Persons and Currie (2011) present a useful and anatomically-rigorous study of tail anatomy in Tyrannosaurus rex and other reptiles. We applaud their careful measurements of muscle masses and their important recognition that such quantitative data from extant Sauria are critical for any reconstructions of anatomy or mechanics in extinct Sauria. We also agree with their approximations of muscle areas and torques, although like all previous efforts, these inevitably have a range of error and assumptions that future sensitivity analyses need to carefully consider (especially for muscle fascicle lengths and maximal isometric stresses or “specific tensions”). It is possible that relative tail muscle sizes increased during tyrannosaur ontogeny but sensitivity analysis will be critical for testing this idea rigorously.
However, we wish to bring attention to what we consider important omissions in their work, and to object to their portrayals and critiques of the contrasting methods and conclusions of several other studies. First, Persons and Currie (2011) state that Hutchinson and Garcia (2002) “assumed a total femoral retractor muscle mass of 297 kg for each leg of a 6,000 kg Tyrannosaurus.” This is incorrect; 297 kg was the mass of the thigh segment (skin, muscles, bone, etc.) used in their inverse dynamics model used to calculate the moments that muscles needed to balance, but it was not the thigh muscle mass input, retractor or otherwise, and was not considered in other calculations. Indeed, Persons and Currie's (2011) critique misses the point of that and nine other follow-up studies (Hutchinson,2004a,2004b; Hutchinson et al.,2005,2007; Allen et al.,2009; Gatesy et al.,2009; Pontzer et al.,2009; also reviews by Hutchinson,2006; Hutchinson and Allen,2009), only one of which is cited (Hutchinson,2004b). The point is that it is important to know how much leg muscle various bipedal dinosaurs needed to run versus how much they actually had. Those studies generally had the required muscle mass as an output result, not as input data. They all agreed that to run quickly, with ground reaction forces of 2.5 times body weight or more, large tyrannosaurs needed more leg muscle mass than could be presumed to have been present. These studies considered a wide range of skeletal and comparative evidence as well as input parameters and assumptions, including tail dimensions and caudofemoral muscle mass as explained further below. Thus, Persons and Currie's (2011) critique is unfounded and quite misleading.
Persons and Currie (2011) also misconstrue the extent to which previous studies by the aforementioned authors have explicitly analyzed caudofemoral muscle mass and tail dimensions in theropod dinosaurs such as Tyrannosaurus. Hutchinson et al. (2007) estimated the contribution of M. caudofemoralis longus muscle mass to the potential hip extensor muscle mass actually present as one-fourth the tail base segment's mass, or 140.8 kg in their initial model. That mass is smaller than the 261 kg estimate of Persons and Currie (2011) for M. caudofemoralis brevis and longus. However, Hutchinson et al. (2007) also considered alternative masses; even increasing that mass to 300 kg would not have qualitatively changed their conclusions. Persons and Currie (2011) do not acknowledge that M. caudofemoralis longus was clearly considered in their interpretations of locomotor and turning abilities. Hutchinson et al. (2007) also estimated the mass of other hip extensors as approximately 297–398 kg in various models (their Table 8), far over the 25 kg that Persons and Currie (2011) portray as an assumption of these previous studies.
Allen et al. (2009) actually showed how tail dimensions based on saurians, as Persons and Currie (2011) later did, differ considerably from those suggested by the skeleton alone. They also showed how these dimensions alter center of mass estimates by the body, as previously predicted by Motani (2001). Although these changes affect estimates of locomotor performance, we caution that many other uncertainties involved in these estimates (especially the variable relationship between skeletal and fleshy dimensions; discussed by the studies cited above and below) render the impact of these changes on conclusions drawn from them far from as straightforward as Persons and Currie (2011) contend.
The above studies falsify Persons and Currie's (2011) critique that M. caudofemoralis's “contribution to locomotion has not been quantitatively analyzed.” Indeed, Hutchinson et al. (2005) also considered M. caudofemoralis muscle moment arms in quantitative detail, arriving at results overlapping those of Persons and Currie (2011). Furthermore, Hutchinson (2004b) and subsequent studies have noted that the ankle extensor muscle mass seems to have been the most stringent limit on running performance in large theropods (especially considering the large M. caudofemoralis; Hutchinson et al.,2007), so a focus on hip muscles may be overlooking a “weaker link” in the limb.
We similarly caution that Sellers and Manning (2007) used the same input data (plus important methodological refinements) as Hutchinson (2004b) to directly estimate running speed in theropod dinosaurs. They achieved results that broadly match those studies' indirect estimates: speeds of approximately 5–11 ms−1 for Tyrannosaurus, as discussed by Hutchinson (2004b) and Hutchinson and Allen (2009), which contrasts rather than agrees [as Persons and Currie (2011) state] with the faster speeds advocated by less quantitative functional studies by Bakker (1986), and Paul (2000).
A more minor point is that consideration of M. caudofemoralis muscle architecture data from Allen et al.'s (2010) study of ontogenetic scaling in Alligator mississippiensis would have been of use to this article, which considers ontogenetic changes of muscles in tyrannosaurs, particularly as M. caudofemoralis masses do not show ontogenetic allometry in Alligator.
Persons and Currie (2011) close their discussion with a criticism of previous reconstructions of theropod tail sizes, with particular mention of another of our previous studies (Bates et al.,2009a). Specifically, Persons and Currie (2011) contend that “the current prevalent fashion among paleoartists” to depict dinosaur tails as relatively “unmuscular and laterally compressed” is reflected in the reconstructions of Bates et al. (2009a) and is “consistent with the more traditional tail muscle restoration technique described.” Once again, this inaccurately portrays our previous work in multiple ways. First, Bates et al. (2009a) did not present a single “lightly built” tail reconstruction of Tyrannosaurus and other theropods, but (similar to Hutchinson et al.,2007 and Allen et al.,2009) produced a suite of differently sized and proportioned reconstructions in an attempt to constrain a plausible range of mass set values based on fossilized skeletal evidence. Thus, this study, like others (Hutchinson et al.,2007; Allen et al.,2009), emphasized the broad range of plausible body mass and tail mass values for extinct dinosaurs, given the absence of exceptional and extensive soft tissue preservation. In suggesting that these reconstructions are “less-than-robust,” Persons and Currie (2011) have missed the most fundamental point demonstrated by all of our earlier studies: a single “robust” mass (or other quantitative, functional) estimate for a dinosaur is currently—and probably never will be—possible.
Persons and Currie (2011) also do not provide any unambiguous data to demonstrate that the tail reconstructions of Bates et al. (2009a) are either “unmuscular or laterally compressed” or “less-than-robust.” Superficial comparison of dorsal views of the tail muscle reconstructions of BHI 3033 (Fig. 9 in Persons and Currie,2011) with our “best-estimate” tail reconstruction (Fig. 5b in Bates et al.,2009a) of the same specimen actually appear quite similar. The proximal width of Persons and Currie's (2010) tail reconstruction may actually appear misleadingly wide because their “digitally sculpted” ilium lacks the lateral curvature/expansion of the postacetabular blade present in BHI 3033, which is captured in our three-dimensional laser scan of the actual specimen (Bates et al.,2009a). Bates et al. (2009a) discussed at the length, the differences between their models and those of previous workers. We will not repeat these here, except to note that where comparative data were available, the tail volumes of Bates et al. (2009a) were considerably larger than previous reconstructions' (e.g., “best estimate” tail volume of Acrocanthosaurus 1.149 m3 versus ∼0.679 m3 by Henderson and Snively, 2003).
Persons and Currie (2011) also contradictorily conclude that the tail volumes of Bates et al. (2009a) are “unmuscular and laterally compressed” (hence underestimating mass); yet they use a body mass of 4,500 kg for BHI 3033, which is 3,000 kg lighter than Bates et al.'s (2009a) initial estimate for this specimen. The estimate of 4,500 kg by Stevens et al. (2008) was produced by digitally wrapping a skin outline directly on to a beautifully crafted skeletal model of BHI 3033 and is thus representative of an almost “skin-tight” body volume. This model served the purpose of the interesting and thoughtful analysis subsequently carried out by Stevens et al. (2008). However, it is difficult to reconcile Persons and Currie's (2011) use of such a small body mass value [considered implausibly low by Bates et al. (2009a)] with their arguments for larger tails and subsequent criticisms of our reconstructions.
In response to Persons and Currie's (2011) claims that many supposed “lightly built” tail reconstructions represent an effort to portray theropods as “more aerodynamic and superficially more athletic,” we direct readers to two further studies by Bates et al. (2009b,2010). Like Hutchinson et al. (2007) for Tyrannosaurus, these two later studies set out to constrain a plausible range of mass set values for Allosaurus (Bates et al.,2009b), and subsequently (Bates et al.,2010) to demonstrate the spectrum of running gaits predicted for this animal considering the breadth of possible body mass, centre of mass and muscle mass values in the previous study. Like the independent studies and methods of Hutchinson et al. cited above and Sellers and Manning (2007), these studies demonstrated first and foremost how a lack of knowledge on soft tissue anatomy limits the accuracy of running speed estimates of extinct dinosaurs. Second, these studies showed that even with large hind limb muscle masses or maximized contractile properties, large bipedal dinosaurs could not have obtained particularly fast speeds, particularly those postulated by some workers cited by Persons and Currie (e.g., Bakker,1986; Paul,2000).
We do not seek to selfishly increase citation metrics for our work, or to embarrass the authors with our corrections to their study. Indeed, we have great respect for the skillful contribution of Persons and Currie (2011) to this area of research and look forward to their future work, particularly as our studies share more common ground in explicit, reproducible scientific practice and quantitative detail than this commentary might suggest. However, we have written this correspondence to clarify what our research has actually done and concluded versus the inaccurate portrayal of our (and others') work by the study of Persons and Currie (2011), and thereby steer the discussion toward more productivity, fairness, and rigor.