Quantifying the Importance of Burial Environment and Bone Characteristics Matthiesen, Henning; Eriksen, Anne Marie Hoier; Hollesen, Jorgen; Collins, Matthew

Bone degradation at five Arctic archaeological sites Quantifying the importance of burial environment and bone characteristics Matthiesen, Henning; Eriksen, Anne Marie Hoier; Hollesen, Jorgen; Collins, Matthew

Published in: Journal of Archaeological Science

DOI: 10.1016/j.jas.2020.105296

Publication date: 2021

Document version Publisher's PDF, also known as Version of record

Document license: CC BY

Citation for published version (APA): Matthiesen, H., Eriksen, A. M. H., Hollesen, J., & Collins, M. (2021). Bone degradation at five Arctic archaeological sites: Quantifying the importance of burial environment and bone characteristics. Journal of Archaeological Science, 125, [105296]. https://doi.org/10.1016/j.jas.2020.105296

Download date: 24. sep.. 2021 Journal of Archaeological Science 125 (2021) 105296

Contents lists available at ScienceDirect

Journal of Archaeological Science

journal homepage: http://www.elsevier.com/locate/jas

Bone degradation at five Arctic archaeological sites: Quantifying the importance of burial environment and bone characteristics

Henning Matthiesen a,*, Anne Marie Høier Eriksen a, Jørgen Hollesen a, Matthew Collins b,c a National Museum of Denmark, Environmental Archaeology and Materials Science, IC Modewegsvej, DK-2800, Kongens Lyngby, Denmark b The GLOBE Institute, University of Copenhagen, Øster Farimagsgade 5, 1353, Copenhagen, Denmark c McDonald Institute for Archaeological Research, Downing Street, Cambridge, UK

ARTICLE INFO ABSTRACT

Keywords: The degradation of archaeological bones is influencedby many variables. The bone material itself is a composite Bone diagenesis of both organic and inorganic components, and their degradation depends on processes occurring both before Burial environment and after burial, and on both intrinsic bone characteristics as well as extrinsic environmental parameters. In this Arctic study we attempt to quantify the effect of some of the variables using a novel approach that includes detailed Oxygen consumption monitoring of the burial environment combined with respirometry studies of bone material from five archaeo Bone moisture Temperature logical sites in West Greenland. First, we compare the state of preservation of excavated bone material with the current burial environment including the soil pH, thawing degree days, soil porosity and soil moisture. Secondly, we investigate oxic degradation of collected bone samples through respirometry and quantify the effects of temperature and moisture on the oxidation rate of individual bones. Finally, we discuss how the oxidation rate is influenced by intrinsic bone parameters. Some of the main conclusions are:

1) There is a significant correlation between the current burial environment and the current state of preservation of the bones. 2) The oxidation rate measured by respirometry increases on average fourfold as temperature in ◦ creases by 10 C, and more than hundred-fold when dry bones are soaked in water. 3) The oxidation rate of different bones varies over two orders of magnitude due to intrinsic variables such as organic content and state of preservation of the bones. ◦ 4) The median oxidation rate of wet bone at 15 C corresponds to a yearly loss of 3.8% of their mean organic content, while the median yearly loss for dry bones at 75% RH is 0.02%. 5) Respirometry is a promising tool for quantitative degradation studies of bone, but more studies are needed in order to obtain a better understanding of the oxidation processes involved.

1. Introduction species and type of bone are also important for the degradation (Lopez-Costas´ et al., 2016; Nicholson 1996). The bone material itself is Bone material is a central component for the interpretation of ac complex, consisting of both organic (mainly collagen) and inorganic tivities at many archaeological sites, but the interpretation not only (mainly bioapatite) components that mutually protect each other, so requires an understanding of what is preserved but also what may have degradation of one of the components influences the degradation disappeared due to pre- and post-depositional degradation. A recent pathway of the other by enhancing the access of e.g. microbes, oxygen review (Kendall et al., 2018) describes the complexity of bone diagen and solvents (Collins et al., 2002). The bioapatite is mainly degraded or esis, and points towards extrinsic environmental parameters such as soil recrystallized through chemical dissolution (Berna et al., 2004) and pH, soil hydrology, oxygen and ambient temperature as important for microbial tunnelling (Kendall et al., 2018), while the collagen is mainly the preservation of skeletal material. Intrinsic parameters such as the degraded by microorganisms even if chemical hydrolysis may also occur

* Corresponding author. E-mail address: [email protected] (H. Matthiesen). https://doi.org/10.1016/j.jas.2020.105296 Received 29 May 2020; Received in revised form 3 November 2020; Accepted 19 November 2020 Available online 16 December 2020 0305-4403/© 2021 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). H. Matthiesen et al. Journal of Archaeological Science 125 (2021) 105296

(Child 1995; Collins et al., 2002). Aerobic microorganisms seem to be study both microbial and chemical degradation of different archaeo especially active in tunnelling and degradation (Kendall et al., 2018) logical materials (Matthiesen 2007; Mortensen and Matthiesen 2013; and during aerobic microbial catabolism amino acids from collagen are Watkinson and Rimmer 2013). It is non-destructive and allows repeated removed for both energy and protein synthesis (Child 1995). measurements on the same bone fragment under different conditions, Even if the key variables of bone degradation are known, it is difficult eliminating the need for grinding and homogenising the bones. Aerobic to quantify their overall importance due to the heterogeneity of both the oxygen consuming microorganisms are important in bone degradation bone material and the environment. Several detailed studies of bone (Child 1995; Kendall et al., 2018), but still, to our knowledge, respi material found at archaeological sites have been carried out (e.g. Hedges rometry has only been used in a single study of archaeological bone et al., 1995; Jans et al., 2004; Nielsen-Marsh and Hedges 2000; Todisco material (Tjellden´ et al., 2016) and a study of fresh bone (Pfretzschner and Monchot 2008), but this is seldom linked to equivalently detailed 2006). studies of the burial environment (Nielsen-Marsh et al., 2007). The studies are often restricted to a description of the current state of pres ervation of the bone and the environment is often described based on a 1.1. Site and methods single soil sample, while there is less focus on the dynamics of the ma terial and the environment. Burial experiments with fresh bone material The study is based on bone samples and environmental data have been carried out in various environments (Crowther 2002; Eriksen collected at five archaeological sites in the Nuuk fjord area in West et al., 2020; Nicholson 1996; Turner-Walker et al., 2008) whereas Greenland (Fig. 1). The sites are located along a transect stretching from controlled degradation experiments using archaeological bone material the sea in the West and approx. 100 km inland towards the Inland Ice are rare (Berna et al., 2004; Ottoni et al., 2009; Stiner et al., 1995), Sheet to the East. The capital Nuuk has an average yearly temperature of ◦ despite the fact that they may behave differently than modern bone. The 1.4 C and receives 781 mm precipitation (climate normal for the use of archaeological bone in burial and degradation experiments is period 1981–2010 - Cappelen 2019). However, the sites are subject to challenging as the material is often very heterogeneous, and in order to different weather conditions, with inland sites having almost twice as isolate the effects of environmental parameters it is advantageous to many thawing degree days (TDD) and receiving significantlyless rain as make repeated measurements on the same bone sample under different sites at the outer coast (Hollesen et al., 2019). conditions. This is sometimes solved by grinding and homogenising the bone material (Berna et al., 2004), however, destroying the physical Table 1 structure of the bone will definitely change its properties and remove Description of sites and test pits. The number of bones cannot be compared any evidence of the original heterogeneity. directly, as the excavated volume of the different test pits varied. Dating of the In this study we try to go beyond a static characterisation of the excavated midden layers was based on typology and former investigations, and environment and bone material, and contribute to a more quantitative the dates are only approximate. understanding of bone degradation. We do this by comparing data from Site Culture Approx. Dates Maximum Total number of a detailed monitoring of the burial environment with systematic in (CE) depth (cm) bones sampled vestigations of bone preservation at five archaeological sites in the Sandnæs Norse 1025–1350 60 59 Arctic. Furthermore, we test the potential of using respirometry to (Kilaarsarfik) obtain a quantitative measure of microbial respiration and aerobic Iffiartarfik Norse? 1025? 35 419 Inuit 1700–1800 degradation of bones. We do this by measuring the oxygen consumption Qoornoq, T1 Inuit 1800–1920 50 97 of archaeological bone samples at different temperatures and moisture Qoornoq, T2 Inuit 1700–1850 90 1073 contents. Respirometry is a well-established method to study biological Ersaa Inuit 1700–1900 45 55 ¨ – activity in soil (Ohlinger et al., 1996) that is increasingly being used to Kangeq Inuit 1800 1950 50 245

Fig. 1. Study region with five visited archaeological sites near the capital Nuuk.

2 H. Matthiesen et al. Journal of Archaeological Science 125 (2021) 105296

The sites contain preserved middens from Norse and/or Inuit in Table 2 habitants as described in Table 1. The middens are placed close to house Description and examples of the fivepreservation classes used in the study, from entrances and consist of different types of household waste. It is likely, excellent to very poor preservation. Decreasing numbers represent decreasing that most of the bones in the middens have been cooked during food state of preservation, as is also the case in e.g. the Oxford Histological Index of preparation. At each site small test excavations were carried out during bone preservation (Hedges et al., 1995). August 2016. The archaeological deposits were found just beneath the State of Description soil surface, capped by only 5–10 cm of turf. The test excavations were preservation dug all the way through the midden layers down to natural soil below, 5 - Excellent Bone surface looks fresh with no sign of cracking or flakingdue except for the sites Sandnæs and Kangeq, where excavations stopped at to weathering, surface details are intact. Bone has high density 50–60 cm depth due to the midden layers being frozen. At Qoornoq, test and is complete 4 - Good Bone shows modest cracking, normally parallel to the fiber pits were made in two different Inuit middens a few hundred meters structure (e.g. longitudinal in long bones). Ends may be missing apart, one of them being slightly older, thicker and rich in mussel shells on long bones. (T2). At Iffiartarfik the test pit presumably contained Norse material at 3 - Medium Partial loosening of surface (less than 50%), cracks parallel to the bottom and Inuit material at the top, but the small test pit did not fibre structure. Often low weight and fragmented bones. 2 – Poor The bone surface is coarsely fibrous and rough in texture, but allow exact dating. The excavations lasted from a few hours to a few still identifiable; more than 50% of the surface is loose; days, and the test pits were closed after installation of monitoring weathering penetrates into inner cavities, cracks are open equipment (described below). 1 – Very poor Surface is missing or is not identifiable. Bone may be falling apart in situ or be present as bone meal 1.2. Burial environment bones were selected to cover different states of preservation, including Environmental parameters were measured in situ for each 5–10 cm two ribs from seal from each of the categories (i.e. 10 bones). A further from the top to the bottom of the test pits. The water content was 14 bones were randomly selected from among the frozen material, to get measured with a WET probe from Delta-T Devices Ltd. (Cambridge, a first impression of the variations in microbial activity. The selection England). The pH was measured using a SS-37 ISFET electrode from included bones that could not be identified to species or type. Finally, Hach Company (Loveland, Colorado) as described in Matthiesen (2004). ◦ two samples of fresh seal bone (macerated at 37 C and brieflyheated to Soil ring samples of 100 cm3 volume were taken for each 10 cm and used boiling) were included as reference material of non-deteriorated bone. A for analysis of water content, organic content and soil porosity. list of the selected samples is shown in the supplementary material Continuous monitoring of soil temperatures was carried out using 107 (Table S1). sensors from Campbell Scientific INC (Logan, Utah) and Tiny Tag The microbial activity was measured using respirometry i.e. by Thermistor Probe PB-5001 from Geminidatalogger (Chichester, En measuring the oxygen consumption under controlled conditions. A sub- gland) and soil water contents were measured using SM300 and Theta sample of 0.5–1.5 g bone was placed in a 4.9 mL glass vial closed with an probes from Delta-T Devices Ltd. The sensors were installed at four airtight lid. Three subsamples were taken from each bone if the size depths in each test pit (except Qoornoq T2) and were connected to allowed (for six of the bones only two subsamples, and for three bones dataloggers programmed to log temperature every hour and water only one subsample could be taken). The bone material was not ground contents every 6 h (see Hollesen et al. (2019) for further details). The soil or homogenized as grinding may cause an artificially high oxygen ring samples were analysed in the laboratory, measuring the water ◦ ◦ consumption compared to whole samples (due to an increased surface content (drying at 105 C) and loss on ignition (after ignition at 450 C) area, increased access for microorganisms, and possibly formation of according to ASTM (2013). The loss on ignition was interpreted as the reactive radicals during grinding - Matthiesen and Mortensen 2011). organic content, and the soil porosity was calculated from the weights of Decreasing oxygen concentration inside each vial was measured with inorganic and organic material in the volume specific samples, using optical oxygen sensors (SensorDishReader from PreSens, Regensburg, estimated particle densities of 2.65 g/cm3 for inorganic and 1.50 g/cm3 Germany) and used to calculate an oxygen consumption rate according for organic material (Rühlmann et al., 2006). to Matthiesen (2007). The vials were kept in a climate chamber at fixed temperature for a week or more, measuring oxygen concentrations every 1.3. Bone preservation 15 min. Measurements were carried out at different water contents of the bone samples: moist samples (i.e. as found, but partially air dried Excavation of test pits was carried out in 10 cm strata or (where before freezing), saturated samples (where the samples had been sub visible) according to the stratigraphy of the midden. Soil samples from merged in deionised water under vacuum), dried samples stabilised at each soil strata were dry sieved (6 mm) and bone material was sampled. 95% RH in a climate chamber, and dried samples at 75% RH. The bone The state of preservation of the bone material was described visually and samples were weighed at all water contents, and their volume was each bone was categorized from 5 (Excellent) to 1 (Very poor) as determined (through buoyancy) as this influencesthe air volume of the described in Table 2. The scale describes typical types of damage on ◦ vials. The oxygen consumption measurements were carried out at 15 C, bones from the study area and examples (photos) of the different bone and for the saturated samples additional measurements were carried out categories have earlier been published in Harmsen et al. (2018). The ◦ at 5 and 10 C. descriptions are inspired by the weathering stages defined by Todisco After the oxygen consumption measurements the bone samples were and Monchot (2008) and by Behrensmeyer (1978) but some adjustments ◦ dried (48 h at 105 C) to determine their dry weight. One subsample were necessary to pick up the variations in our material. In total, 1948 from each bone was used to determine the weight losses (loss on igni bones were retrieved and categorized from the fivesites, and the number ◦ tion) when heated to 600 C for 12 h. The rest of the subsamples were of bones in the different preservation categories were summarised for used to determine the C and N content, where 2–3 mg ground bone each soil layer at each site. material was analysed on an elemental analyser (Flash, 2000; Thermo Scientific, Bremen, Germany). 1.4. Oxygen consumption

Parts of the bone material were frozen a few days after excavation ◦ and kept at 20 C until further analysis. In the laboratory 24 of the frozen bones were used for studying the microbial respiration and aer obic degradation at different temperatures and moisture contents. The

3 H. Matthiesen et al. Journal of Archaeological Science 125 (2021) 105296

2. Results variation with depth is presented in supplementary Figure S3, showing also several preservation categories at each depth, but overall with a 2.1. Burial environment better preservation in the deeper deposits. In order to analyse these data in more detail, a weighted average Some of the key parameters describing the burial environment are preservation category was calculated for each site and depth presented in Fig. 2. The soil porosity and water contents shown are ∑ 5 |Cat | ⋅ Number of bones in Cat based on soil ring samples. Soil porosities vary between 44 and 94% vol Average pres.cat. = i=1 i i Total number of bones with the highest values found in the upper soil. On the day of sampling the water contents were below 50% vol in all samples. The pore space In this way, the categorical data were transformed into numerical volume was partly filled with air with a calculated air content (soil data that are easier to visualize and analyse (Fig. 4). The preservation porosity minus water content) between 25 and 83% vol on the day of scale is not necessarily linear, and thus the average preservation cate sampling. Time series for the water contents (at four depths for all sites) gories are not true mean values, but may still be seen as representing the are presented in supplementary Figure S1. Overall, the time series show typical preservation at this site and depth. the largest temporal variations in the uppermost soil layers, with more The average preservation category varies between 1 and 3.8, with a stable conditions in the deeper deposits. At no point were the deposits tendency towards better preservation in the deeper deposits (Fig. 4). completely water saturated, except for a few days at the site Kangeq. The soil pH shows two different patterns: in four of the middens the 2.3. Bone analysis and oxygen consumption pH in the top soil is around 4–5, and increases with depth to pH 5–6.5 at 40 cm depth. At the two other middens (Iffiartarfikand Qoornoq T2) soil ◦ The results from loss on ignition at 600 C (LOI) of the 26 bones are conditions are more basic with pH 7–8 in most of the midden layers, presented in Table S1 in the supplementary material, and following with lower values at the very top and bottom of Qoornoq T2. Zioupos et al. (2008) and Mkukuma et al. (2004) this is interpreted as Time series of soil temperature are presented in supplementary the organic content of the bone. The LOI varies between 23 and 40% Figure S2, showing that the uppermost soil layers at most sites are frozen w/dw (weight loss relative to dry weight of the bones), with some of the from October to April, while the deeper soil layers are frozen for even higher contents (35%) found in the fresh seal ribs. Some of these organic longer periods. These have been used to model soil temperatures for the contents are relatively high, and for the archaeological bones it is likely period 2012–2016, as described in Hollesen et al. (2019). Thawing de that a partial leaching of the bone mineral has led to a relative increase gree days (TDD) are then calculated by summarising the daily mean in the organic content, or that there is a contribution from soil organic temperatures for a whole year, including only days with thaw (i.e. with matter or small roots. No reference values for fresh seal ribs have been positive mean temperatures), and Fig. 2 presents the average yearly TDD found, but Tont et al. (1977) have investigated fresh fin whale ribs, for the period 2012–2016. In the upper 10 cm the thawing degree days sperm whale and sea lion, and found organic contents of 36, 35 and 33% at the different sites vary between 400 and 700 TDD, and there is a rapid w/dw (recalculated from their density values for wet bones), which are decrease with depth to only 100–400 TDD at 40 cm depth and less than very close to the 35% found here. 100 TDD below 60 cm depth. Iffiartarfik and Ersaa only had relatively The results from C:N analyses are presented in Table S1 in the sup thin deposits and the TDD were only calculated for the upper 40 cm. The plementary material showing values between 3.5 and 4.9 mol/mol. Most thawing degree days are influenced by parameters such as air temper of the organic material in bone consists of collagen that has a low C:N ature and heat conductivity of the soil. value. The amino acid composition of seal collagen has earlier been analysed by one of the authors (MC) showing an average elemental 2.2. Bone preservation composition of C1H1.56N0.32O0.33S0.001 which corresponds to a molar C: N ratio of 3.1. The values measured here for bones were slightly higher The bone preservation was described for each site and each stratum indicating that they contain some material with a higher C:N ratio such by counting the number of bones in the different preservation categories as for instance soil organic matter, thin roots, or lipids from the bone (Supplementary Table S2). In Fig. 3 all bones from each site have been marrow. compiled in six histograms. There is a large variation with several The oxygen consumption rates at different temperatures and mois preservation categories being present at each site, but still it is evident ture contents are summarised in Table S3 in the supplementary material. that Qoornoq T1 has a relatively high share of very poorly preserved The consumption varies within four orders of magnitude (from 0.0001 bones (cat 1) while Qoornoq T2 has most bones in category 3 and 4. The to 1 mg O2/g dry weight/day) with large variations both between

Sandnæs Iffiartarfik Sandnæs Iffiartarfik Sandnæs (lab) Iffiartarfik Sandnæs Iffiartarfik Qoornoq, T1 Qoornoq, T2 Qoornoq, T1 Qoornoq, T2 Qoornoq,T1 Qoornoq,T2 Qoornoq Ersaa Ersaa Kangeq Kangeq Ersaa Kangeq Kangeq 0 0 0 0 -10 -10 -10 -10 -20 -20 -20 -20 -30 -30 -30 )

) -30 ) ) m m m m c -40 -40 -40 -40 -50 -50 -50 -50 Depth ( c Depth ( Depth ( c Depth ( c -60 -60 -60 -60 -70 -70 -70 -70 -80 -80 -80 -80 -90 -90 -90 -90 0 50 100 0 50 100 456789 0 500 1000 Porosity (% vol) Water content (% vol) pH Thawing degree days

Fig. 2. Key environmental parameters. The results for soil porosity and water content are based on ring samples taken together with the bones. The pH measurements are based on in situ measurements on the day of sampling (except for Sandnæs, where pH was measured in the laboratory on soil samples). The thawing degree days are average values for the period 2012–2016.

4 H. Matthiesen et al. Journal of Archaeological Science 125 (2021) 105296

Ersaa Iffiartarfik Kangeq

Qoornoq, T1 Qoornoq, T2 Sandnes

1 2 3 4 5 1 2 3 4 5 1 2 3 4 5

Fig. 3. Histograms showing the number of bones in different states for preservation for each of the six middens.

decay rate? Sandnæs Iffiartarfik Qoornoq, T1 Qoornoq, T2 3.1. Current state of preservation Ersaa Kangeq 0 The current state of preservation of the bone material at different sites depends on a large number of variables, including both deposi -10 tional history, type of bone and environmental conditions. Here we focus specificallyon the possible effects of the environmental conditions -20 by making a detailed monitoring of the burial environment surrounding ) -30 the bones. We pool different types of bones in order to have a sufficient

cm number of bones for statistical analysis. We use material from different (

h -40 sites in order to have a larger range and variation in the environmental t conditions. Finally, we focus on those environmental parameters that

Dep -50 earlier studies have pointed out as most important, i.e. the soil pH, the soil water content, the soil porosity, and the thawing degree days -60 (Damann and Carter 2014; Hedges and Millard 1995; Kendall et al., -70 2018). A number of multiple linear regression models are set up and tested -80 in R (ver 3.5.3), where the average preservation category of bones from 1 2 345 each site and layer (weighted with the number of bones in the layer) is Average preservation category correlated to the different variables, trying to find a good compromise (1: very poor; 5: excellent) between complexity and explanatory power of the model. A simple model including only the pH and thawing degree days (Supplementary Fig. 4. Average preservation category for the bone material. Figure S4 – left) can explain 47% of the variance in state of preservation (r2). Here the soil pH comes out very significant (P = 0.004) showing samples and between different temperatures and moisture conditions - that a high pH value gives a good preservation, as is well known from the background of this large variation is discussed below. There are only other studies (Crowther 2002; Gordon and Buikstra 1981; Pokines and few literature values of oxygen consumption of bone for comparison, but Baker 2014). The effect of the thawing degree days is best modelled by e.g. Pfretzschner (2006) measures a consumption of 0.1–0.15 mg including a second order term (P = 0.01), where a large number of ◦ ´ O2/g/day for saturated fresh bone at 20 C, and Tjellden et al. (2016) thawing degree days gives a poor preservation. This is in accordance report consumptions of 0.03–0.1 mg O2/g dw/day for saturated with studies from more temperate regions, where the Accumulated ◦ archaeological bone at 5 C, both within the ranges measured here Degree Days are often used as a key parameter to describe bone degra (Table S3). dation in forensics (Damann and Carter 2014). An alternative model including soil porosity and thawing degree days (Supplementary 3. Discussion Figure S4 – middle) can explain 57% of the variance, where the effect of the degree days are as described above and where a high soil porosity Three main issues are addressed: Can the current state of preserva gives a poor preservation (P = 0.0003) possibly because both air and tion of bone be linked to the current burial environment at the sites? water may move more freely in high-porosity soils (Hedges and Millard How does oxygen consumption relate to environmental conditions and 1995). Combining soil porosity, pH and thawing degree days in a single to bone characteristics? And how does oxygen consumption reflectbone model (Supplementary Figure S4 – right) explains slightly more of the variance (59%) but also makes the model more complex. Changing or

5 H. Matthiesen et al. Journal of Archaeological Science 125 (2021) 105296 expanding the models to include soil water content or air content does et al., 2019). The increased oxidation rate with increasing temperature is not improve the models (not shown). The lack of effect from soil mois also in line with the observation that the current state of preservation of ture on the bone preservation in this dataset is probably because the the bone material is correlated to the thawing degree days in the soil moisture conditions are not extreme in any of the pits: the soil is suffi (above). The fresh seal bones show some of the highest consumptions ciently wet so the microbes are not limited by lack of water, and at the and a weaker temperature response, but comparison to the archaeo same time the soil is sufficiently dry to avoid anoxic conditions, which logical bones should be done with caution due to possible differences in typically occur when the soil air content is lower than 10–15% vol the microbial communities on the bones and due to possible presence of (Matthiesen et al., 2016b). more reactive components (such as fat/oils) in the fresh bones. Overall, the effects from pH, soil porosity and degree days shown Correspondingly, Fig. 6 shows the oxygen consumption at different ◦ here are in line with observations from other studies in temperate re water contents of the samples, but at a fixed temperature (15 C). The gions (Crowther 2002; Gordon and Buikstra 1981; Hedges and Millard consumption rates cover four orders of magnitude, but with a clear 1995; Pokines and Baker 2014). The amount of rainwater or melting tendency that the oxygen consumption is lower for dried bones, which is water percolating down through the different soil layers could also be in line with the practical experience that archaeological bone material relevant to include in the models, however, it is very difficultto measure must be dried to avoid microbial growth (mould) on the bones (Grant or estimate on these sites where most of the precipitation comes as snow 2007). There is a tendency that the oxygen consumption plateaus at high that is unevenly distributed across the middens. Still, it is noteworthy water contents in Fig. 6, presumably due to a limited oxygen access that the effects from the current environment are so significant in our inside the water saturated bones, meaning that aerobic microbial material despite the fact that we have included different types of bones degradation will mainly take place at the bone surface. The low porosity from different middens with varying site formation histories. The visual and small pores already causes a steric hindrance to microbial degra bone characterisation used here does not allow us to distinguish be dation inside the bone (Kendall et al., 2018) and this effect may be tween different types of degradation, but the effects from both pH, soil enhanced if water in the pores reduces the rate of oxygen diffusion. porosity and thawing degree days indicates that both leaching of bone Besides the extrinsic environmental parameters, the oxygen con mineral and microbial decay of the organic part of the bone may be sumption may also be influenced by a large number of intrinsic pa important. The remainder of the paper will focus on aerobic microbial rameters regarding the bones themselves, such as e.g. species, bone type, decay.

) Saturated Moist 95% RH 75% RH 3.2. Oxygen consumption in relation to environmental conditions and y

da 10 bone characteristics / w d /g As mentioned previously bone degradation is highly complex and 2 1 O

depends on both environmental parameters and the characteristics of g m ( individual bones. The bone material is heterogeneous, and it can be n 0.1 o difficult to isolate effects from different parameters. However, with i pt respirometry it is possible to make repeated, non-destructive measure m 0.01 ments on individual bone samples, which makes it possible to isolate onsu c effects from environmental parameters such as temperature and water content: n 0.001 ge y

Fig. 5 shows the oxygen consumption of saturated bone samples at x O three different temperatures. The oxygen consumptions vary within 0.0001 three orders of magnitude, but for all archaeological bones there is a 5% 10% 20% 40% 80% 160% clear tendency towards an increase in the oxygen consumption with Water content of bone (%w/dw) increasing temperatures. The average Q10 value for the archaeological bones is 4.1 ± 2.1 (n = 24), i.e. the reaction rate increases four-fold Fig. 6. Oxygen consumption of 66 subsamples of bone, each measured with 4 ◦ ◦ when the temperature is increased by 10 C. This is in line with different conditions at 15 C. The water content (in % of dry weight) of each typical Q10 values observed for the oxidation of archaeological wood subsample was determined as the difference in weight before and after oven drying. Samples of fresh bone are shown with open symbols. Note logarithmic from the Arctic (Q10 = 4.2 ± 0.8, Matthiesen et al., 2014) but slightly scale on both axes. higher than measured for soil organic matter (Q10 = 2.3 ± 0.4, Hollesen

1 ) 539_1 539_2

/ da y 549_1 549_2 w

d 549_3 549_4 / g

2 540_1 540_2 O 0.1 540_3 540_4 g m

( 538_1 538_2

549_8 549_11

ptio n 549_13 548_1 m

u 0.01 548_2 548_3 549_5 549_6 on s c 549_7 549_9

gen 549_10 549_12 MOD_1 MOD_2 Ox y 0.001 2.557.5 10 12.5 15 17.5 Temperature (°C)

◦ Fig. 5. Temperature dependence of oxygen consumption for 26 bones (saturated) measured at 5, 10 and 15 C, results are given as mg O2/g dry weight/day. Results are mean values for 1–3 subsamples for each bone. Samples of fresh bone are shown with open symbols and dashed lines (MOD_1,2).

6 H. Matthiesen et al. Journal of Archaeological Science 125 (2021) 105296 pre-depositional treatment (e.g. heating), their state of preservation, speculative and require further studies to validate or reject. For instance, organic content, C:N ratio etc. It is not the purpose of the current study to it has not been possible to evaluate effects of different bone types, ani evaluate all these variables in detail, but still it is of interest to give a mal species or pre-depositional treatment, based on this limited sample preliminary discussion of the possible effect of some parameters, which material (Table S1). may then be elucidated through further studies using carefully selected bone material. ◦ 3.3. Oxygen consumption and bone decay rate Fig. 7 shows the oxygen consumption at 5 C for saturated bone (sub) samples. Overall, the oxygen consumption varies within 2 orders of Oxygen consumption studies allows a more detailed and quantitative magnitude, from approx. 0.002 to 0.25 mg O2/g dw/day, i.e. there is a study of the effects of different parameters (Matthiesen et al., 2014; large variation that cannot be explained simply by differences in the Hollesen and Matthiesen 2015). If we assume that the oxygen is used by extrinsic parameters (temperature, moisture), but needs to consider microorganisms to oxidise the organic content of the bone (which intrinsic parameters. Taking the loss-on-ignition (LOI, Fig. 7 left) first, mainly consists of collagen) a theoretical loss rate for the organic con there is a tendency that the oxygen consumption rate increases with tent may be calculated: increasing LOI. The LOI is normally interpreted as the organic content of The composition of seal collagen is C1H1.56N0.32O0.33S0.001. If we the sample (Mkukumi et al., 2004) and it seems reasonable that the assume that the organic content of the bones has a similar composition oxygen consumption increases when more organic material is present. A and that it is oxidised all the way to CO2, H2O, NO2 and SO2, it requires similar correlation has earlier been observed for e.g. archaeological soil 3.11 O-atoms per unit, giving a weight ratio of 2.13 g oxygen per g samples (Hollesen et al., 2016, supplementary S14). organic matter. The median oxygen consumption of saturated archae ◦ In soil science, C:N ratios are used as a measure for the reactivity and ological bones at 15 C is 0.067 mg O2/g dw/day (or 24.4 mg O2/g dw/ microbial activity in soils, where e.g. barley straw with a high C:N ratio year), which is equivalent to 11.5 mg organic matter/g dw/year being of 94 is considered a low-quality resource and cow manure with a low C: oxidised. For comparison the mean loss on ignition of the archaeological N ratio of ca 18 is considered a high-quality resource (Carter et al., bones is 307 mg/g dw. If we assume that the loss-on-ignition represents 2007). Organic material in bone is protein (i.e. nitrogen) dominated, organic matter it is approx. 3.8% of the organic matter that is oxidised in ◦ resulting in extremely low C:N ratios. The bones in this study had a a year at constant temperature of 15 C. Looking at the full range of – narrow range of C:N ratios (3.5 4.9) and there does not seem to be a measurements, the oxygen consumption of the different bones varied ◦ correlation between the C:N value and the bone reactivity (Fig. 7, between 0.012 and 0.97 mg/g dw/day at 15 C (supplementary middle). Table S2), which corresponds to annual losses of 0.7–54% of the (mean) Finally, the state of preservation of the bones seem to influence the organic content of the bones. The corresponding values for saturated ◦ oxygen consumption (Fig. 7, right). There is a large variation within bones at 5 C is a median loss of 0.7% of the organic content (range each category, which is not surprising as they include different types of 0.2–13%) and for the dry bones at 75% RH the median loss is 0.02% of bone. However, if we focus on the seal ribs (filled diamonds) there is a the organic content (ranging from below the detection limit and up to tendency that the highest consumption rates are found in category 2 0.11%). (poor preservation), while the consumption rates are lower both for The order of magnitudes of these calculated annual losses are – ◦ category 1 (very poor preservation) as well as category 3 5 (medium, considered realistic: They indicate that at 15 C organic matter in wet good and excellent preservation). In the best preserved bones (category bones under full oxic conditions may survive for a few years and up to 5) it is expected that the internal structure is least damaged. This means some decades, whereas for the dry bones the organic matter may survive ◦ that the bone mineral (apatite) can help protect the collagen fibres for several millennia. The rates at 5 C for wet bones indicate survival for ◦ against microbial attack, as the microbial enzymes as well as the mi decades and up to several hundred years. A constant temperature of 5 C crobes themselves are too large to fit into the collagen-apatite porosity in the laboratory corresponds to 1825 TDD per year which is much (Kendall et al., 2018; Turner-Walker 2007). The category 2 bones are higher than what is being observed at the study sites where temperatures ◦ partly degraded, and at the same time show some of the highest organic of 5 C are only reached during summer and only in the uppermost soil contents (supplementary figureS5 ). This indicates that some of the bone layers (Fig. 2 and Supplementary Figure S2). Consequently, the yearly mineral has been leached, which may increase the bone porosity and the loss of organic matter in situ is expected to be significantlylower than in accessibility of the collagen fibres. At the lowest end of the scale (cate the laboratory, indicating that the organic content may survive for gory 1) one may speculate if most of the easily degradable organic centuries or millennia even under full oxic conditions. This corresponds material is already gone, making the bones less attractive to microbial well with the fact that there are still significant amounts of organic decay. matter (LOI 23–40% - Supplementary Table S1) left in the bones after At this point the effects from intrinsic bone parameters are still some centuries in the ground, and also with the observation that the ) ) ) y y y 1 1 1 da da da / / / Ribs w w w d d d

g g Other /g / / 2 2 2 O O O

g g g 0.1 0.1 0.1 m m m ( ( (

on ion t ti ption p p m m m u u u s s s 0.01 0.01 0.01 on on on c c c

gen gen gen y y y x x x O O O 0.001 0.001 0.001 20% 30% 40% 2 3456 0 1 234 5 6 Loss-on-ignition (%w/dw) C:N ratio (mol/mol) Preservation category

◦ Fig. 7. Correlation between oxygen consumption and different bone parameters. All measurements are carried out with saturated bone samples at 5 C. Results are shown for individual sub-samples, and open symbols represent fresh bone.

7 H. Matthiesen et al. Journal of Archaeological Science 125 (2021) 105296 most degraded bones are found in the upper soil layers experiencing the Declaration of competing interests highest temperatures and most TDD (Figs. 2 and 4). The upper soil layers also experience the lowest pH and most leaching, giving further degra The authors declare that they have no known competing financial dation of the bones. In principle it is possible to calculate site specific interests or personal relationships that could have appeared to influence and depth specific theoretical yearly oxidation rates based on the the work reported in this paper. monitored environmental data – this has been done in the case of wood (Matthiesen et al., 2014) and for soil organic matter (Hollesen and Acknowledgements Matthiesen 2015; Hollesen et al., 2019). However, the bone material shows a much larger variation in oxygen consumption rate compared to The authors wish to thank Christian Koch Madsen and Hans Harmsen the wood and soil organic matter and we need a better understanding of from Greenland National Museum for archaeological fieldwork, inter the bone material itself and how and why the decay rate varies between pretation, and permission to carry out studies on the bone material, different bones before reliable estimates of bone decay rates in situ can Nanna Bjerregaard Pedersen for help with bone characterisation, Kris be produced. tian Murphy Gregersen from Natural History Museum for identifying Going from oxygen consumption to decay rate of bone is not trivial, bones, Sille Petersen for preparation of bones for oxygen consumption as we are not sure exactly which processes or organisms consumes the measurements, Per Lennart Ambus from CENPERM for carrying out C/N oxygen, and as there may be other decay processes (such as leaching at analyses, and David Gregory and two anonymous reviewers for com low pH or anaerobic microbial decay) that are not oxygen consuming. ments on the manuscript. VELUX FONDEN (33813) and Danish National Still, respirometry is considered a promising tool that can shed light on Research Foundation DNRF (128) are thanked for financial support. some of the bone decay processes, even if further studies are necessary to release its full potential. For instance, it is necessary to verify what the Appendix A. Supplementary data consumed oxygen is actually used for: we suggest that at least some of it is used to oxidise collagen, but the pathway must be studied in more Supplementary data to this article can be found online at https://doi. detail, focussing not only on the initial proteolytic steps of collagen org/10.1016/j.jas.2020.105296. degradation but also on the polypeptides and individual amino acids, to determine whether they are used by the microorganisms for protein References synthesis or oxidised and used as an energy source (Child 1995; Dippold and Kuzyakov 2013). Ideally, this should be combined with identifica ASTM, 2013. ASTM D2974-13: Standard Test Method for Moisture, Ash and Organic Matter for Peat and Other Organic Soils. ASTM International, West Conshohocken, tion of the aerobic microbial fauna in the bone (Child 1995) for instance PA. through DNA studies (Eriksen et al., 2020), and with a better under Behrensmeyer, A.K., 1978. Taphonomic and ecologic information from bone weathering. standing of the content and distribution of organic components in the Paleobiology 4, 150–162. Berna, F., Matthews, A., Weiner, S., 2004. Solubilities of bone mineral from bones in question for instance through chemical and histological ana archaeological sites: the recrystallization window. J. Archaeol. Sci. 31, 867–882. lyses (Kontopoulos et al., 2019). Alternative oxygen consuming pro Cappelen, J., 2019. Climatological Standard Normals 1981-2010 - Denmark, the Faroe cesses such as e.g. pyrite oxidation (Fellowes and Hagan 2003) need to Islands and Greenland - Based on Data Published in DMI Reports 18-02, 18-04 and be examined for the bones under study – accumulation of pyrite is 18-05. Danish Meteorological Institute. DMI Report 18-19. Carter, D.O., Yellowlees, D., Tibbett, M., 2007. Cadaver decomposition in terrestrial considered unlikely in the bones from this study (that come from oxic ecosystems. Naturwissenschaften 94, 12–24. environments), but it could be highly relevant at other sites. Finally, it is Child, A., 1995. Towards an understanding of the microbial decomposition of – necessary to combine the laboratory studies and theoretical decay rates archaeological bone in the burial environment. J. Archaeol. Sci. 22, 165 174. Collins, M., Nielsen-Marsh, C.M., Hiller, J., Smith, I., Roberts, J.P., Prigodich, R.V., with field studies where the degradation over time is monitored on Wess, T.J., Csapo, J., Millard, A., Turner-Walker, G., 2002. The survival of organic either modern or archaeological samples. matter in bone: a review. Archaeometry 44, 383–394. Crowther, J., 2002. The experimental earthwork at Wareham, Dorset after 33 years: retention and leaching of phosphate released in the decomposition of buried bone. 4. Conclusions J. Archaeol. Sci. 29, 405–411. Damann, F.E., Carter, D.O., 2014. Human decomposition ecology and postmortem Analysis of the state of preservation of bone material from fiveArctic microbiology. In: Pokines, J.T., Symes, S.A. (Eds.), Manual of Forensic Taphonomy. CRC Press, Boca Raton, pp. 37–49. sites showed a significant correlation between the current burial envi Dippold, M.A., Kuzyakov, Y., 2013. Biogeochemical transformations of amino acids in ronment and the preservation of the bone material. Multiple linear soil assessed by position-specific labelling. Plant Soil 373, 1–17. regression showed that especially the soil pH, the thawing degree days, Eriksen, A.M.H., Nielsen, T.K., Carøe, C., Matthiesen, H., Hansen, L.H., Gregory, D.J., Turner-Walker, G., Collins, M.J., Gilbert, M.T.P., 2020. Biodeterioration of bone - the and the soil porosity had an impact on the preservation. The soil mois effect of four marine and terrestrial depositional environments on bacterial ture content seemed to be of less importance at these sites, probably community of pig bone fragments. PloS One 15, 24. because the soil moisture contents were not extreme in situ, i.e. there Fellowes, D., Hagan, P., 2003. Pyrite oxidation: the conservation of historic shipwrecks – was sufficient moisture to allow microbial decay at all sites, but at the and geological and palaeontological specimens. Reviews in Conservation 4, 26 38. Gordon, C.C., Buikstra, J.E., 1981. Soil pH, bone preservation, and sampling bias at same time there wasn’t sufficient moisture to induce anoxic conditions mortuary sites. Am. Antiq. 46, 566–571. at any of the sites. Grant, T., 2007. Conservation of wet faunal remains: bone, antler and ivory. Canadian The degradation of bone under oxic conditions was studied further Conservation Institute. Canada. CCI Notes 4/3. Harmsen, H., Hollesen, J., Madsen, C.K., Albrechtsen, B., Myrup, M., Matthiesen, H., using respirometry, i.e. measurement of the oxygen consumption of 2018. A ticking clock? Assessing the deterioration of Greenlands archaeology in the bone samples. The method allows repeated, non-destructive measure 21st century. Conserv. Manag. Archaeol. Sites 20, 175–198. ments on individual samples under controlled conditions and gives a Hedges, R.E.M., Millard, A.R., 1995. Bones and groundwater: towards the modelling of diagenetic processes. J. Archaeol. Sci. 22, 155–164. quantitative understanding of the effects of parameters such as tem Hedges, R.E.M., Millard, A.R., Pike, A.W.G., 1995. Measurements and relationships of perature and bone moisture content. Furthermore, it allows a direct diagenetic alteration of bone from three archaeological sites. J. Archaeol. Sci. 22, comparison of the reactivity of different bone samples. The oxygen 201–209. Hollesen, J., Matthiesen, H., 2015. The influence of soil moisture, temperature, and consumption can be used to calculate theoretical rates for loss of organic oxygen on the decay of organic archaeological deposits. Archaeometry 57, 362–377. content of the bones, giving yearly losses that are realistic when Hollesen, J., Matthiesen, H., Fenger-Nielsen, R., Abermann, J., Westergaard-Nielsen, A., compared to field observations. Overall, the results show that respi Elberling, B., 2019. Predicting the loss of organic archaeological deposits at a regional scale in Greenland. Sci. Rep. 9 (9097). rometry is a promising tool for studying and quantifying degradation of Hollesen, J., Matthiesen, H., Møller, A.B., Westergaard-Nielsen, A., Elberling, B., 2016. bone organic content, but more studies are needed in order to obtain a Climate change and the loss of organic archaeological deposits in the Arctic. Sci. better understanding of the oxidation processes involved. Rep. 6, 28690.

8 H. Matthiesen et al. Journal of Archaeological Science 125 (2021) 105296

¨ Jans, M., Nielsen-Marsh, C.M., Smith, C.I., Collins, M., Kars, H., 2004. Characterisation of Ohlinger, R., Beck, T., Heilmann, B., Beese, F., 1996. Soil respiration. In: Schinner, F., ¨ microbial attack on archaeological bone. J. Archaeol. Sci. 31, 87–95. Ohlinger, R., Kandeler, E., Margesin, R. (Eds.), Methods in Soil Biology. Springer Kendall, C., Eriksen, A.M.H., Kontopoulos, I., Collins, M.J., Turner-Walker, G., 2018. Berlin Heidelberg, Berlin, Heidelberg, pp. 93–110. Diagenesis of archaeological bone and tooth. Palaeogeogr. Palaeoclimatol. Ottoni, C., Koon, H., Collins, M., Penkman, K., Rickards, O., Craig, O., 2009. Preservation Palaeoecol. 491, 21–37. of ancient DNA in thermally damaged archaeological bone. Naturwissenschaften 96, Kontopoulos, I., Penkman, K., McAllister, G.D., Lynnerup, N., Damgaard, P.B., 267–278. Hansen, H.B., Allentoft, M.E., Collins, M.J., 2019. Petrous bone diagenesis: a multi- Pfretzschner, H.-U., 2006. Collagen gelatinization: the key to understand early bone- analytical approach. Palaeogeogr. Palaeoclimatol. Palaeoecol. 518, 143–154. diagenesis. Palaeontographica Abt.A 278, 135–148. ´ Lopez-Costas,´ O., Lantes-Suarez,´ O., Martínez Cortizas, A., 2016. Chemical compositional Pokines, J.T., Baker, J.E., 2014. Effects of burial environment on osseous remains. In: changes in archaeological human bones due to diagenesis: type of bone vs soil Pokines, J.T., Symes, S.A. (Eds.), Manual of Forensic Taphonomy. CRC Press, Boca environment. J. Archaeol. Sci. 67, 43–51. Raton. Matthiesen, H., 2004. In situ measurement of soil pH. J. Archaeol. Sci. 31, 1373–1381. Rühlmann, J., Korschens,¨ M., Graefe, J., 2006. A new approach to calculate the particle Matthiesen, H., 2007. A novel method to determine oxidation rates of heritage materials density of soils considering properties of the soil organic matter and the mineral in vitro and in situ. Stud. Conserv. 52, 271–280. matrix. Geoderma 130, 272–283. Matthiesen, H., Hollesen, J., Dunlop, R., Seither, A., de Beer, J., 2016. Monitoring and Stiner, M.C., Kuhn, S.L., Weiner, S., Bar-Yosef, O., 1995. Differential burning, mitigation works in unsaturated archaeological deposits. Conserv. Manag. Archaeol. recrystallization, and fragmentation of archaeological bone. J. Archaeol. Sci. 22, Sites 18, 86–98. 223–237. Matthiesen, H., Jensen, J.B., Gregory, D., Hollesen, J., Elberling, B., 2014. Degradation of Tjelld´en, A.K.E., Matthiesen, H., Petersen, L.M.M., Emil Søe, N., Kristiansen, S.M., 2016. archaeological wood under freezing and thawing conditions - effects of permafrost In situ preservation solutions for deposited iron age human bones in alken enge, and climate change. Archaeometry 56, 479–495. Denmark. Conserv. Manag. Archaeol. Sites 18, 126–138. Matthiesen, H., Mortensen, M.N., 2011. Oxygen measurements in conserved Todisco, D., Monchot, H., 2008. Bone weathering in a periglacial environment: the archaeological wood. p. 123-135. In: Strætkvern, K., Williams, E. (Eds.), Proceedings Tayara site (KbFk-7),Qikirtaq island, Nunavik (Canada). Arctic 61, 87–101. from the 11th ICOM-CC Conference on Waterlogged Organic Archaeological Tont, S.A., Pearcy, W.G., Arnold, J.S., 1977. Bone-structure of some marine vertebrates. Materials, Greenville, May 2010. International Council of Museums, USA. Mar. Biol. 39, 191–196. Mkukuma, L.D., Skakle, J.M.S., Gibson, I.R., Imrie, C.T., Aspden, R.M., Hukins, D.W.L., Turner-Walker, G., 2007. The Chemical and Microbial Degradation of Bones and Teeth. 2004. Effect of the proportion of organic material in bone on thermal decomposition In: Pinhasi, R., Mays, S. (Eds.), Advances in Human Palaeopathology. Wiley, of bone mineral: an investigation of a variety of bones from different species using pp. 3–29. thermogravimetric analysis coupled to mass spectrometry, high-temperature X-ray Turner-Walker, G., Peacock, E., Gilbert, T., Koon, H.E.C., 2008. An experimental study of diffraction, and fourier transform infrared spectroscopy. Calcif. Tissue Int. 75, morphological and chemical degradation of bone in wetlands: potential for DNA 321–328. extraction and amplification. In: Kars, H., van Heeringen, R.M. (Eds.), Proceedings Mortensen, M.N., Matthiesen, H., 2013. Oxygen consumption by conserved from the Conference "Preserving Archaeological Remains in Situ 3, pp. 75–84. archaeological wood. Anal. Bioanal. Chem. 405, 6373–6377. Amsterdam December 2006, Amsterdam. Nicholson, R.A., 1996. Bone degradation, burial medium and species representation: Watkinson, D., Rimmer, M., 2013. Quantifying effectiveness of chloride desalination debunking the myths, an experiment-based approach. J. Archaeol. Sci. 23, 513–533. treatments for archaeological iron using oxygen measurement. Conference Nielsen-Marsh, C.M., Hedges, R.E.M., 2000. Patterns of diagenesis in bone I: the effects of Proceedings from the Metal 2013 Conference. Edinburgh, Scotland from September site environments. J. Archaeol. Sci. 27, 1139–1150. 16 - 20, 2013. Nielsen-Marsh, C.M., Smith, C.I., Jans, M., Nord, A.G., Kars, H., Collins, M., 2007. Bone Zioupos, P., Cook, R.B., Hutchinson, J.R., 2008. Some basic relationships between diagenesis in the European Holocene II: taphonomic and environmental density values in cancellous and cortical bone. J. Biomech. 41, 1961–1968. considerations. J. Archaeol. Sci. 34, 1523–1531.