The analysis of stable oxygen isotope ratios (δ18O) is a tool that is frequently applied in bioarchaeology for researching human remains. Oxygen isotopic values of animal calcified tissues are correlated with the isotopic composition of local environmental water (D’Angela and Longinelli, 1990; Daux
The main source of drinking water for historical and prehistorical human populations as well as animals coexisting with a certain group is environmental water. People used to populate areas close to lakes and rivers; they tended to take water from springs and streams. In isotope research conducted upon human and animal skeletons, a positive correlation between the isotopic composition of bone phosphates and the isotope ratio of environmental (drinking) water was observed (D’Angela and Longinelli, 1990; Daux
One of the methods of reconstructing the origins and migration paths of prehistoric peoples is the reconstruction of isotopic ranges of the environment inhabited by a given individual. To this end, information on the δ18O level of contemporary precipitation water in a given area is obtained, and if skeletal remains of animals are available on the archaeological site, investigators use isotopic data obtained from bone tissue carefully selected from the fauna that inhabited the area with the studied human population at the same time (Bentley and Knipper, 2005). Attempts to reconstruct past environmental conditions in the context of isotopic analysis give rise to certain issues that may affect the credibility of the interpretation of the outcomes. This is due to the fact that the average annual δ18O of local rainfall varies over time. Arguably, pollution is also significant in this respect. On the other hand, the conversion of the isotopic composition of oxygen in bone phosphates to δ18O of water generates a relatively large error margin that widens the isotopic range of the environment. Pellegrini
Apart from environmental issues, one of the most frequent doubts pertaining to the isotope research of oxygen isotopes in the context of migration is the fact that humans insert some liquids into the organisms with the isotopic composition possibly differentiated, which is determined not only by many geographical and climatic factors but also cultural (social) issues. Humans have been consuming thermally processed food ever since the Neolithic period (Kirsanow and Tuross, 2011; McGlynn, 2007). In the Middle Ages, brews, beers and alcoholic beverages were prepared in different ways, specifically by cooking, boiling or fermentation, and were frequently given to children (Brettell
Most historical sources, including those pertaining to food-making, come from the Middle Ages. These show that the favourable method for preparing meals was to cook them slowly on low heat, which is characteristic for stewing under a lid (Brettell
Therefore, some pieces of literature emphasise that, during the interpretation of data, one must consider the potential influence of cultural practices, such as boiling water and cooking food, on the oxygen isotope ratios in bone apatite (Brettell
Conducting research explaining this issue with regard to the human body tissues for obvious ethical reasons is not possible; therefore, an animal model was used in this research, which was a laboratory rat.
The primary goal of the research was to determine the influence of thermal processing of consumed water on δ18O in bone phosphates of rats. First, we determined the degree of fractionation on the level of tap water
Due to obvious reasons, it is not possible to learn how bone tissue would change its isotopic composition in response to a factor introduced in an
Animal subjects in the experiment were divided into two groups according to the type of consumed water. The first type was tap water, poured directly into drinking troughs. Water samples were obtained from a water supply network over a short period of time at the same time of the year. The water was always drawn from the same point (tap) after 5 minutes of continuous flow. The temperature of the water was controlled and each time it was 12 ºC. As a result, thanks to a lack of variation in air temperature and the stable temperature of the tap water, the fractionation process was limited to a minimum. Thanks to the small variability of tap water, only a few controlled supply water measurements were taken.
The second type was thermally processed water. Thermal processing of water consisted of pouring tap water into a dish with 2.5 L volume. The vessel was covered with a lid to 2/3 of its area. Water was heated until boiling (for about 20 minutes), and then it was boiled on low heat for about 2.5 hours. Some water evaporated and left the system, and some of it condensed on a lid and went back into the container. Next, it was cooled for 15 minutes, poured into bottles (which were then sealed), and transferred to the animal quarters. Due to the fact that boiling water may cause higher variability of isotopic values of oxygen, samples of thermally processed water were obtained from each boiled batch (31 samples).
The experiment included 8 female rats of the Wistar Cmd:(WI)WU strain. The rat breeding was conducted in stable humidity (55%), lighting (artificial light 60 lx, 12 h/12 h), temperature (22°C) and availability of feed and water
The femoral diaphysis of each rat was the research material in this experiment. After cleaning any soft tissue and marrow, the femoral diaphyses were rinsed with deionized water by an ultrasonic cleaner, dried up and ground in a ball mill (Retsch MM 200) and divided into rations with a similar weight of about 0.4 g.
The analysis of isotopic composition of oxygen was conducted for the phosphates of bone apatite. The analytical procedures of isolation of bone phosphates were conducted according to the procedure suggested by O’Neil and collaborators (1994) as described by Vennemann
The isotopic composition of oxygen in the extracted phosphates was determined in the Department of Radioisotopes, Institute of Physics-CSE, Silesian University of Technology, Gliwice, Poland. Continuous-flow isotope ratio mass spectrometry was used. The results of isotopic analysis is presented in delta notation δ={[R(sample)-R(standard)]/R(standard)} ×1000), where R refers to the measured isotopic ratio of 18O/16O (Grimes and Pellegrini, 2013). A method similar to one described by Lécuyer (2004) was used for stable oxygen isotope analysis. About 250 μg of a sample of Ag3PO4 was wrapped in a silver capsule. Analysis for each sample was performed in triplicate. Samples were subjected to pyrolysis in a Eurovector EA3000 elemental analyser. Pyrolysis took place in a reactor filled with glassy carbon and equipped with graphite crucible heated to 1300°C. During pyrolysis, carbon monoxide was formed, which was then taken by helium flow to a gas chromatograph where it was separated from residual atmospheric nitrogen. A sample of CO was introduced into the IsoPrime isotope ratio mass spectrometer through an open-split interface. Before each sample, the monitoring CO gas (Linde, 4.7 grade) was analysed. In addition to the three NIST-120C (Lécuyer
This paper encompasses an attempt to examine how the isotopic deltas are being changed during long-term boiling (
For this purpose, the first step included analysis of obtained values of isotope ratios for tap water and thermally processed water.
Detailed data of the analysis are given in Supplementary Material. Due to disproportion in the number of samples of thermally processed water and tap water (31
Fig. 1
The values of oxygen isotopic delta of tap water and thermally processed water with the measurement errors.

The composition of stable oxygen isotopes in tap water and thermally processed water was used to calculate the values of isotopic deltas of bone phosphates, with the application of an equation developed for laboratory rats by Luz and Kolodny (1985) (
Fig. 2
The isotopic composition of water given to laboratory rats, recalculated to values of the isotopic composition of apatite phosphates in specimens potentially consuming water with a certain composition.

This was followed by an analysis of the variability of the isotopic composition of oxygen in the bone tissue of rats that drank tap water and rats that consumed thermally processed water. The analysis of the results of the isotopic analysis of bone phosphates began with checking if there was a significant difference in the isotopic composition of apatites in the bones and teeth. It turned out that there were no statistically significant differences between the analysed tissues (Mann-Whitney test: p=0.5736). Similar results concerning the lack of differences between the bone and tooth tissues of rats were obtained by Luz and Kolodny (1985). This is probably due to the continuous mineralisation of constantly growing teeth (incisors), whose remodelling rate might be similar to that of bones. Due to the lack of significant differences between the analysed tissues, further analyses were based on combined bone and teeth results.
In order to verify whether there were any differences between young (weaned) individuals and their adult mothers, isotope ratios were compared with the use of Mann-Whitney's nonparametric test. It turned out that there were no statistical differences between females and their offspring (W=320, p=0.08). Hence, adult females and their weaned offspring were combined into a single group for further analyses.
The next stage consisted of verifying whether there were any statistically significant differences between the isotopic composition of water calculated into phosphates and the isotopic delta of apatite phosphates in rat bones. The results for water were calculated based on a regression equation proposed by Luz and Kolodny (1985), which is the only equation dedicated to rats available in the literature. In the case of tap water, there were no statistically significant differences between the median δ18O of tap water calculated to phosphates and the δ18O of phosphates in the bones of rats drinking tap water (W=19.0, p=0.082). However, there was a significant difference in the case of boiled water. The median δ18O of thermally processed water converted to phosphates was lower than the median δ18O of phosphates in rats drinking thermally processed water (W=180.0, p=0.000). It was noted that the variability observed in the bone apatites of the studied animals overlapped with the variability of water calculated to bone phosphates within its lower variability ranges.
In order to verify the hypothesis about a lack of significant differences between the groups of rats drinking boiled and tap water, a nonparametric Mann-Whitney test was conducted, which revealed that the difference between the medians was statistically significant (W=13.0; p=0.000) and amounted to nearly 4.0‰. This means that the rats consumed boiled water were characterised by an approximately 29% higher δ18O value compared to the rats consuming tap water (
Fig. 3
The difference in the values of isotopic delta of bone phosphates in animals consuming two different types of water.

It needs to be emphasised that the fractionation observed during water evaporation in the course of the experiment resulted in an increase of the water’s oxygen isotopic delta by 6.1‰.
It is not possible to generalise and develop a universal plan for the issue of preparing beverages and meals in human history. Cooking, whether short and intensive or slow, is inevitably connected to evaporation and therefore to isotopic fractionation. During cooking, the water temperature increases, thus intensifying the molecule movement. Heavy isotopes (18O) have larger bonding energy, lower diffusion speed and move slower than lighter molecules (16O). Therefore, the frequency of collision between the 18O isotopes is lower. Such differences in physio-chemical properties of isotopes translate to general reactivity of chemical compounds consisting of such isotopes. In the process of boiling water, relatively more lighter isotopes (16O) are eliminated, together with steam, under the influence of temperature, which makes the water in the container isotopically heavier.
During their research, Brettel and collaborators (2012) conducted a similar experiment where water was cooked in pans for 10 minutes and 180 minutes, respectively. In the mentioned studies, the time of cooking/evaporation was a significant factor affecting the results of the isotopic composition of water and meals subjected to thermal processing. The change of isotopic delta values in the aforementioned experiment was +0.4‰ and +26.2‰, respectively, as compared to the starting level of δ18Ow. The result obtained in this research was significantly different from the results obtained in the Brettel paper (2012), largely due to methodological reasons. One might assume that the containers used to cook water or liquid food in the past were not entirely uncovered but (depending on the historical period, existence and development of pottery techniques) could have been entirely closed, as in stewing and cooking in leather containers shaped like a conical pot (narrower at the top) or at least partially covered. In each of the above cases, water evaporation was being limited. Steam escaped into the atmosphere to a large extent; however, some water condensed on the container walls or the lid and returned to the solution. In the experiment herein, the pot filled with water was covered 2/3 with a lid; therefore, some water escaped the system, and some of it condensed on the lid and the container. As may be concluded from the comparison of both experiments, slow cooking on low heat and partially covering the container limited the intensity of evaporation and therefore the degree of isotopic fractionation of oxygen in the course of thermal processing of water.
It should be noted that the variability of isotopic values in water that has been cooked for about 2.5 hours is quite high (
The existence of such fluctuations did not, however, significantly affect the research issue being considered. It would be illogical and unimaginable to assume that the preparation of meals by human populations occurred under the same pressures and temperatures over the centuries. A significant result of the conducted analyses is, therefore, the observation of change of isotopic composition of oxygen in thermally processed water, consisting of an increase of its value by +6.1‰, which may constitute a reference point for further analyses.
Isotopic fractionation of oxygen in the course of thermal processing of food and its effect in the form of change of isotopic delta were already observed (Brettell
The influence of thermal processing of food on isotopic ratios of oxygen in cooked food according to different authors.
Paper | Material | Cooking (min.) time | Temperature (°C) | Isotopic effect of cooking process |
---|---|---|---|---|
Food (vegetables) water vs initial water | 20 | min. 100 | +6.2 ‰ | |
+2.2 ‰ (chicken) | ||||
Daux | Food (meat) water vs initial water | 20 | min. 100 | +2.8 ‰ (beef) |
+3.7 ‰ (mackerel) | ||||
Food (rice, lentils) water vs initial water | 20 | min. 100 | +2.6 ‰ | |
Brettell | Vegetables and meat | 60 | min. 100 | +4.0 ‰ |
Brettell | Vegetables and meat | 180 | min. 100 | +10.1 ‰ |
Sweet potato | 300 | 125 | +3.0 ‰ | |
Tuross | Sweet potato | 300 | 150 | +6.0 ‰ |
Meat | 270 | 125 | +1.1 ‰ |
In the paper by Daux and collaborators (2008), the isotopic analysis encompassed the ingredients of a meal (vegetables, meat and fish), the water in which the food was cooked, and the water from the aforementioned products upon 20 minutes of cooking. The result of that research was the conclusion that water contained in meat cooked for a short time and the water in which such meat was cooked did not differ significantly in terms of isotopic composition. A slightly larger difference in isotopic delta was observed for the water isolated from vegetables in comparison to the water in which such vegetables were cooked. It was also noted that the isotopic value of cooked vegetables altered by less per mille than the water such vegetables were submerged in, and the less hydrated vegetable was, the higher the change of isotopic value.
Bone phosphates enter the state of isotopic balance with the body’s water very fast. The isotopic composition of body water is correlated to the composition of δ18O of drinking water (Levinson
It just so happens, however, there are differences between the environmental variability of δ18Ow and δ18Op, because in some geographic regions, the values of oxygen isotopes within phosphates do not correspond to those observed in the environmental water. This may be a sign that there is some other factor than a physiological or environmental one that plays an important role in shaping the relationship at the level: local environmental water - skeleton (Pellegrini
It needs to be emphasised that, within the premises of the experiment, the only factor differentiating the two groups of rats was the type of water they consumed. Other factors which might have affected the conditions of the experiment, including air temperature and humidity, were controlled on a regular basis.
The source literature emphasises that, in the case of analysing the mobility of a group of humans with the application of oxygen isotopes, there is a need to evaluate a potential effect of culinary practices on the values of isotopic ratios in bone phosphates (Brettell
As has already been mentioned, the range of the isotopic variability of an environment is obtained by analysing contemporary water in a given area or the use of isotopic calculators (Daux
In order to determine the range of isotopic variability of a given living environment of a prehistoric population, we use the observation of the existing relationship between the isotopic composition of bone phosphates and environmental (drinking) water (Daux
The available regression equations, developed based on the natural variability of water in the natural environment (Daux
Fig. 4
δ18Op mean values (± sd) of bone samples from rats drinking tap water (square) and thermally processed water (diamond) plotted versus the δ18O average of tap water (–10.011‰) and thermally processed water (–3.7894‰) respectively. Horizontal lines determine the range of δ18O variability in the environmental water from the area of Kraków. Scope no. 1 was determined based on the results obtained for rats drinking tap water ± 1 sd. Scope no. 2 was determined as the minimum and maximum values determined on the basis of the presented regression lines. The results obtained for rats drinking boiled water were shifted to the same level as the results for rats drinking tap water to better visualize the relationship to the compared groups and reference to ranges showing the local environmental level.

When extrapolating the results obtained from the animal model described above to the interpretations of results for prehistoric human populations, similar trends are likely to be expected. What is more, the enormous biocultural diversity of the human race translates into differentiation of corresponding mathematical models for our species (Daux
Fig. 5
Variability of oxygen isotopic composition based on the relationship between δ18Op and δ18Ow in human studies.

As it has been mentioned above, in order to analyse the mobility of human groups and determine potential immigrants, the results obtained from the human osteological material need to be related to and confronted with the isotopic environmental background characteristics of the studied area.
The analysed individuals exhibited values below the lower limit of the isotopic variability of the environmental background are potential allochthons, whereas those whose results exceeded the upper limit of the background did not necessarily so. This results from the fact that there is a range of both physiological and biocultural factors that may increase isotope values in the body, including culinary practices (Brettell
The analyses described in this paper have shown that the range of δ18Op variability (between Q1-Q2 quantiles) in rats drinking tap water was 13.07–14.44‰, which may be deemed as the environmental background for the isotopic ratio of water for this area. On the other hand, when considering the δ18Op level of the environmental background according to the regression line for rodents, we obtain a range of 11.61–15.81‰ (assuming δ18Ow=–10‰), which confirms the previous statement that the variability of the δ18O level in environmental water is higher than the variability of δ18O in bone phosphates (
Based on the result obtained for rats, it could be surmised that those individuals who exceeded the upper δ18O limit of the environmental range by up to 3–4‰ may be considered representatives of the studied group, provided the group consumed thermally processed food from the biocultural point of view. Raising the upper value of the environmental background estimated based on the analyses of fauna or environmental water by the value obtained on the basis of this model seems to be justifiable for most of the populations, as the factor connected to processing food, which is a unique and almost immanent feature of our species, is known to have occurred at least since Neolithic biocultural transformations (Kirsanow and Tuross, 2011; McGlyn, 2007).
The analyses described in this paper have contributed to expanding knowledge about the influence of the thermal processing of consumed water on the isotopic composition of apatite phosphates.
The experiment conducted on the laboratory rat has yielded the following results:
Raising the values of the isotopic composition of water by
The isotopic values of tap water calculated to phosphates fall within the broad range of variability defined by regression equations for rodents (D’Angela and Longinelli, 1990; Luz and Kolodny, 1985; Navarro
Consuming thermally processed water (potentially boiled meal) significantly raises the isotopic level in a rat's body by approximately 3–4 per mille (29%). It may be assumed that regular consumption of isotope-heavy drinks and foods by humans may cause a similar effect. Exceeding the upper limit of the isotopic range of environmental variability by just a few per mille may mean that the individual belonged to the studied local population, assuming that the drinks and food consumed by the group were thermally processed.
Fig. 1

Fig. 2

Fig. 3

Fig. 4

Fig. 5

The influence of thermal processing of food on isotopic ratios of oxygen in cooked food according to different authors.
Paper | Material | Cooking (min.) time | Temperature (°C) | Isotopic effect of cooking process |
---|---|---|---|---|
Food (vegetables) water vs initial water | 20 | min. 100 | +6.2 ‰ | |
+2.2 ‰ (chicken) | ||||
Food (meat) water vs initial water | 20 | min. 100 | +2.8 ‰ (beef) | |
+3.7 ‰ (mackerel) | ||||
Food (rice, lentils) water vs initial water | 20 | min. 100 | +2.6 ‰ | |
Vegetables and meat | 60 | min. 100 | +4.0 ‰ | |
Vegetables and meat | 180 | min. 100 | +10.1 ‰ | |
Sweet potato | 300 | 125 | +3.0 ‰ | |
Sweet potato | 300 | 150 | +6.0 ‰ | |
Meat | 270 | 125 | +1.1 ‰ |