I decided to focus this blog on these three articles because of the problems they raise in my own data. Here I hope to dive a little deeper into these issues so that I may gain a bit more perspective on the limitations of my results. Although these authors focus their studies on methodological constraints in isotopic research, my goal is to remove the excess soft tissues and pinpoint the main purpose of their study, what would otherwise be considered the bare bones of their intent. I would like to take this blog opportunity to address these issues now, since I will be confronting the following problems in my final article at the end of this class.
Daux V, Lécuyer C, Héran MA, Amiot R, Simon L, Fourel F, Martineau F, Lynnerup N, Reychler H, and Escarguel G. 2008. Oxygen Isotope Fractionation Between Human Phosphate and Water Revisited. Journal of Human Evolution 55: 1138-1147.
The oxygen isotope composition of bones and teeth are used to identify regional origins, long-term residency, and migration of historic and prehistoric human populations. Since the oxygen isotope composition of teeth fix following enamel mineralization, the signatures locked in teeth provide evidence regarding the water source consumed as children. Bone tissue offers a different perspective because the signatures preserved in bone apatite-carbonate reflect the average oxygen isotope composition of the water source consumed before death. This provides information pertaining to long-term residency into (hopefully) late adulthood.
Since many palaeo-climate and environmental scientist rely on these data to reconstruct prehistoric temperatures (the same is true for deep ice-core oxygen signatures), statistical regressions that replicate valid relationships are key. There exists a deep relationship here between the isotope geochemists involved in palaeo-climate reconstruction and the physical anthropologists that use these regression formulae to provenance origin and/or track human migration. For this reason, Daux et al. (2008) regenerate the slope and intersection results via regression formulae first articulated by Luz and colleagues, Levinson, and Longinelli and colleagues. The regression formulae from these authors have been extensively used to provide information concerning both historic and prehistoric human mobility. Daux et al. (2008) call into question these parameters due to several factors by highlighting their small sample sizes and temporal contexts. With the oxygen isotope data generated by their sample, Daux et al. (2008) refine these earlier attempts by lumping all four samples together, then regressing the results. The result of this ‘super’ sample provided an equation that fit with the geographic variation of mean meteoric isotope values from various water-catchment facilities across the planet. This also means that the Daux et al. (2008) formula more accurately predicts the oxygen isotope ratios (18O/16O) of animal tissues via the natural parameters of the heavy oxygen (18O) composition of precipitation across the globe (i.e. temperature, distance from the coast, and altitude). One more important dimension to mention from this study is the results obtained from cooking certain types of food. Here the authors test whether boiling or heat-drying meat and grain affected the isotope composition of specific foods. The result of this pilot study showed that as light oxygen (16O) evaporated first, the food quickly reached isotopic equilibrium with the more enriched surrounding water, elevating the oxygen isotope signature of the food by almost +2 per mil.
The reason for this long rant is because my oxygen results, obtained from early 19th century soldiers, is likely affected by this cooking enrichment effect. It is however unknown to what extent body water is composed of water ingested from food (some estimates place this value between 10% and 30%). Bearing this in mind I will have to look a little more carefully at my tooth and bone oxygen results. Sorry, but I had to get this all out. My next two article summaries are much shorter, I promise.
Apart from being comprehensive and humble about their successful contribution(s), Daux et al. (2008) gently disintegrate one layer of canonical incumbency at a time. It is an interesting read precisely because the authors knew the scale and scope of their work. They removed what had previously held sway over the scientific community for almost 20 years. They understood the scope of this impact, too, singling out archaeologists interested in these issues, advocating more caution when undertaking this kind of research.
What interests me the most however are the oxygen isotope results obtained from cooking and heat-drying food. This has great potential in explaining why some of my oxygen isotope results (a significant fraction of them, in fact) were so 18O heavy. Heavy oxygen values suggest residency or origins at lower levels of latitude. In the United States, this means anywhere from the Carolinas to the Florida panhandle. However, Daux et al.’s (2008) study points out that this could be wrong. Instead, it is possible that some of these individuals lived at more northerly latitudes yet consumed significant quantities of stew, soups, and other cooked foods. And they did. This study is of major cross-disciplinary importance, and it is nice to see that from time to time one article can make such a drastic impact. The authors knew it; now everybody knows it – back to the drawing board on this issue.
Bataille CP and Bowen GJ. 2012. Mapping 87Sr/86Sr Variations in Bedrock and Water for Large Scale Provenance Studies. Chemical Geology 304-305: 39-52.
Strontium is incorporated into plants from the regional bedrock, which is then passed on unaltered to consumers. Since internal mass fractionation factors that affect the 87Sr result in tooth enamel are corrected by the mass spectrometer itself, strontium isotopes can determine the location of the foods consumed during childhood. Although similar to oxygen isotopes in application, strontium can also be used as a dietary indicator provided that the local (or baseline) strontium variation in animal tissues is known. Until recently there has been no attempt to document the 87Sr composition of the United States and Canada.
Lucky for me, however, Bataille and Bowen (2012) released this research right around the time I received my enamel strontium results (in April). Unfortunately, though, their study terminates at the Canada-U.S. border. There is nothing particularly spectacular about this study (other than the fact that I and others desperately needed it), being a rudimentary data driven study to produce variation maps for forensic and archaeological purposes. The authors use both lithic-specific and water-catchment parameters (and samples) to generate their maps, two important variables covering both the strontium composition of plants (lithic-specific) and water (water-catchment). In this respect, their contribution is another attempt to determine the baseline variation of this isotope across the coterminous North American continent. I have included this article in my blog because there is one fairly large problem in using the two variables mentioned above. I will elaborate on this below.
I suspect Bataille and Bowen (2012: 40) knew exactly what they were saying when they noted that: “Bedrock weathering is the main source of Sr to biological systems, however, these authors proposed that with improved understanding the model patterns could be used to interpret the geographic origin of biological materials.” They fall short of telling what might have otherwise been misinterpreted as a perfect relationship between outcrop weathering processes and the soils generated to produce fertile land. They also fail to mention anything about glacial processes forming many of the moraine-dominated landscapes we see today, especially along the northern border of the United States. In relation to my study then, water-catchment data may indeed be erroneous since the major source of Sr in river and lake systems derive from elevated regions, in this case the Appalachian mountain range. The Appalachian mountain range is quite old, forming during the early Paleozoic. This means that farmers growing crop on adjacent valley regions will likely yield biologically derived Sr consistent with those values found in the Appalachian Formation itself. It is here that I think the authors failed to address any of the major, well-known geological processes, substituting their water-catchment data (which is transported Sr) and lithic-time specific models instead (which is the actual geological substrate of the farm land, not the lithic age of the Appalachian mountains). In my opinion, the authors knew that generating these maps would be problematic. I am not saying that it is not a good study, it is. However it seems apparent to me that Bataille and Bowen (2012) deliberately chose not to include soil formation processes here. In the end, it is more likely that they were uncomfortable with the uncertainty that random processes would have on their mathematical models.
Hedges REM. 2003. On Bone Collagen – Apatite – Carbonate Isotopic Relationships. International Journal of Osteoarchaeology 13: 66-79.
A systematic relationship exists between the carbon isotope composition of bone apatite and collagen, frequently denoted in the literature as Δ13Cap-col. This relationship assumes that carbon from different dietary nutrients (protein, lipids, and carbohydrates) are incorporated into the skeleton in linear proportions. The main reason why this constant offset varies among different animal trophic levels is however poorly understood. Lee Thorp and colleagues were the first to discover that the dietary spacing of collagen and mineral apatite varies throughout food web depending on the trophic level of an organism. That is, between carnivores (small offset), omnivores (intermediate offset) and herbivores (large offset). Hedges’ (2003) attempt to question the main physiological and dietary mechanisms influencing these spacing differences calls into focus many unexplored hypotheses that have so far (and still to this day) eluded methodological verification. This is a very complicated topic (much of which I still don’t quite understand) that requires a solid background in physiology, biophysics and chemistry. Nonetheless, Hedges (2003) constructs a computer-generated model integrating many of these physical systems, including many known dietary parameters. His final results indicate that his computer generated model accounts for only half of the variation observed in natural systems. The details of this require a lot of explanation, however it is important to keep in mind that this study is the first theoretical attempt at explaining these patterns. The methodological difficulties in obtaining this type of data (e.g. the relationship between gut methanogenesis, expired CO2, and carbon isotope mass balance) are not small.
It is also important for me to think critically about this problem because I have calculated the Δ13Cap-col spacing for the Smith’s Knoll soldiers in my study. Overall, the data suggest that these soldiers fall within a ‘herbivorous’ diet, with meat constituting only a small fraction of the total diet (10% to 20%). This fits nicely with other data obtained for 19th century individuals living in North America, notably the isotopic data generated by Anne Katzenberg.
Hedges’ (2003) main intent here was to focus isotopic anthropologists on outside, cross-disciplinary sources of information. So far, major attempts to solve this question among the anthropological community has relied on minor quantitative research. This usually takes the form of calculating total dietary proportion of one photosynthetic plant type (i.e. C3) over another (i.e. C4). Since it is basically impossible to determine the proportional amount of dietary admixture of these two plant types with any certainty, Hedges (2003) breaks away from this traditional approach to come up with his own model. However, unlike Bataille and Bowen’s (2012) omission of important sedimentary variables, Hedges (2003) himself indicates that his model is likely too narrow to account for the stochastic nature of atomic routing in the human body. A humble assertion, no doubt, and an important trait to highlight from his study because I think some authors knowingly craft their results to fit the criteria set by their hypotheses. In this case, then, hypothesis testing should undoubtedly be subject to change. For me, this article is a good example of the aforementioned ‘stray from the pack’ mentality outlined by Luker. As a result, I have no formal criticism of his work. What this article has taught me, though, is that my apatite-collagen results are a product of metabolic and physiological mechanisms that are, for the meantime, poorly understood. For now I plan to use the comparative data to interpret my own results, albeit with caution.