers any locomotory advantage over a bipedal pose, because neither mobility nor stability is increased. To the contrary, locomotion is limited to extremely slow speeds, and energetically ineffective due to high flexion angles in the hind limb joints in most tested postures. The problem of limited femur protraction at low abduction angles due to a collision with the pubes is aggravated compared to a bipedal pose. Additionally, a quadrupedal pose either does not place a significant part of the weight on the forelimbs, or exposes them to high compressive loads while inhibiting the step cycle in the hindlimbs through extreme flexion. Furthermore, the potential feeding envelope in a quadrupedal stance is smaller, and circumferential vision is limited to an extreme neck posture, while a bipedal stance offers easy 360° sight.

It seems counterintuitive that Plateosaurus should have used a posture and gaits that offer no advantage, while suffering from several potentially severe restrictions. For very slow locomotion over an unusual substrate, a slippery, muddy river bank, at least one ornithischian adopted a very clumsy quadrupedal gait (Wilson et al. 2009), and for such circumstances or for climbing steep inclines Plateosaurus may have similarly used a crawl on all fours. However, the pedal morphology of Plateosaurus is more similar to the theropod discussed by Wilson et al. (2009) that did not use a wide-gauge, crawling quadrupedal gait on the same substrate, so that even this possibility seems unlikely. For normal locomotion, it appears unreasonable for Plateosaurus to have adopted a quadrupedal posture, even if one assumes that manus pronation was somehow possible without strong humeral abduction.