LAUSR

This comment addresses several inferences made by Bocca et al. (2018) to assess the nutrition of middle-aged inhabitants from four Sardinian localities (Alghero, Bisarcio, Geridu, and Sassari). Bocca et al. analyzed 72 human bone specimens dating from the twelfth to the eighteenth century for metal contents (Ba, Ca, Cd, Cu, Hg, Pb, Sr, and Zn) by SC-ICPMS and applied a univariate and multivariate approach, using element-Ca ratios, to examine the eating habits and environmental exposure of the individuals. While we agree with their interpretation of the Hg/Ca and Pb/Ca ratios, attributable to various forms of environmental exposure, we disagree with some of their paleodiet claims. Based on certain elemental indexes of bone tissue (i.e., Ba/Ca, Cu/Ca, Sr/Ca, and Zn/ Ca), Bocca et al. suggest that the individuals from the four sites consumed different foods (e.g., vegetables, cereals, red meat, and marine food). For instance, the individual from Sassari according to Bocca et al. would have consumed a high-protein diet because of their significantly higher level of Cu in bones compared to other sites. Similar dietary inferences have been argued by Guede et al. (2017) from trace element content of human bones from the Tauste medieval necropolis in Spain. However, a commentary by Lugli and Cipriani (2017) has recently debunked the misconceptions and the improper use of trace elements to assess eating habits. They contend the results of Guede et al., in particular the misinterpretation of the individual diet based on bio-essential element data (e.g., Zn, Mg, and Na). Unfortunately, similarly to Guede et al. (2017), Bocca et al. have ignored a big portion of the literature on trace element analyses for paleodiet purposes (e.g., Klepinger, 1990; Ezzo, 1994; Burton et al., 1999; Balter et al., 2002; Burton and Price, 2002; Mays, 2003; Burton, 2008; Humphrey et al., 2008; Balter et al., 2012; Lösch et al., 2014; Reynard and Balter, 2014; Lugli et al., 2017; Tacail et al., 2017), and rely on literature which is based on erroneous assumptions (i.e., Giorgi et al., 2005 and references therein), and does not provide a scientific basis for their conclusions. In fact, Bocca et al. (2018) (and a number of other studies, e.g., Francalacci 1989; Busetto et al., 2008; Corti et al., 2013; Guede et al., 2017) claim that the bone elemental content is directly correlated to the diet ignoring several physiological mechanisms that regulate the actual abundance of these elements in bones. Among others, homeostasis, gut absorption and pathways from blood serum to bones (e.g., Burton et al., 1999; Burton and Wright, 1995). The lack of theoretical bases in using trace elements as a proxy of diet has been shown since the 1990s. In particular, Ezzo (1994) has discussed the problematics of zinc as a paleodiet marker in archeological studies. The mineral essential nutrients (e.g., Zn, Fe, Mn, and Cu) in the body are controlled by homeostasis and their level is physiologically kept constant, in a sort of Bsupply-anddemand^ system, where a steady and sufficient supply of the micronutrient is ensured and a simultaneous excess is prevented (Scheiber et al., 2013). For this reasons, biological essential elements can be used, at best, to unravel nutritional deficiencies (Reynard and Balter, 2014), and not to detect the Bincidence of vegetables, cereals, and animal foods in the diet^ (Bocca et al., 2018). The inadequacy of essential elements for human diet reconstruction has been also demonstrated by Klepinger (1990) Responsible editor: Philippe Garrigues