We are responding to the submission by Oates et al., https://doi.org/10.1017/S0003598X00096885 Equids and an "acrobat": closure rituals at Tell Brak. The paper describes the interpretation of a partially preserved skeleton with bone changes that indicate ‘habitual energetic activity’ (p. 398), or a person who was ‘physically active, using jumping and twisting movements in a very disciplined way with the feet pointed downwards during leaps, as can be seen in modern dancers’ (p. 396). While we acknowledge that the skeleton was deposited in an unusual way, and that documentary sources at Elba and seals at Tell Brak suggest people were present who were characterised as 'always jumping about' (translation of a Semitic word which can also infer an 'entertainer, juggler or horseman'), we would like to explore further what is described as the main subject of this article, namely the 'occupational' interpretation of an incomplete skeleton as that of an 'acrobat.'

In bioarchaeology some readers will be aware that there have been numerous conference presentations and published papers focusing on how skeletal lesions are used to determine 'occupation', especially in the last 10 years since the International Journal of Osteoaearchaeology’s special issue in 1998. A recent survey has also provided a critical retrospective of research since then (Reference HendersonHenderson et al. 2008). As bioarchaeologists with a very long experience working with archaeologically derived skeletal remains, and with expertise in the particular subject of reconstructing behavioral attributes from human skeletal remains, we are especially concerned about this paper and its misleading aspects, especially for those readers with no experience of this type of study.

There are several omissions in the osteological analyses, particularly notable given the central importance this skeleton is accorded. Firstly, there is no diagram or complete listing of preserved elements, but only a brief summary indicating the skeleton is quite fragmentary (p. 393), and in Table 1 a list of measurements taken. When attempting to infer meaning from a skeleton from an archaeological site, firstly a complete skeleton is desirable so that distribution patterns of lesions can be seen and a most likely 'diagnosis' presented; secondly, it is also preferable to have a large sample of skeletons at a site to determine the range of variation in lesion expression, especially for those lesions that may be related to increasing age (this is to see if indeed a particular skeleton is unique). Figure 3 also states that the skeleton in question is the ‘most complete of the three specimens’ but the image is not very clear and an arrow would have been helpful on the figure. There is also a comment at the foot of p. 391 that the upper part of the skeleton has already been reported elsewhere, but it is impossible to tell if any of those observations are included in this current paper. For example, one humerus is reported to have osteoarthritis, but there is no metrical data for that humerus in Table 1. Of note is the attribution of ‘mature adult’ to the remains (p. 393), but no citations relevant to methods used are provided; such documentation is both standard and relevant to supporting those particular age-related changes to the bones of the skeleton evaluated by the authors.

In terms of methodological concerns, specifically, there is only superficial description of lesions in the feet and no differential diagnosis is offered; Figure 4b does not give clear enough detail to support the description. Nor is there any consideration of a differential diagnosis which might help explain the unusually short stature of this individual. Lastly, apparently no radiographs were taken, and certainly no radiological information is reported; this may have helped in interpretation.

The weakest part of the article is also its core, that is, the interpretation of presumed behavior based on skeletal/morphological changes which are mainly lesions associated with muscle and ligament attachment/insertion sites on the bones preserved. Other than some determination of acute trauma in the lower limb (which we do not have adequate grounds to dispute), this analysis focuses almost entirely on those changes in muscle and ligamentous insertions, or what bioarchaeologists commonly refer to as 'entheses', 'enthesopathies', or one of a suite of 'musculoskeletal markers of stress/activity'. There is a surprising lack of reference to published literature on the methods used to analyse this skeleton and how data may be interpreted; this would normally provide readers a context to assess the validity of this approach. Without providing this crucial groundwork, most readers will be left completely in the dark without any possibility of assessing the very basis of the claims made by the authors. Indeed, only three citations that concern relevant skeletal methods/interpretations are provided, with just one from the last decade; furthermore, one of the earlier ones was authored by one of the authors, where a similarly unfounded behavioral interpretation from this same site was made.

As might be expected, there is indeed a considerable literature on this subject with numerous older publications supporting the notion that enthesopathies provide a useful indicator of, at least, general levels of activity (e.g. Reference DutourDutour 1986; Reference KennedyKennedy 1989; Reference HawkeyHawkey & Merbs 1995). Nevertheless, over the last several years there have also been critiques emphasising the multifactorial etiology, oversimplification, and resulting over-interpretation of these types of bone changes and their ostensible relationship to activity (Reference WaldronWaldron 1994; Reference JurmainJurmain 1999). Ageing, biological sex, genetics, work/activity, obesity, diet, and disease can all lead to the development of MSMs together or separately (Resnick & Niwayama 1983; Reference RogersRogers et al. 1997).

More recent attempts have controlled for the confounding factors of age, sex and body size (Reference MolnarMolnar 2006; Reference WeissWeiss 2007; Reference CardosoCardoso 2008a & Reference Cardosob) and, in so doing, the correlations with activity have become yet weaker. Cardoso’s research (Reference CardosoCardoso 2008a) has also indicated that, even with historical documentation about 'activity' associated with nineteenth and early twentieth-century Portuguese skeletal remains, there was no correlation between 'activity' with either enthesophytes or with osteoarthritis (Reference CardosoCardoso 2008b; Reference HendersonHenderson & Cardoso 2008). In fact, entesopathies and osteoarthritis were not necessarily associated with demanding activities. Furthermore, recent research has indicated that the anatomical basis for enthesophyte development is poorly understood in the bioarchaeological field and that bioarchaeologists would be wise to develop a better understanding of the anatomy of the soft tissues in which the lesions develop (Reference HendersonHenderson 2008; Reference HendersonHenderson & Cardoso 2008).

Finally, recent experimental research using rigorous controls on nonhuman animals, has shown virtually no relationship of activity and any identifiable enthesial bone change (Reference ZumwaltZumwalt 2006). From these results the researcher concluded, ‘In spite of decades of assumption otherwise, there appears to be no direct causal relationship between muscle size or activity and attachment site morphology, and reconstructions of behavior based on these features should be viewed with caution’ (p. 444).

The lack of correlation of activity with distinct morphological bone change is, in reality, not that surprising. What determines the timing, degree of remodeling, and ultimately development of an enthesial reaction is ultimately regulation of bone turnover physiology. In turn, bone turnover is governed by gene action, including identified specific loci that are variable between individuals (Spector & McGregor 2004). Recent experimental results have identified specific genes influencing enthesial bone formation (Reference ChenChen et al. 2007). In most prehistoric societies the majority of individuals were likely highly physically active, but not all show similar bone changes; in fact, many show very little bone change, even at those anatomical sites where they should be most expected.

Why is this the case? Simply put, most of the variance in physical expression at enthesial sites is not the result of activity. Most of the variance is likely due to genetic influence (including sex difference in hormonal activity), with further substantial contributions due to age and body size. The portion of variance explained by activity is, regrettably, quite small. These factors explain why more controlled research has yielded ever-more disappointing results.

In fact, while in clinical literature there are indeed reports of people developing bone lesions such as osteoarthritis in specific occupations, not all people with the same occupation will develop lesions or even the same lesions! Moreover, if a particular 'activity' is being inferred from a skeleton, one should always be referring to clinical literature to understand how specific 'activities' affect the bones of the skeleton. For example, how does whiplash affect the skeleton and do acrobats, dancers, gymnasts or any other presumably identifiable 'at risk' contemporary group of people develop bone changes such as those described in this paper?

In recent years the scientific literature has included a number of such now-unsupportable studies that attempted to reconstruct general levels of activity in past societies. However, endeavors to divine specific occupations have been much rarer. Even at the peak of enthusiasm for using skeletal 'markers of activity' (such as osteoarthritis and enthesopathies), most researchers understood the monumental obstacles of attempting such suppositions from analysis solely of skeletal remains (Reference JurmainJurmain 1999; Reference WaldronWaldron 2006). When definitive artefactual associations are present, the possibilities of concluding something reasonable about 'occupation' may be better, but 'jumping' to conclusions about an individual skeleton is dangerous.

If the data, particularly those controlled most rigorously, are so lacking in support, what has led these researchers (as well as many others) to be so compelled to make hasty, superficial conclusions? In following what Tony Reference WaldronWaldron (1994) has termed the 'alluring prospect', every bioarchaeologist at some point has been highly tempted to make similar suppositions. We, of course, want to know interesting details about the lives of those people whose skeletons we study, and we can imagine all sorts of fascinating things they may have done. Nevertheless, ultimately imagination must yield to the inescapable constraints of our science.

The general public naturally also wants to hear good stories about the past. This sort of popular demand can exert further pressure on researchers. Speculation is healthy and, oftentimes, amusing. What we choose to publish is quite another matter. The expectations of the general public regarding what can be learned from skeletal analysis often are exaggerated and may not be based in sound research. Much of this unrealistic view has been magnified by the media’s misrepresentation of forensic science research. We feel that unsupportable claims that would be eagerly received in some circles must not be made, merely because they are 'fashionable'. More generally, respectable scholarly publications do not serve the public interest by furthering such a 'CSI' mentality.